Split (1)

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

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2152270649	http://journals.iucr.org/d/issues/2012/08/00/nj5116/nj5116.pdf	English	null	Nanolitre-scale crystallization using acoustic liquid-transfer technology	Acta crystallographica. Section D, Biological crystallography	2,012	cc-by	6,253	research papers Nanolitre-scale crystallization using acoustic liquid-transfer technology Acta Crystallographica Section D Biological Crystallography ISSN 0907-4449 Armando G. Villasen˜or,a‡ April Wong,a Ada Shao,a Ankur Garg,b Timothy J. Donohue,b Andreas Kuglstatterc and Seth F. Harrisd* Armando G. Villasen˜or,a‡ April Wong,a Ada Shao,a Ankur Garg,b Timothy J. Donohue,b Andreas Kuglstatterc and Seth F. Harrisd* Focused acoustic energy allows accurate and precise liquid transfer on scales from picolitre to microlitre volumes. This technology was applied in protein crystallization, successfully transferring a diverse set of proteins as well as hundreds of precipitant solutions from custom and commercial crystal- lization screens and achieving crystallization in drop volumes as small as 20 nl. Only higher concentrations (>50%) of 2-methyl-2,4-pentanediol (MPD) appeared to be system- atically problematic in delivery. The acoustic technology was implemented in a workﬂow, successfully reproducing active crystallization systems and leading to the discovery of crystallization conditions for previously uncharacterized proteins. The technology offers compelling advantages in low-nanolitre crystallization trials by providing signiﬁcant reagent savings and presenting seamless scalability for those crystals that require larger volume optimization experiments using the same vapor-diffusion format. aDepartment of Discovery Technologies, Roche Palo Alto LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA, bEDC Biosystems, 49090 Milmont Drive, Fremont, CA 94538, USA, cDepartment of Discovery Technologies, F. Hoffmann-La Roche, 4070 Basel, Switzerland, and dDepartment of Structural Biology, Genentech, 1 DNA Way, South San Francisco, CA 94080, USA Correspondence e-mail: harris.seth@gene.com ‡ Present address: Department of Structural Chemistry, Gilead Sciences Inc., 333 Lakeside Drive, Foster City, CA 94404, USA. Received 20 January 2012 Accepted 16 April 2012 aDepartment of Discovery Technologies, Roche Palo Alto LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA, bEDC Biosystems, 49090 Milmont Drive, Fremont, CA 94538, USA, cDepartment of Discovery Technologies, F. Hoffmann-La Roche, 4070 Basel, Switzerland, and dDepartment of Structural Biology, Genentech, 1 DNA Way, South San Francisco, CA 94080, USA Acta Cryst. (2012). D68, 893–900 Correspondence e-mail: harris.seth@gene.com research papers wells. We designed our acoustic setups to deliver a ﬁrst pass transferring the precipitant solutions (typically 96 wells, each delivering 20 nl) followed by a second pass delivering the protein (typically 20 nl) on top of each of the just-dispensed 96 precipitant solutions. The volumetric ratio of precipitant to protein was predeﬁned by the user with standard text-ﬁle inputs (i.e. ‘ejection maps’) designating volume and destina- tion wells for each liquid. Similarly, ‘energy maps’, used in some of our experiments, speciﬁed the amplitude and time base of the acoustic energy pulse for each separate well. In contrast, for the ﬁxed-energy benchmark used in most of the experiments described here the instrument was set to an amplitude of 1.6 and a burst of 350 (manufacturer’s scale) for all wells. All crystallization drops were assembled with the instrument’s volume calibration at 5 nl delivery per energy pulse. The entire setup of one 96-well crystallization plate was achieved in 2.5 min. The completed plate was immediately sealed with CrystalSeal Film (Hampton Research, Aliso Viejo, California, USA) to minimize evaporation. Source-plate and target-plate geometries were entered using well-spacing dimensions from the manufacturers with distal corners as ﬁducial markers. Additionally, a ﬂat-surface micro-plate lid (Corning 3950) was used as an alternative target when testing individual liquids and their ability to transfer. liquid meniscus up onto inverted target destinations with no physical contact with the sample. In prior review, acoustic dispensing showed unmatched precision (<2% CV) at sub- 100 nl volumes compared with several of the technologies described above (Comley, 2004). Furthermore, the lack of ﬂuidic lines and tips greatly reduces consumables, main- tenance, washing steps, clogs and cross-contamination. To achieve acoustic liquid transfer, a small transducer bathed in a coupling ﬂuid (e.g. water) to effect efﬁcient energy transfer is positioned under a micro plate (‘source plate’). Acoustic reﬂections can determine the liquid level in the source wells and thereby the appropriate focal distance to position the transducer. The focused energy leads to defor- mations at the liquid–air interface and the ejection of nano- litre-scale droplets on a millisecond time scale. These droplets travel upwards to contact and accumulate on an inverted target plate, building up to the prescribed volume. research papers The tech- nology has rapidly expanded in the realm of small-molecule compound library management and high-throughput screening groups, where the compounds of interest are typi- cally in a well characterized organic solvent such as dimethyl sulfoxide (DMSO; Ellson et al., 2005; Olechno et al., 2006). Further research has also demonstrated its utility in micro- array setups (Wong & Diamond, 2009) and MALDI imaging mass spectroscopy (Aerni et al., 2006). 1. Introduction Crystallization experiments, as with many scientiﬁc ﬁelds, have followed a trend towards increased automation and minia- turization to improve efﬁciency and accelerate discovery (Stevens, 2000). Tip-based liquid dispensing, the natural extension of the most common manual tools, has been a robust method for setting up crystallization trials well into the sub- microlitre range, with typical volumes of 100–200 nl (e.g. Mosquito by TTP LabTech, Royston, England, Phoenix by Art Robbins, Sunnyvale, Califonia, USA and Crystal Creator by Cybio AG, Jena, Germany, among others). Yet smaller volumes are enabled by inkjet-type solenoid dispensers (Rose, 1999), which aspirate liquid into a tubing line and dispense from a tip without making contact with the destination surface. Such non-contact dispensers still contact the sample at the dispensing head and as such are still prone to surface-contact artifacts, clogs and possible cross-contamination from viscous solutions (Walter et al., 2003). Microﬂuidic methods also achieve small-volume experiments (10 nl), but are constrained by the physical parameters of the ﬂuidics channels and the resulting impact on crystal harvesting and challenges in translation from free-interface diffusion, for example, to more conventional methods (Fluidigm, South San Francisco, California, USA and Plugmaker, Emerald BioSystems, Bain- bridge Island, Washington, USA). Acoustic liquid transfer, as implemented in the ATS-100 (EDC Biosystems, Fremont, California, USA) and the Echo series instruments (Labcyte, Sunnyvale, California, USA), uses focused sound energy to eject single-digit nanolitre droplets directly from a source- 893 Acta Cryst. (2012). D68, 893–900 doi:10.1107/S0907444912016617 2.2. Transfer-energy adjustments based on acoustic liquid-level measurements Motivated by the advantages described above and a potential saving of an order of magnitude in protein consumption relative to other vapor-diffusion methods, we explored the application of this technology in protein crys- tallography. A primary concern was whether the instrument, which was calibrated for performance with DMSO, would work acceptably with the diverse range of aqueous solutions involved in typical crystallization screening. We anticipated that the well-by-well energy tuning available on the ATS-100 instrument would permit us to mitigate the impact of such distinct physicochemical properties of the crystallization cocktails (e.g. viscosity and surface tension). The instrument uses acoustic reﬂection to measure liquid levels in each source well. We used these data to establish liquid levels before and after dispensing 200 droplets for each well and thereby the average change in volume per single drop ejection. We tested dispensing in ﬁve different screens: Nextal JCSG+, Hampton Research Index HT, Nextal Classics, Emerald BioSystems Wizards I and II, and an in-house medium-weight polyethylene glycol grid [‘PEG MW’, sampling 5, 15, 20 and 25%(w/v) PEG 3350, PEG 5000 MME and PEG 6000 across the pH range 3.5–9.8 using the buffer systems sodium lactate, sodium acetate, 2-(N-morpholino)- ethanesulfonic acid (MES), imidazole, potassium phosphate, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 2-[(2-aminoacetyl)amino]acetic acid (Gly-Gly) and glycine]. Each sample was run six times (two source-plate replicates, each run three times) and the average drop volume was determined. The residual difference from the target value of 5 nl was calculated (x in the equations below) and was used to set a revised burst energy from an empirically derived adjustment factor as a proof-of-concept gauge of improved accuracy of delivery. Liquids that lacked a liquid-level reading were excluded. For liquids that delivered more than 5 nl (i.e. x < 0) the new burst was given by Villasen˜or et al. Acoustic liquid transfer Table 1 Table 1 Formulations of 33 crystallization solutions (28 unique) that failed to dispense from a total of 480 from the JCSG+, Classics, Index, Wizards and PEG MW screens. MW screens. The ten paired duplicates in this list are marked in bold. 26 of these solutions (21 unique) contained MPD (2-methyl-2,4-pentanediol). JCSG+ No. 79, Classics No. 45 and Wizards No. 88 are marked with an asterisk as these conditions were transferred measurably in another screen. nm, not measured. Solution No. Table 1 Formulation Viscosity (mPa s) Density (g cm3) JCSG+ 41 (137) 70%(v/v) MPD, 0.1 M HEPES pH 7.5 21.3 0.98 JCSG+ 11 (107) 50%(v/v) MPD, 0.2 M ammonium phosphate, 0.1 M Tris pH 8.5 10.2 1.01 JCSG+ 88 (184) 45%(v/v) MPD, 0.2 M calcium chloride, 0.1 M bis-tris pH 5.5 8.2 1.02 JCSG+ 89 (185) 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M bis-tris pH 5.5 7.7 1 JCSG+ 96 (192) 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M HEPES pH 7.5 7.8 1 JCSG+ 17 (113) 40%(v/v) MPD, 0.1 M cacodylate pH 6.5 14.8 1.01 JCSG+ 44 (140) 40%(v/v) MPD, 0.1 M Tris pH 8 5.6 1 JCSG+ 53 (149) 40%(v/v) MPD, 0.1 M CAPS pH 10.5 6.1 1 JCSG+ 61 (157) 30%(w/v) Jeffamine M-600, 0.05 M cesium chloride, 0.1 M MES pH 6.5 5.9 1.03 JCSG+ 37 (133) 24%(w/v) PEG 1500, 20%(v/v) glycerol 11.04 1.1 JCSG+ 79 (175)* 15%(w/v) PEG 3350, 0.1 M succinic acid pH 7.0* 4.83 1.03 Classics 22 (22) 70%(v/v) MPD, 0.1 M HEPES pH 7.5 19 0.98 Classics 21 (21) 50%(v/v) MPD, 0.2 M ammonium phosphate, 0.1 M Tris pH 8.5 9.8 1 Classics 17 (17) 30%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M sodium citrate pH 5.6 4.4 1.02 Classics 18 (18) 30%(v/v) MPD, 0.2 M magnesium acetate, 0.1 M cacodylate pH 6.5 5 1.02 Classics 19 (19) 30%(v/v) MPD, 0.2 M sodium citrate, 0.1 M HEPES pH 7.5 5.3 1.04 Classics 20 (20) 30%(v/v) MPD, 0.5 M ammonium sulfate, 0.1 M HEPES pH 7.5 4.8 1.04 Classics 45 (45)* 1.6 M trisodium citrate* 7.3 1.25 Index 48 45%(v/v) MPD, 0.2 M calcium chloride dihydrate, 0.1 M bis-tris pH 5.5 nm nm Index 49 45%(v/v) MPD, 0.2 M calcium chloride dihydrate, 0.1 M bis-tris pH 6.5 nm nm Index 50 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M bis-tris pH 5.5 nm nm Index 51 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M bis-tris pH 6.5 nm nm Index 52 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M HEPES pH 7.5 nm nm Index 53 45%(v/v) MPD, 0.2 M ammonium acetate, 0.1 M Tris pH 8.5 nm nm Index 58 45%(v/v) polypropylene glycol P 400, 0.1 M bis-tris pH 6.5 nm nm Index 56 35%(v/v) pentaerythritol propoxylate, 0.2 M potassium chloride, 0.05 M HEPES pH 7.5 nm nm Wizards 4 35%(v/v) MPD, 0.2 M magnesium chloride, 0.1 M imidazole pH 8.0 nm nm Wizards 24 35%(v/v) MPD, 0.2 M sodium chloride, 0.1 M Tris pH 7.0 nm nm Wizards 43 35%(v/v) MPD, 0.1 M sodium/potassium phosphate pH 6.2 nm nm Wizards 50 35%(v/v) MPD, 0.2 M lithium sulfate, 0.1 M MES pH 6.0 nm nm Wizards 69 35%(v/v) MPD, 0.1 M sodium acetate pH 4.5 nm nm Wizards 73 35%(v/v) MPD, 0.2 M sodium chloride, 0.1 M HEPES pH 7.5 nm nm Wizards 88* 20%(w/v) PEG 3000, 0.2 M zinc acetate, 0.1 M imidazole pH 8.0* nm nm A linear slope extrapolation was used for simplicity, antici- pating more sophisticated ﬁts based on knowledge learned. Table 1 The coefﬁcients of variation were calculated by standard methods, 2.4. Protein reagents Six proteins with known crystallization conditions were employed to evaluate the reproduction of crystallization in acoustically dispensed drops. These proteins were hepatitis C virus helicase (HCV helicase), human serum albumin (HSA), hepatitis C virus RNA-dependent RNA polymerase (HCV polymerase), human immunodeﬁciency virus reverse tran- scriptase (HIV RT), human IL2-inducible T-cell kinase (ITK) and chicken egg-white lysozyme. Visible protein crystals were obtained as follows. For HCV helicase (puriﬁcation modeled as described in Kim et al., 1998), we assembled crystallization drops with 15 or 25 nl 15%(w/v) PEG 6000, 0.2 M lithium sulfate, 0.02 M magnesium chloride, 0.1 M Tris pH 8.0 and 15 or 25 nl 10 mg ml1 protein. The same volumes were used for HSA (catalog No. pro369; ProSpecBio, Rehovot, Israel), where we mixed 27%(v/v) PEG 550 monomethyl ether, 0.05 M magnesium chloride, 0.1 M HEPES pH 7.5 and 25 mg ml1 protein in 0.05 M potassium phosphate pH 7.2. HCV polymerase drops were assembled with 20 nl 24%(w/v) PEG 4000, 7.5%(v/v) glycerol, 0.05 M sodium citrate pH 4.6 and 40 nl 10 mg ml1 protein puriﬁed as described by Le Pogam et al. (2006). HIV RT drops were formed with 15 nl CV ¼ ðstandard deviation=average valueÞ 100: 2.1. Hardware and consumables Standard tip-based liquid dispensing (Cybiwell liquid handler from Cybio AG, Jena, Germany) was employed in advance to ﬁll the large reservoirs in crystallization plates (Corning 3785 96-well sitting-drop vapor-diffusion plate; Corning Life Sciences, Lowell, Massachusetts, USA) with 25 ml per well and also to ﬁll the source plates (384 IQ-LV, 200 mm; Aurora Biotechnologies, Carlsbad, California, USA) with custom and commercial screens with 15 ml per well. All acoustic liquid transfers were performed on the EDC Biosystems ATS-100. The protein sample to be transferred was pipetted into a single well (15 ml) in the source plate just before the start of the acoustic run. A gripper inverts the crystallization plate (target) over the source plate during transfer, with each plate free to translate relative to the other such that any source well can be addressed to any of the target burst ¼ 30x þ 350; burst ¼ 30x þ 350; while for liquids that delivered less than 5 nl (x > 0) the new burst was given by while for liquids that delivered less than 5 nl (x > 0) the new burst was given by burst ¼ 140x þ 350: burst ¼ 140x þ 350: Acta Cryst. (2012). D68, 893–900 894 Villasen˜or et al. Acoustic liquid transfer Figure 1 g Measurement of viscosity (green circles) and density (red squares) of solutions from the Classics (a) and JCSG+ (b) screens. The larger symbols represent solutions that failed to dispense and show no clear trend toward either extreme. 1.2 M sodium malonate, 5%(v/v) ethylene glycol, 0.1 M potassium phosphate pH 7.2 and 15 nl 10 mg ml1 protein (puriﬁed as described previously; Sweeney et al., 2008) inhib- ited with 0.25 mM nevirapine. ITK drops were set up with 30 nl 25%(w/v) PEG 3350, 0.2 M ammonium acetate, 0.1 M HEPES pH 7.5 and 30 nl 10 mg ml1 triple mutant C477S/ E614A/E617A (puriﬁed as described previously; Kutach et al., 2010). Lysozyme drops were formed with 15 nl 0.9 M sodium chloride, 0.1 M sodium acetate pH 5.0 and 15 nl 10 mg ml1 protein (catalog No. L6876; Sigma, Saint Louis, Missouri, USA) dissolved in 0.1 M sodium acetate pH 4.6. The ATS-100 instrument was set at ﬁxed energy for all experiments with these proteins. All vapor-diffusion trials featured large reservoirs ﬁlled with only 15 or 25 ml of the corresponding precipitant to prevent liquids from dripping from inverted plates during acoustic dispensing. The experi- ments were sealed with Clear Seal tape (Hampton Research, Aliso Viejo, California, USA) immediately after setup. The drops were then incubated at 293 K and digitally imaged in a Crystal Farm imager (Nexus Biosystems, Poway, California, USA) on days 0, 3, 5, 10, 15 and 30. We wished to characterize the successful liquid transfers for reproducibility and accuracy. The standard calibration method of ﬂuorescein-doped deliveries was impractical for the scale of this work. We used the acoustic reﬂection measurements of liquid heights (included in the instrument run logs) before and after a series of ejections to determine the volume delivered per droplet compared with the calibrated setting (5 nl). While the effect of the distinct solutions on the speed of sound confounds the absolute accuracy of these measurements, the differential nature of the pre- and post-delivery determination provides a relative difference that ought to be robust. Using six repeat experiments, Fig. 2 plots the number of conditions in each screen within CV categories binned by 5% increments. Each screen demonstrates that the majority of conditions have good reproducibility (CV < 10%). In addition, the PEG MW 2.3. Viscosity measurements The viscosities and densities of solutions from the JSCG+ and Classics screens were determined using a 30-tube sampler to feed an Anton Paar SP3-V viscosity meter in tandem with a DMA4500 density meter (Anton Paar, Graz, Austria). A total of 10 ml of each solution was loaded into 12 ml glass vials in the auto sampler. Average values were calculated from a total of four individual measurements per solution at 293 K. Each viscosity measurement was based on the free-fall duration of a 1.5 mm diameter gold-plated steel ball in a 1.6 mm diameter capillary 100 mm in length with an inclination of 60. Free-fall durations were marked by an inducible coil sensor at both ends of the capillary. 895 Acta Cryst. (2012). D68, 893–900 Villasen˜or et al. Acoustic liquid transfer research papers Figure 1 Measurement of viscosity (green circles) and density (red squares) of solutions from the Classics (a) and JCSG+ (b) screens. The larger symbols represent solutions that failed to dispense and show no clear trend toward either extreme. 284 unique conditions; Newman et al., 2010) as well as a 96-condition medium- weight PEG grid. Using a ﬁxed energy, 93% (447/480) of the precipitant solu- tions transferred successfully and the remaining 7% (33 conditions) failed to dispense. We were curious whether the viscosity of the liquid sample was a predictor of poor transfer capabilities. Fig. 1 shows the viscosities and densities of difﬁcult and well behaved solutions from the Classics and JCSG+ screens, demonstrating that there is no correla- tion linking viscosity or density alone to inability to dispense. Surface tension is a dominant factor in the mechanism of droplet ejection, but we could not readily characterize this property. A full description of the contents of the 33 problematic solutions is shown in Table 1. It was immediately obvious that the majority of the non-transferring conditions (26 out of 33) contained high concentrations of 2-methyl-2,4-penta- nediol (MPD). Conditions with MPD at less than 30% concentration could be transferred. Three samples including 30% MPD with observed transfer all contained 0.1 M sodium acetate pH 4.6 A full description of the contents of the 33 problematic solutions is shown in Table 1. It was immediately obvious that the majority of the non-transferring conditions (26 out of 33) contained high concentrations of 2-methyl-2,4-penta- nediol (MPD). Conditions with MPD at less than 30% concentration could be transferred. 2.3. Viscosity measurements Three samples including 30% MPD with observed transfer all contained 0.1 M sodium acetate pH 4.6 and a chloride salt, while other buffered 30% MPD solutions failed, as seen in Table 1. All samples with greater than 30% MPD failed to transfer. Further work has suggested that MPD attenuates acoustic energy and that stronger input enables the delivery of solutions up to 50% MPD (J. Bramwell, Labcyte, personal communication). Of the seven non-MPD-containing failed wells, three are represented in other screens, where measurable transfer was observed (marked with asterisks in Table 1). A few other unusual constituents such as branched polymers such as pentaerythritol propoxylate and poly- propylene glycol are present in two non-transferring wells, but as these are less common our data do not inform whether or not their impact is systematic. s from the Classics (a) dispense and show no Figure 1 Table 2 Aggregate and ‘by screen’ statistics of droplet volume distribution. Mean volume (nl) and standard deviation (SD) are shown to indicate bias and variance (conditions that did not dispense were excluded). Root-mean-square deviations (r.m.s.d.s) of the residuals (relative to 5 nl target volume) are calculated as a gauge of overall accuracy. Figure 2 Figure 2 The distribution of coefﬁcients of variation (CVs) for reproducibility in ﬁve crystallization screens calculated from six repeat dispenses. Each block of color shows the count of conditions from that screen within the prescribed range. The brown segment tallies wells that did not transfer and hence have no calculated CV. Screen Energy proﬁle Mean volume SD (r.m.s.d.) JCSG+ Fixed 4.81 1.17 (1.18) Tuned 5.41 0.52 (0.66) Classics Fixed 5.46 1.69 (1.74) Tuned 5.59 0.91 (1.08) Wizards Fixed 5.14 1.22 (1.22) Tuned 5.94 0.85 (1.26) PEG MW Fixed 5.46 0.82 (0.94) Tuned 5.75 0.59 (0.95) Aggregate Fixed 5.23 1.28 (1.30) Tuned 5.67 0.76 (1.01) Screen Energy proﬁle Mean volume SD (r.m.s.d.) in the commercial screens (Newman et al., 2010) also provided 84 duplicate solutions from different screens with observed transfer. As an additional measure of reproducibility, the mean difference between the measured volume per droplet for these 42 pairs was only 0.22 0.21 nl. Figure 2 g The distribution of coefﬁcients of variation (CVs) for reproducibility in ﬁve crystallization screens calculated from six repeat dispenses. Each block of color shows the count of conditions from that screen within the prescribed range. The brown segment tallies wells that did not transfer and hence have no calculated CV. p y We used the measured residual volume differences (versus a 5 nl target) to calculate semi-empirical adjusted energy para- meters for individual wells to gauge potential improvements versus ﬁxed-energy operation. Fig. 3 plots these residuals from four screens, showing a much tighter distribution when well- by-well energy adjustments were applied. This is quantitated in Table 2, where the standard deviation over all wells of the absolute volume (i.e. not the residual) relative to the mean of the distribution improves signiﬁcantly from 1.28 to 0.76. Naturally, the more chemically diverse sparse-matrix screens have a higher absolute deviation than the limited chemistry of the PEG MW screen, but the salient metric is in the improvement (smaller standard deviation) in each screen upon adjusted-energy dispenses (Table 2). 3. Results Our panel of test screens included several sparse-matrix kits to ensure a good sample of the conditions encountered in crys- tallization trials. Four 96-well commercial screens were tested (JCSG+, Index, Classics and Wizards I and II, representing 896 896 Villasen˜or et al. Acoustic liquid transfer Acta Cryst. (2012). D68, 893–900 Table 2 Aside from proﬁling the absolute volume delivered, we gauged accuracy by calculating the root-mean-square deviation of the residuals (relative to the 5 nl target volume). These r.m.s.d.s showed slightly less improvement than the standard deviations of the absolute volumes (residual r.m.s.d. of 1.30 A˚ at ﬁxed energy to 1.01 A˚ with ‘tuned’ energies). It is apparent that our coarsely calculated energy adjustments led to a slightly larger average volume delivered (5.23 nl per droplet mean volume at ﬁxed energy, 5.67 nl mean with tuned energies) and this systematic bias, which is purposefully captured in the residual r.m.s.d., partly offsets the gains of the tighter distribution relative to the mean. Adjusting the constant parameter in our admittedly simplistic ﬁrst-pass energy-adjustment calculations would presumably improve the accuracy. Furthermore, Fig. 3(b) highlights a segment of the PEG MW screen residual plot in which each set of four points represents increasing concen- trations of PEG (5, 15, 20 and 25%). Overall, the energy adjustments yield a much ﬂatter range within each group of four data points. Of note, however, is that those highest PEG concentrations (the fourth point of each set) with the most reduced delivery volumes at ﬁxed energy resulted in over-compensation, a phenomenon that accounts for many of Figure 3 Figure 3 research papers screen, with a moderate diversity of components, had superior precision, with 93 of 96 conditions showing CVs of less than 10% (and many of less than 2%). The overlap of conditions screen, with a moderate diversity of components, had superior precision, with 93 of 96 conditions showing CVs of less than 10% (and many of less than 2%). The overlap of conditions research papers the higher scatter points in Fig. 3(a) and contributes to the upward bias in average volumes. The adjustment energies would be likely to improve with an asymptotic curve instead of our linear extrapolation. These studies investigating the rela- tionship between energy parameters and volume delivery of different chemical series helped to inform more sophisticated developments for the acoustic methodologies that are now the higher scatter points in Fig. 3(a) and contributes to the upward bias in average volumes. The adjustment energies would be likely to improve with an asymptotic curve instead of our linear extrapolation. These studies investigating the rela- tionship between energy parameters and volume delivery of different chemical series helped to inform more sophisticated developments for the acoustic methodologies that are now forthcoming. However, given the complexity of determining and implementing the energy maps in the current format versus the practical functionality and emerging successes of the ﬁxed-energy operation, we performed all our subsequent crystallization trials using the simpler ﬁxed-energy setting. tionship between energy parameters and volume delivery of different chemical series helped to inform more sophisticated developments for the acoustic methodologies that are now crystallization tria We continued samples would c 898 Vill ˜ l A i li id f Figure 4 Images collected on day 15 from crystallization drops of various protein samples. The total drop volume is double the protein value given. The volumes are representative and were not optimized as similar crystals appeared at the various volumes explored. (a) HCV helicase (50 nl). (b) Human serum albumin (50 nl). (c) HCV polymerase (30 nl). (d) HIV RT (15 nl). (e) ITK (30 nl). (f) Lysozyme (15 nl). We continued experiments to verify that diverse protein samples would crystallize using this new application of acoustic technology. In addition to a trivial control (lysozyme), we were able to reproduce the crystallization of HCV helicase, HCV polymerase, HSA, HIV- RT and ITK (ﬁve in-house targets). Crystals of all six proteins were visible by day 5 and most stopped growing by day 15. Fig. 4 shows typical crystals; various representative volumes were used in these trials but were not a necessary part of optimization. y p p Finally, we experimented with the delivery of a set of oils to look at alternative ways to reduce evaporation from the small crystallization drops and to consider whether acoustic delivery could enable batch-under-oil methods. research papers We used high energy settings in these tests, typically with amplitude at 2.5–3.5 and burst settings greater than 1000. Silicone oil, DMS (polydimethyl- siloxane terminated with trimethyl- siloxy), FMS [poly(3,3,3-triﬂuoropropyl- methylsiloxane)] and Al’s oil (a 50:50 mixture of silicone and parafﬁn oil) failed to transfer at these energies. However, we did observe transfer of parafﬁn oil and mineral oil, but did not quantitate these deliveries. Challenges to using oils in such ﬂying-drop setups remain, as there was noticeable spatter upon aqueous drop and oil surface contact. Figure 3 g Plots of the residual error in measured volume delivered relative to a 5 nl target for experiments at ﬁxed energy throughout (light triangles) versus individually adjusted energies (dark circles). (a) Distribution of residuals across four screens (color-coded). (b) Enlarged view of conditions 13–28 of the PEG MW grid. Each set of four linked points is a progression of PEG concentration (5, 15, 20, 25%), illustrating the nature of the effect of the diminished measured volumes delivered as PEG increases (triangles). Application of a coarse energy correction (circles) demonstrates better delivery proﬁles (smaller range in residuals across PEG concentrations) and highlights further improvements to our energy-correction calcula- tions to reduce overcompensation of the lowest measured residuals (e.g. points 20 and 24). 897 Villasen˜or et al. Acoustic liquid transfer Acta Cryst. (2012). D68, 893–900 4. Discussion Smaller crystallization-drop volumes increase the number of conditions that can be investigated per volume of protein sample. Nanocrystallization thus provides a more efﬁcient use of protein reagents and may afford a thorough crystallization screening of low-expressing targets that were previously intractable with traditional techniques. On the other hand, the altered biophysical environment of such small drops is complex, with a possible detrimental impact on nucleation and thus crystallization. Whether lessened research papers A sealed 1536-well plate presents a compelling environment in terms of the issues mentioned above: minimal evaporation, short ﬂight path for accurate drop placement and efﬁcient reagent usage (the acoustic ‘source’ plate becomes the reser- voir). Indeed, this provides a very facile non-invasive method for staged addition of precipitant. This real-time modulation could help investigate the trajectory of a droplet on the phase diagram and chart optimal nucleation versus post-nucleation crystal-growth regimes (Saridakis & Chayen, 2000). It was unclear whether the diverse aqueous solutions common to crystallography would transfer acceptably using this technology, which was previously established for well characterized and consistent organic solvents (e.g. CV < 2% for DMSO solutions). However, the large majority of crys- tallization solutions (93%) transferred successfully without any energy adjustments. Additionally, reproducibility was robust as 73% of conditions across ﬁve screens (349 of 480 conditions) had CV < 10% and 87% of conditions had CV < 15%. Ultimately, viable crystallization was our desired read-out. We successfully used acoustic droplet ejection to crystallize proteins from our active structure-based drug-discovery projects. This included broad screening to identify a novel crystal form of the protein kinase ITK with characteristics suitable for a structure-based drug-design campaign (Kutach et al., 2010) and a Fab–protein target complex. In the latter case, a crystal from trials using only 20 nl protein solution per drop was harvested and diffracted sufﬁciently to solve the structure, with the entire campaign using less than 10 ml protein solution (Harris et al., manuscript in preparation). These examples demonstrate dramatic savings in protein consumption, reagent cost and manpower allocation, allowing concomitant shifts in protein expression and puriﬁcation methodologies to higher throughput small-scale systems. In this context, a few microlitres of protein, on a par with the amounts often used in sample characterization, become sufﬁcient for a thorough full crystallization screen to scan directly for likelihood to crystallize. The versatile technology is being explored in other creative applications. Rather than transferring microcrystal suspen- sions to seed crystallization, as we have described elsewhere (Villasen˜or et al., 2010), recent work transferred microcrystals to mesh supports for data collection (Soares et al., 2011). Furthermore, structure-based drug-discovery or structure- based fragment-screening efforts routinely require signiﬁcant numbers of small-molecule compounds to be included in cocrystallization or crystal-soaking experiments. research papers homogeneous nucleation from bulk solution may be offset by increased surface-effect contributions and heterogeneous nucleation is a relevant discussion (Bodenstaff et al., 2002). In our practice, reproduction of known crystallization systems was relatively straightforward, suggesting that at least at these volumes (20–60 nl) the various biophysical considerations had not shifted too greatly from more traditional scales (i.e. 300 nl to single microlitre). Additionally, the delivered volume in an acoustic instrument is provided in a series of millisecond time- scale pulses resulting in ejections of a series of droplets (e.g. 5 nl, though even picolitre volumes are technically achievable) that rapidly build to the desired target volume, thus providing continuous access to the whole regime from nanolitre to microlitre volumes. This consistent format mitigates issues that have been encountered in microﬂuidic free-interface diffusion methods, where reproduction of crystallization in larger, more accessible systems has not always been straightforward. lization trials using 10 nl precipitant solution and 10 nl protein solution were successful, but there was visible evaporation at the perimeter of the precipitant solution prior to the addition of the protein in a second pass. Also, such small volumes highlighted the fact that the more intractable solutions showed higher variability in placement. Optimization of several vari- ables such as enclosure humidiﬁcation, short droplet ﬂight paths and minimizing the plate volume airspace with new plate geometries should allow the practical use of 5–10 nl protein consumption per experiment or less. This also obviates the need for larger reservoirs. In test cases, we achieved crystal growth even with reservoir volumes of only 1–2 ml (20–40 times larger than the crystallization droplet) set alongside the 60 nl crystallization droplet (data not shown); this reduction allows all the experimental liquids including the ‘reservoir’ to be delivered by acoustic liquid transfer. To crystallographers, the inverted orientation of the target plate immediately suggests the familiar hanging-drop experi- ment. Given a suitable method for accurately positioning a lid or ﬁlm seal relative to the ‘source’ plate below, an array of protein droplets could be rapidly set and sealed as hanging drops. The soundwave technology we describe here has the unique capability to eject droplets of the reservoir solutions up to the protein droplet even after the plate has been sealed. 1 Supplementary material has been deposited in the IUCr electronic archive (Reference: NJ5116). Services for accessing this material are described at the back of the journal. Figure 4 g Images collected on day 15 from crystallization drops of various protein samples. The total drop volume is double the protein value given. The volumes are representative and were not optimized as similar crystals appeared at the various volumes explored. (a) HCV helicase (50 nl). (b) Human serum albumin (50 nl). (c) HCV polymerase (30 nl). (d) HIV RT (15 nl). (e) ITK (30 nl). (f) Lysozyme (15 nl). 898 Villasen˜or et al. Acoustic liquid transfer Acta Cryst. (2012). D68, 893–900 research papers Villasen˜or et al. Acoustic liquid transfer research papers With these compounds typically in DMSO a few nanolitres of each can be precisely and accurately added to crystallization drops in under a minute for 100 compounds. With the acoustic technology already more common in compound-management groups, shared use of source plates makes efﬁcient use of small-molecule stocks as structural biology groups work through hit lists from high-throughput or fragment-screening campaigns. Our study helps to establish the practical application of acoustic liquid transfer to protein crystallization, but also identiﬁes several possible avenues to streamline the workﬂow towards even smaller volumes. Drop placement is rapid [45 s for a single pass over 96 wells at 40 nl delivery (see Supplementary Videos S1 and S21)], but mitigation of evaporation will become crucial at smaller volumes. Crystal- We provide a pioneering practical demonstration of the implementation of acoustic liquid transfer in protein crystal- lization, a particularly challenging application owing to the sheer diversity of the screening space. Our exploration of tuning transfer energies to match liquid properties was here a coarse proof of concept designed to inspire a thorough treatment; more recently both EDC Biosystems and Labcyte 899 Acta Cryst. (2012). D68, 893–900 Villasen˜or et al. Acoustic liquid transfer research papers have built on this, making progress in more sophisticated acoustic interrogation of each well to establish appropriate energy proﬁles automatically (Forbush et al., 2006). These emerging developments obviate the need for users to calibrate and characterize their ﬂuids, providing enhanced accessibility to the technology. Given the excellent reliability of the tech- nology at very small volumes with well characterized liquids such as DMSO (Comley, 2004), exciting further possibilities and opportunities are poised to be realised as crystallization reagents approach similar performance benchmarks. Throughput Screen. 8, 489–498. Throughput Screen. 8, 489–498. Forbush, M., Chow, H., Chiao, J. & Rose, A. (2006). J. Assoc. Lab. Autom. 11, 188–194. Kim, J. L., Morgenstern, K. A., Grifﬁth, J. P., Dwyer, M. D., Thomson, J. A., Murcko, M. A., Lin, C. & Caron, P. R. (1998). Structure, 6, 89–100. Kutach, A. K., Villasen˜or, A. G., Lam, D., Belunis, C., Janson, C., Lok, S., Hong, L.-N., Liu, C.-M., Deval, J., Novak, T. J., Barnett, J. W., Chu, W., Shaw, D. & Kuglstatter, A. (2010). Chem. Biol. Drug Des. 76, 154–163. Le Pogam, S. et al. (2006). J. Virol. 80, 6146–6154. Newman, J., Fazio, V. J., Lawson, B. & Peat, T. S. (2010). Cryst. Growth Des. 10, 2785–2792. Olechno, J., Shieh, J. & Ellson, R. (2006). J. Assoc. Lab. Autom. 11, 240–246. We would like to thank James Bull (Biosero Inc.) for productive interactions aiding this work and the Roche Palo Alto protein-production group for the preparation of protein reagents. Conﬂict of interest statement: TJD and AG are afﬁliates of EDC Biosystems Inc. Rose, D. (1999). Drug Discov. Today, 4, 411–419. Saridakis, E. & Chayen, N. E. (2000). Protein Sci. 9, 755–757. Soares, A. S., Engel, M. A., Stearns, R., Datwani, S., Olechno, J., Ellson, R., Skinner, J. M., Allaire, M. & Orville, A. M. (2011). Biochemistry, 50, 4399–4401. y Stevens, R. C. (2000). Curr. Opin. Struct. Biol. 10, 558–563. Aerni, H. R., Cornett, D. S. & Caprioli, R. M. (2006). Anal. Chem. 78, 827–834. Bodenstaff, E. R., Hoedemaeker, F. J., Kuil, M. E., de Vrind, H. P. M. & Abrahams, J. P. (2002). Acta Cryst. D58, 1901–1906. Comley, J. (2004). Drug Discov. World, Summer 2004, 43–54. Ellson, R., Stearns, R., Mutz, M., Brown, C., Browning, B., Harris, D., Qureshi, S., Shieh, J. & Wold, D. (2005). Comb. Chem. High 900 Villasen˜or et al. Acoustic liquid transfer Acta Cryst. (2012). D68, 893–900 References Sweeney, Z. K. et al. (2008). Bioorg. Med. Chem. Lett. 18, 4352– 4354. Villasen˜or, A. G., Wong, A., Shao, A., Garg, A., Kuglstatter, A. & Harris, S. F. (2010). Acta Cryst. D66, 568–576. Walter, T. S., Diprose, J., Brown, J., Pickford, M., Owens, R. J., Stuart, D I & H l K (2003) J A l C 36 308 314 Walter, T. S., Diprose, J., Brown, J., Pickford, M., Owens, R. J., Stuart, D. I. & Harlos, K. (2003). J. Appl. Cryst. 36, 308–314. D. I. & Harlos, K. (2003). J. Appl. Cryst. 36, 308–314. Wong, E. Y. & Diamond, S. L. (2009). Anal. Chem. 81, 509– 514. 900 Acta Cryst. (2012). D68, 893–900 900 Villasen˜or et al. Acoustic liquid transfer
https://openalex.org/W4387027271	https://fermet.misis.ru/jour/article/download/2574/1827	Russian	null	Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium	Izvestiâ vysših učebnyh zavedenij. Černaâ metallurgiâ	2,023	cc-by	5,228	Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками углеродфторсодержащего материала и титана А. А. Усольцев 1, Н. А. Козырев 2 , Л. П. Бащенко 1, Р. Е. Крюков 1, А. В. Жуков 1 А. А. Усольцев 1, Н. А. Козырев 2 , Л. П. Бащенко 1, Р. Е. Крюков 1, А. В. Жуков 1 Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Металлургические технологии Metallurgical technologies УДК 621.762.04:624.567.1:669 DOI 10.17073/0368-0797-2023-4-403-409 УДК 621.762.04:624.567.1:669 DOI 10.17073/0368-0797-2023-4-403-409 Оригинальная статья Original article Оригинальная статья Original article n.kozyrev@chermet.net Аннотация. В работе исследуется качество электродугового покрытия, полученного с использованием порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками углеродфторсодержащего материала и титана. Формирование электродугового покрытия осуществляется с помощью аппарата для автоматической дуговой сварки ASAW-1250 с применением новой хромсодержащей порошковой проволоки на пластины из стали марки Ст3. Для исключения перемешивания наплавляемого металла со сталью подложки проводят многослойную наплавку. Режим наплавки рассчитывается и уточняется экспериментальным путем. Авторы исследовали состав и свойства поверхности электродугового покрытия после наплавки. В качестве заменителя аморфного углерода используется углеродфторсодержащий материал (пыль газоочистки алюминиевого производства). Наплавку осуществляли под флюсом, изготовленным из шлака производства силикомарганца с повышенным содержанием серы. Проведенный регрессионный анализ показывает влияние химического состава наплавленного слоя на его твердость и скорость износа. В работе получены математические модели исследуемых эксплуатационных характеристик электродугового покрытия. При увеличении содержания хрома, вольфрама, углерода и кремния повышаются твердость наплавленного металла и устойчивость его к абразивному износу. Результаты проведенных исследований позволяют выработать мероприятия для обеспечения требуемого уровня эксплуатационных характеристик электродугового покрытия и могут использоваться для составления прогноза твердости наплавленного слоя и его износостойкости при изменении химического состава металла, прогнозировать эксплуатационную стойкость прокатных валков, наплавленных проволоками типа ПП-Нп-35В9Х3СФ. Математические модели твердости наплавленного слоя и его износостойкости позволяют уточнить механизм упрочнения и формирования защитных свойств поверхностных слоев прокатных валков посредством электродуговых покрытий, наплавленных порошковыми проволоками. Ключевые слова: порошковая проволока, электродуговое покрытие, многослойная наплавка, прокатные валки, т Ключевые слова: порошковая проволока, электродуговое покрытие, многослойная наплавка, прокатные валки, твердость, скорость износа Для цитирования: Усольцев А.А., Козырев Н.А., Бащенко Л.П., Крюков Р.Е., Жуков А.В. Разработка порошковой проволоки систе мы Fe – C – Si – Mn – Cr – W – V с присадками углеродфторсодержащего материала и титана. Известия вузов. Черная металлургия. ковая проволока, электродуговое покрытие, многослойная наплавка, прокатные валки, твердость, скорость износа Для цитирования: Усольцев А.А., Козырев Н.А., Бащенко Л.П., Крюков Р.Е., Жуков А.В. Разработка порошковой проволоки систе мы Fe – C – Si – Mn – Cr – W – V с присадками углеродфторсодержащего материала и титана. Известия вузов. Черная металлургия. 2023;66(4):403–409. https://doi.org/10.17073/0368-0797-2023-4-403-409 Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium A. A. Usol’tsev 1, N. A. Kozyrev 2 , L. P. Bashchenko 1, R. E. Kryukov 1, A. V. Zhukov 1 Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium A. A. Usol’tsev 1, N. A. Kozyrev 2 , L. P. Bashchenko 1, R. E. Kryukov 1, A. V. Zhukov 1 A. A. Usol’tsev 1, N. A. Kozyrev 2 , L. P. Bashchenko 1, R. E. Kryukov 1, A. V. Zhukov 1 1 Siberian State Industrial University (42 Kirova Str., Novokuznetsk, Kemerovo Region – Kuzbass 654007, Russian Federation) 2 I.P. Bardin Central Research Institute of Ferrous Metallurgy (23/9 Radio Str., Moscow 105005, Russian Federation) n.kozyrev@chermet.net Abstract. The paper considers research of quality of the electric arc coating obtained using flux-cored wire of the Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium. The formation of an electric arc coating was carried out using an automatic arc 403 © А. А. Усольцев, Н. А. Козырев, Л. П. Бащенко, Р. Е. Крюков, А. В. Жуков, 2023 Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками ... Известия вузов. Черная металлургия. 2023;66(4):403–409. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W Известия вузов. Черная металлургия. 2023;66(4):403–409. д б Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – C welding machine ASAW-1250 with a new chromium-containing flux-cored wire on plates made of St3 steel. To exclude mixing of the deposited metal with the substrate steel, multilayer surfacing was conducted. The surfacing mode was calculated and refined experimentally. The authors studied the composition and properties of the surface of the electric arc coating after surfacing. As a substitute for amorphous carbon they used a carbon- fluorine-containing material (dust of gas purification of aluminum production). Surfacing was carried out under a flux made from slag produced by silicomanganese with a high content of sulfur. A regression analysis of influence of the deposited layer’s chemical composition on its hardness and wear rate was carried out and mathematical models of the investigated performance characteristics of the electric arc coating were obtained. With an increase in the content of chromium, tungsten, carbon and silicon, hardness of the deposited metal and its resistance to abrasive wear increase. The results of the conducted research make it possible to develop measures ensuring the required level of performance characteristics of the electric arc coating and can be used to make a forecast of hardness of the deposited layer and its wear resistance when the chemical composition of the metal changes, to predict the operational resistance of rolling rolls deposited with wires of the PP-Np-35V9Kh3SF type. n.kozyrev@chermet.net Mathematical models of hardness of the deposited layer and its wear resistance help to clarify the mechanism of hardening and formation of protective properties of the surface layers of rolling rolls by means of electric arc coatings deposited with flux-cored wires. Keywords: flux-cored wire, electric arc coating, multilayer surfacing, rolling rolls, hardness, wear rate For citation: Usol’tsev A.A., Kozyrev N.A., Bashchenko L.P., Kryukov R.E., Zhukov A.V. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium. Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. https://doi.org/10.17073/0368-0797-2023-4-403-409 For citation: Usol’tsev A.A., Kozyrev N.A., Bashchenko L.P., Kryukov R.E., Zhukov A.V. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives of carbon-fluorine-containing material and titanium. Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. https://doi.org/10.17073/0368-0797-2023-4-403-409 Введение для ремонта валков, работающих в наиболее тяжелых условиях при эксплуатации, чаще всего используются порошковые проволоки типа ПП-Нп-35В9Х3СФ по ГОСТ 26101 – 84 [19 – 21]. При изменении состава шихты этих порошковых проволок и введении в их состав ряда элементов возможно увеличение износо стойкости наплавленного слоя и повышение эксплуа тационной стойкости наплавляемых прокатных валков. Совершенствование и изменение химического состава используемых при наплавке порошковых проволок является сложной научно-производственной задачей, решение которой связано с выполнением требований экономичности и экологичности процесса [21 – 23]. Основной технологический инструмент в прокат ном переделе металлургических предприятий – прокат ные валки. В процессе эксплуатации при пластической деформации металла валки испытывают значитель ные удельные давления, термические воздействия и, соответственно, подвергаются интенсивному изнаши ванию [1 – 3]. В связи с этим обостряется проблема качества проведения ремонтов валков. В последнее время широкое распространение получила восстано вительная электродуговая наплавка прокатных вал ков с использованием порошковых проволок [2; 4; 5]. Однако использование порошковых проволок для наплавки не лишено ряда недостатков, и для обеспе чения качественных ремонтных операций необходимо совершенствование состава шихты для этих проволок и технологии их использования [6 – 8]. В связи с этим проведение теоретических и экспериментальных исследований физической природы, процессов и меха низмов упрочнения и формирования защитных свойств поверхностных слоев прокатных валков посредством электродуговых покрытий, наплавленных порошко выми проволоками, актуально и имеет большое научно- практическое значение. Целью настоящей работы является установление закономерностей повышения износостойкости и твер дости электродуговых покрытий прокатных валков, наплавленных порошковыми проволоками системы Fe – C – Si – Mn – Cr – W – V, с введением в их состав титана и углеродфторсодержащего материала. Материалы и методы исследования Процессы наплавки и изготовления порошковой проволоки, а также состав наполнителя для исследуе мой порошковой проволоки и сварочного флюса опи саны в работах [20 – 22]. Следует отметить, что использование применяе мых в настоящее время порошковых проволок при наплавке прокатных валков связано с рядом образую щихся при наплавке дефектов, высокой стоимостью и дефицитностью наплавочных материалов, а также несовершенством технологий наплавки [9 – 11]. При этом большое значение имеют выявление и устране ние дефектной структуры покрытий, приводящей к их преждевременному разрушению [12 – 14]. Представ ляет интерес разработка технологичных наплавочных материалов [15 – 17], обеспечивающих в наплавленном металле структуры низкоуглеродистого мартенсита. В исследованиях при разработке новой порошко вой проволоки в качестве сравнения за аналог прини мали проволоку ПП-Нп-35В9Х3СФ, изготовленную с использованием графита марки ГЛ-1 (образец 1). Далее осуществляли изменение концентрации титана (образцы 2 – 4) и углеродфторсодержащего материала (образцы 5 – 9). Перед изготовлением порошковых проволок про водили предварительный расчет количества порошко вых материалов. Взвешивание материалов осуществ ляли на лабораторных аналитических весах AUX 120. Смешение порошков проводили на лабораторных вращательных смесителях в течение не менее 30 мин. Изготовление порошковой проволоки проводилось на Для наплавки прокатных валков широкое распро странение получили порошковые проволоки сис темы Fe – C – Si – Mn – Cr – W – V [18 – 20]. При этом 404 Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – ром 20×55×14 мм. Измерение твердости выполняли по методу Роквелла на твердомере модели ТК-14-250 в соответствии с требованиями ГОСТ 9013 – 59 путем вдавливания конического алмазного наконечника с углом при вершине 120°. лабораторном станке: сворачивание ленты путем воло чения через фильеру и намотка полученной порошко вой проволоки на барабан. ром 20×55×14 мм. Измерение твердости выполняли по методу Роквелла на твердомере модели ТК-14-250 в соответствии с требованиями ГОСТ 9013 – 59 путем вдавливания конического алмазного наконечника с углом при вершине 120°. Материалы и методы исследования Наплавку электродуговых покрытий проводили сварочным трактором ASAW-1250 с использованием изготовленной порошковой проволоки на пластины из стали в пять слоев для исключения перемешивания наплавляемого металла со сталью подложки. Режим наплавки рассчитывали и уточняли экспериментально. После наплавки исследовали состав и свойства электро дугового покрытия. Испытания на износ в настоящее время прово дят согласно ГОСТ 23.208 – 79. Стандарт распрост раняется на металлические материалы и покрытия и регламентирует метод их испытаний на абразивное изнашивание при трении о нежестко закрепленные абразивные частицы. Сущность метода состоит в том, что при одинаковых условиях осуществляют трение образцов исследуемого и эталонного материалов об абразивные частицы, подаваемые в зону трения и при жимаемые к образцу вращающимся резиновым роли ком. Измеряют износ образцов испытуемого и эта лонного материалов, износостойкость испытуемого материала оценивают путем сравнения износа эталон ного и исследуемого образцов. Обработку результа тов проводят по данным взвешивания образцов до и после испытаний, определяя среднеарифметические значения потерь массы эталонного образца и образцов исследуемого материала. Химический состав наплавленных покрытий опре деляли рентгенофлюоресцентным методом на спек трометре XRF-1800 и атомно-эмиссионным методом на спектрометре ДФС-71. В ряде образцов определе ние химического состава металла проводили хими ческими методами: на содержание углерода по ГОСТ 12344 – 2003, серы по ГОСТ 12345 – 2001 и фос фора по ГОСТ 12347 – 77. Образцы для проведения исследований макро- и микроструктуры, твердости, износостойкости были подготовлены по методике, включающей вырезку образцов на отрезном станке KKS 315L, шлифование на плоскошлифовальном станке 3Д725, полирование на полировальном станке FROMMIA 835 SE. Для анализа оценки влияния химического состава порошковых проволок на скорость (степень) износа и твердость электродуговых покрытий использовали многофакторный корреляционный анализ, который позволяет изучить закономерности изменения того или иного показателя в зависимости от поведения различных факторов. Для этого определяли факторы, которые оказывают воздействие на изучаемый пока затель, и отбирали наиболее существенные из них, после чего выполняли проверку исходной информа ции на достоверность, однородность и соответствие закону нормального распределения. Далее строили модель факторной системы. Поскольку в приведен ных системах имеют место независимые факторные признаки, использовали детерминированный фактор ный анализ. Схема вырезки образцов для исследования пред ставлена на рис. 1. Для определения механических свойств из выре занных образцов изготавливали макрошлифы разме Рис. 1. Схема вырезки образцов для проведения испытаний и анализа: 1 и 2 – для изготовления макро- и микрошлифов; 3 и 4 – для испытаний на твердость и износостойкость; 5 – для определения содержания водорода Fig. 1. Fig. 1. Scheme of samples cutting for testing: 1 and 2 – for obtaining macro- and micro-plates; 3 and 4 – for tests on hardness and wear resistance; 5 – for hydrogen content determination Материалы и методы исследования Scheme of samples cutting for testing: 1 and 2 – for obtaining macro- and micro-plates; 3 and 4 – for tests on hardness and wear resistance; 5 – for hydrogen content determination Скорость истирания наплавленного слоя опыт ных образцов определяли путем испытаний на износ на машине 2070 СМТ-1, принципиальная кинемати ческая схема которой приведена на рис. 2. Диапазон измерения частоты вращения вала нижнего образца (диапазон А) составлял 75 – 750 мин–1; диапазон измерения момента трения (диапазон I ) 1 – 10 Н·м. Машина трения 2070 СМТ-1 может работать по схеме с замкнутым и открытым силовым контуром и имеет следующее устройство: от электродвигателя 2 посред ством ременной передачи 10 вращение одновременно передается на нижний 5 и верхний 6 образцы. Обра зец 6 устанавливается на валу откидной каретки 7. Каретка уравновешивается пружинным механизмом 8. На приводе вала нижнего образца установлен упругий торсион датчика 9 момента трения с бесконтактным токосъемом, сигнал с которого выводится на пульт. Рис. 1. Схема вырезки образцов для проведения испытаний и анализа: 1 и 2 – для изготовления макро- и микрошлифов; 3 и 4 – для испытаний на твердость и износостойкость; 5 – для определения содержания водорода 405 Известия вузов. Черная металлургия. 2023;66(4):403–409. . и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками .. Известия вузов. Черная металлургия. 2023;66(4):403–409. . и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – C Известия вузов. Черная металлургия. 2023;66(4):403–409. д б Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками ... Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками ... Частота вращения измеряется с помощью тахогене ратора 3, размещенного на валу двигателя, а число обо ротов нижнего образца 5 – с помощью бесконтактного датчика 1. Вал-колодка, которым проводилось истира ние образцов, изготовлен из стали марки Р18. Результаты и их обсуждение Химический состав наплавленного слоя, получен ного с применением опытной порошковой проволоки, представлен в табл. 1, значения скорости износа образ цов приведены в табл. 2. Зависимости твердости и скорости износа от содер жания различных элементов показаны на рис. 3. Применение математико-статистических методов позволило построить математическую модель влияния химического состава наплавленного слоя на его твер дость и скорость износа. Рис. 2. Принципиальная кинематическая схема машины 2070 СМТ-1: Рис. 2. Принципиальная кинематическая схема машины 2070 СМТ-1: Полученные зависимости проверяли по показателю средней ошибки аппроксимации 1 – бесконтактный датчик количества оборотов; 2 – электродвигатель; 3 – тахогенератор; 4 – узел нагружения (пружинный механизм); 5 и 6 – нижний и верхний образцы; 7 – каретка; 8 – пружинный механизм; 9 – датчик момента трения; 10 – плоскозубая ременная передача где m – количество наблюдений; i – вычисленное зна чение результирующего показателя; i – фактическое значение результирующего показателя. Fig. 2. Basic kinematic scheme of the 2070 СМТ-1 machine: 1 – contactless speed sensor; 2 – electric motor; 3 – rate generator; 4 – loading unit (spring mechanism); 5 and 6 – lower and upper samples; 7 – carriage; 8 – spring mechanism; 9 – friction torque sensor; 10 – flat-toothed belt drive Регрессионный анализ влияния химического состава наплавленного слоя на его твердость и скорость износа представлен следующими уравнениями: – твердость HRC: – твердость HRC: – твердость HRC: Нагружение образцов проводится с помощью пру жинного механизма 4. Величина нормальной силы регу лируется поворотом рукоятки узла нагружения, а пере дача показаний на пульт осуществляется через гибкую связь от резистора, размещенного в данном узле. –39,056 + 58,725C + 4,983Si + 37,87Mn + + 6,058Cr – 7,096Cu – 107,503Mo – 0,341V – 0,435W (ошибка аппроксимации составляет 0,0012 %); –39,056 + 58,725C + 4,983Si + 37,87Mn + + 6,058Cr – 7,096Cu – 107,503Mo – 0,341V – 0,435W (ошибка аппроксимации составляет 0,0012 %); Т а б л и ц а 1 Т а б л и ц а 1 Химический состав наплавленных слоев Table 1. Chemical composition of the deposited layers Table 1. Chemical composition of the deposited layers Образец Содержание элементов, % (по массе) HRC С Si Mn Cr Cu Mo V W Ni P S Al Co Nb Ti 1 0,39 1,26 1,69 3,43 0,14 0,11 0,21 8,76 0,12 0,037 0,022 0,060 0,04 0,010 0,020 58,2 2 0,37 1,31 1,74 3,25 0,13 0,11 0,20 8,19 0,12 0,034 0,021 0,110 0,04 0,010 0,030 58,4 3 0,39 1,23 1,76 3,35 0,17 0,10 0,20 8,38 0,10 0,032 0,022 0,100 0,03 0,020 0,030 54,5 4 0,35 1,18 1,73 3,03 0,18 0,09 0,19 7,42 0,13 0,031 0,021 0,090 0,04 0,020 0,030 57,0 5 0,17 0,92 1,76 1,43 0,06 0,05 0,06 3,68 0,07 0,018 0,029 0,020 0,02 0,010 0,006 43,4 6 0,18 0,83 1,55 0,84 0,06 0,03 0,04 2,23 0,06 0,014 0,033 0,009 0,01 0,010 0,007 34,8 7 0,15 0,96 1,66 1,12 0,07 0,03 0,06 3,31 0,07 0,017 0,036 0,008 0,01 0,006 0,007 39,0 8 0,09 0,76 1,46 0,60 0,06 0,02 0,03 1,50 0,06 0,013 0,033 0,004 0,01 0,006 0,006 25,7 9 0,10 1,03 1,75 1,12 0,07 0,05 0,05 3,48 0,07 0,019 0,046 0,004 0,02 0,007 0,006 37,6 406 Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Т а б Скорость износа образцов Table 2. Wear rate of the samples Образец Масса образца, г Количество оборотов (Vоб ) Износостойкость Δm·10–5 до изнашивания m1 после изнашивания m2 разница до и после 1 86,0384 85,9819 0,0565 2500 2,26 2 90,1120 90,0561 0,0559 2445 2,29 3 102,6680 102,5870 0,0810 2780 2,91 4 105,5680 105,5010 0,0668 2580 2,59 5 85,6461 85,4218 0,2243 2300 9,75 6 100,0260 99,8323 0,1939 2300 8,43 7 115,5400 115,4650 0,0757 3380 2,24 8 94,4399 94,3375 0,1024 2570 3,98 9 112,6090 112,3920 0,2170 2500 8,68 Рис. 3. Зависимости свойств наплавленного металла от содержания углерода (а), кремния (б), хрома (в) и вольфрама (г): – твердость; – скорость износа; 1 – линейная твердость; 2 – линейная скорость износа Fig. 3. Dependences of the deposited metal properties on content of carbon (a), silicon (б), chromium (в) and tungsten (г): hardness; wear rate; 1 linear hardness; 2 linear wear rate Т а б л и ц а 2 орость износа, г/об.: При увеличении содержания хрома, вол Т а б Скорость износа образцов Table 2. Список литературы / References 17. Тылкин М.А. Повышение долговечности деталей металлургического оборудования. Москва: Металлургия; 1971:608. 1. Metlitskii V.A. Flux-cored wires for arc welding and surfac ing of cast iron. Welding International. 2008;22(11):796–800. https://doi.org/10.1080/09507110802593646 18. Данильченко Б.В. Выбор износостойкого наплавленного металла для работы в условиях абразивного изнашива ния. Сварочное производство. 1992;(5):31–35. Danil’chenko B.V. Selection of wear-resistant deposited metal for operation under abrasive wear. Svarochnoe proiz vodstvo. 1992;(5):31–35. (In Russ.). 2. Filippov M.A., Shumyakov V.I., Balin S.A., Zhilin A.S., Lehchilo V.V., Rimer G.A. Structure and wear resistance of deposited alloys based on metastable chromium-carbon aus tenite. Welding International. 2015;29(10):819–822. https://doi.org/10.1080/09507116.2014.986891l 19. Гусев А.И., Романов Д.А., Козырев Н.А., Громов В.Е., Иванов Ю.Ф. Структура сварного шва износостойкой наплавки порошковой проволокой системы Fe – C – Si – Mn – Cr – Ni – Mo. Вестник Сибирского государствен ного индустриального университета. 2021;(2(36)):3–10. 19. Гусев А.И., Романов Д.А., Козырев Н.А., Громов В.Е., Иванов Ю.Ф. Структура сварного шва износостойкой наплавки порошковой проволокой системы Fe – C – Si – Mn – Cr – Ni – Mo. Вестник Сибирского государствен ного индустриального университета. 2021;(2(36)):3–10. Gusev A.I., Romanov D.A., Kozyrev N.A., Gromov V.E., Ivanov Yu.F. Structure of weld of wear-resistant surfacing by flux-cored wire of Fe – C – Si – Mn – Cr – Ni – Mo system. Bulletin of the Siberian State Industrial University. 2021;(2(36)):3–10. (In Russ.). 3. Liu D.S., Liu R.P., Wei Y.H. Influence of tungsten on micro structure and wear resistance of iron base hardfacing alloy. Materials Science and Technology. 2014;30(3):316–322. https://doi.org/10.1179/1743284713Y.0000000359 Gusev A.I., Romanov D.A., Kozyrev N.A., Gromov V.E., Ivanov Yu.F. Structure of weld of wear-resistant surfacing by flux-cored wire of Fe – C – Si – Mn – Cr – Ni – Mo system. Bulletin of the Siberian State Industrial University. 2021;(2(36)):3–10. (In Russ.). 4. Kejžar R., Grum J. Hardfacing of wear-resistant depo sits by MAG welding with a flux-cored wire having gra phite in its filling. Materials and Manufacturing Processes. 2005;20(6):961–976. https://doi.org/10.1081/AMP-200060424 20. Козырев Н.А., Титов Д.А., Кибко Н.В. Изучение влияния состава порошковой проволоки ПП-Нп-35В9Х3СФ на свойства наплавляемого валика. Актуальные проблемы в машиностроении. 2016;3(3):60–64. Kozyrev N.A., Titov D.A., Kibko N.V. The study of the influence of the composition of flux-cored wire ПП-Нп- 35В9Х3СФ on the properties surfaced roller. Actual Prob lems in Machine Building. 2016;3(3):60–64. (In Russ.). 5. Li R., He D.Y., Zhou Z., Wang Z.J., Song X.Y. Wear and high temperature oxidation behavior of wire arc sprayed iron based coatings. Surface Engineering. 2014;30(11):784–790. https://doi.org/10.1179/1743294414Y.0000000331 6. Выводы 8. Zhuk Yu. Super-hard wear-resistant coating systems. Mate rials Technology. 1999;14(3):126–129. https://doi.org/10.1080/10667857.1999.11752827i Проведен регрессионный анализ влияния хими ческого состава наплавленного слоя на его твердость и скорость износа и получены математические модели исследуемых эксплуатационных характеристик элек тродугового покрытия. При увеличении содержания хрома, вольфрама, углерода и кремния повышаются твердость наплавленного металла и устойчивость его к абразивному износу. 9. Hardell J., Yousfi A., Lund M., Pelcastre L., Prakash B. Abra sive wear behaviour of hardened high strength boron steel. Tribology – Materials, Surfaces and Interfaces. 2014;8(2): 90–97. https://doi.org/10.1179/1751584X14Y.0000000068 10. Deng X.T., Fu T.L., Wang Z.D., Misra R.D.K., Wang G.D. Epsilon carbide precipitation and wear behaviour of low alloy wear resistant steels. Materials Science and Techno logy. 2016;32(4):320–327. https://doi.org/10.1080/02670836.2015.1137410 Результаты проведенных исследований позволяют выработать мероприятия для обеспечения требуемого уровня эксплуатационных характеристик электродуго вого покрытия и могут использоваться для составления прогноза твердости наплавленного слоя и его износо стойкости при изменении химического состава наплав ленного металла и прогнозировать эксплуатационную стойкость прокатных валков, наплавленных проволо ками типа ПП-Нп-35В9Х3СФ. 11. Kirchgaßner M., Badisch E., Franek F. Behaviour of iron- based hardfacing alloys under abrasion and impact. Wear. 2008;265(5-6):772–779. https://doi.org/10.1016/j.wear.2008.01.004 https://doi.org/10.1016/j.wear.2008.01.004 12. Хасуи А., Моригаки О. Наплавка и напыление. Москва: Машиностроение; 1985:240. 13. Патон Б.Е. Технология электрической сварки металлов и сплавов плавлением. Москва: Книга по Требованию; 2021:769. Математические модели твердости наплавленного слоя и его износостойкости позволяют уточнить меха низм упрочнения и формирования защитных свойств поверхностных слоев прокатных валков посредством наплавленных порошковыми проволоками электроду говых покрытий. 14. Гуляев А.П. Металловедение. Учебник для вузов. Москва: Металлургия; 1986:544. 15. Хрущов М.М., Бабичев М.А. Абразивное изнашивание. Москва: Наука; 1970:251. 16. Елагина О.Ю. Технологические методы повышения износостойкости деталей машин. Москва: Логос; 2009:488. Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с прис Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками ... Известия вузов. Черная металлургия. 2023;66(4):403–409. . и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Cr – W – V с присадками ... Известия вузов. Черная металлургия. 2023;66(4):403–409. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Известия вузов. Черная металлургия. 2023;66(4):403–409. Усольцев А.А., Козырев Н.А. и др. Разработка порошковой проволоки системы Fe – C – Si – Mn – Table 1. Chemical composition of the deposited layers Wear rate of the samples Образец Масса образца, г Количество оборотов (Vоб ) Износостойкость Δm·10–5 до изнашивания m1 после изнашивания m2 разница до и после 1 86,0384 85,9819 0,0565 2500 2,26 2 90,1120 90,0561 0,0559 2445 2,29 3 102,6680 102,5870 0,0810 2780 2,91 4 105,5680 105,5010 0,0668 2580 2,59 5 85,6461 85,4218 0,2243 2300 9,75 6 100,0260 99,8323 0,1939 2300 8,43 7 115,5400 115,4650 0,0757 3380 2,24 8 94,4399 94,3375 0,1024 2570 3,98 9 112,6090 112,3920 0,2170 2500 8,68 Рис. 3. Зависимости свойств наплавленного металла от содержания углерода (а), кремния (б), хрома (в) и вольфрама (г): – твердость; – скорость износа; 1 – линейная твердость; 2 – линейная скорость износа Fig. 3. Dependences of the deposited metal properties on content of carbon (a), silicon (б), chromium (в) and tungsten (г): – hardness; – wear rate; 1 – linear hardness; 2 – linear wear rate Скорость износа образцов Table 2. Wear rate of the samples симости свойств наплавленного металла от содержания углерода (а), кремния (б), хрома (в) и вольфрама (г): – твердость; – скорость износа; 1 – линейная твердость; 2 – линейная скорость износа симости свойств наплавленного металла от содержания углерода (а), кремния (б), хрома (в) и вольфрама (г): – твердость; – скорость износа; 1 – линейная твердость; 2 – линейная скорость износа Fig. 3. Dependences of the deposited metal properties on content of carbon (a), silicon (б), chrom Fig. 3. Dependences of the deposited metal properties on content of carbon (a), silicon (б), chromium (в) and tungsten (г): – hardness; – wear rate; 1 – linear hardness; 2 – linear wear rate Fig. 3. Dependences of the deposited metal properties on content of carbon (a), silicon (б), chromium (в) and tungsten (г): – hardness; – wear rate; 1 – linear hardness; 2 – linear wear rate g p p p p ( ) ( ) – hardness; – wear rate; 1 – linear hardness; 2 – linear wear rate – скорость износа, г/об.: – скорость износа, г/об.: При увеличении содержания хрома, вольфрама, углерода и кремния повышаются твердость и устой чивость к абразивному износу наплавленного металла Построенные зависимости позволяют осуществлять прогноз по получению результатов твердости и скоро сти износа при изменении химического состава наплав ленных слоев. 407 Список литературы / References Ma H.R., Chen X.Y., Li J.W., Chang C.T., Wang G., Li H., Wang X.M., Li R.W. Fe-based amorphous coating with high corrosion and wear resistance. Surface Engineering. 2017;33(1):56–62. 21. 21. Уманский А.А., Козырев Н.А., Титов Д.А. Экспери ментальные исследования влияния состава порошко вых проволок систем легирования C–Si–Mn–Cr–V–Mo и C–Si–Mn–Cr–W–V на структуру наплавленного слоя прокатных валков. Черная металлургия. Бюллетень https://doi.org/10.1080/02670844.2016.1176718 7. Lim S.C., Gupta M., Goh Y.S., Seow K.C. Wear resistant WC – Co composite hard coatings. Surface Engineering. 1997;13(3):247–250. https://doi.org/10.1179/sur.1997.13.3.247 408 Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – V system with additives ... Izvestiya. Ferrous Metallurgy. 2023;66(4):403–409. Usol’tsev A.A., Kozyrev N.A. etc. Development of flux-cored wire of Fe – C – Si – Mn – Cr – W – научно-технической и экономической информации. 2016;(4):74–78. Umanskii A.A., Kozyrev N.A., Titov D.A. The experi mental investigations into the influence of the composi tions of the cored wires of the C–Si–Mn–Cr–V–Mo and C–Si–Mn–Cr–W–V alloying systems on the build-up mill roll layer structure. Ferrous Metallurgy. Bulletin of Scientific, Tech nical and Economic Information. 2016;(4):74–78. (In Russ.). научно-технической и экономической информации. 2016;(4):74–78. ние эксплуатационных характеристик наплавленного металла. Производство проката. 2016;(5):43–47. Umanskii A.A., Kozyrev N.A., Titov D.A. Research and development of new compositions of flux-cored wire for sur facing of rolling rolls providing an increase in operational characteristics of the deposited metal. Proizvodstvo prokata. 2016;5:43–47. (In Russ.). ние эксплуатационных характеристик наплавленного металла. Производство проката. 2016;(5):43–47. Umanskii A.A., Kozyrev N.A., Titov D.A. Research and development of new compositions of flux-cored wire for sur facing of rolling rolls providing an increase in operational characteristics of the deposited metal. Proizvodstvo prokata. 2016;5:43–47. (In Russ.). Umanskii A.A., Kozyrev N.A., Titov D.A. The experi mental investigations into the influence of the composi tions of the cored wires of the C–Si–Mn–Cr–V–Mo and C–Si–Mn–Cr–W–V alloying systems on the build-up mill roll layer structure. Ferrous Metallurgy. Bulletin of Scientific, Tech nical and Economic Information. 2016;(4):74–78. (In Russ.). Umanskii A.A., Kozyrev N.A., Titov D.A. Research and development of new compositions of flux-cored wire for sur facing of rolling rolls providing an increase in operational characteristics of the deposited metal. Proizvodstvo prokata. 2016;5:43–47. (In Russ.). 23. Пат. 2641590 РФ. Список литературы / References МПК8 B23 К35/36 В 23 К35/36 Порош ковая проволока / Н.А. Козырев, А.И. Гусев, Г.В. Галевс кий, Р.Е. Крюков, И.В. Осетковский и др. Заявл. 22.06.2016; опубл. 18.01.2018. 22. Уманский А.А., Козырев Н.А., Титов Д.А. Исследование и разработка новых составов порошковой проволоки для наплавки прокатных валков, обеспечивающих повыше Information about the Authors Сведения об авторах Александр Александрович Усольцев, к.т.н., доцент кафедры металлургии черных металлов, Сибирский государственный индустриальный университет ORCID: 0000-0001-6220-7910 Aleksandr A. Usol’tsev, Cand. Sci. (Eng.), Assist. Prof. of the Chair of Fer rous Metallurgy, Siberian State Industrial University ORCID: 0000-0001-6220-7910 l @ bl E-mail: a.us@rambler.ru E-mail: a.us@rambler.ru Nikolai A. Kozyrev, Dr. Sci. (Eng.), Deputy Director of the Scientific Cen ter for High-Quality Steels, I.P. Bardin Central Research Institute of Fer rous Metallurgy ORCID: 0000-0002-7391-6816 E-mail: n kozyrev@chermet net Николай Анатольевич Козырев, д.т.н., заместитель директора научного центра качественных сталей, Центральный научно- исследовательский институт черной металлургии им. И.П. Бар дина ORCID: 0000-0002-7391-6816 ORCID: 0000-0002-7391-6816 E-mail: n.kozyrev@chermet.net E-mail: n.kozyrev@chermet.net E-mail: n.kozyrev@chermet.net Lyudmila P. Bashchenko, Cand. Sci. (Eng.), Assist. Prof. of the Chair “Thermal Power and Ecology”, Siberian State Industrial University ORCID: 0000-0003-1878-909X E-mail: luda.baschenko@gmail.com Людмила Петровна Бащенко, к.т.н., доцент кафедры тепло энергетики и экологии, Сибирский государственный индуст риальный университет ORCID: 0000-0003-1878-909X E-mail: luda.baschenko@gmail.com Людмила Петровна Бащенко, к.т.н., доцент кафедры тепло энергетики и экологии, Сибирский государственный индуст риальный университет Людмила Петровна Бащенко, к.т.н., доцент кафедры тепло энергетики и экологии, Сибирский государственный индуст риальный университет ORCID: 0000-0003-1878-909X E-mail: luda.baschenko@gmail.com E-mail: luda.baschenko@gmail.com Roman E. Kryukov, Dr. Sci. (Eng.), Assist. Prof. of the Chair of Ferrous Metallurgy, Siberian State Industrial University ORCID: 0000-0002-3394-7941 E-mail: rek_nzrmk@mail.ru Роман Евгеньевич Крюков, д.т.н., доцент кафедры металлургии черных металлов, Сибирский государственный индустриальный университет E-mail: rek_nzrmk@mail.ru ORCID: 0000-0002-3394-7941 E-mail: rek_nzrmk@mail.ru E-mail: rek_nzrmk@mail.ru E-mail: rek_nzrmk@mail.ru Andrei V. Zhukov, Postgraduate of the Chair of Ferrous Metallurgy, Siberian State Industrial University E-mail: Svarka42@mail.ru Андрей Владимирович Жуков, аспирант кафедры металлургии черных металлов, Сибирский государственный индустриальный университет E-mail: Svarka42@mail.ru А. А. Усольцев – формирование основной идеи исследований, разработка методологии работы, анализ результатов исследова ний. A. A. Usol’tsev – formation of the main research idea, development of the work methodology, analysis of research results. N. A. Kozyrev – formation of the main research idea, development of the research plan, setting the tasks, analysis of the research results. Поступила в редакцию 20.04.2022 После доработки 31.10.2022 Принята к публикации 12.12.2022 409
https://openalex.org/W4387719050	https://www.intertrends.ru/jour/article/download/276/128	Russian	null	The “managed competition” concept in the US–China security nexus. Theoretical analysis and practical implementation	Meždunarodnye processy	2,022	cc-by	18,755	Международные процессы, Том 20, № 4 (71), сс. 93–118 DOI 10.17994/IT.2022.20.4.71.6 Международные процессы, Том 20, № 4 (71), сс. 93–118 DOI 10.17994/IT.2022.20.4.71.6 Международные процессы, Том 20, № 4 (71), сс. 93–118 DOI 10.17994/IT.2022.20.4.71.6 Статья подготовлена в рамках научно-исследовательского проекта РФФИ/АНО ЭИСИ «Управ ление конкурентной средой во взаимодействии США, КНР, ЕС и России в XXI веке. Последствия для России». Договор № 21-011-31129\21 от 02.06.2021 г. Регистрационный номер: 121090300108-1. КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ ТЕОРЕТИЧЕСКОЕ ОСМЫСЛЕНИЕ И ПРАКТИЧЕСКОЕ ВОПЛОЩЕНИЕ ВЛАДИМИР ЩЕРБУНОВ МГИМО МИД России, Москва, Россия Статья подготовлена в рамках научно-исследовательского проекта РФФИ/АНО ЭИСИ «Управ ление конкурентной средой во взаимодействии США, КНР, ЕС и России в XXI веке. Последствия для России». Договор № 21-011-31129\21 от 02.06.2021 г. Регистрационный номер: 121090300108-1. Дата поступления рукописи в редакцию: 01.02.2022 Дата принятия к публикации: 26.10.2022 Для связи с автором / Corresponding author: Email: vshcherbunov@yandex.ru Резюме С Резюме С приходом к власти президента Дж. Байдена в лексиконе американских политиков и экспертов всё чаще стали появляться идеи «управляемости» и «регулируемости» соперничества на глобаль- ном и региональном уровнях, прежде всего в отношении конкуренции США и КНР. Большинство работ по конкуренции двух держав подчёркивают её невоенные аспекты, что не отражает усилия, которые США прикладывают к удержанию превосходства в военно-политической сфере, а также значимости её регулирования для предотвращения вооружённых конфликтов между двумя стра- нами. Настоящая статья стремится решить двуединую задачу: во-первых, выявить смысловое содержание концепции «управляемой конкуренции» (managed competition) и, во-вторых, опреде- лить подходы к ограничению рисков эскалации в американской политике на азиатском направ- лении. Проанализировав академические и аналитические работы последних лет, посвящённые феномену «управляемости», автор приходит к выводу, что содержательно концепция «управляе- мой конкуренции» включает в себя три компонента: 1) регулирования конкуренции посредством институционализации отношений с целью повышения их предсказуемости, усиления мер кри- зисного регулирования и снижения конфликтогенного потенциала, 2) опережающего програм- мирования действий международных игроков путём создания стимулов и сдерживающих факто- ров для принятия внешнеполитических решений, 3) определения мер, необходимых для достиже- ния успеха в конкуренции (англ. outcompeting). На практике управление конкуренцией в военно- политических отношениях США и КНР реализуется в 1) создании двусторонних и многосторон- них институтов для снижения вероятности инцидентов между вооружёнными силами сторон (регулирования взаимодействия между вооружёнными силами США и союзников и Китаем в Южно-Китайском море и Западной части Тихого океана), 2) определении повестки в Азиатско- Тихоокеанском регионе (концепции «индо-тихоокеанского региона» и «свободного и открытого Тихого океана»), консолидации и расширении военно-политических институтов США (QUAD, AUKUS) с приданием им антикитайской направленности, 3) действиях США, затрудняющих реализацию таких военно-политических целей КНР, как воссоединение с Тайванем, контроль над 94 ВЛАДИМИР ЩЕРБУНОВ Южно-Китайским морем, расширение военно-политических контактов КНР со странами Азиатско-Тихоокеанского региона. Ключевые слова: управляемая конкуренция; конкуренция великих держав; международная безопасность; Азиатско-Тихоокеанский регион; Китай; США. тику на китайском направлении. Вместе с тем, ввиду инертности мышления Вашингтона [Drezner 2017; Porter 2018; Шаклеина 2020: 13], обновлённый курс ло- гически продолжает стратегию сдержива- ния КНР в Индо-Тихоокеанском регионе команды Д. Трампа4. В феврале 2021 г. Дж. Байден заявил, что его администрация готова к «бескомпромиссной конкурен- ции» с Пекином, в основе которой лежит подход, отличающийся от принципов веде- ния дел республиканцев5. В сентябре по итогам встречи с китайским посланником Яном Цзечи Советник по национальной безопасности Дж. Салливан заявил, что США будут прилагать все усилия для до- стижения «ответственной конкуренции»6. Это заявление повторяло слова торгового представителя США К. Тай в том же меся- це7. 1 Под мировым порядком понимается «система межгосударственных отношений, регулируемых совокупностью принципов внешнеполитического поведения; согласованных на их основе конкретных установлений; набора признаваемых моральными и допустимыми санкций за их нарушения; потен- циала уполномоченных стран или институтов эти санкции осуществить; политической воли стран- участниц этим потенциалом воспользоваться [Богатуров 1997: 40]. 3 Ещё в 1993 г. американский специалист А. Фридберг характеризовал Азиатско-Тихоокеанский регион как «потенциально готовый к противостоянию и чреватый соперничеством» [Friedberg 1993]. Прогнозы американо-китайских трений см. также в: [Баталов 2006; Posen, Ross 1997; Layne 1998; Friedberg 2005]. [ ] у ц ц ур 2 National Security Strategy of the United States of America 2017. [Электронный ресурс] // White House Archive. December 2017. URL: http://nssarchive.us/wp-content/uploads/2020/04/2017.pdf (accessed: 15.12.2022). 4 US Strategic Framework for the Indo-Pacific. [Электронный ресурс] // White House Archive. January 2021. URL: https://trumpwhitehouse.archives.gov/wp-content/uploads/2021/01/IPS-Final- Declass.pdf (accessed: 15.12.2022). p 5 Biden: China should expect `extreme competition’ from US. [Электронный ресурс] // AP News. 7 February 2021. URL: https://apnews.com/article/joe-biden-xi-jinping-china-8f5158c12eed14e002bb1 c094f3a048a (accessed: 15.12.2022). Резюме С Двумя неделями позднее в телефонном разговоре с лидером Китая Си Цзиньпином американский президент заявил о необхо- На протяжении XXI века усилия США направлены на реализацию большой стра- тегии (Grand Strategy), суть которой заклю- чается в выстраивании благоприятного соотношения сил в ключевых регионах мира. Преимущественно они выражаются в военно-политическом и экономическом сдерживании, а также в ослаблении не только России, но и Китая с целью постро- ения мирового либерального порядка (аме- риканоцентричного порядка) [Шаклеина 2021]1. Ещё в период правления Дональда Трампа (2017–2021) Пекин был причислен к державам-ревизионистам и назван стра- тегическим конкурентом2. Прогнозируе мое рядом исследователей нарастание на- пряжённости между Соединёнными Шта тами и КНР к началу 2020-х годов стало реальностью3. Приход к власти в США в 2021 г. Джозефа Байдена привнёс элементы управляемости и ответственности в американскую поли- g 4 US Strategic Framework for the Indo-Pacific. [Электронный ресурс] // White House Archive. January 2021. URL: https://trumpwhitehouse.archives.gov/wp-content/uploads/2021/01/IPS-Final- Declass.pdf (accessed: 15.12.2022). 5 Biden: China should expect `extreme competition’ from US. [Электронный ресурс] // AP News. 7 February 2021. URL: https://apnews.com/article/joe-biden-xi-jinping-china-8f5158c12eed14e002bb1 c094f3a048a (accessed: 15.12.2022). 6 The U.S. and China must manage ‘intense competition,’ top Biden advisor says. [Электронный ресурс] // CNBC. 7 September 2021. URL: https://www.cnbc.com/2021/10/08/us-china-must-manage- intense-competition-top-biden-advisor-says.html (accessed: 15.12.2022). 6 The U.S. and China must manage ‘intense competition,’ top Biden advisor says. [Электронный ресурс] // CNBC. 7 September 2021. URL: https://www.cnbc.com/2021/10/08/us-china-must-manage- intense-competition-top-biden-advisor-says.html (accessed: 15.12.2022). p p y 7 The Transcript of A Conversation with Ambassador Katherine Tai, U.S. Trade Representative. [Электронный ресурс] // Center for Strategic and International Studies. 4 October 2021. URL: https:// www.csis.org/analysis/conversation-ambassador-katherine-tai-us-trade-representative (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 95 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ димости установить «защитные ограждения и барьеры» (англ. guardrails and parameters), предотвращающие возникновение кон- фликта между двумя странами8. Идея управ- ления конкуренцией была конкретизирова- на в мае 2022 г. в программном выступ лении государственного секретаря США Э. Блинкена в Джорджтаунском универси- тете. В нём повторялись тезисы о стрем лении «избежать конфликта» и «холодной войны», упоминалась цель превзойти Китай в конкурентной борьбе (англ. outcompete)9. В июне 2022 г. советник по национальной безопасности Джейк Салливан и министр обороны Ллойд Остин отметили необхо димость «сохранения открытых каналов коммуникации для регулирования кон куренции»10. На доктринальном уровне оформление концепции «управляемой кон- куренции» также произошло в 2022 г. в клю- чевых документах администрации и мини стерства обороны. В Индо-Тихоокеанской стратегии США от февраля 2022 г. 8 Readout of President Biden’s Virtual Meeting with President Xi Jinping of the People’s Republic of China. [Электронный ресурс] // The White House. 16 November 2021. URL: https://www.whitehouse. gov/briefing-room/statements-releases/2021/11/16/readout-of-president-bidens-virtual-meeting-with- president-xi-jinping-of-the-peoples-republic-of-china/ (accessed: 15.12.2022). p 10 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс] // White House. 13 June 2022. URL: https://www.whitehouse.gov/briefing-room/ statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans-meeting-with- politburo-member-yang-jiechi-3/ (accessed: 15.12.2022); Garamone J. Austin Meets With Chinese Counterpart in Singapore. [Электронный ресурс] // U.S. Department of Defence. 10 June 2022. URL: https://www.defense.gov/News/News-Stories/Article/Article/3058994/austin-meets-with-chinese- counterpart-in-singapore/ (accessed: 15.12.2022). p j p g p p p / 9 The Administration’s Approach to the People’s Republic of China. [Электронный ресурс] // White House. 26 May 2022. URL: https://www.state.gov/the-administrations-approach-to-the-peoples- republic-of-china/ (accessed: 15.12.2022). 8 Readout of President Biden’s Virtual Meeting with President Xi Jinping of the People’s Republic of China. [Электронный ресурс] // The White House. 16 November 2021. URL: https://www.whitehouse. gov/briefing-room/statements-releases/2021/11/16/readout-of-president-bidens-virtual-meeting-with- president-xi-jinping-of-the-peoples-republic-of-china/ (accessed: 15.12.2022). 9 The Administration’s Approach to the People’s Republic of China. [Электронный ресурс] // White House. 26 May 2022. URL: https://www.state.gov/the-administrations-approach-to-the-peoples- republic-of-china/ (accessed: 15.12.2022). 10 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс] // White House. 13 June 2022. URL: https://www.whitehouse.gov/briefing-room/ statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans-meeting-with- politburo-member-yang-jiechi-3/ (accessed: 15.12.2022); Garamone J. Austin Meets With Chinese Counterpart in Singapore. [Электронный ресурс] // U.S. Department of Defence. 10 June 2022. URL: https://www.defense.gov/News/News-Stories/Article/Article/3058994/austin-meets-with-chinese- counterpart-in-singapore/ (accessed: 15.12.2022). 11 Indo-Pacific Strategy of the United States // White House. February 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/02/U.S.-Indo-Pacific-Strategy.pdf (accessed: 15.12.2022). 12 National Security Strategy 2022 // White House. 12 October 2022. URL: https://www.whitehouse. gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security-Strategy-10.2022. pdf (accessed: 15.12.2022). 13 Под эпистемическими сообществами понимаются объединения специалистов в какой-либо области, отстаивающие общие критерии валидности знания и политические рецепты решения воз- никающих проблем. Подробнее о влиянии мозговых центров (эпистемических сообществ) на внеш- неполитический курс см.: [Adler, Haas 1992; Legro 2000; Истомин 2009; Шаклеина 2018]. p g p / 11 Indo-Pacific Strategy of the United States // White House. February 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/02/U.S.-Indo-Pacific-Strategy.pdf (accessed: 15.12.2022). 8 Readout of President Biden’s Virtual Meeting with President Xi Jinping of the People’s Republic of China. [Электронный ресурс] // The White House. 16 November 2021. URL: https://www.whitehouse. gov/briefing-room/statements-releases/2021/11/16/readout-of-president-bidens-virtual-meeting-with- president-xi-jinping-of-the-peoples-republic-of-china/ (accessed: 15.12.2022). 9 The Administration’s Approach to the People’s Republic of China. [Электронный ресурс] // White House. 26 May 2022. URL: https://www.state.gov/the-administrations-approach-to-the-peoples- republic-of-china/ (accessed: 15 12 2022) Резюме С гово рилось, что задача США – «сформировать стратегическую среду, выстраивая баланс влияния в мире, максимально благопри ятный для США, наших союзников и парт нёров, общих интересов и ценностей. Мы также будем стремиться ответственно управлять конкуренцией с КНР» (выделено автором. – В.Щ.)11. Схожие идеи были отра жены в Стратегии национальной безопас- ности США, опубликованной 12 октября 2022 г., и в Стратегии национальной обо- роны от 27 октября 2022 г. В СНБ-2022 в части, посвящённой вызовам со стороны Китая, говорилось о том, что, «несмотря на активную конкуренцию, мы будем вести конкуренцию ответственно»12. Приведённые заявления и документы свидетельствуют о необходимости углу- блённого изучения концепции «управляе- мой конкуренции» как значимого элемента обновлённой стратегии США по отноше- нию к Китаю. Автор настоящей статьи опирается на исследования, обосновываю- щие влияние эпистемических сообществ на внешнеполитический процесс13. Выска зывания американских политических дея- телей отражают результаты деятельности этого сообщества, что определяет обосно- ванность использования работ его пред ставителей для анализа управляемой кон- куренции, а также соотношения концеп туальных разработок с действиями Соеди нённых Штатов. Возникает вопрос о сферах вышеупо мянутой конкуренции. Многие аналити ческие и академические труды выделяют преимущественно невоенное измерение: p 13 Под эпистемическими сообществами понимаются объединения специалистов в какой-либо области, отстаивающие общие критерии валидности знания и политические рецепты решения воз- никающих проблем. Подробнее о влиянии мозговых центров (эпистемических сообществ) на внеш- неполитический курс см.: [Adler, Haas 1992; Legro 2000; Истомин 2009; Шаклеина 2018]. p 13 Под эпистемическими сообществами понимаются объединения специалистов в какой-либо области, отстаивающие общие критерии валидности знания и политические рецепты решения воз- никающих проблем. Подробнее о влиянии мозговых центров (эпистемических сообществ) на внеш- неполитический курс см.: [Adler, Haas 1992; Legro 2000; Истомин 2009; Шаклеина 2018]. p 13 Под эпистемическими сообществами понимаются объединения специалистов в какой-либо области, отстаивающие общие критерии валидности знания и политические рецепты решения воз- никающих проблем. Подробнее о влиянии мозговых центров (эпистемических сообществ) на внеш- неполитический курс см.: [Adler, Haas 1992; Legro 2000; Истомин 2009; Шаклеина 2018]. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 96 ВЛАДИМИР ЩЕРБУНОВ является допустимым подходом к научным исследованиям [Sil, Katzenstein 2010]. является допустимым подходом к научным исследованиям [Sil, Katzenstein 2010]. экономическую, технологическую и ди- пломатическую области. Между тем США предпринимают интенсивные усилия по сдерживанию Китая и в военно-политиче- ской сфере. Они проводят модернизацию системы альянсов, развивают военно-тех- ническое сотрудничество с государствами Азиатско-Тихоокеанского региона (АТР) и используют широкий арсенал методов военной дипломатии. 15 Конкуренция рассматривается как процесс и результат построения отношений между двумя социальными субъектами, которые одновременно желают заполучить некий объект, воспринимае- мый каждой из сторон как дефицитный в рамках определённой системы ранжирования. В данной работе разграничиваются понятия конкуренции (competition) и соперничества (rivalry) на том допу- щении, что соперничество сопряжено с угрозами применения или применением военной силы, а конкуренция преимущественно носит мирный характер [Diehl, Goertz 2001: 22]. О конкуренции в международных отношениях см.: Mazarr M.J. et al. Understanding the Emerging Era of International Competition: Theoretical and Historical Perspectives. [Электронный ресурс] // RAND Corporation, 2018. URL: https://www.rand.org/pubs/research_reports/RR2726.html (accessed: 15.12.2022) и Mazarr M.J. Understanding Competition Great Power Rivalry in a Changing International Order – Concepts and Theories. [Электронный ресурс] // RAND Corporation. March 2022. URL: https://www. rand.org/pubs/perspectives/PEA1404-1.html (accessed: 15.12.2022). О соперничестве в междуна- родных отношениях см.: [Hensel 1999; Diehl, Goertz 2001; Valeriano 2012]. 14 The Department of Defense. Indo-Pacific Strategy Report. Preparedness, Partnerships, and Promoting a Networked Region. [Электронный ресурс] // U.S. Department of Defense. June 2019. URL: https://media.defense.gov/2019/Jul/01/2002152311/-1/-1/1/DEPARTMENT-OF-DEFENSE-INDO- PACIFIC-STRATEGY-REPORT-2019.PDF (accessed: 15.12.2022). Концепция «управляемой конкуренции»: регулирование, управление, достижение успеха 1) Подход «управление посредством регулирования» подразумевает упорядочи- вание отношений между конкурентами с целью повышения их прозрачности и предсказуемости путём создания междуна- родных институтов, понижающих вероят- ность эскалации и перехода к соперниче- ству или военному конфликту. 1) Подход «управление посредством регулирования» подразумевает упорядочи- вание отношений между конкурентами с целью повышения их прозрачности и предсказуемости путём создания междуна- родных институтов, понижающих вероят- ность эскалации и перехода к соперниче- ству или военному конфликту. В научный дискурс термин «управляемая конкуренция» ввёл американский исследо- ватель рынка медицинских услуг А. Энтхо вен, предложивший модель их предостав- ления, при которой «спонсор» (например, группа компаний по предоставлению меди- цинского страхования) управляет функци- онированием рынка, устанавливает спра- ведливые правила во избежание ценовых войн [Enthoven 1993: 24]. На политическом уровне данное словосочетание применя- лось президентом США У. Клинтоном (1993–2001) для описания предлагаемых реформ в области здравоохранения16. б 2) Подход «управление посредством канализирования» означает опережающее программирование действий международ- ных субъектов с целью создания такой международной среды, в которой конку- рент вынужден действовать реактивно в наиболее выгодном для инициатора на- правлении, а союзники оказываются во- влечены в инициативы, ограничивающие их свободу манёвра по отношению к кон- куренту. В работах по международным отноше- ниям словосочетания «управляемая конку- ренция», «управляемая конкуренция между великими державами» или «управляемые отношения» начали появляться в начале 2000-х годов, отражая возрастающее бес- покойство американских аналитиков отно- сительно перспектив развития отношений США с другими великими державами, в особенности с КНР [Lampton 2001; Ross 2005]17. Тем не менее в широкий оборот в аналитических и академических трудах они вошли во второй половине 2010-х годов, когда отношения Соединённых Штатов с Китаем и Россией обострились и более отчётливо проявилась тенденция к росту конкуренции между великими державами [Goldstein 2015; Porter 2019; George 2019; Шаклеина 2021; Rudd 2022]. 3) Подход «управление посредством совместного решения проблем» связан с самоограничением и отказом сторон от тотального соперничества, то есть со- трудничеством с конкурентом в областях, представляющих взаимный интерес для сохранения ресурсов и минимизации из- держек. Главной целью для сторон тем не менее остаётся победа в долгосрочной кон- куренции, или в конкуренции, которая может не иметь временного горизонта, – то есть сохранение или достижение более высокой позиции по отношению к некоему благу в военно-политической, дипломати- ческой или экономической сфере18. Поскольку конкуренция представляет собой многогранное явление, многие ана- литические или академические труды предполагают сочетание обозначенных подходов. КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ https://www.nytimes.com/1992/09/27/opinion/clinton-waffles-on-health.html (accessed: 15.12.2022). 17 О применении термина “great power competition” в XXI веке см.: Friedman U. The New Concept Everyone in Washington Is Talking About. [Электронный ресурс] // The Atlantic, 6 August 2019. URL: https://www.theatlantic.com/politics/archive/2019/08/what-genesis-great-power-competition/595405/ (accessed: 15.12.2022). 18 Данный подход сочетания сдерживания, конкуренции и сотрудничества во многом схож с иде- ями “congagement” (от англ. containment и engagement) А. Фридберга [Friedberg 2011], “responsible competition” Т. Райта [Wright 2017] и “asymmetric competition” Р. Доши [Doshi 2021]. p y p 17 О применении термина “great power competition” в XXI веке см.: Friedman U. The New Concept Everyone in Washington Is Talking About. [Электронный ресурс] // The Atlantic, 6 August 2019. URL: https://www.theatlantic.com/politics/archive/2019/08/what-genesis-great-power-competition/595405/ (accessed: 15.12.2022). 18 Данный подход сочетания сдерживания, конкуренции и сотрудничества во многом схож с иде- 16 Clinton Waffles on Health. [Электронный ресурс] // New York Times. 27 September 1992. URL: https://www.nytimes.com/1992/09/27/opinion/clinton-waffles-on-health.html (accessed: 15.12.2022). 18 Данный подход сочетания сдерживания, конкуренции и сотрудничества во многом схож с иде- ями “congagement” (от англ. containment и engagement) А. Фридберга [Friedberg 2011], “responsible competition” Т. Райта [Wright 2017] и “asymmetric competition” Р. Доши [Doshi 2021]. 16 Clinton Waffles on Health. [Электронный ресурс] // New York Times. 27 September 1992. URL: https://www.nytimes.com/1992/09/27/opinion/clinton-waffles-on-health.html (accessed: 15.12.2022). 17 О применении термина “great power competition” в XXI веке см.: Friedman U. The New Concept Everyone in Washington Is Talking About [Электронный ресурс] // The Atlantic 6 August 2019 URL: Резюме С В этой связи пред- ставляется целесообразным рассмотреть практическое воплощение стратегии Сое динённых Штатов по управлению военно- политической конкуренцией в Азии как наиболее конфликтогенной области в от- ношениях двух стран и «наиболее значимо- го региона для будущего США»14. Неореализм (структурный реализм) рас- сматривает межгосударственные отноше- ния с точки зрения выживания и обеспече- ния собственной безопасности в условиях анархии, сдерживания других центров си- лы [Waltz 1978] путём балансирования или примыкания [Walt 1985]. Неореалистский подход фокусирует внимание на военно- политической и экономической конкурен- ции США и Китая15, то есть попытках каждой из сторон создать благоприятное для себя соотношение сил в регионе. Традиционно конкуренция в военной сфе- ре, выражающаяся в наращивании воору- жённых сил, порождает дилемму безопас- ности [Jarvis 1978; Liff, Ikenberry 2014]. Определение теоретико-методологиче- ской основы исследования сопряжено с проблемой выбора и использования раз- личных направлений и уровней теории, учи- тывая наличие трёх подходов к управлению конкуренцией, речь о которых пойдёт даль- ше, со своими специфическими целями и средствами. Тем не менее настоящая работа по своему замыслу не стремится к выявле- нию причинно-следственных связей, кото- рые возможны только в рамках строгих тео- ретических допущений, и ориентируется в первую очередь на выводы описательного характера [King, Keohane, Verba 1994: 7]. С учётом вышесказанного концепция «управляемой конкуренции» рассматрива- ется главным образом в рамках сочетания неореализма и неоинституционализма, что Тем не менее вопросы сотрудничества и снижения конфликтности рассматривают- ся посредством донесения собственных мотивов конкуренту и создания механиз- мов контроля её ограничения [Lindley 2007; Glasner 2010]. Анализ вышеуказанных ме- ханизмов осуществляется на основе неоин- ституционализма. В усилиях по снижению вероятности конфликта представители этого направления уделяют значительное внимание роли международных институ- тов – «набору имплицитных или экспли- цитных принципов, норм, правил и про- цедур принятия решений, вокруг которых происходит конвергенция ожиданий субъ- ектов, действующих в данной сфере меж- дународных отношений» [Krasner 1982]. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 97 Концепция «управляемой конкуренции»: регулирование, управление, достижение успеха Примечательно, что в одной из первых работ начала 2000-х годов, посвя- щённых регулированию конкуренции К началу 2020-х годов можно говорить о становлении трёх подходов к концепции «управляемой конкуренции» (табл.), кото рые различаются как по конкретным целям, так и по средствам их достижения. 18 Данный подход сочетания сдерживания, конкуренции и сотрудничества во многом схож с иде- ями “congagement” (от англ. containment и engagement) А. Фридберга [Friedberg 2011], “responsible competition” Т. Райта [Wright 2017] и “asymmetric competition” Р. Доши [Doshi 2021]. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 98 ВЛАДИМИР ЩЕРБУНОВ Таблица Подходы к концепции «Управляемая конкуренция» Подходы к управлению конкуренцией Цель Средства Примеры 1. Регулирование повышение предсказуемости отношений и снижение конфликтности создание и модернизация международных институтов различных типов Международные соглашения по снижению вероятности инцидентов в воздухе и на море, «горячие линии», двусторонние и многосторонние режимы регулирования безопасности 2. Канализирование программирование действий субъектов Предложение повестки, сигнализирование, конкретные внешнеполитические шаги в рамках конкурентных стратегий Стратегическая оборонная инициатива, обвинение правительства Ирака в хранении ОМП, выход из Договора по ПРО, заявления о «красных линиях» 3. Совместное решение проблем сохранение или достижение высокой конкурентной позиции в ключевых сферах симметричное или асимметричное сдерживание с элементами сотрудничества внешнее и внутреннее балансирование конкурента Источник: составлено автором. Источник: составлено автором. следователь предлагает меры по сдержива- нию Китая (например, в рамках тайвань- ского вопроса). В глобальном плане укре- пление и сохранение позиций Вашингтона должно опираться на создание ситуации, при которой «Восточная Азия будет разде- лена между великими державами» (вклю- чая США. – Прим. авт.), что не позволит такой региональной державе, как Китай, добиться гегемонии и создать в перспекти- ве угрозу территории самих Соединённых Штатов. США и КНР, использовались все три под- хода. В 2005 г. в докладе аналитического центра Stanley Foundation американский политолог Р. Росс применял определение «управление конкуренцией между велики- ми державами» в контексте рекомендуемой им стратегии Соединённых Штатов в от- ношении Китая [Ross 2005: 3]. Главная цель предложенного варианта управления за- ключается в стремлении «избежать высо- ких экономических и военных издержек противостояния прошлого» в условиях «неизбежного» роста влияния КНР. Тем самым она соответствует третьему подходу. При этом суть стратегии должна сводиться к созданию стратегической среды управля- емой конкуренции между США и Китаем с целью минимизации региональной не- стабильности и недопущения эскалации из-за конфликта интересов двух стран, что в большей степени согласуется уже с пер- вым и вторым подходами. 19 Под аккомодацией понимается процесс и результат признания критическим числом великих держав интересов и легитимности претензии державы, претендующей на статус великой, а также распределение лидерских ролей в соответствии с претензиями [Paul 2016: 3–4]. 20 Под «влиянием» в данной работе понимается «способность воздействовать на убеждения дру- гих, то есть на их знание или мнение либо о том, что есть, либо о том, что должно быть, о том, что (эмпирически) истинно или ложно, что (нормативно) правильно или неправильно, хорошо или плохо, желательные или нежелательные» [Zimmerling 2005: 141]. Подробнее о влиянии в международных отношениях см.: Mazarr M.J. et al. Understanding Influence in the Strategic Competition with China. [Электронный ресурс]. RAND Corporation, 2021. URL: https://www.rand.org/content/dam/rand/pubs/ research_reports/RRA200/RRA290-1/RAND_RRA290-1.pdf (accessed: 15.12.2022). О борьбе за влияние между США и Китаем см.: Moyer J.D. et al. China-Us Competition: Measuring Global Influence. [Электронный ресурс] // Atlantic Council. May 2021. URL: https://www.atlanticcouncil.org/wp-content/ uploads/2021/06/China-US-Competition-Report-2021.pdf (accessed: 15.12.2022). рр р д р ц p j g 22 См., например: Hass R. Principles For Managing U.S.-China Competition. [Электронный ресурс] // Brookings Foundation. August 2018. URL: https://www.brookings.edu/wp-content/uploads/2018/08/ FP_20180817_managing_competition.pdf (accessed: 25.12.2021); Allen J.R., Hass R., Jones B. Rising to The Challenge: Navigating Competition, Avoiding Crisis, And Advancing Us Interests In Relations With China. [Электронный ресурс] // Brookings Foundation. November 2021. URL: https://www. brookings.edu/wp-content/uploads/2021/11/FP_20211105_us_china_rivalry_hass_jones_allen.pdf (accessed: 25.12.2021); Paul C. et al. A Guide to Extreme Competition with China. [Электронный ресурс] // RAND Corporation. 2021. URL: https://www.rand.org/pubs/research_reports/RRA1378-1. html (accessed: 25.12.2021); Heath T. U.S. Strategic Competition with China. A RAND Research Primer. [Электронный ресурс] // RAND Corporation. 2021. URL: https://www.rand.org/pubs/research_ reports/RRA1378-1.html (accessed: 15.12.2022); Jones B. Major Power Rivalry and the Management of Global Threats. [Электронный ресурс] // Council on Foreign Relations. Center for Preventive Action. Discussion Paper Series on Managing Global Disorder. No. 7. November 2021. URL: https://cdn.cfr.org/ sites/default/files/report pdf/Jones MajorPowerRivalry.pdf (accessed: 15.12.2022). p p p p 21 От ред.: 29 июля 2019 г. Атлантический совет (Atlantic Сouncil) внесён в Перечень иностранных и международных неправительственных организаций, деятельность которых признана нежелатель- ной на территории Российской Федерации (см.: https://minjust.gov.ru/ru/documents/7756/). p _p _ j y p 23 С точки зрения учётов Китая выделяется работа Л. Голдстейна, в которой предлагается пере- чень мер для КНР и для США, направленных на снижение степени конкуренции в военной сфере (например, ограничение строительства американского авианосного флота). Подробнее см.: [Goldstein 2019: 344]. Концепция «управляемой конкуренции»: регулирование, управление, достижение успеха Тем не менее вместо способов или условий аккомода- ции19 интересов Пекина, создания меха- низмов предотвращения инцидентов ис- Если учесть, что идеи «стратегического соперничества» с КНР существовали уже в период администрации Дж. Буша-мл. (2001–2009) [Green 2017: 484], и принять во внимание усилия США в АТР по институ- ционализации и активизации сотрудниче- ства с союзниками, можно предположить, что появлявшиеся аналитические докла- ды отражали американские представления о формате взаимодействия с Китаем. Тем не менее полноценный переход к стратеги- ческому соперничеству оставался невоз- можным из-за военных кампании на Ближ Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 99 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ нем и Среднем Востоке в 2000-х годах, а также надежд на то, что стратегия вовле- чения КНР принесёт плоды и Пекин ста- нет «ответственным участником междуна- родной системы» [Zoellick 2021: 550]. так и в форме академических статей и мо- нографий специалистов-международников [Thomas, Montogmery 2012; Rosen 2012; Christensen 2015; Goldstein 2015; Doshi 2021, Hass 2021 и др.]22. так и в форме академических статей и мо- нографий специалистов-международников [Thomas, Montogmery 2012; Rosen 2012; Christensen 2015; Goldstein 2015; Doshi 2021, Hass 2021 и др.]22. По мере роста опасений США по поводу военно-экономического возвышения Китая и активизации внешней политики Пекина на ряде направлений (например, в отноше- нии территориальных споров в Южно- Китайском море или на границе с Индией) наметился отход Вашингтона от политики вовлечения. Ей на смену пришли подходы, предлагавшие американским политикам конкретные меры по созданию среды управляемой конкуренции, стратегической конкуренции или конкуренции за влия- ние20 между двумя державами. Работы по этой тематике выходили как в формате аналитических докладов из-под пера со- трудников западных аналитических цен- тров (таких как Атлантической совет21, Институт Брукингса, Корпорация РЭНД), Перечисленные работы преимуще- ственно ориентировались на невоенные аспекты конкуренции. Кроме того, в них отсутствовали объяснения по поводу того, чем именно возвышение КНР будет угро- жать национальной безопасности Соеди нённых Штатов, а также практически пол- ностью отсутствовали упоминания акко- модации с Китаем23. Предлагались меры по достижению успеха в конкуренции, что отражало преемственность американской большой стратегии, направленной на под- держание гегемонии и создание америка- ноцентричного мирового порядка. Тем самым они демонстрировали инертность мышления внешнеполитического сооб щества США, о чём писали российские и западные эксперты [Шаклеина 2021: 7; p _p _ j y p 23 С точки зрения учётов Китая выделяется работа Л. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 24 McDevitt M. U.S.-China Military Relations. [Электронный ресурс] // Sweet and Sour. Candid Talks to Improve U.S.-China Relations. NCAFP. November 2010. URL: https://www.ncafp.org/sweet-and-sour- candid-talks-to-improve-u-s-china-relations/ (accessed: 15.12.2022). p 25 China-US Joint Statement (October 29,1997). [Электронный ресурс] // Embassy of the People’s Republic of China in the UNitesed States of America. 30 October 1997. URL: http://us.china-embassy. gov.cn/eng/zmgx/zywj/lhsm3/200310/t20031023_4917631.htm (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 Концепция «управляемой конкуренции»: регулирование, управление, достижение успеха Голдстейна, в которой предлагается пере- чень мер для КНР и для США, направленных на снижение степени конкуренции в военной сфере (например, ограничение строительства американского авианосного флота). Подробнее см.: [Goldstein 2019: 344]. p _p _ j y p 23 С точки зрения учётов Китая выделяется работа Л. Голдстейна, в которой предлагается пере- чень мер для КНР и для США, направленных на снижение степени конкуренции в военной сфере (например, ограничение строительства американского авианосного флота). Подробнее см.: [Goldstein 2019: 344]. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 100 ВЛАДИМИР ЩЕРБУНОВ Drezner 2017]. Вместе с тем они указывали на сложность, а подчас и невозможность конвергенции интересов США и КНР в решении таких проблем, как, например, будущее Тайваня. нализации военно-политических отноше- ний, сопровождавшаяся созданием новых институтов, которые с точки зрения регу- лирования конкуренции можно разделить на три группы: 1) формализованные прави- ла поведения вооружённых сил или ис- пользования военных платформ, 2) диало- говые форматы в сфере обороны, 3) каналы кризисного общения («горячие линии»). Появление и развитие данных институтов с конца 1990-х годов стало, с одной сторо- ны, отражением глобального дрейфа в сто- рону более институционализированной мировой политики [Баталов 2016: 16]. С другой – было связано с количествен- ным ростом взаимодействия вооружённых сил двух стран и соответствующими инци- дентами с их участием. 24 McDevitt M. U.S.-China Military Relations. [Электронный ресурс] // Sweet and Sour. Candid Talks to Improve U.S.-China Relations. NCAFP. November 2010. URL: https://www.ncafp.org/sweet-and-sour- candid-talks-to-improve-u-s-china-relations/ (accessed: 15.12.2022). 25 China-US Joint Statement (October 29,1997). [Электронный ресурс] // Embassy of the People’s Republic of China in the UNitesed States of America. 30 October 1997. URL: http://us.china-embassy. gov cn/eng/zmgx/zywj/lhsm3/200310/t20031023 4917631 htm (accessed: 15 12 2022) Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 26 Agreement Between the Department of Defense of the United States of America and the Ministry of National Defense of The People's Republic of China on Establishing a Consultation Mechanism to Strengthen Military Maritime Safety. [Электронный ресурс] // US Department of State. Treaties and Other International Acts Series 12924. 19 January 1998. URL: https://www.state.gov/wp-content/ uploads/2019/02/12924-China-Maritime-Matters-Misc-Agreement-1.19.1998.pdf (accessed: 15.12.2022). 27 1 апреля 2001 г. американский самолёт-разведчик EP-3E проводил разведывательный вылет в районе Южно-Китайского моря в 110 км от китайской провинции Хайнань. В ходе перехвата двумя китайскими истребителями J-8 произошло столкновение американского и китайского летательных аппаратов, в ходе которого китайский истребитель был разрушен и его пилот погиб. EP-3E совершил вынужденную посадку на китайской базе Линшуй. Экипаж был задержан и допрошен, но вскоре воз- вращён в США после письма американской стороны с извинениями. [ Регулирование конкуренции: институционализация отношений и проблемы кризисного управления и проблемы кризисного управления Идея регулирования межгосударствен- ных отношений не является новой в науч- ном дискурсе или в политике, учитывая опыт создания международных институ- тов и режимов с целью снижения вероят- ности конфликтов между сверхдержавами ХХ века или регулирования других аспек- тов их отношений [George 2019]. Создан ные неформальные институты невмеша- тельства в сферы влияния между Советским Союзом и США в годы «холодной войны» [Levy, Young, Zurn 1995: 272] наряду с более формальными институализированными практиками в области нераспространения ядерного оружия и стратегической ста бильности, предотвращения столкновений на море или создания горячей линии с це- лью недопущения ядерных инцидентов или полномасштабных конфликтов спо- собствовали повышению предсказуемости отношений в условиях угрозы столкно вения двух стран [Батюк 2008]. Следо вательно, управление заключалось преи- мущественно в институционализации и регулировании взаимодействия для пони- жения вероятности военных конфликтов высокой интенсивности (что соответству- ет первому подходу из описанных выше), а также для препятствования свободной деятельности конкурента в сфере, где у со- перника может быть преимущество (вто- рой подход). Н б ХХ XXI В частности, в октябре 1994 г. амери- канские противолодочные самолёты в ходе учений обнаружили китайскую подвод ную лодку в Жёлтом море и начали сбра- сывать звуковые буи для сбора информа- ции акустической сигнатуры, что было стандартной практикой для американских военных со времён «холодной войны» [Thurston 2021: 312]. Ответом китайской стороны стала отправка истребителей для перехвата. Представители КНР также зая- вили, что «если подобный инцидент про- изойдет снова, то Китай попытается сбить американский самолёт»24. Последовавший за этим в марте 1996 г. третий тайваньский кризис, в котором были замешаны амери- канские авианосные группы, аналогич- ным образом поставил вопрос о необходи- мости снижения вероятности конфлик- тов и кризисных ситуаций. В 1998 г. была создана горячая линия на уровне президен- тов25. В январе 1998 г. министр обороны США У. Коэн в ходе визита в Китай под- писал Военно-морское консультативное соглашение (Martitime Military Consultative Agreement, MMCA), направленное на На рубеже ХХ и XXI веков во взаимо- действии Вашингтона и Пекина наблюда- лась тенденция к повышению институцио- Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 101 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ стане. В августе 2005 г. по инициативе КНР был создан Диалог о стратегическом партнёрстве, предусматривающий ежегод- ные встречи делегаций высокого уровня и обсуждение вопросов в сфере безопас ности с участием двух стран, получившие название Переговоры по координации обо- ронной политики (Defense Policy Coordi nation Talks)29. В июле 2006 г. р р р 28 Rosental E., Sanger D.E. U.S. Plane in China after it collides with Chinese Jet. [Электронный ресурс] // The New York Times. 2 April 2001. URL: https://www.nytimes.com/2001/04/02/world/ us-plane-in-china-after-it-collides-with-chinese-jet.html (accessed: 15.12.2022). 29 [ p j 29 Remarks at U.S. Embassy Beijing of Robert Zoellick, Deputy Secretary of State. [Электронный ресурс] // U.S. Department of State Archive. 2 August 2005. URL: https://2001-2009.state.gov/s/d/ former/zoellick/rem/50498.htm (accessed: 15.12.2022). 31 Beck L. China and U.S. sign accord on defense hotline. [Электронный ресурс] // Reuters. 29 Feb ruary 2008. URL: https://www.reuters.com/article/us-china-us-defence-idUSPEK7130320080229 (accessed: 15.12.2022). 30 Kucera J. Beijing military chief meets Rumsfeld. [Электронный ресурс] // South China Morning Post. 20 July 2006. URL: https://www.scmp.com/article/557292/beijing-military-chief-meets-rumsfeld (accessed: 15.12.2022). 32 U.S.-China Military Contacts: Issues for Congress. [Электронный ресурс] // Congressional Research Service. 5 January 2015. P.14. URL: https://www.everycrsreport.com/files/20150105_ RL32496_c568174a7b783237c7d89d015f864210b0046888.pdf (accessed: 15.12.2022). 33 Economic Dialogue. [Электронный ресурс] // U.S. Department of Treasury. Press Releases. 28 July 2009. URL: https://home.treasury.gov/news/press-releases/tg242 (accessed: 15.12.2022). 34 Opening Statement of U.S. Secretary of Defense Robert M. Gates at Joint Press Conference with Chinese Minister of National Defense Liang Guanglie. [Электронный ресурс] // DefenceIQ. 10 January 2011. URL: https://www.defenceiq.com/air-land-and-sea-defence-services/press-releases/us-defense- secretary-opens-dialogue-with-chinese-m 35 U.S., China Conduct Asia-Pacific Security Dialogue. [Электронный ресурс] // U.S. Department of Defence. DOD News. 11 March 2015. URL: https://www.defense.gov/News/News-Stories/Article/ Article/604252/ (accessed: 15.12.2022). Регулирование конкуренции: институционализация отношений и проблемы кризисного управления Октябрь–декабрь / 2022 102 ВЛАДИМИР ЩЕРБУНОВ и ликвидации последствий стихийных бед- ствий, борьбе с пиратством и для проведе- ния контртеррористических операций. Американская сторона также выступила с предложением об установлении страте гического диалога по ядерному оружию, противоракетной обороне, операциям в космосе и кибербезопасности34. Китай не продемонстрировал интереса к предло- жениям США. продажи вооружений Тайваню наложили негативный отпечаток на военно-полити- ческое сотрудничество двух стран, став причиной крупного сокращения числа контактов между представителями воен- ных ведомств (с 22 в 2006 г. до 12 и 7 в 2007 и 2010 годах соответственно) [Saunders, Bowie 2016: 13]. В период президентства Б. Обамы (2009– 2017), особенно второго срока, несмотря на сохранявшиеся противоречия, к кото- рым дополнительно добавились террито риальные претензии Китая к союзникам Соединённых Штатов, Филиппинам и Японии в Южно-Китайском и Восточно- Китайском морях, институционализация продолжилась. Были расширены контакты на уровне вооружённых сил, оформлен ди- алог для предотвращения инцидентов, КНР впервые приняла участие в многосторон- них учениях РИМПАК32, которые будут рассмотрены далее. Наметившаяся тенден- ция была обусловлена возобновлением аме риканского курса на встраивание Пекина в американоцентричный порядок и отсут- ствием крупных поставок вооружений Тай ваню в 2012–2014 годах. Последние часто осложняли отношения двух стран. Во время второго срока президентства Б. Обамы существенно увеличилось число контактов по линии военных министерств. В 2014 г. в рамках СЭД была достигнута до- говорённость о создании двустороннего Диалога по безопасности в АТР (Asia- Pacific Security Dialogue (APSD)), который был запущен в 2015 г. и затрагивал, в част- ности, ситуацию на Корейском полуостро- ве и другие вопросы безопасности в регио- не35. В результате была заложена более прочная основа для военных контактов, которые неуклонно развивались с 2011 года. Общее число визитов представителей аме- риканской стороны выросло с 10 в 2011 г. до 39 в 2014 г. [Saunders, Bowie 2016: 13]. Контакты включали: 1) визиты на высшем уровне министров обороны, председате- ля объединенного комитета начальников штабов, глав родов войск и региональ- ных командующих, отвечающих за Азию; 2) встречи на высшем уровне на между народных мероприятиях; 3) регулярные обмены с участием офицеров, молодых лидеров и представителей учреждений про фессионального военного образования; 4) функциональные и академические обме- ны между представителями профессио нального военного образования; 5) заходы В 2009 г. Стратегический диалог был совмещён с существовавшим параллель- но Экономическим диалогом33. Стал дей- ствовать Стратегический и экономический диалог (СЭД), охвативший функционал двух предшествующих площадок. В 2011 г. в ходе визита министра обороны США Р. Регулирование конкуренции: институционализация отношений и проблемы кризисного управления в ходе встречи министра обороны Дональда Рамсфелда и старшего вице-председателя Коммунисти ческой партии Китая (КПК) генерала Го Боксюна по предложению американ- ской стороны была достигнута догово рённость о проведении совместных воен- ных поисково-спасательных учений, кото- рые, впрочем, не состоялись30. Соеди нённые Штаты также получили доступ к архивам Народно-освободительной армии Китая (далее – НОАК) с информа- цией об американских военнопленных, пропавших без вести во время Корей ской войны (1950–1953). Заключительным шагом углубления отношений при адми- нистрации Дж.Буша-мл стало соглашение от февраля 2008 г. об открытии горячей линии между Министерством обороны США и Министерством национальной обороны КНР 31. Последовавшие в 2007–2008 годах снижение вероятности инцидентов между силами Соединённых Штатов и КНР на море и заложившее основу ежегодных консультаций по вопросам взаимодей- ствия на море26. р В конце лета 2000 г. американской сто- роной была произведена переоценка тре- бований к разведывательным полётам, в соответствии с которой выросла интен- сивность наблюдения за целями на пост советском пространстве. Следовательно, было зафиксировано увеличение числа наблюдательных полётов у границ КНР, в связи с чем китайская сторона выразила протест в декабре 2000 года [Thurston 2021: 318]. Столкновение над Южно-Китайским морем в апреле 2001 г. американского раз- ведывательного самолёта EP-3 с китай- ским истребителем J-827 продемонстриро- вало необходимость создания институтов предотвращения инцидентов в воздушном пространстве, а также выявило проблемы в сфере кризисного управления28. Во время второго президентства Дж. Буша-мл. военно-политические отно- шения начали стабилизироваться ввиду понижения напряжённости вокруг Тай ваня и возросшей вовлечённости США в военные операции в Ираке и Афгани р р р 28 Rosental E., Sanger D.E. U.S. Plane in China after it collides with Chinese Jet. [Электронный ресурс] // The New York Times. 2 April 2001. URL: https://www.nytimes.com/2001/04/02/world/ us-plane-in-china-after-it-collides-with-chinese-jet.html (accessed: 15.12.2022). p j 29 Remarks at U.S. Embassy Beijing of Robert Zoellick, Deputy Secretary of State. [Электронный ресурс] // U.S. Department of State Archive. 2 August 2005. URL: https://2001-2009.state.gov/s/d/ former/zoellick/rem/50498.htm (accessed: 15.12.2022). p j 29 Remarks at U.S. Embassy Beijing of Robert Zoellick, Deputy Secretary of State. [Электронный ресурс] // U.S. Department of State Archive. 2 August 2005. URL: https://2001-2009.state.gov/s/d/ former/zoellick/rem/50498.htm (accessed: 15.12.2022). 30 Kucera J. Beijing military chief meets Rumsfeld. [Электронный ресурс] // South China Morning Post. 20 July 2006. URL: https://www.scmp.com/article/557292/beijing-military-chief-meets-rumsfeld (accessed: 15.12.2022). 31 Beck L. China and U.S. sign accord on defense hotline. [Электронный ресурс] // Reuters. 29 Feb ruary 2008. URL: https://www.reuters.com/article/us-china-us-defence-idUSPEK7130320080229 (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). y p g 35 U.S., China Conduct Asia-Pacific Security Dialogue. [Электронный ресурс] // U.S. Department of Defence. DOD News. 11 March 2015. URL: https://www.defense.gov/News/News-Stories/Article/ Article/604252/ (accessed: 15.12.2022). _ g p 39 Memorandum of Understanding On the Rules of Behavior for Safety of Air and Maritime Encounters Between the Department of Defense of the United States of America And the Ministry of National Defense of the People's Republic of China. [Электронный ресурс] // U.S. Department of Defense. 9–10 November 2014. URL: https://dod.defense.gov/Portals/1/Documents/pubs/141112_Memorandu mOfUnderstandingRegardingRules.pdf (accessed: 15.12.2022). g 37 Проведение учений отчасти продемонстрировало специфику восприятия китайской стороны американских инициатив. После учений 2014 г. представители ВМФ США заявили, что, помимо четырёх приглашённых судов ВМФ НОАК, неподалёку от учений присутствовал также китайский разведывательный корабль. Подобная практика наблюдения за электронными сигнатурами судов в годы «холодной войны» являлась типичной для советской и американской сторон. 38 Memorandum of Understanding Between the United States of America the Department of Defense фnd the People's Republic of China Ministry of National Defense on Notification of Major Military Activities Confidence-Building Measures Mechanism. [Электронный ресурс] // U.S. Department of Defense. 31 October – 4 November 2014. URL: https://dod.defense.gov/Portals/1/Documents/pubs/ 141112 MemorandumOfUnderstandingOnNotification.pdf (accessed: 15.12.2022). 36 Military and Security Developments Involving the People’s Republic of China 2021. A Report to Congress. [Электронный ресурс] // US Department of Defense. 2021. URL: https://media.defense. gov/2021/Nov/03/2002885874/-1/-1/0/2021-CMPR-FINAL.PDF (accessed: 15.12.2022). g g g p 40 Supplement to the Memorandum of Understanding On the Rules of Behavior for Safety of Air and Maritime Encounters Between … [Электронный ресурс] // University of Southern California. US-China Institute. 15 September 2015. URL: https://china.usc.edu/sites/default/files/article/attachments/ US-CHINA_AIR_ENCOUNTERS_ANNEX_SEP_2015.pdf (accessed: 15.12.2022). 36 Military and Security Developments Involving the People’s Republic of China 2021. A Report to Congress. [Электронный ресурс] // US Department of Defense. 2021. URL: https://media.defense. gov/2021/Nov/03/2002885874/-1/-1/0/2021-CMPR-FINAL.PDF (accessed: 15.12.2022). 37 Проведение учений отчасти продемонстрировало специфику восприятия китайской стороны американских инициатив. После учений 2014 г. представители ВМФ США заявили, что, помимо четырёх приглашённых судов ВМФ НОАК, неподалёку от учений присутствовал также китайский разведывательный корабль. Подобная практика наблюдения за электронными сигнатурами судов в годы «холодной войны» являлась типичной для советской и американской сторон. 38 Memorandum of Understanding Between the United States of America the Department of Defense фnd the People's Republic of China Ministry of National Defense on Notification of Major Military Activities Confidence-Building Measures Mechanism. [Электронный ресурс] // U.S. Department of Defense. 31 October – 4 November 2014. URL: https://dod.defense.gov/Portals/1/Documents/pubs/ 141112_MemorandumOfUnderstandingOnNotification.pdf (accessed: 15.12.2022). 39 Memorandum of Understanding On the Rules of Behavior for Safety of Air and Maritime Encounters Between the Department of Defense of the United States of America And the Ministry of National Defense of the People's Republic of China. [Электронный ресурс] // U.S. Department of Defense. 9–10 November 2014. URL: https://dod.defense.gov/Portals/1/Documents/pubs/141112_Memorandu mOfUnderstandingRegardingRules.pdf (accessed: 15.12.2022). 40 Supplement to the Memorandum of Understanding On the Rules of Behavior for Safety of Air and Maritime Encounters Between … [Электронный ресурс] // University of Southern California. US-China Institute. 15 September 2015. URL: https://china.usc.edu/sites/default/files/article/attachments/ US-CHINA_AIR_ENCOUNTERS_ANNEX_SEP_2015.pdf (accessed: 15.12.2022). 41 Annual Report to Congress: Military and Security Developments Involving the People’s Republic of China 2015. [Электронный ресурс] // Office of the Secretary of Defense. April 2015. P.62. URL: https://dod. defense.gov/Portals/1/Documents/pubs/2015_China_Military_Power_Report.pdf (accessed: 15.12.2022). 42 Document: Code for Unplanned Encounters at Sea [Электронный ресурс] // US Naval Institute News. 17 June 2013. URL: https://news.usni.org/2014/06/17/document-conduct-unplanned- encounters-sea (accessed: 15.12.2022). g / / / /p / _ _ y_ _ p p 42 Document: Code for Unplanned Encounters at Sea [Электронный ресурс] // US Naval Institute News. 17 June 2013. URL: https://news.usni.org/2014/06/17/document-conduct-unplanned- encounters-sea (accessed: 15.12.2022). _ _ _ _ _ p 41 Annual Report to Congress: Military and Security Developments Involving the People’s Republic of China 2015. [Электронный ресурс] // Office of the Secretary of Defense. April 2015. P.62. URL: https://dod. defense.gov/Portals/1/Documents/pubs/2015_China_Military_Power_Report.pdf (accessed: 15.12.2022). _ _ _ _ _ p 41 Annual Report to Congress: Military and Security Developments Involving the People’s Republic of China 2015. [Электронный ресурс] // Office of the Secretary of Defense. April 2015. P.62. URL: https://dod. defense.gov/Portals/1/Documents/pubs/2015_China_Military_Power_Report.pdf (accessed: 15.12.2022). 42 Document: Code for Unplanned Encounters at Sea [Электронный ресурс] // US Naval Institute News. 17 June 2013. URL: https://news.usni.org/2014/06/17/document-conduct-unplanned- d Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 Annual Report to Congress: Military and Security Developments Involving the People s Republic of China 2015. [Электронный ресурс] // Office of the Secretary of Defense. April 2015. P.62. URL: https://dod. defense.gov/Portals/1/Documents/pubs/2015_China_Military_Power_Report.pdf (accessed: 15.12.2022). 42 Document: Code for Unplanned Encounters at Sea [Электронный ресурс] // US Naval Institute News. 17 June 2013. URL: https://news.usni.org/2014/06/17/document-conduct-unplanned- encounters-sea (accessed: 15.12.2022). Регулирование конкуренции: институционализация отношений и проблемы кризисного управления Гейтса в Пекин стороны договорились создать рабочую группу для разработки механизма проведения учений по поиску и спасанию на море, гуманитарной помощи Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 103 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ а также Меморандум о взаимопонимании о правилах поведения для обеспечения без- опасности воздушных и морских столкно- вений (air and maritime encounters)39. В сен- тябре 2015 г. представители оборонных ведомств США и КНР подписали дополни- тельное приложение к МОВ о правилах по- ведения, в котором оговаривались правила- поведения сторон для обеспечения безо- пасности в воздухе40. Меморандумы имеют возможность добавления новых механиз- мов путём подписания новых приложений, что оставляет окно возможностей для про- должения курса на укрепление мер дове- рия, чем на 2022 г. стороны так и не вос- пользовались. В пресс-релизе по поводу подписания меморандума американская сторона заявила о заинтересованности во включении нового приложения по вопро- су уведомлений о пусках баллистических ракет41, что, однако, не было реализовано. На многостороннем уровне стороны под- писали Кодекс незапланированных столк новений в море, который, не имея юри дически обязывающей силы, применялся военными судами США и КНР в 2015 г.42 военных судов в порты стран и совместные учения36. В 2014 и 2016 годах по приглашению американской стороны ВМФ КНР приня- ли участие в международных учениях РИМПАК37, в 2014 г. – в многосторонних учениях Кобра Голд, на которых отраба тывалось предоставление гуманитарной помощи и спасение при стихийных бед- ствиях. Примечательно, что большинство учений США и КНР затрагивали вопросы неконвенциональной безопасности – сферы, которая обладает наибольшим потенциа- лом для сотрудничества двух стран даже с учётом нарастающей конкуренции. З Заключительным элементом институцио нализации стало подписание двух мемо- рандумов и дополнений к ним, направ ленных на активизацию мер доверия и сни- жение вероятности инцидентов и конфлик- тов. В 2014 г. стороны заключили Мемо рандум о взаимопонимании (МОВ) об уве- домлении о крупных военных действиях, предусматривающий добровольный меха- низм уведомления о военной политике, стратегии двух стран и военных учениях38, Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 104 ВЛАДИМИР ЩЕРБУНОВ Несмотря на то что к концу 2010-х годов сформировался комплекс военно-политиче- ских институтов, которые должны способ- ствовать понижению вероятности инциден- тов и регулированию кризисных ситуаций, ключевой проблемой оставалось кризисное управление, которое американские эксперты привязывали к систематическому нежела- нию Китая использовать созданные каналы связи для разрешения кризисных ситуаций43. Как отмечали в мемуарах бывшие сотруд- ники администрации Дж.Буша-ст. 43 Kine P. ‘Spiral into crisis’: The U.S.-China military hotline is dangerously broken. [Электронный ресурс] // Politico. 9 January 2021. URL : https://www.politico.com/news/2021/09/01/us-china-military- hotline-508140 (accessed: 15.12.2022). 44 Ibid. 45 Ibid 46 U.S. Department of Defense Hosts First Crisis Communications Working Group With the People's Republic of China People's Liberation Army. [Электронный ресурс] // U.S. Department of Defense. 29 October 2020. URL: https://www.politico.com/news/2021/09/01/us-china-military-hotline-508140 (accessed: 15.12.2022). 47 Antwood K. Biden administration looks to set up ‘red phone’ to China for emergency communica tions. [Электронный ресурс] // CNN. 14 July 2021. URL: https://edition.cnn.com/2021/07/14/politics/ biden-red-phone-china-xi/index.html (accessed: 15.12.2022). 43 Kine P. ‘Spiral into crisis’: The U.S.-China military hotline is dangerously broken. [Электронный ресурс] // Politico. 9 January 2021. URL : https://www.politico.com/news/2021/09/01/us-china-military- hotline-508140 (accessed: 15.12.2022). 44 Ibid. 45 Ibid. 46 U.S. Department of Defense Hosts First Crisis Communications Working Group With the People's Republic of China People's Liberation Army. [Электронный ресурс] // U.S. Department of Defense. 29 October 2020. URL: https://www.politico.com/news/2021/09/01/us-china-military-hotline-508140 (accessed: 15.12.2022). 47 Antwood K. Biden administration looks to set up ‘red phone’ to China for emergency communica tions. [Электронный ресурс] // CNN. 14 July 2021. URL: https://edition.cnn.com/2021/07/14/politics/ biden-red-phone-china-xi/index.html (accessed: 15.12.2022). 48 Ali I. U.S. Defense Secretary says committed to stable, constructive relationship with China. [Электронный ресурс] // Reuters. 27 July 2021. URL: https://www.reuters.com/world/china/us-defense- secretary-says-committed-stable-constructive-relationship-with-china-2021-07-27/ (accessed: 15.12.2022). 49 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс]. URL: The White House. 13 June 2022. URL: https://www.whitehouse.gov/ briefing-room/statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans- meeting-with-politburo-member-yang-jiechi-3/?mc_cid=7dbbb91296&mc_eid=627c47469b (accessed: 15.12.2022). p 48 Ali I. U.S. Defense Secretary says committed to stable, constructive relationship with China. [Электронный ресурс] // Reuters. 27 July 2021. URL: https://www.reuters.com/world/china/us-defense- secretary-says-committed-stable-constructive-relationship-with-china-2021-07-27/ (accessed: 15.12.2022). 49 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс]. URL: The White House. 13 June 2022. URL: https://www.whitehouse.gov/ briefing-room/statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans- meeting-with-politburo-member-yang-jiechi-3/?mc_cid=7dbbb91296&mc_eid=627c47469b (accessed: 15.12.2022). 43 Kine P. ‘Spiral into crisis’: The U.S.-China military hotline is dangerously broken. [Электронный ресурс] // Politico. 9 January 2021. URL : https://www.politico.com/news/2021/09/01/us-china-military- hotline-508140 (accessed: 15.12.2022). 44 Ibid. 45 Ibid. 46 U.S. Department of Defense Hosts First Crisis Communications Working Group With the People's Republic of China People's Liberation Army. [Электронный ресурс] // U.S. Department of Defense. 29 October 2020. URL: https://www.politico.com/news/2021/09/01/us-china-military-hotline-508140 (accessed: 15.12.2022). 47 A t d K Bid d i i t ti l k t t ‘ d h ’ t Chi f i 55 Под «серыми зонами» понимаются области, где правительство КНР проводит геополитические, экономические, военные, кибер- и информационные операции (кибер), выходящие за рамки обыч- ной дипломатической и экономической деятельности и ниже применения кинетической военной силы. Аналогичный смысл западными экспертами вкладывается в понятия «все действия, кроме войны» (all measures short of war) или «операции на низком уровне» (low-level operations). Подробнее см.: Lin B. et al. A New Framework for Understanding and Countering China’s Gray Zone Tactics. [Электронный ресурс] // Rand Corporation. 2022. 8 p. URL: https://www.rand.org/pubs/research_ briefs/RBA594-1.html (accessed: 15.12.2022). 52 The Ministry of Foreign Affairs Announces Countermeasures in Response to Nancy Pelosi’s Visit to Taiwan. [Электронный ресурс] // Ministry of Foreign Affairs of the People’s Republic of China. 05 August 2022. URL: https://www.fmprc.gov.cn/eng/zxxx 662805/202208/t20220805 10735706.html КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ ных и стратегических документах США, опубликованных в 2022 г.: Индо-Тихооке анской стратегии, Стратегии националь- ной обороны и Стратегии национальной безопасности. ных и стратегических документах США, опубликованных в 2022 г.: Индо-Тихооке анской стратегии, Стратегии националь- ной обороны и Стратегии национальной безопасности. Тайвань привело к заморозке некоторых форматов в августе 2022 г.: консультаций в рамках военно-морского консультатив- ного соглашения, прямых военных кон- тактов глав региональных сил (например, INDOPACOM), а также переговоров по координации оборонной политики50. Тем не менее в ноябре 2022 г. после встречи лидеров двух стран и переговоров на уров- не министров обороны американская сто- рона заявила о вероятном перезапуске механизмов и уверенности к возвращению за стол переговоров о кризисном управ лении, в том числе о создании дополни- тельных контактов на уровне глав началь- ников штабов51. Реализация первой задачи призвана пре дотвратить ситуацию, при которой Сое динённые Штаты и Китай окажутся втя нуты в конфликты в «серых зонах»53 или в инциденты, связанные с обязательствами США по отношению к Тайваню, Филиппи нам, Японии, Республике Корея. Её пред- полагается решать путём обеспечения пере- веса, который достаточен для сдерживания Китая от применения вооружённой силы. Тем не менее расширение военной актив- ности США и их союзников и партнёров в западной части Тихого океана в условиях количественного и качественного роста вооружённых сил НОАК при отсутствии эффективных механизмов координации может привести к большой вероятности столкновений в морском или воздушном пространствах54. Помимо этого, возникает вопрос об институционализации контактов для кризисного управления не только между двумя конкурентами, но и среди американ- ских союзников, на которых распространя- ются гарантии безопасности Вашингтона, и их партнёров. В опубликованной Стра тегии национальной обороны 2022 г. в ок- тябре 2022 г. отмечалась роль коммуника- Вопрос кризисного регулирования при- обретает всё большую актуальность с учё- том задач, которые ставит перед собой администрация Дж. Байдена. Во Времен ном стратегическом руководстве по нацио- нальной безопасности 2021 г. перед США ставились две ключевые цели52: 1) создание благоприятного соотношения сил для сдерживания и предотвращения непосред- ственной угрозы Соединённым Штатам и их союзникам, объединённое с планами по доминированию в ключевых регионах; 2) поддержка стабильной и открытой меж- дународной системы, опирающейся на многосторонние институты и нормы. Дан ные цели были повторены в доктриналь- 54 Существует точка зрения, согласно которой кризисные ситуации на море легче контролировать, чем на суше, в силу специфики соприкосновения вооружённых сил (одиночные военные платформы, а не войсковые соединения) и пространственной разрежённости. Подробнее см.: [O’Hanlon 2020]. Mеждународные процессы. Том 20. / 54 Interim National Security Strategic Guidance [Электронный ресурс] // White House. March 2021. URL : https://www.whitehouse.gov/wp-content/uploads/2021/03/NSC-1v2.pdf (accessed: 15.12.2022). 53 Detsch J. U.S. Seeks to Allay Fears of Military Crisis With China. [Электронный ресурс] // Foreign Policy. 22 November 2022. URL: https://foreignpolicy.com/2022/11/22/us-china-military-talks-asean- taiwan/ (accessed: 15.12.2022). 54 Существует точка зрения, согласно которой кризисные ситуации на море легче контролировать, чем на суше, в силу специфики соприкосновения вооружённых сил (одиночные военные платформы, а не войсковые соединения) и пространственной разрежённости. Подробнее см.: [O’Hanlon 2020]. 52 The Ministry of Foreign Affairs Announces Countermeasures in Response to Nancy Pelosi’s Visit to Taiwan. [Электронный ресурс] // Ministry of Foreign Affairs of the People’s Republic of China. 05 August 2022. URL: https://www.fmprc.gov.cn/eng/zxxx_662805/202208/t20220805_10735706.html 53 Detsch J. U.S. Seeks to Allay Fears of Military Crisis With China. [Электронный ресурс] // Foreign Policy. 22 November 2022. URL: https://foreignpolicy.com/2022/11/22/us-china-military-talks-asean- taiwan/ (accessed: 15.12.2022). 54 Interim National Security Strategic Guidance [Электронный ресурс] // White House. March 2021. URL : https://www.whitehouse.gov/wp-content/uploads/2021/03/NSC-1v2.pdf (accessed 15.12.2022). 55 Под «серыми зонами» понимаются области, где правительство КНР проводит геополитические экономические, военные, кибер- и информационные операции (кибер), выходящие за рамки обыч- ной дипломатической и экономической деятельности и ниже применения кинетической военной силы. Аналогичный смысл западными экспертами вкладывается в понятия «все действия, кроме войны» (all measures short of war) или «операции на низком уровне» (low-level operations). Подробнее см.: Lin B. et al. A New Framework for Understanding and Countering China’s Gray Zone Tactics [Электронный ресурс] // Rand Corporation. 2022. 8 p. URL: https://www.rand.org/pubs/research_ briefs/RBA594-1.html (accessed: 15.12.2022). 54 Существует точка зрения, согласно которой кризисные ситуации на море легче контролировать чем на суше, в силу специфики соприкосновения вооружённых сил (одиночные военные платформы а не войсковые соединения) и пространственной разрежённости. Подробнее см.: [O’Hanlon 2020]. Регулирование конкуренции: институционализация отношений и проблемы кризисного управления и ди- пломаты, во время событий на площади Тяньаньмэнь, когда вооружённые силы НОАК подавили протестные выступления студентов в Пекине, президент Дж. У. Буш- ст. не смог связаться с лидером КНР Дэн Сяопином. Ситуация повторилась в апреле 2001 г. во время кризиса после воздушного столкновения EP-3 и J-8, когда посол в Пекине Дж. Пруэр не мог дозвониться до китайского военного ведомства в течение 12 часов44. В мае 1999 г. американским военным не удалось установить контакт с китайской стороной после ошибочного нанесения бомбового удара по китайскому посольству в Белграде во время операции НАТО45. лизации ситуации, сохранявшейся до кон- ца 2010-х. Однако приход администрации Д. Трампа, с его жёсткой антикитайской риторикой, сокращением контактов, отзы вом приглашения для китайской стороны участвовать в учениях РИМПАК, а также нарастанием интенсивности американ- ских операций по обеспечению свободы мореходства в Южно-Китайском море, снова обострил данную проблему. С целью её решения в октябре 2020 г. оборонными министерствами двух стран была создана Рабочая группа по кризисной коммуника- ции с целью «обсуждения концептов кри- зисного общения, предотвращения кризи- сов и регулирования кризисов»46. С избранием Дж. Байдена проблема кри- зисного управления стала упоминаться чаще. В начале 2021 г. об игнорировании китай- ской стороной звонков по горячей линии между оборонными ведомствами заявил член Совета национальной безопасности К. Кэмпбелл47. В июле 2021 г. в ходе вы ступления данную проблему подчеркнул министр обороны Л. Остин48. Впослед ствии она поднималась на двусторонних встречах на уровне официальных пред ставителей, например, в июне 2022 г.49 Охлаждение отношений между двумя стра- нами на фоне визита спикера Палаты представителей США Н. Пелоси на Последующее сближение во время адми нистраций Дж.Буша-мл и Б.Обамы, о ко- тором говорилось ранее, привело к норма- 47 Antwood K. Biden administration looks to set up ‘red phone’ to China for emergency communica tions. [Электронный ресурс] // CNN. 14 July 2021. URL: https://edition.cnn.com/2021/07/14/politics/ biden-red-phone-china-xi/index.html (accessed: 15.12.2022). p 48 Ali I. U.S. Defense Secretary says committed to stable, constructive relationship with China. [Электронный ресурс] // Reuters. 27 July 2021. URL: https://www.reuters.com/world/china/us-defense- secretary-says-committed-stable-constructive-relationship-with-china-2021-07-27/ (accessed: 15.12.2022). 49 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс]. URL: The White House. 13 June 2022. URL: https://www.whitehouse.gov/ briefing-room/statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans- meeting-with-politburo-member-yang-jiechi-3/?mc_cid=7dbbb91296&mc_eid=627c47469b (accessed: 15.12.2022). 49 Readout of National Security Advisor Jake Sullivan’s Meeting with Politburo Member Yang Jiechi. [Электронный ресурс]. URL: The White House. 13 June 2022. URL: https://www.whitehouse.gov/ briefing-room/statements-releases/2022/06/13/readout-of-national-security-advisor-jake-sullivans- meeting-with-politburo-member-yang-jiechi-3/?mc_cid=7dbbb91296&mc_eid=627c47469b (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 105 52 The Ministry of Foreign Affairs Announces Countermeasures in Response to Nancy Pelosi’s Visit to Taiwan. [Электронный ресурс] // Ministry of Foreign Affairs of the People’s Republic of China. 05 August 2022. URL: https://www.fmprc.gov.cn/eng/zxxx_662805/202208/t20220805_10735706.html 53 Detsch J. U.S. Seeks to Allay Fears of Military Crisis With China. [Электронный ресурс] // Foreign Policy. 22 November 2022. URL: https://foreignpolicy.com/2022/11/22/us-china-military-talks-asean- taiwan/ (accessed: 15.12.2022). 54 Interim National Security Strategic Guidance [Электронный ресурс] // White House. March 2021. URL : https://www.whitehouse.gov/wp-content/uploads/2021/03/NSC-1v2.pdf (accessed: 15 12 2022) p g y y 59 Statement by Pentagon Press Secretary Peter Cook on Incident in South China Sea. [Электронный ресурс] // U.S. Department of Defense. 16 December 2016. URL: https://www.defense.gov/News/ Releases/Release/Article/1032611/statement-by-pentagonpress-secretary-peter-cook-on-incident-in- south-china-sea/ (accessed: 15.12.2022). 60 О С h l 55 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). 58 Bennett B. Trump’s erratic style could undermine China’s agreement to stop hacking U.S. businesses [Электронный ресурс] // Los Angeles Times. 3 April 2017. URL: https://www.latimes.com/ politics/la-fg-us-china-cyber-20170403-story.html 56 Annual Threat Assessment of the US Intelligence Community. [Электронный ресурс] // Office of the Director of National Intelligence. 7 February 2022. P.8. URL: https://www.dni.gov/files/ODNI/documents/ assessments/ATA-2022-Unclassified-Report.pdf (accessed: 15.12.2022) p p 57 Jones S. US and China set up ‘space hotline’. [Электронный ресурс] // Financial Times. 20 Novem ber 2015. URL: https://www.ft.com/content/900870f4-8f9f-11e5-a549-b89a1dfede9b (accessed: 15.12.2022). 60 О восприятии КНР и США кризисных ситуаций см.: Morris L.J., Marcrum K. Another 'Hotline' with China Isn't the Answer. [Электронный ресурс] // RAND Corporation. 27 July 2022. URL: https:// www.rand.org/blog/2022/07/another-hotline-with-china-isnt-the-answer.html (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 55 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). 56 Annual Threat Assessment of the US Intelligence Community. [Электронный ресурс] // Office of the Director of National Intelligence. 7 February 2022. P.8. URL: https://www.dni.gov/files/ODNI/documents/ assessments/ATA-2022-Unclassified-Report.pdf (accessed: 15.12.2022) 57 Jones S. US and China set up ‘space hotline’. [Электронный ресурс] // Financial Times. 20 Novem ber 2015. URL: https://www.ft.com/content/900870f4-8f9f-11e5-a549-b89a1dfede9b (accessed: 15.12.2022). 58 Bennett B. Trump’s erratic style could undermine China’s agreement to stop hacking U.S. businesses [Электронный ресурс] // Los Angeles Times. 3 April 2017. URL: https://www.latimes.com/ politics/la-fg-us-china-cyber-20170403-story.html 59 Statement by Pentagon Press Secretary Peter Cook on Incident in South China Sea. [Электронный ресурс] // U.S. Department of Defense. 16 December 2016. URL: https://www.defense.gov/News/ Releases/Release/Article/1032611/statement-by-pentagonpress-secretary-peter-cook-on-incident-in- south-china-sea/ (accessed: 15.12.2022). 60 О восприятии КНР и США кризисных ситуаций см.: Morris L.J., Marcrum K. Another 'Hotline' with China Isn't the Answer. [Электронный ресурс] // RAND Corporation. 27 July 2022. URL: https:// www.rand.org/blog/2022/07/another-hotline-with-china-isnt-the-answer.html (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 p Подробнее о деятельности департамента см.: Spevacek A.M. Net Assessment: Overview and Resources. [Электронный ресурс] // USAID Knowledge Services Center (KSC). 28 June 2017. URL: https://pdf.usaid.gov/pdf_docs/PA00SXMH.pdf (accessed 15.12.2022). Подробную общую оценку потенциала США, Японии по отношению к Китаю, выполненную Центром Карнеги в 2017 г., можно посмотреть в: [Swaine et al. 2017]. 61 Mazarr M. J. et al. Understanding the Emerging Era of International Competition: Theoretical and Historical Perspectives. [Электронный ресурс] // RAND Corporation, 2018. URL: https://www.rand.org/ pubs/research_reports/RR2726.html (accessed: 15.12.2022). 62 Согласно Директиве №5111.11 Министерства обороны США от 22 августа 2001 г. и 23 дека- бря 2009 г. под общими оценками (net assessment) «понимается сравнительный анализ военных, технологических, политических, экономических и других факторов, влияющих на относительную военную мощь наций». Цель департамента – «разрабатывать и координировать общие оценки состо- яния, тенденций и будущих перспектив военного потенциала США по сравнению с другими странами или группами стран, чтобы выявить возникающие или будущие угрозы или возможности для США. Взято из: Directive 5111.11. [Электронный ресурс] // Cryptome. 23 December 2009. URL: http:// www.dtic.mil/whs/directives/corres/pdf/511111 (accessed: 15.12.2022). 61 Mazarr M. J. et al. Understanding the Emerging Era of International Competition: Theoretical and Historical Perspectives. [Электронный ресурс] // RAND Corporation, 2018. URL: https://www.rand.org/ pubs/research_reports/RR2726.html (accessed: 15.12.2022). 61 Mazarr M. J. et al. Understanding the Emerging Era of International Competition: Theoretical and Historical Perspectives. [Электронный ресурс] // RAND Corporation, 2018. URL: https://www.rand.org/ pubs/research_reports/RR2726.html (accessed: 15.12.2022). 62 Согласно Директиве №5111.11 Министерства обороны США от 22 августа 2001 г. и 23 дека- бря 2009 г. под общими оценками (net assessment) «понимается сравнительный анализ военных, технологических, политических, экономических и других факторов, влияющих на относительную военную мощь наций». Цель департамента – «разрабатывать и координировать общие оценки состо- яния, тенденций и будущих перспектив военного потенциала США по сравнению с другими странами или группами стран, чтобы выявить возникающие или будущие угрозы или возможности для США. Взято из: Directive 5111.11. [Электронный ресурс] // Cryptome. 23 December 2009. URL: http:// www.dtic.mil/whs/directives/corres/pdf/511111 (accessed: 15.12.2022). Подробнее о деятельности департамента см.: Spevacek A.M. Net Assessment: Overview and Resources. [Электронный ресурс] // USAID Knowledge Services Center (KSC). 28 June 2017. URL: https://pdf.usaid.gov/pdf_docs/PA00SXMH.pdf (accessed 15.12.2022). Подробную общую оценку потенциала США, Японии по отношению к Китаю, выполненную Центром Карнеги в 2017 г., можно посмотреть в: [Swaine et al. 2017]. КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ Октябрь–декабрь / 2022 107 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ Разработчиком конкурентных стратегий выступал Э. Маршалл, занявший в 1973 г. должность главы созданного в том же году Управления общих оценок Министерства обороны (Office of Net Assessment)62. Как отмечал американский исследователь С. Розэн, по своей сути, «конкурентные стратегии пытаются побудить конкурен- тов играть в нашу игру. Игру, которую мы, вероятно, выиграем… заставить их совер- шать ошибки, к которым они склонны, отда вая предпочтение подходам, провоцирующим соперника и побуждающим его к неэффек- тивному распределению ресурсов» [Rosen 2012: 12]. В контексте «холодной войны» речь шла о создании стратегических воору- жений и удержании СССР в гонке воору- жений с целью его истощения. Данная стратегия должна была использовать воен- ный потенциал и его наращивание в усло- виях мирной конкуренции, опираясь на четыре компонента: 1) долговременный характер планирования; 2) использование собственных сильных сторон и компенса- ция слабых сторон; 3) упор на слабости оппонента; 4) стремление отвлечь внима- ние оппонента с целью совершения дей- ствий и усилий, которые наименее опасны или наиболее выгодны для инициатора [Thomas, Montgomery 2012: 257]. струмента выражения неприятия той или иной деятельности США, нивелируя их эф- фективность. В-третьих, планы США по усилению институционального контроля над теми или иными военно-политически- ми сферами китайского ВПК в силу коли- чественного и качественного перевеса аме- риканской стороны, а также в условиях взаимного недоверия, с большой вероятно- стью, будут отвергнуты китайской сторо- ной. Главным аргументом в пользу согласия китайской стороны на углубление диалога может быть желание получить переговор- ный рычаг в отношениях с США. КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ Номер 4 (71). Октябрь–декабрь / 2022 106 ВЛАДИМИР ЩЕРБУНОВ ций в кризисных ситуациях с союзниками, партнёрами и конкурентами55. американской третьей стратегии компен- сации [Hass 2021: 128]. В декабре 2016 г. в международных водах Южно-Китайского моря китайские военно-морские силы за- хватили американский подводный разве- дывательный дрон. Хотя он и был возвра- щён, но случай заложил прецедент на буду- щее59. В данной связи возникают вопросы относительно потенциала сдерживания военных сил конкурента смертоносными автономными системами и тем, как гибель или захват автономных платформ с участи- ем страны-соперника будет воспринимать- ся конкурентом в условиях отсутствия пра- вил регулирования подобных ситуаций. В контексте второй задачи Стратегии всё более остро встаёт вопрос по регулирова- нию отношений в сферах со слабо развиты- ми правилами или институтами, таки- ми как киберпространство, космос, новые ударные системы, основанные в том числе на подрывных технологиях (disruptive tech nology). В докладе Разведывательного управления США 2022 г. космическое про- странство было выделено в качестве одной из ключевых сфер, где Китай представляет угрозу Соединённым Штатам56. Его пози- ционирование в таком качестве обусловле- но наличием противоспутникового оружия и темпами развития китайской космиче- ской программы. В 2015 г. в СМИ появи- лись публикации о создании канала связи между двумя странами по предотвращению инцидентов в космическом пространстве, но на официальном уровне эта информа- ция не подтверждалась57. Существовавший с 2015 г. между сторонами канал по взаимо- действию по киберпреступлениям продол- жает функционировать, но его эффектив- ность остаётся неизвестной58. Управление конкуренцией посредством её регулирования продолжит оставаться од- ной из наиболее обсуждаемых тем двусто- ронних отношений в обозримом будущем. Вместе с тем американская стратегия повы- шения институционализации отношений может в итоге оказаться неэффективной в силу ряда причин. Во-первых, наличие каналов кризисного регулирования не устраняет противоречий великодержавной конкуренции и в теории может способство- вать её обострению, создавая ложное чув- ство контроля над ситуацией и повышая вероятность провокаций. Во‑вторых, китай- ская сторона может продолжить курс на не- приятие расширения этих каналов с целью сохранения фактора непредсказуемости60 либо на их использование в качестве ин- Всё более актуальным становится регу- лирование развёртывания автономных смер- тоносных систем, использующих искус- ственный интеллект, например беспилот- ных подводных аппаратов, которые ста новятся одним из ключевых элементов 60 О восприятии КНР и США кризисных ситуаций см.: Morris L.J., Marcrum K. Another 'Hotline' with China Isn't the Answer. [Электронный ресурс] // RAND Corporation. 27 July 2022. URL: https:// www.rand.org/blog/2022/07/another-hotline-with-china-isnt-the-answer.html (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Управление конкуренцией: конкурентные стратегии и программирующее лидерство Управление конкуренцией: конкурентные стратегии и программирующее лидерство Помимо регулирования военно-полити- ческой конкуренции, с 2010-х годов в ака- демических кругах стали активнее выска- зываться идеи о возможности влиять на внешнеполитический курс Китая посред- ством сочетания конкуренции и сотрудни- чества [Friedberg 2011], а также о возмож- ности направлять конкуренцию в нужном для США направлении61 [Campbell, Sullivan 2019]. Истоки данного подхода в отноше- ниях великих держав связаны с конкурент- ными стратегиями, применявшимися США с 1970-х годов для воздействия на военное строительство и внешнеполитиче- ский курс СССР, а также с идеями про- граммирующей стратегии начала 2000-х годов, направленными как на оппонента, так и на союзников. В отношении Китая эксперты в середине 2010-х годов предлагали две конкурентные стратегии: максималистскую и минима- листскую. Максималистская, построенная на опыте соперничества с СССР, предпола- Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 108 ВЛАДИМИР ЩЕРБУНОВ Объектами вышеуказанной стратегии могут быть союзники, партнёры или кон- куренты. Внешнеполитические цели США и союзников либо полностью отождеств ляются, либо американская позиция пре- обладает над ними при некоторой степени слияния. Подобная конвергенция дости гается в условиях объективной близости национальных интересов (например, со- хранение благоприятного баланса сил или выгодных международных институтов в региональной подсистеме – для союзни- ков и партнёров либо решение глобальных проблем – для соперников). Помимо пере- крёстных интересов между ними, большую роль играют разница потенциалов и уча- стие в военно-политических союзах. Как отмечает И.А. Истомин, одной из ключе- вых функций, которую выполняют такого рода альянсы, является именно ограниче- ние связей малых партнёров с потенциаль- ными конкурентами [Истомин 2021: 31]. Данное программирование подкрепляется инструментальным путём, то есть поста- новкой опережающей повестки дня и опре- делением дискурса, дипломатическим дав- лением, апеллированием к общим ценно- стям, а также предоставлением материаль- ных ресурсов (ударных или оборонитель- ных платформ, военной инфраструктуры, доступа к технологиям) и информации [Троицкий 2004: 49]. гала, что страна-инициатор должна стре- миться положить конец конкуренции, вы- зывая экономическое истощение конкурен- та. При такой логике США могут и должны поощрять усилия Китая по вложению ре- сурсов в дорогостоящие средства проециро- вания силы на большие расстояния, вклю- чая авианосцы, десантные корабли, круп- ные надводные боевые корабли и средства снабжения. Минималистская стратегия бы- ла направлена на ограничение масштабов конкуренции, сохранение наиболее благо- приятного баланса сил и создание стратеги- ческих вариантов, которые можно исполь- зовать в будущем как возможности для удержания КНР в рамках «первой/второй островной цепи»63, провоцируя его инве- стиции в оборонные системы ограничения и воспрещения доступа и манёвра (англ. 63 Концепция первой и второй островной цепи, предложенная госсекретарем Дж.Ф. Даллесом, связана с использованием США и их союзниками островов, простирающихся от Камчатки до Индонезии, для сдерживания угрозы в Евразии вообще и от Китая и России в частности. Подробнее о важности островных цепей см.: [Haddick 2014] Управление конкуренцией: конкурентные стратегии и программирующее лидерство anti-access and area denial A2\AD), например системы эшелонированной ПВО [Thomas, Montgomery 2012: 257]. С точки зрения влияния на поведение как союзников, так и оппонентов наибо- лее ярко этот феномен описал российский исследователь М.А. Троицкий, определив- ший его как «программирование». Кон цепция «программирующего лидерства (или партнёрства)», а в обобщенном смыс- ле программирующая стратегия была пред- ложена для описания политики США времён администрации Дж. Буша-мл. – «действий, направленных на опережающее установление повестки дня, то есть круга проблем, определяющих спектр и направ- ленность совместных действий США с го- сударствами – объектами данной страте- гии» [Троицкий 2002: 87]. Опережающий характер позволял Соединённым Штатам принуждать страны-объекты к реактив ному внешнеполитическому курсу. Напри мер, в конце 1990-х – начале 2000-х годов Россия оказалась вынуждена реагировать на пересмотр и последующий выход США из соглашения по ПРО. В отношениях с соперниками США опираются на «вовлечение в решение про- блем и военное сдерживание» [Christensen 2015: 286–287], что дополняется, с одной стороны, навязыванием стране-объекту проблемы, которая будет ограничивать активность и нейтрализовать ресурсы, а с другой стороны, цели в рамках повест- ки, которая в перспективе может оказаться затратной для страны-объекта. Напри мер, демократизация может оказать кар динальное воздействие на лояльность по- литических элит, решимость населения Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 109 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ вторив данный тезис в июне 2022 г.66 Однако к концу 2022 г. риторика сдвинулась к про- тивостоянию «демократий против автокра- тий», что было закреплено в СНБ-2022, в которой отмечалась идеологическая при- рода конкуренции США и КНР, а также либерально-демократическое устройство главных партнёров и союзников США67. продолжать соперничество с Соединённы ми Штатами и внутреннюю стабильность страны в целом. Российский исследователь Т.А. Шаклеина трансформировала эту идею, выделив два трека мировой полити- ки: либерально-демократический, включа- ющий темы, в которых Запад имеет преи- мущество (например, зелёная повестка), а также традиционный, ориентированный на вопросы стабилизации международных отношений [Шаклеина 2021: 7]. продолжать соперничество с Соединённы ми Штатами и внутреннюю стабильность страны в целом. Российский исследователь Т.А. Шаклеина трансформировала эту идею, выделив два трека мировой полити- ки: либерально-демократический, включа- ющий темы, в которых Запад имеет преи- мущество (например, зелёная повестка), а также традиционный, ориентированный на вопросы стабилизации международных отношений [Шаклеина 2021: 7]. Отношения с союзниками также говорят о консолидации и конвергенции общих интересов. 67 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). P. 8. Управление конкуренцией: конкурентные стратегии и программирующее лидерство Создание военно-технологиче- ского блока Австралии, Великобритании и США AUKUS в октябре 2021 года, хоть и являющееся логическим развитием воен- но-политических отношений Соединён ных Штатов и Австралии на протяжении XXI века [Щербунов 2021], продолжает процесс консолидации системы союзов, а фактически способствует их переходу от системы «оси и спиц», завязанной на США, на «федеративную» [Green, Hicks, Cooper 2014], или «узловую», оборону [Simón, Lanoszka, Meijer 2021]. В данной системе Вашингтон отводит Австралии, Японии и Индии роль узлов в рамках воз- родившегося форума Четырёхсторонний диалог по безопасности (QUAD) [Худай кулова, Рамич 2020], вокруг которых будет происходить военно-политическая консо- лидация антикитайских усилий. Дополне нием QUAD и существовавшего ранее трёхстороннего диалога (США–Япония– Австралия) стало Соглашение о взаимном доступе Австралии и Японии от 6 января 2022 года, заложившее основу для совмест- ных учений в Индо-Пацифике (ИТР) и углубления сотрудничества в сфере без Если посмотреть на стратегию програм- мирования, то к началу 2020-х годов её положения продолжали использоваться Вашингтоном. Активное отстаивание кон- цепции «Открытого и свободного Индо- Тихоокеанского региона» с её явной анти- китайской направленностью задавало вектор для стран-союзников [Торкунов 2019: 26]. При этом оно вынуждало Китай реагиро- вать, препятствуя перехвату Пекином ини- циативы. На доктринальном уровне это особенно заметно по Стратегии нацио- нальной безопасности 2022 года, где гово- рилось, что «США и наши союзники и партнёры имеют возможность формиро- вать внешнюю среду КНР и России таким образом, чтобы это влияло на их поведе- ние, даже когда мы с ними конкурируем»64. Если посмотреть на стратегию програм- мирования, то к началу 2020-х годов её положения продолжали использоваться Вашингтоном. Активное отстаивание кон- цепции «Открытого и свободного Индо- Тихоокеанского региона» с её явной анти- китайской направленностью задавало вектор для стран-союзников [Торкунов 2019: 26]. При этом оно вынуждало Китай реагиро- вать, препятствуя перехвату Пекином ини- циативы. На доктринальном уровне это особенно заметно по Стратегии нацио- нальной безопасности 2022 года, где гово- рилось, что «США и наши союзники и партнёры имеют возможность формиро- вать внешнюю среду КНР и России таким образом, чтобы это влияло на их поведе- ние, даже когда мы с ними конкурируем»64. Примечателен подход Вашингтона к вза имодействию с партнёрами, а именно про- тиворечивость риторики. В первые месяцы президентства Дж. Байдена министр оборо ны США Л.Остин, находясь с визитом в Син гапуре в июле 2021 г.65, заявил о нежелании принуждать региональных игроков к выбо- ру между США или КНР, впоследствии по- Примечателен подход Вашингтона к вза имодействию с партнёрами, а именно про- тиворечивость риторики. В первые месяцы президентства Дж. 64 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). 64 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). 65 Secretary of Defense Remarks at the 40th International Institute for Strategic Studies Fullerton Lecture (As Prepared). [Электронный ресурс] // U.S. Department of Defense. 27 July 2021. Singapore. URL: https://www.defense.gov/News/Speeches/Speech/Article/2708192/secretary-of-defense- remarks-at-the-40th-international-institute-for-strategic/ (accessed: 15.12.2022). 66 Garamone J. Austin Can Point to Growing Ties With Asia During Visit to Singapore. [Электронный ресурс] // U.S. Department of Defense. 9 June 2022. URL: https://www.defense.gov/News/News- Stories/Article/Article/3057206/austin-can-point-to-growing-ties-with-asia-during-visit-to-singapore/ (accessed: 15.12.2022). 67 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. hit h / t t/ l d /2022/10/Bid H i Ad i i t ti N ti l S it 66 Garamone J. Austin Can Point to Growing Ties With Asia During Visit to Singapore. [Электронный ресурс] // U.S. Department of Defense. 9 June 2022. URL: https://www.defense.gov/News/News- Stories/Article/Article/3057206/austin-can-point-to-growing-ties-with-asia-during-visit-to-singapore/ (accessed: 15.12.2022). 65 Secretary of Defense Remarks at the 40th International Institute for Strategic Studies Fullerton Lecture (As Prepared). [Электронный ресурс] // U.S. Department of Defense. 27 July 2021. Singapore. URL: https://www.defense.gov/News/Speeches/Speech/Article/2708192/secretary-of-defense- remarks-at-the-40th-international-institute-for-strategic/ (accessed: 15.12.2022). 65 Secretary of Defense Remarks at the 40th International Institute for Strategic Studies Fullerton Lecture (As Prepared). [Электронный ресурс] // U.S. Department of Defense. 27 July 2021. Singapore. URL: https://www.defense.gov/News/Speeches/Speech/Article/2708192/secretary-of-defense- remarks-at-the-40th-international-institute-for-strategic/ (accessed: 15.12.2022). 66 Garamone J. Austin Can Point to Growing Ties With Asia During Visit to Singapore. [Электронный ресурс] // U.S. Department of Defense. 9 June 2022. URL: https://www.defense.gov/News/News- Stories/Article/Article/3057206/austin-can-point-to-growing-ties-with-asia-during-visit-to-singapore/ (accessed: 15.12.2022). 67 National Security Strategy 2022. // White House. 12 October 2022. URL: https://www. whitehouse.gov/wp-content/uploads/2022/10/Biden-Harris-Administrations-National-Security- Strategy-10.2022.pdf (accessed: 15.12.2022). P. 8. Управление конкуренцией: конкурентные стратегии и программирующее лидерство Байдена министр оборо ны США Л.Остин, находясь с визитом в Син гапуре в июле 2021 г.65, заявил о нежелании принуждать региональных игроков к выбо- ру между США или КНР, впоследствии по- Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 110 ВЛАДИМИР ЩЕРБУНОВ опасности. Создание военно-технологиче- ского блока AUKUS во многом цементиру- ет конфронтационный характер отноше- ний между Китаем и Австралией. с КНР связаны, как отмечалось выше, с не- военной областью. В докладе Атлантиче ского совета 2019 года68 в качестве основ- ных направлений выделялись три: эко номическое, дипломатическое, а также в сфере безопасности. В отчёте Корпора ции РЭНД, посвящённом борьбе за влия- ние в ИТР между Вашингтоном и Пекином, было представлено шесть областей: эконо- мическая, дипломатическая, информацион- ная (охват СМИ), культурная, социальная (people-to-people) и сфера безопасности. В первом докладе в военно-политиче- ском измерении предлагался относительно широкий перечень мер по повышению управляемости конфронтации, суть кото- рых заключалась в сдерживании китайских позиций минимальными усилиями, что контрастировало с выводами относительно слабого влияния КНР в вопросах безопас- ности в АТР. Управление конкуренцией посредством канализирования в теории повышает шан- сы США на успех в конкуренции с Китаем. С помощью выстраивания более благо приятной для себя международной среды и баланса сил, делая поведение конкурента бо- лее предсказуемым, США сужают количе- ство вариантов и вызовов, с которыми они могут столкнуться. Тем не менее успеш- ность предпринимаемых мер зависит от материального потенциала США и их союзников. В условиях ухудшения эконо- мической ситуации для многих «колеблю- щихся» стран встанет вопрос в первую оче- редь о сохранении экономической и соци- альной стабильности. В данной связи вы- бор для малых и средних стран между углу- блением экономических связей с США и с Китаем может сыграть решающую роль, фиксируя ориентацию на ту или иную дер- жаву. Дополнительную устойчивость пози- ции США дают союзники с высоким уров- нем социально-экономического развития вроде Австралии и Японии. Что касается влияния на поведение Китая напрямую, то наибольшая проблема заключается в со- поставимости потенциалов двух стран и возможности Китая отвечать симметрич- но, а также в отсутствии экономических ограничений, стоявших перед СССР. Соединённым Штатам предлагалась стратегия асимметричного сдерживания и сочетания конкуренции и сотрудничества с целью сдерживания Китая в военно- политической и экономической сферах при наименьших издержках, то есть экс- перты указывали на необходимость избе- гать полномасштабного соперничества [Doshi 2021: 313]. Управление конкуренцией: конкурентные стратегии и программирующее лидерство Основные задачи связы- вались с модернизацией потенциала США и их ближайших союзников и партнёров, сдерживанием деструктивного поведения КНР и налаживанием сотрудничества с Пе кином по глобальным проблемам, под ко- торым подразумеваются вопросы климата, окружающей среды, водных ресурсов и здравоохранения [Christensen 2015: 296]69. Схожие идеи «вовлечения и сдерживания» поддерживали и другие авторы70. Вашинг тону не стоило вставать на пути у Китая, 68 От ред.: 29 июля 2019 г. Атлантический совет (Atlantic council) внесён в Перечень иностранных и международных неправительственных организаций, деятельность которых признана нежелатель- ной на территории Российской Федерации (см.: https://minjust.gov.ru/ru/documents/7756/). 69 Американский политолог А.Л. Фридберг в работе 2011 г. высказывал точку зрения, что нали- чие этих проблем не является фактором, который приведёт к сотрудничеству между США и КНР. См.: [Friedberg 2011: 38]. 70 См., например: Heath T.R. U.S. Strategic Competition with China. A RAND Research Primer. [Электронный ресурс] // RAND Corporation. June 2021. URL: https://www.rand.org/pubs/perspectives/ PEA290-3.html (accessed: 15.12.2022). g 70 См., например: Heath T.R. U.S. Strategic Competition with China. A RAND Research Primer. [Электронный ресурс] // RAND Corporation. June 2021. URL: https://www.rand.org/pubs/perspectives/ PEA290-3.html (accessed: 15.12.2022). и международных неправительственных организаций, деятельность которых признана нежелатель- ной на территории Российской Федерации (см.: https://minjust.gov.ru/ru/documents/7756/). 69 Американский политолог А.Л. Фридберг в работе 2011 г. высказывал точку зрения, что нали- чие этих проблем не является фактором, который приведёт к сотрудничеству между США и КНР. См.: [Friedberg 2011: 38]. 70 См., например: Heath T.R. U.S. Strategic Competition with China. A RAND Research Primer. 68 От ред.: 29 июля 2019 г. Атлантический совет (Atlantic council) внесён в Перечень иностранных и международных неправительственных организаций, деятельность которых признана нежелатель- ной на территории Российской Федерации (см.: https://minjust.gov.ru/ru/documents/7756/). ной на территории Российской Федерации (см.: https://minjust.gov.ru/ru/documents/7756/). 69 Американский политолог А.Л. Фридберг в работе 2011 г. высказывал точку зрения, что нали- чие этих проблем не является фактором, который приведёт к сотрудничеству между США и КНР. См.: [Friedberg 2011: 38]. 70 С H h T R U S S C h Ch A RAND R h P Kim P.M. Managing U.S. China Relations in an Era of Peer Competition. [Электронный ресурс] // Сouncil on Foreign Relations. November, 2018. URL: https://www.cfr.org/blog/managing-us-china- relations-era-peer-competition (accessed: 15.12.2022). 72 Kramer F.D. Managed Competition: Meeting China’s Challenge in a Multi-vector World. [Элек тронный ресурс]. Atlantic Council. December 2019. URL: https://www.atlanticcouncil.org/wp-content/ uploads/2019/12/Meeting-Chinas-Challenges-Report-WEB.pdf (accessed: 25.12.2021); 73 Например: «С 2012 года Китай провел более 100 совместных учений и тренировок с более чем 30 странами… [и] направил более 1700 военнослужащих для обучения в более чем 50 стран. Более 20 китайских военных учебных заведений наладили и поддерживают межвузовские обмены со своими коллегами из более чем 40 стран. Между тем более 10 000 иностранных военнослужа- щих из более чем 130 стран прошли обучение в военных университетах Китая». В 2018 г. Китай принимал Китайско-африканский форум по обороне и безопасности, Форум высокого уровня по обороне КНР и Латинской Америки и Форум для высокопоставленных должностных лиц мини стерства обороны из стран Карибского бассейна и южной части Тихого океана. China’s National Defense in a New Era. [Электронный ресурс] // The State Council Information Office of the People’s Republic of China, July 2019: P. 31–32. URL: https://english.www.gov.cn/archive/ whitepaper/201907/24/content_WS5d3941ddc6d08408f502283d.html (accessed: 15.12.2022). 74 Mazarr M.J. et al. Understanding Influence in the Strategic Competition with China. [Электронный ресурс] // RAND Corporation, 2021. URL: https://www.rand.org/content/dam/rand/pubs/research_ reports/RRA200/RRA290-1/RAND_RRA290-1.pdf (accessed: 15.12.2022). 71 Kim P.M. Managing U.S.-China Relations in an Era of Peer Competition. [Электронный ресурс] // Сouncil on Foreign Relations. November, 2018. URL: https://www.cfr.org/blog/managing-us-china- relations-era-peer-competition (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 Достижение успеха в конкуренции: курс на сдерживание Китая и использование американских преимуществ О С Основные элементы стратегии США по обеспечению управляемой конкуренции Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 111 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ ров. Специалисты РЭНД также указывали на дефицит опыта военного сотрудниче- ства с третьими странами у КНР. Связи Китая были сфокусированы на нескольких традиционных партнёрах, например на Камбодже и Африканском союзе. Объём военной помощи последнему со стороны КНР за 2013–2018 годы совокупно соста- вил около 120 млн долларов74. С точки зрения проведения международных уче- ний Китай начал приближаться к амери- канским показателям, но их охват и пред- ставительность оставались относительно низкими. превращая его во врага. Предпочтительным было бы победить его в экономической конкуренции, используя свои сравнитель- ные преимущества [Hаss 2021]. Оптималь ный вариант взаимодействия с КНР требо- вал не прямого сдерживания, сопряжённо- го с высокими издержками, а параллельно- го существования и конкуренции с опорой на сильные стороны американского поло- жения, главным образом на союзников и партнёров71. Стратегия, разработанная американ- скими экспертами, строилась вокруг отве- та на вызывавшие американские опасения шаги Китая72: кибершпионаж, агрессив- ные действия в Восточном и Южно-Ки тайском морях и военную дипломатию в двух её компонентах: 1) контакты пред- ставителей НОАК с военнослужащими военных ведомств других стран (военные обмены, назначение военных атташе, соз- дание механизмов для консультаций)73; 2) обучение и подготовка вооружённых сил других стран. Тем не менее в докладе РЭНД 2021 г. отмечались относительно слабые позиции Пекина в этой сфере по сравнению с Вашингтоном и его союзни- ками. В качестве оценки используется объём военной помощи и безвозмездных поставок вооружений на 2018 год: США – 13 млрд долларов, их союзники – 7 млрд долларов, Китай – около 100 млн долла- В ответ на действия Пекина американ- ские специалисты предлагали сделать упор на обеспечение гарантий безопасно- сти союзникам и на сдерживание КНР как в гибридной сфере, так и в конвенцио- нальной. В первом случае предлагалось усиление мер в киберпространстве (повы- шение устойчивости, создание Между народного совета по кибербезопасности из государств-единомышленников, уста- новление ограничений на использование китайских технологий), а также противо- действие применению силы «на низком уровне» в ВКМ и ЮКМ (дипломатическая поддержка, обеспечение свободы судо ходства и расширение морских инициа- тив для предоставления разведыватель- ной информации, обучение и тренировка 74 Mazarr M.J. et al. Understanding Influence in the Strategic Competition with China. [Электронный ресурс] // RAND Corporation, 2021. URL: https://www.rand.org/content/dam/rand/pubs/research_ reports/RRA200/RRA290-1/RAND_RRA290-1.pdf (accessed: 15.12.2022). Mеждународные процессы. Том 20. Номер 4 (71). 81 Talisman Sabre 2019, Largest Ever Bilateral Defense Exercise in Australia Opens. [Электронный ресурс] // U.S. Indo-Pacific Command. 9 July 2019. URL: https://www.pacom.mil/Media/News/News- Article-View/Article/1899433/talisman-sabre-2019-largest-ever-bilateral-defense-exercise-in-australia- opens/ 80 Greene A. US marines to join Australian warships deployed in the Pacific as anxiety over China grows. [Электронный ресурс] // ABS News Australia. 4 May 2018. URL https://www.abc.net.au/ news/2018-05-04/us-marines-join-australian-warship-pacific-china-anxiety-grows/9729652 (accessed: 15.12.2022). 75 Kramer F.D. Managed Competition: Meeting China’s Challenge in a Multi-vector World. [Электронный ресурс] // Atlantic Council. December 2019. URL: https://www.atlanticcouncil.org/ wp-content/uploads/2019/12/Meeting-Chinas-Challenges-Report-WEB.pdf (accessed: 15.12.2022). 76 McDevitt M. The South China Sea: Assessing U.S. Policy and Options for the Future Arlington. [Электронный ресурс] // CNA Corporation, November 2014. URL: https://www.cna.org/cna_files/pdf/ IOP-2014-U-009109.pdf (accessed: 15.12.2022). 77 Indo-Pacific Endeavour 2018 returns to Australia. [Электронный ресурс] // Navaltoday.com. 23 August 2018. URL: https://www.navaltoday.com/2018/08/23/indo-pacific-endeavour-2018-returns- to-australia/ (accessed: 15.12.2022). 78 Strating B. A sea ride with Australia’s Indo-Pacific Endeavour. [Электронный ресурс] // Lowy Institute. 6 January 2019. URL: https://www.lowyinstitute.org/the-interpreter/sea-ride-australia-s-indo- pacific-endeavour (accessed: 15.12.2022). 79 Indo-Pacific Endeavour returns to Australia. [Электронный ресурс] // Australian Government. Defense. 2 December 2022. URL: https://www.defence.gov.au/news-events/releases/2022-12-02/indo- pacific-endeavour-returns-australia (accessed: 15.12.2022). 80 Greene A. US marines to join Australian warships deployed in the Pacific as anxiety over China grows. [Электронный ресурс] // ABS News Australia. 4 May 2018. URL https://www.abc.net.au/ news/2018-05-04/us-marines-join-australian-warship-pacific-china-anxiety-grows/9729652 (accessed: 15.12.2022). 81 Talisman Sabre 2019, Largest Ever Bilateral Defense Exercise in Australia Opens. [Электронный ресурс] // U.S. Indo-Pacific Command. 9 July 2019. URL: https://www.pacom.mil/Media/News/News- Article-View/Article/1899433/talisman-sabre-2019-largest-ever-bilateral-defense-exercise-in-australia- opens/ p 79 Indo-Pacific Endeavour returns to Australia. [Электронный ресурс] // Australian Government. Defense. 2 December 2022. URL: https://www.defence.gov.au/news-events/releases/2022-12-02/indo- pacific-endeavour-returns-australia (accessed: 15.12.2022). 75 Kramer F.D. Managed Competition: Meeting China’s Challenge in a Multi-vector World. [Электронный ресурс] // Atlantic Council. December 2019. URL: https://www.atlanticcouncil.org/ wp-content/uploads/2019/12/Meeting-Chinas-Challenges-Report-WEB.pdf (accessed: 15.12.2022). 78 Strating B. A sea ride with Australia’s Indo-Pacific Endeavour. [Электронный ресурс] // Lowy Institute. 6 January 2019. URL: https://www.lowyinstitute.org/the-interpreter/sea-ride-australia-s-indo- pacific-endeavour (accessed: 15.12.2022). 75 Kramer F.D. Managed Competition: Meeting China’s Challenge in a Multi-vector World. [Электронный ресурс] // Atlantic Council. December 2019. URL: https://www.atlanticcouncil.org/ wp-content/uploads/2019/12/Meeting-Chinas-Challenges-Report-WEB.pdf (accessed: 15.12.2022). 76 McDevitt M. The South China Sea: Assessing U.S. Policy and Options for the Future Arlington. [Электронный ресурс] // CNA Corporation, November 2014. URL: https://www.cna.org/cna_files/pdf/ IOP-2014-U-009109.pdf (accessed: 15.12.2022). 77 Indo-Pacific Endeavour 2018 returns to Australia. [Электронный ресурс] // Navaltoday.com. 23 August 2018. URL: https://www.navaltoday.com/2018/08/23/indo-pacific-endeavour-2018-returns- to-australia/ (accessed: 15.12.2022). 78 Strating B. A sea ride with Australia’s Indo-Pacific Endeavour. [Электронный ресурс] // Lowy Institute. 6 January 2019. URL: https://www.lowyinstitute.org/the-interpreter/sea-ride-australia-s-indo- pacific-endeavour (accessed: 15.12.2022). 79 Indo-Pacific Endeavour returns to Australia [Электронный ресурс] // Australian Government [Электронный ресурс] // CNA Corporation, November 2014. URL: https://www.cna.org/cna_files/pdf/ IOP-2014-U-009109.pdf (accessed: 15.12.2022). 77 Indo-Pacific Endeavour 2018 returns to Australia. [Электронный ресурс] // Navaltoday.com. 23 August 2018. URL: https://www.navaltoday.com/2018/08/23/indo-pacific-endeavour-2018-returns- to-australia/ (accessed: 15.12.2022). 78 Strating B. A sea ride with Australia’s Indo-Pacific Endeavour. [Электронный ресурс] // Lowy Institute. 6 January 2019. URL: https://www.lowyinstitute.org/the-interpreter/sea-ride-australia-s-indo- pacific-endeavour (accessed: 15.12.2022). 79 Indo-Pacific Endeavour returns to Australia. [Электронный ресурс] // Australian Government. Defense. 2 December 2022. URL: https://www.defence.gov.au/news-events/releases/2022-12-02/indo- 23 August 2018. URL: https://www.navaltoday.com/2018/08/23/indo pacific endeavour 2018 returns to-australia/ (accessed: 15.12.2022). 78 Strating B. A sea ride with Australia’s Indo-Pacific Endeavour. [Электронный ресурс] // Lowy Institute. 6 January 2019. URL: https://www.lowyinstitute.org/the-interpreter/sea-ride-australia-s-indo- pacific-endeavour (accessed: 15.12.2022). 79 Indo Pacific Endeavour returns to Australia [Электронный ресурс] // Australian Government p 77 Indo-Pacific Endeavour 2018 returns to Australia. [Электронный ресурс] // Navaltoday.com. 23 August 2018. URL: https://www.navaltoday.com/2018/08/23/indo-pacific-endeavour-2018-returns- to-australia/ (accessed: 15.12.2022). 78 [ ] Достижение успеха в конкуренции: курс на сдерживание Китая и использование американских преимуществ О С Октябрь–декабрь / 2022 112 ВЛАДИМИР ЩЕРБУНОВ вым службам действовать на основании принципов операционной совместимости, оспаривая территориальные претензии КНР. В этой связи нельзя не вспомнить об усилиях Вашингтона по наращиванию числа участников учений, организуемых как самими США, так и их союзниками. Например, с 2017 г. австралийские ВМС проводят учения Пацифик Индэвор (Pacific Endeavor)77, ориентированные на вовле чение стран АСЕАН и Океании, в рамках которых австралийские военные экспер- ты выступали проводниками западных, а фактически американских ценностей78. В 2022 г. учения включали взаимодействие с вооружёнными силами 14 стран Индо- Тихоокеанского региона79. В июле 2018 г. в составе тактической группы австралий ских вооружённых сил принимали участие морские пехотинцы США, однако данная практика не получила распространения80. Параллельно с этим происходило углубле- ние взаимодействия с привлечением новых стран к существующим форматам учений. В частности, в 2019 г. Канада, Новая Зеландия и Япония начали принимать уча- стие в учениях Талисман Сейбр (Talisman Saber), которые были впервые организова- ны Соединёнными Штатами и Австралией в начале 2000-х годов81. вооружённых сил заинтересованных стран)75. Проводимые Китаем операции в «серых зонах» или применение силы на низком уровне, названные американ- ским адмиралом М. МакДевиттом поли- тикой «нарезания салями», представля- ли собой подход, основанный на малых шагах, исключающих войну. Вместе с тем в совокупности они обеспечивали стра тегический результат, как произошло с милитаризацией и де-факто установле- нием Китаем контроля над ЮКМ76. Ряд авторов полагали, что именно выработка скоординированного подхода в духе «сдер живания посредством наказания» совмест- но с другими странами региона, которых затрагивают упомянутые операции КНР, должна стать одной из основ предотвра- щения военной конфронтации [Wright 2021: 176–177; Hass 2021: 144; McKinney, Harris 2021: 24]. В частности, Р. Хасс от- верг идею абсолютного военного сдержи- вания, активно выступая за использова- ние мер невоенного характера (экономи- ческих санкций) в ответ на действия Пекина [Hаss 2021: 131]. Одной из ключевых целей США, по мнению экспертов, должно стать создание плотной сети интегрированных парт нёрств, которые позволят флотам и берего- Mеждународные процессы. Том 20. Номер 4 (71). 82 Kramer F.D. Managed Competition: Meeting China’s Challenge in a Multi-vector World. [Электронный ресурс] // Atlantic Council. December 2019. URL: https://www.atlanticcouncil.org/ wp-content/uploads/2019/12/Meeting-Chinas-Challenges-Report-WEB.pdf (accessed: 15.12.2021); 83 Allison G., Glick-Unterman J. The Great Military Rivalry: China vs the U.S. [Электронный ресурс] // Belfer Center for Science and International Affairs, Harvard Kennedy School. 16 December 2021. P.10. URL: https://www.belfercenter.org/sites/default/files/GreatMilitaryRivalry_ChinavsUS_211215.pdf (accessed: 15.12.2022). Достижение успеха в конкуренции: курс на сдерживание Китая и использование американских преимуществ О С Октябрь–декабрь / 2022 113 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ В области конвенциональных вооруже- ний предлагалось сделать акцент на предо- ставлении союзникам платформ для созда- ния зон ограничения доступа вооружённых сил оппонентов, проведении дополнитель- ной модернизации военного потенциала (упор на противокорабельные ракеты уве- личенной дальности, платформы по проти- водействию системам управления войска- ми, сенсоры для обнаружения беспилот- ных подводных и воздушных объектов), а также ускорение развития передовых тех- нологий военной направленности (гипер звуковые ударные системы, искусственный интеллект, квантовые вычисления, робото- техника)82. Подобный перечень мер фикси- ровал ориентацию Китая на вложение в средства обороны и ограничения доступа в рамках первой и второй островной гряды, но последние оценки существующего по- ложения дел и физических ограничений технологий ограничения доступа говорят о том, что США утратят [Biddle, Oelrich 2016] либо уже утратили возможность сдерживать КНР в западной части Тихого океана83. средних игроков Индо-Тихоокеанского ре- гиона различными вооружениями может вместо сдерживания иметь обратный эф- фект, повышая воинственность стран по отношению к Китаю, а значит, и вероят- ность столкновения двух держав. * * * Отстаиваемая США концепция «управ- ляемой конкуренции» представляет собой альтернативу фронтальному соперниче- ству, существовавшему в годы «холодной войны». В условиях экономической взаи- мозависимости двух оппонентов и невоз- можности сдерживать Китай в одиночку подход, основанный на сочетании конку- ренции и сотрудничества, представляется Вашингтону способом направлять борьбу за будущее наиболее выгодного для себя мирового порядка. Доктринально оформ- ленная к 2022 г. концепция «управляемой конкуренции» включает три компонента: 1) регулирование конкуренции посред- ством институционализации отношений с целью повышения их предсказуемости, усиления мер кризисного регулирования и снижения конфликтогенного потенциала; 2) опережающее программирование дей- ствий международных субъектов путём создания стимулов и сдерживающих фак- торов для принятия внешнеполитических решений; 3) описание мер, необходимых для достижения успеха в конкуренции. В работах американских аналитиков наи- более широко представлен именно третий подход с предложениями по асимметрич- ному сдерживанию. Управление конкуренцией посредством совместного решения проблем (то есть ограниченная конкуренция) страдает от той же слабости, что и остальные подходы – не- адекватного учёта ответных мер усиливаю- щегося Китая. С одной стороны, при усло- вии успешной реализации США курса на антикитайское сдерживание Пекин может начать воспринимать двусторонние отно- шения как «игру с нулевой суммой», где любая форма сотрудничества будет воспри- ниматься как уступка сопернику. Кроме этого, даже в условиях меньших успехов американской стороны Китай может на- чать обусловливать участие в инициативах США уступками по тем или иным вопро- сам. Список литературы Баталов Э.Я. Мировое развитие и мировой порядок. М.: РОССПЭН, 2005. 376 с. Б Э Я Н й // М Я. Мировое развитие и мировой порядок. М.: РОССПЭН, 2005. 376 с. Баталов Э.Я. Мировое развитие и мировой порядок. М.: РОССПЭН, 2005. 376 с. Баталов Э.Я. Новая институционализация мировой политики // Международные процессы. 2016. Т. 14. №. 1. С. 6–25. DOI 10.17994/IT.2016.14.1.44.1. Баталов Э.Я. Новая институционализация мировой политики // Международные процессы. 2016. Т. 14. №. 1. С. 6–25. DOI 10.17994/IT.2016.14.1.44.1. Батюк В.И. Советско-американские и российско-американские двусторонние режимы (1945– 2005 гг.): Дисс. … д-ра полит. наук. М.: Ин-т Соединенных Штатов Америки и Канады РАН, 2008. 530 с. Богатуров А.Д. Великие державы на Тихом океане. История и теория международных отношений в Восточной Азии после Второй мировой войны (1945–1995). М.: Конверт. МОНФ, 1997. 352 с. Истомин И.А. Механизм научно-аналитического обеспечения внешнеполитического процесса в США // Вестник МГИМО Университета 2009 №6 С 97 110 DOI 10 24833/2071 8160 Богатуров А.Д. Великие державы на Тихом океане. История и теория международных отношений в Восточной Азии после Второй мировой войны (1945–1995). М.: Конверт. МОНФ, 1997. 352 с. И И А М ур Д д р р р ду р д в Восточной Азии после Второй мировой войны (1945–1995). М.: Конверт. МОНФ, 1997. 352 с. Истомин И.А. Механизм научно-аналитического обеспечения внешнеполитического процесса в США // Вестник МГИМО-Университета. 2009. №. 6. С. 97–110. DOI 10.24833/2071-8160- 2009-6-9-97-110. Истомин И.А. Механизм научно-аналитического обеспечения внешнеполитического процесса в США // Вестник МГИМО-Университета. 2009. №. 6. С. 97–110. DOI 10.24833/2071-8160- 2009-6-9-97-110. Истомин И.А. Управление обязательствами в асимметричных альянсах // Международные процессы. 2021. Т. 19. №. 1. С. 26–55. DOI 10.17994/IT.2021.19.1.64.1. Торкунов А.В. Стратегия администрации Трампа в Азиатско-Тихоокеанском регионе // Мировая эко- номика и международные отношения. 2019. Т. 63. № 6. С. 25–37. DOI: 10.20542/0131-2227- 2019-63-6-25-37. Троицкий М.А. Концепция «программирующего лидерства» в евроатлантической стратегии США // Pro et Contra. 2002. № 4. С. 86–103. Троицкий М.А. Трансатлантический союз. 1991–2004. Трансформация системы американо-европей- ского партнёрства после распада биполярности. М.: НОФМО, 2004. 252 с. Худайкулова А.В., Рамич М.С. «КВАД 2.0»: Четырёхсторонний диалог для контрбалансирования КНР в Индо-Тихоокеанском регионе // Polis: Journal of Political Studies. 2020. № 3. С. 23–43. DOI: 10.17976/jpps/2020.03.03. jpp Шаклеина Т.А. «Мозговые центры» и их роль в формировании внешней политики США // Введение в прикладной анализ международных ситуаций. М.: Аспект Пресс, 2018. С. 109–132. Ш Т А П США Р // Шаклеина Т.А. Политика США в отношении России: конкуренция, сдерживание и управление // Вестник РГГУ. Достижение успеха в конкуренции: курс на сдерживание Китая и использование американских преимуществ О С С другой стороны, насыщение малых и Тем не менее каждый из элементов управления оказывается уязвимым в пер- вую очередь перед ответными мерами КНР, которая на фундаментальном уровне не заинтересована в предлагаемом США под- ходе. Пекин исходит из того, что с каж- Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 114 ВЛАДИМИР ЩЕРБУНОВ Нет оснований полагать, что Китай со сво- ей стороны не будет принимать меры в духе конкурентных стратегий, вынуждающие Соединённые Штаты к реакции. дым годом его позиции в конкуренции с Вашингтоном укрепляются, в результате чего его переговорные возможности по бу- дущему мирового порядка усиливаются. Список литературы Серия: Политология. История. Международные отношения. 2020. №. 4. С.10–26. https://doi.org/10.28995/2073-6339-2020-4-10-26. Шаклеина Т.А. Новый этап в формировании мирового порядка. Планы США по управлению междер- жавной конкуренцией // Международные процессы. 2021. Т. 19. № 3. С. 6–22. DOI 10.17994/ IT.2021.19.3.66.6. Щербунов В.О. Военно-политические отношения Австралии и США в условиях роста влияния Китая в АТР // Вестник РГГУ. Серия: Политология. История. Международные отношения. 2021. № 4. С. 32–46 DOI: 10.28995/2073-6339-2021-4-32-46. Adler E., Haas P.M. Conclusion: epistemic communities, world order, and the creation of a reflective research program // International organization. 1992. Vol. 46. No. 1. P. 367–390. DOI: 10.1017/ S0020818300001533. g Rising Powers: Past, Present, and Future / ed. by T. V. Paul. Cambridge University 6. 335 p. Accommodating Rising Powers: Past, Present, and Future / ed. by T. V. Paul. Cam Press, 2016. 335 p. Biddle S., Oelrich I. Future Warfare in the Western Pacific: Chinese Antiaccess/Area Denial, US AirSea Battle, and Command of the Commons in East Asia // International Security. 2016. Vol. 41. No. 1. P. 7–48. https://doi.org/10.1162/ISEC_a_00249. p g _ _ M., Sullivan J. Competition Without Catastrophe: How American Can Both Challenge a With China // Foreign Affairs. 2019. Vol. 98. No. 5. P. 96–111. T J Th Chi Ch ll Sh i h Ch i f Ri i P WW N & C _ _ Campbell K.M., Sullivan J. Competition Without Catastrophe: How American Can Bo Coexist With China // Foreign Affairs. 2019. Vol. 98. No. 5. P. 96–111. Christensen T.J. The China Challenge: Shaping the Choices of a Rising Power. WW Norton & Company, 2015. 408 p. Diehl P.F., Goertz G. War and Peace in International Rivalry. University of Michigan Press, 2001. 327 p. Doshi R. The Long Game: China's Grand Strategy to Displace American Order. Oxford University Press, Diehl P.F., Goertz G. War and Peace in International Rivalry. University of Michigan Press, 2001. 327 p. Doshi R. The Long Game: China's Grand Strategy to Displace American Order. Oxford University Press, 2021. 432 p. p Drezner D.W. The Ideas Industry. How Pessimists, Partisans, and Plutocrats are Transforming the Marketplace of Ideas. NY: Oxford University Press, 2017. 360 p. E h A C Th H d P l f M d C // H l h Aff 1993 V l 12 Enthoven A.C. The History and Principles of Managed Competition // Health Affairs. 1993. Vol. 12. No. suppl 1. P. 24-48. https://doi.org/10.1377/hlthaff.12.suppl_1.24 Mеждународные процессы. Том 20. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 Список литературы P. 185–205. https://www.jstor.org/stable/2706520. Lampton D.M. Same Bed, Different Dreams: Managing U.S.- China Relations, 1989–2000. University of California Press, 2001. 510 p. Layne C. Rethinking American Grand Strategy: Hegemony or Balance of Power in Century? //World Policy Journal. 1998. Vol. 15. No. 2. P. 8–28. https://www.jstor.org/ Layne C. Rethinking American Grand Strategy: Hegemony or Balance of Power in the Twenty-First Century? //World Policy Journal. 1998. Vol. 15. No. 2. P. 8–28. https://www.jstor.org/stable/40209580 Legro J.W. Whence American Internationalism // International Organization. 2000. Vol. 54. No. 2. P. 253–289. Century? //World Policy Journal. 1998. Vol. 15. No. 2. P. 8 28. https://www.jstor.org/stable/40209580 Legro J.W. Whence American Internationalism // International Organization. 2000. Vol. 54. No. 2. P. 253–289. Levy M. A., Young O. R., Zürn M. The Study of International Regimes // European Journal of International Relations. 1995. Vol. 1. No. 3. P. 267–330. https://doi.org/10.1177/1354066195001003001. Liff A.P., Ikenberry G.J. Racing toward tragedy?: China's Rise, Military Competition in the Asia Pacific, and the Security Dilemma // International Security. 2014. Vol. 39. No. 2. P. 52–91. https://doi. org/10.1162/ISEC_a_00176. g _ _ Lindley L. Promoting Peace through Information: Transparency as a Tool of Security Regimes. Princeton, N.J.: Princeton University Press, 2007. 296 p. McKinney J.M., Harris P. Broken Nest: Deterring China from Invading Taiwan //The US Army War College Quarterly: Parameters. 2021. Vol. 51. No. 4. P. 23–36. O'Hanlon M. The Senkaku Paradox: Risking Great Power War Over Small Stakes. Brookings Institution Press, 2019. 272 p. h d h d b d h l Porter P. Why America's Grand Strategy Has Not Changed: Power, Habit, and the US Foreign Policy Establishment // International Security. 2018. Vol. 42. No. 4. P. 9–46. https://doi.org/10.1162/ isec_a_00311. _ _ Porter P. Advice for a dark age: Managing great power competition //The Washington Quarterly. 2019. Vol. 42. No. 1. P. 7–25. Posen B.R., Ross A.L. Competing Visions for US Grand Strategy // International Securit No. 3. P. 5–53. https://doi.org/10.2307/2539272. ss A.L. Competing Visions for US Grand Strategy // International Security. 1997. Vol. 21. –53. https://doi.org/10.2307/2539272. Rosen S.P. “Competitive Strategies: Theoretical Foundations, Limitations, and Extensions” in Thomas Mahnken, ed., Competitive Strategies for the 21st Century: Theory, History, and Practice. Stanford, CA: Stanford Security Studies, 2012. P. 12–27. https://doi.org/10.1515/9780804783187-005. Saunders, P.C., Bowie, J.G. US–China Military Relations: Competition and Cooperation. // J tegic Studies. 2016. Vol. 39. No. 5–6. P. 662–684. https://doi.org/10.1080/01402390.20 tegic Studies. 2016. Vol. 39. No. 5 6. P. 662 684. https://doi.org/10.1080/01402390.2016. 1221818. Список литературы Номер 4 (71). Октябрь–декабрь / 2022 115 КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ США И КИТАЯ КОНЦЕПЦИИ «УПРАВЛЯЕМОЙ КОНКУРЕНЦИИ» В ВОЕННО-ПОЛИТИЧЕСКИХ ОТНОШЕНИЯХ Friedberg A.L. Ripe for rivalry: prospects for peace in a multipolar Asia // International security. 1993. Vol. 18. No. 3. P. 5–33. https://doi.org/10.2307/2539204. Friedberg A.L. The future of US-China relations: Is conflict inevitable? //International security. 2005. Vol. 30. No. 2. P. 7–45. https://www.jstor.org/stable/4137594 George A.L. Managing US-Soviet Rivalry: Problems of Crisis Prevention. Routledge, 2019. 415 p. Glaser C.L. Rational Theory of International Politics. Princeton University Press, 2010. 328 p. g g g y g , p Glaser C.L. Rational Theory of International Politics. Princeton University Press, 2010. 328 p. Goldstein L J Meeting China Halfway: How to Defuse the Emerging US China Rivalry Georgetown g g g y g p Glaser C.L. Rational Theory of International Politics. Princeton University Press, 2010. 328 p. Goldstein L.J. Meeting China Halfway: How to Defuse the Emerging US-China Rivalr University Press, 2015. 400 p. y p Green M.J. By More Than Providence. Columbia University Press, 2017. 760 p. G M J Hi k K H C Z F d d D f i A i R CSIS D Green M.J., Hicks K.H., Cooper Z. Federated Defense in Asia. Report. CSIS, December 2 H ddi k R J Fi h W Chi A i d h F f h P ifi N l I Haddick R.J. Fire on the Water: China, America, and the Future of the Pacific. Naval Institute Press, 2014. 274 p. Hass R. Stronger: Adapting America’s China Strategy in an Age of Competitive Interdependence. Yale University Press, 2021. 240 p. Hensel P.R. An Evolutionary Approach to the Study of Interstate Rivalry // Conflict Management and Peace Science. 1999. Vol. 17. No. 2. P. 175–206. https://doi.org/10.1177/07388942990170020. Hensel P.R. An Evolutionary Approach to the Study of Interstate Rivalry // Conflict Management and Peace Science. 1999. Vol. 17. No. 2. P. 175–206. https://doi.org/10.1177/07388942990170020. Jervis R. Cooperation Under the Security Dilemma // World politics. 1978. Vol. 30. No. 2. P. 167–214. https://doi.org/10.2307/2009958. Jervis R. Cooperation Under the Security Dilemma // World politics. 1978. Vol. 30. No. 2. P. 167–214. https://doi.org/10.2307/2009958. G h O b S S l 99 26 King G., Keohane R.O., Verba S. Designing Social Inquiry. Princeton University Press, 1994. 264 p. Krasner S.D. Structural causes and regime consequences: regimes as intervening variables // Interna tional organization. 1982. Vol. 36. No. 2. Abstract With the inauguration of President John Biden, the ideas of “managing” competition between the United States and China increasingly began to appear in the lexicon of American politicians and pundits. Most works on US-China competition emphasize its non-military aspects, which does not reflect the efforts that the US puts into competition in the military-political domain, as well as the importance of its regulation in preventing armed conflict between the two powers. This study seeks to solve a dual task: firstly, to identify the ideational content of the concept of "managed competition", and secondly, to analyze the existing formats of managing US-China military-political competition in the Asia-Pacific region in the 21st century. After ana- lyzing the academic and analytical works of recent years, the author concludes that the concept of "managed competition" comprises three components: 1) regulation of competition through institutionalization in order to increase the predictability, strengthen crisis management mechanisms and reduce conflict potential, 2) anticipatory programming of the actions of international actors by creating incentives and deterrents for making foreign policy decisions, 3) describing the measures necessary to achieve success in competition (outcompeting). At a practical level, the management of competition in the military-political relations between the United States and China implies 1) the creation of international institutions to reduce the like- lihood of military incidents in the South China Sea and the Western Pacific, 2) setting the agenda in the Asia-Pacific region (the concept of the "Indo-Pacific region"), the consolidation and expansion of US mili- tary-political institutions (QUAD, AUKUS) with giving them the anti-Chinese focus, 3) US actions that impede the implementation of PRC’s military-political goals: reunification with Taiwan, control over the South China Sea, expansion of military-political contacts with the countries of the Asia-Pacific region. K d Keywords: y managed competition; great power; security; Asia-Pacific region; China; USA. y managed competition; great power; security; Asia-Pacific region; China; USA. y managed competition; great power; security; Asia-Pacific region; China; USA. Список литературы Sil R., Katzenstein P.J. Analytic Eclecticism in the Study of World Politics: Reconfiguring Problems and Mechanisms across Research Traditions // Perspectives on Politics. 2010. Vol. 8. No. 2. P. 411–431. doi:10.1017/S1537592710001179. Simón L., Lanoszka A., Meijer H. Nodal defence: the changing structure of US alliance systems in Europe and East Asia // Journal of Strategic Studies. 2021. Vol. 44. No. 3. P. 360–388. https://doi.org/10. 1080/01402390.2019.1636372. Swaine M.D. et al. China's military & the US-Japan alliance in 2030: A Strategic net assessment. Washington: Carnegie Endowment for International Peace, 2017. 424 p. Thomas J.P., Montgomery E.B. Developing a Strategy for a Long-Term Sino-American Competition // Competitive Strategies for the 21st Century. Stanford University Press, 2012. P. 257–274. https:// doi.org/10.1515/9780804783187-018. Mеждународные процессы. Том 20. Номер 4 (71). Октябрь–декабрь / 2022 116 VLADIMIR SHCHERBUNOV Valeriano B. Becoming Rivals: The Process of Interstate Rivalry Development. Routledge, 2013. 168 p. Walt S.M. The Origins of Alliance. Cornell University Press, 1987. 336 p. g y , p Zimmerling R. Influence and Power: Variations on a Messy Theme. Springer, 2006. Vol. 68. 308 p. Zoellick R. America in the World: A History of US Diplomacy and Foreign Policy. Twevle, 2020. 560 p. THEORETICAL ANALYSIS AND PRACTICAL IMPLEMENTATION VLADIMIR SHCHERBUNOV MGIMO University, Moscow, 119454, Russia Adler E., Haas P.M. (1992) Conclusion: epistemic communities, world order, and the creation of a reflective research program // International organization. Vol. 46. No. 1. P. 367–390. DOI: 10.1017/ S0020818300001533. Batalov E. (2006). Mirovoe razvitie i mirovoj poryadok [World Development and World Order]. Moscow: ROSSPEN. 376 p. Batalov E. (2016). Novaya institucionalizatsiya mirovoj politiki [New Institutionalization of World Politics]. Mezhdunarodnye protsessy. Vol. 14. No. 1. P. 6–25. DOI 10.17994/IT.2016.14.1.44. Adler E., Haas P.M. (1992) Conclusion: epistemic communities, world order, and the creation of a reflective research program // International organization. Vol. 46. No. 1. P. 367–390. DOI: 10.1017/ S0020818300001533. S0020818300001533. Batalov E. (2006). Mirovoe razvitie i mirovoj poryadok [World Development and World Order]. Moscow: ROSSPEN. 376 p. Batalov E. (2016). Novaya institucionalizatsiya mirovoj politiki [New Institutionalization of World Politics]. Mezhdunarodnye protsessy. Vol. 14. No. 1. P. 6–25. DOI 10.17994/IT.2016.14.1.44. Batalov E. (2016). Novaya institucionalizatsiya mirovoj politiki [New Institutionalization of World Politics]. Mezhdunarodnye protsessy. Vol. 14. No. 1. P. 6–25. DOI 10.17994/IT.2016.14.1.44. References (1993). The History and Principles of Managed Competition. Health Af No. suppl 1. P. 24–48. https://doi.org/10.1377/hlthaff.12.suppl_1.24. pp p g pp _ Friedberg A.L. (1993). Ripe for rivalry: prospects for peace in a multipolar Asia. International security. Vol. 18. No. 3. P. 5–33. https://doi.org/10.2307/2539204. Friedberg A.L. (2005). The future of US-China relations: Is conflict inevitable? International security. Vol. 30. No. 2. P. 7–45. https://www.jstor.org/stable/413759. G (20 9) S S l bl f l d George A.L. (2019). Managing US-Soviet rivalry: Problems of crisis prevention. Routledge. 415 p. Glaser C.L. (2010). Rational Theory of International Politics. Princeton University Press. 328 p. George A.L. (2019). Managing US Soviet rivalry: Problems of crisis prevention. Routledge. 415 p. Glaser C.L. (2010). Rational Theory of International Politics. Princeton University Press. 328 p. Goldstein L.J. (2015). Meeting China Halfway: How to Defuse the Emerging US-China Rivalry. Georgetown University Press. 400 p. g y p Green M.J. (2017). By More Than Providence. Columbia University Press. 760 p. G M J Hi k K H C Z (2014) F d d D f i A i R CSIS Green M.J., Hicks K.H., Cooper Z. (2014). Federated Defense in Asia. Report. CSIS. 42 H dd k R J (2014) F h W Ch A d h F f h P f Haddick R. J. (2014). Fire on the Water: China, America, and the Future of the Pacific. Naval Institute Press. 274 p. Hass R. (2021). Stronger: Adapting America’s China Strategy in an Age of Competitive Interdepen dence. Yale University Press. 240 p. Hensel P.R. (1999). An Evolutionary Approach to the Study of Interstate Rivalry. Conflict Management and Peace Science. Vol. 17. No. 2. P. 175-206. https://doi.org/10.1177/07388942990170020. Istomin I. (2009). Mekhanizm nauchno-analiticheskogo obespecheniya vneshnepoliticheskogo processa v SSHA. [Analytical and scientific support mechanism of the US foreign policy process.]. Vestnik MGIMO-Universiteta. Vol. 6. No. 9. P. 97–110. [ Istomin I. (2021). Upravlenie obyazatel'stvami v asimmetrichnyh al'yansakh [Management of Security Commitments in Asymmetric Alliances]. Mezhdunarodnye protsessy. Vol. 19. No. 1. P. 26–55. DOI: 10.17994/IT.2021.19.1.64.1. Jervis R. (1978). Cooperation Under the Security Dilemma. World politics. Vol. 30. No. 2. P. 167–214. https://doi.org/10.2307/2009958. Khudaykulova A.V. and Ramich M.S. (2020). "KVAD 2.0": Chetyrekhstoronnij dialog dlya kontrbalansirovaniya KNR v Indo-Tihookeanskom regione [“Quad 2.0”: Quadrilateral Dialogue for Counterbalancing China in the Indo-Pacific]. Polis. Political Studies. No. 3. P. 23–43. DOI: 10.17976/jpps/2020.03.03. References Adler E., Haas P.M. (1992) Conclusion: epistemic communities, world order, and the creation of a reflective research program // International organization. Vol. 46. No. 1. P. 367–390. DOI: 10.1017/ S0020818300001533. Batalov E. (2006). Mirovoe razvitie i mirovoj poryadok [World Development and World Order]. Moscow: ROSSPEN. 376 p. [ ] Batalov E. (2016). Novaya institucionalizatsiya mirovoj politiki [New Institutionalization of World Politics]. Mezhdunarodnye protsessy. Vol. 14. No. 1. P. 6–25. DOI 10.17994/IT.2016.14.1.44. International Trends. Volume 20. No. 4 (71). October–December / 2022 117 THE “MANAGED COMPETITION” CONCEPT IN THE US-CHINA SECURITY NEXUS THE “MANAGED COMPETITION” CONCEPT IN THE US-CHINA SECURITY NEXUS Batyuk V. (2005). Sovetsko-amerikanskie i rossijsko-amerikanskie dvustoronnie rezhimy (1945– 2005 gg.) [Soviet-American and Russian-American Bilateral Regimes]. PHD thesis. Moscow: In-t Soedinennyh shtatov Ameriki i Kanady RAS, 2008. 580 p. y y p Biddle S., Oelrich I. (2016). Future Warfare in the Western Pacific: Chinese Antiaccess/Area Denial, US AirSea Battle, and Command of the Commons in East Asia. International Security. 2016. Vol. 41. No. 1. P. 7–48. https://doi.org/10.1162/ISEC_a_00249. p g _ _ Bogaturov A. (1997). Velikiye derzhavi na Tikhom okeane. Istoria i teoriya mezhdunarodnykh otnosheniy v Vostochnoy Azii (1945–1995) [Great Powers in the Pacific. History and Theory of International Relations in East Asia (1945–1995)]. Moscow: Konvert – MONF. 352 p. p Campbell K.M., Sullivan J. (2019). Competition without Catastrophe: How American Can Both Challenge and Coexist with China. Foreign Affairs. Vol. 98. No. 5. P. 96–111. h h h h ll h h h f g J. (2015). The China challenge: Shaping the choices of a rising power. WW Norton & 08 p. Christensen T. J. (2015). The China challenge: Shaping the choices of a rising power. WW Norton & Company. 408 p. Diehl P F Goertz G (2001) War and Peace in International Rivalry University of Michigan Press 327 p p y p Diehl P.F., Goertz G. (2001). War and Peace in International Rivalry. University of Michigan Press. 327 p. Doshi R. (2021). The Long Game: China's Grand Strategy to Displace American Order. Oxford University Press. 432 p. y y g p Doshi R. (2021). The Long Game: China's Grand Strategy to Displace American Order. Oxford University Press. 432 p. Drezner D.W. (2017). The Ideas Industry. How Pessimists, Partisans, and Plutocrats are Transforming the Marketplace of Ideas. NY: Oxford University Press. 360 p. (1993). The History and Principles of Managed Competition. Health Affairs. Vol. 12. P. 24–48. https://doi.org/10.1377/hlthaff.12.suppl_1.24. Enthoven A.C. References M., Harris P. (2021). Broken Nest: Deterring China from Invading Taiwan // The US Army War College Quarterly: Parameters. Vol. 51. No. 4. P. 23–36. O'Hanlon M. (2019). The Senkaku Paradox: Risking Great Power War Over Small Stakes. Brookings Institution Press. 272 p. p (2016). Accommodating Rising Powers: Past, Present, and Future. Cambridge University p. Paul T. V. (ed.). (2016). Accommodating Rising Powers: Past, Present, and Future. Cambr Press. 335 p. p Porter P. (2018). Why America's Grand Strategy Has Not Changed: Power, Habit, and the US Foreign Policy Establishment. International Security. Vol. 42. No. 4. P. 9–46. https://doi.org/10.1162/isec_a_0031 Establishment. International Security. Vol. 42. No. 4. P. 9–46. https://doi.org/10.1162/isec_a_0031 Porter P. (2019) Advice for a dark age: Managing great power competition // The Washington Quarterly. Vol. 42. No. 1. P. 7–25. Posen B.R., Ross A.L. (1997). Competing Visions for US Grand Strategy. International Security. Vol. 21. No. 3. P. 5–53. https://doi.org/10.2307/2539272. Rosen S.P. (2012). Competitive Strategies: Theoretical Foundations, Limits, and Extensions. In Competitive Strategies for the 21st Century. Stanford University Press. P. 12–27. https://doi. org/10.1515/9780804783187-005. g Saunders P.C., Bowie, J.G. (2016). US–China Military Relations: Competition and Cooperation. Journal of Strategic Studies. Vol. 39 No. 5–6. P. 662–684. https://doi.org/10.1080/01402390.2016.122 1818. Shakleina T. (2018). “Mozgovye centry” i ih rol' v formirovanii vneshnej politiki SSHA [Think Tanks” and their role in shaping US foreign policy]. Vvedenie v prikladnoj analiz mezhdunarodnyh situacij. 2nd edition. Moscow: Aspekt Press. P.109–133. Shakleina T. (2020). Politika SSHA v otnoshenii Rossii: konkurentsiya, sderzhivanie i upravlenie [American Policy Towards Russia: Competition, Deterrence and Governing Control]. Vestnik RGGU. Seriya: Politologiya. Istoriya. Mezhdunarodnye otnosheniya. No. 4. P.10–26. https://doi.org/10.28995/2073- 6339-2020-4-10-26. Shakleina T.A. (2021). Novyj etap v formirovanii mirovogo poryadka. Plany SSHA po upravleniyu mezhderzhavnoj konkurenciej [New Stage in the Formation of Global Order. US Plans on Management of Great Power Rivalries]. Mezhdunarodnye protsessy. Vol 19. No. 3. P. 6–22. DOI: 10.17994/ IT.2021.19.3.66.6. Shcherbunov V.O. (2021). Voenno-politicheskie otnosheniya Avstralii i SSHA v usloviyah rosta vliyaniya Kitaya v ATR [Military-Political Relations Between Australia and the United States Amid the Growing Influence of China in the Asia-Pacific]. Vestnik RGGU. Seriya: Politologiya. Istoriya. Mezhdunarodnye otnosheniya. No. 4. P. 32–46. DOI: 10.28995/2073-6339-2021-4-32-46. Sil R., Katzenstein P.J. (2010). Analytic Eclecticism in the Study of World Politics: Reconfiguring Problems and Mechanisms across Research Traditions. Perspectives on Politics. Vol. 8. No. 2. P. 411–431. doi:10.1017/S1537592710001179. Simón L., Lanoszka A., Meijer H. (2021). References Ki G K h R O V b S (1994) D i i S i l I i P i U i i P 264 jpp King G., Keohane R. O., Verba S. (1994). Designing Social Inquiry. Princeton University Press. 264 p. Krasner S D (1982) Structural causes and regime consequences: regimes as intervening variables King G., Keohane R. O., Verba S. (1994). Designing Social Inquiry. Princeton University Press. 264 p. Krasner S.D. (1982). Structural causes and regime consequences: regimes as intervening variables. International organization. Vol. 36. No. 2. P. 185–205. https://www.jstor.org/stable/2706520. Lampton D.M. (2001). Same Bed, Different Dreams: Managing U.S.-China Relations, 1989–2000. University of California Press. 510 p. International organization. Vol. 36. No. 2. P. 185–205. https://www.jstor.org/stable/2706520. Lampton D.M. (2001). Same Bed, Different Dreams: Managing U.S.-China Relations, 1989–2000. University of California Press. 510 p. C ( 998) h k G d S l f h Layne C. (1998). Rethinking American Grand Strategy: Hegemony or Balance of Power in the Twenty- First Century? World Policy Journal. Vol. 15. No. 2. P. 8–28. https://www.jstor.org/stable/40209580 Legro J.W. (2000). Whence American Internationalism // International Organization. Vol. 54. No. 2. P. 253–289. https://www.jstor.org/stable/2601298. p j g Levy M.A., Young O.R., Zürn M. (1995). The Study of International Regimes. European journal of international relations. Vol. 1. No. 3. P. 267–330. https://doi.org/10.1177/1354066195001003001. Liff A.P., Ikenberry G.J. (2014). Racing toward tragedy?: China's Rise, Military Competition in the Asia P ifi d th S it Dil I t ti l S it V l 39 N 2 P 52 91 htt //d i Levy M.A., Young O.R., Zürn M. (1995). The Study of International Regimes. European journal of international relations. Vol. 1. No. 3. P. 267–330. https://doi.org/10.1177/1354066195001003001. Liff A.P., Ikenberry G.J. (2014). Racing toward tragedy?: China's Rise, Military Competition in the Asia Pacific, and the Security Dilemma. International Security. Vol. 39. No. 2. P. 52–91. https://doi. org/10.1162/ISEC_a_00176. Levy M.A., Young O.R., Zürn M. (1995). The Study of International Regimes. European journal of international relations. Vol. 1. No. 3. P. 267–330. https://doi.org/10.1177/1354066195001003001. Liff A.P., Ikenberry G.J. (2014). Racing toward tragedy?: China's Rise, Military Competition in the Asia Pacific, and the Security Dilemma. International Security. Vol. 39. No. 2. P. 52–91. https://doi. org/10.1162/ISEC_a_00176. International Trends. Volume 20. No. 4 (71). October–December / 2022 118 VLADIMIR SHCHERBUNOV Lindley L. (2007). Promoting Peace through Information: Transparency as a Tool of Security Regimes. Princeton, N.J.: Princeton University Press. 296 p. McKinney J. International Trends. Volume 20. No. 4 (71). October–December / 2022 International Trends. Volume 20. No. 4 (71). October–December / 2022 References Nodal defence: the changing structure of US alliance systems in Europe and East Asia. Journal of Strategic Studies. Vol. 44. No. 3. P. 360–388. https://doi.org/ 10.1080/01402390.2019.1636372. Swaine M. D. et al. (2017). China's military & the US-Japan alliance in 2030: A Strategic net assessment. Washington: Carnegie Endowment for International Peace. 424 p. Thomas J. P., Montgomery E. B. (2012). Developing a Strategy for a Long-Term Competition in Competitive Strategies for the 21st Century. Stanford University Press Thomas J. P., Montgomery E. B. (2012). Developing a Strategy for a Long-Term Sino-American Competition in Competitive Strategies for the 21st Century. Stanford University Press. P. 257–274. Torkunov A.V. (2019). Strategiya administracii Trampa v Aziatsko-Tihookeanskom regione [Strategy of the Trump Administration in the Asia-Pacific region]. World Economy and International Relations. Vol. 63. No. 6. P. 25–37. DOI: 10.20542/0131-2227-2019-63-6-25-37. p p g y y Torkunov A.V. (2019). Strategiya administracii Trampa v Aziatsko-Tihookeanskom regione [Strategy of the Trump Administration in the Asia-Pacific region]. World Economy and International Relations. Vol. 63. No. 6. P. 25–37. DOI: 10.20542/0131-2227-2019-63-6-25-37. Troitskiy M.A. (2002). Kontseptsiya “programmiruschego liderstva” v evroatlanticheskoy strategii SSHA [The “Programming Leadership” Concept in the Euro-Atlantic Strategy of the United States]. Pro et Contra. Vol. 7. No. 4. P. 86–103. Troitskiy M.A. (2004). Transatlanticheskiy soyuz. 1991-2004. Transformatsia sistemi amerikanoevro peiskogo partnerstva posle raspada bipoliarnosti [Transatlantic Alliance. 1991–2004. Transformation of American-European Partnership After the End of Bipolarity]. Мoscow: NOFMO. 252 p. V l B (2013) B R l Th P f I R l D l R l d 168 g y p Walt S.M. (1987). The Origins of Alliance. Cornell University Press. 336 p. Waltz K. N. (1979). Theory of International Politics. Boston: McGraw Hill. 251 p. W h T J (2017) All M Sh f W Y l U P 288 Wright T. J. (2017). All Measures Short of War. Yale University Press. 288 p. Zimmerling R. (2006). Influence and Power: Variations on a Messy Theme. Springer. Vol. 68. 308 p. Zoellick R. (2020). America in the World: A History of US Diplomacy and Foreign Policy. Twevle. 560 p.
https://openalex.org/W3018091849	https://www.e3s-conferences.org/articles/e3sconf/pdf/2020/26/e3sconf_icsf2020_02005.pdf	English	null	Justification of the method of determination of the border overburden ratio	E3S web of conferences	2,020	cc-by	4,171	1 Introduction to the rapid depreciation of the loading faces of the open pits on the final designed boundaries in the near future. And this will entail a significant increase in the cost of the opencast mining due to the reactivation of non-working faces, if the open-pit boundary will be redefined in future. The economic efficiency of the enterprise and the operation period, the full utilization of mineral reserves and the total cost of operation depend on the selected depth and boundaries of the opencast mining. At present, there are a lot of methods in the theory and practice of design to determine the border overburden ratio and open-pit boundaries that develop steep-dipping banks. Today, the open pit boundaries are determined on the basis of a comparison of one of the overburden ratio of the designed open pit with the border overburden ratio. At the same time, to determine the border overburden ratio, economic indicators of the cost of ore underground mining method or the wholesale price of the mineral are most often used. However, in the world market of commodity iron ore products the only indicator of its competitiveness is price. Production costs and profitability are commercial secret. Therefore, the main condition for a comparative assessment of the operational efficiency of mining and processing plant in the world market is to compare the price of marketable products with the price of the same marketable products in the world market. Nowadays, in the development of steeply falling iron ore deposits, their working contours on the surface have reached design marks on the most open pits. Mining development occurs only when they are lowered. In such conditions of the deposit development it is necessary to evaluate the capabilities of the raw material base for the further development of the mining enterprise. When choosing opencast mining alternatives, their main parameters – open pit boundaries, mining mode and ore productivity – are assessed by economic indicators. These indicators are: profit from sales of saleable output; production costs of the saleable output; profitability, etc. As practice shows, the depth and position of the final boundaries of the most large open pits as the mineral deposits are worked out are repeatedly revised and adjusted. However, it is imperative to determine the final boundaries of the open pit mining in which the development of open pit deposits will be effective. Justification of the method of determination of the border overburden ratio Serhii Joukov1, Serhii Lutsenko1, Yulian Hryhoriev1,, Maxim Martyniuk2, and Volodymyr Peregudov3 1Kryvyi Rih National University, Mining and Metallurgical Faculty, 11 Vitalii Matusevuch Str., Kryvyi Rih, 50027, Ukraine 2SOUTH Mining and Processing Plant, Technical department, Kryvyi Rih, 50026, Ukraine 3SDI “Кrivbassproekt”, 40 Poshtoviy Ave., Kriviy Rih, 50000, Ukraine hii Joukov1, Serhii Lutsenko1, Yulian Hryhoriev1,, Maxim Martyniuk2, and Volodymyr Peregudov3 1Kryvyi Rih National University, Mining and Metallurgical Faculty, 11 Vitalii Matusevuch Str., Kryvyi Rih, 50027, Ukraine 2SOUTH Mining and Processing Plant, Technical department, Kryvyi Rih, 50026, Ukraine 3SDI “Кrivbassproekt”, 40 Poshtoviy Ave., Kriviy Rih, 50000, Ukraine Abstract. The economic efficiency of the enterprise and the length of its operation, the full utilization of mineral reserves and the total cost of operation depend on the selected depth and boundaries of opencast mining. In such conditions of the deposits’ development it is necessary to evaluate the capabilities of the raw material base for the further development of the mining enterprise. The basis for determining of the border overburden ratio is on the condition that the price of manufactured marketable output of the designed enterprise should not exceed the price of the same marketable output in the world market. The realization of this condition is achieved analytically. This takes into account the impact of the rate of return on the final depth of the designed open pit. A new method of determining of the border overburden ratio is developed, which ensures the competitiveness of iron ore concentrate in the world market. The value of the border overburden ratio is developed for the Pershotravnevij open pit of the Northern mining and processing plant, depending on the situation of the iron-containing products’ prices, which ensures the competitiveness of marketable iron-ore output in the domestic and world markets. 1 Introduction There is a concern about the designing a new enterprise, when similar mining companies are already operating. * Corresponding author: juliangrig@gmail.com © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1051/e3sconf/202016602005 https://doi.org/10.1051/e3sconf/202016602005 E3S Web of Conferences 166, 02005 (2020) ICSF 2020 EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 icenses/by/4.0/). 2 Object and methods of research Recent applied and scientific studies on this issue clearly indicate the need to re-evaluate the final boundaries of Ukraine’s open pits [1-3]. The development in the approved by existing designs boundaries can lead A great number of scientists were engaged in the questions of determining the open pit boundaries. The largest contribution [4-11] in the theory of designing the boundaries of open-cast mining is owned by E3S Web of Conferences 166, 02005 (2020) ICSF 2020 https://doi.org/10.1051/e3sconf/202016602005 А. I. Arsentyev, V. V. Rzhevskij, B. P. Yumatov, V. S Hohryakov, V. G. Bliznyukov, A. K. Polishuk, Yu. I. Anistratov. At present, the strategic assessment of the interaction of the company with the external environment we must first take into account the qualitative and quantitative impact of competitors and demand on pricing [19], and, accordingly, the profit of the enterprise [20]. With different levels of the competitive environment and the supply-demand relationship, different options for choosing mining company strategy and, accordingly, a suitable methodology for selecting the final open pit outline. Designers and scientists use the main criteria for evaluating [9-10] the boundaries of effective mining of open-pit deposits use the cut-off (economically viable) overburden ratio as the main criterion, which is the maximum permissible overburden ratio, according to which open-pit mining in these specific conditions economically feasible. Therefore, the goal was to develop a method of determining the cut off overburden ratio, which will allow while the formation of the price for iron ore products on the world market to determine the boundaries of the open pit, that ensures the competitiveness of mining and processing plant. There are several well-known principles to justify the depth of the open pits, based on the comparison of one of the overburden ratio (average, initial, contour, current, mid-operational) or the sum of these coefficients with the cut-off (economically feasible) overburden ratio. The economic basis of all calculation principles for determining the boundaries of the open pits is to compare the possible cost of ore mining with the expected over the designed open pit [12], and the cut-off overburden ratio of the overburden is calculated by the technical and economic indicators achieved at the mining enterprises at the time of design. 3 Results The only indicator of the competitiveness of the marketable iron ore products is its price. Therefore, the main condition for a comparative assessment of the designed mining and processing plant in the world market is to compare the price of finished product with the price of the same product in the world market. Hence, the basis of the determination of the cut-off overburden ratio will be the condition that the price of the finished product of the designed enterprise does not exceed the prices of the same marketable products in the world market, which is represented by inequality: Recently, studies have been conducted on the determination of prospective mining boundaries of the open pits: the Scientific and Research Mining Institute (SRMI) [1, 13]; The Academy of Mining Sciences of Ukraine [2] and Kryvyi Rih National University (KNU) [3]. The determination of the cut-off overburden ratio was performed by these institutes on the basis of comparison of indicators of work of the investigated enterprise with indicators of enterprises with the underground method of mining, as well as with indicators of similar mining and processing plants. The price of the marketable products (concentrate), which depends on its quality, has always determined the boundaries of opencast mining. However, at different mining enterprises, marketable products of the same quality may have different selling prices [14], which in turn may make an error in determining the final boundaries of the open pits. Therefore, in the paper [15] it was proved contractual factors influence on the price of commodity products, which makes inaccurate application of existing methods for determining the final open pit boundaries. ܲ௖ ௗ≤ܲ௖ ௐ (1) (1) where ܲ௖ ௗ – the price of the concentrate of the designed mining and processing plant, UAH/t; ܲ௖ ௐ the price of the concentrate in the world market, UAH/t. At any mining enterprise we plan the profit margin, which is expressed by the final sum of production profitability of the enterprise. Therefore, the price of the concentrate of the designed mining and processing plant can be represented as: ܲ௖ ௗ= С௖ ௗ⋅(1 + ߰), UAH/t, (2) (2) where С௖ ௗ – the prime cost of the concentrate of the designed mining and processing plant, UAH/t;  – target profit marginу, parts un. 2 Object and methods of research To achieve this goal, a complex research method was used in the work, which includes theoretical and instrumental methods: analysis and theoretical generalization of enterprises reports, scientific and technical publications and other information sources in the field of determining the boundaries of careers in substantiating the problem and task of research; system analysis in researching career indicators; analytical, graphical and graphoanalytical methods of the research of technological and economic interconnections of the mining mode and the productivity by minerals in determining the boundaries of the career; economic and mathematical modeling of career options. In this case, to determine the cut-off overburden ratio, the most commonly used economic indicators of the cost of ore mining by underground method or the wholesale price of minerals. However, it should be understood that comparing the cost of open and underground mining is not correct due to the different production volumes; the open pit productivity is much higher than the productivity of mines, which to a different extent covers the needs of the market. In addition, the cost of mining by underground method remains equally high throughout all the service life, while the open pit reaches a production cost peak in only one of the operating periods at which the surface boundaries reach the final pit. 3 Results However, the existing methods of determining the cut- off overburden ratio [16-18] do not take into account the competitiveness of the iron ore products of the designed enterprise in the world market. 2 https://doi.org/10.1051/e3sconf/202016602005 E3S Web of Conferences 166, 02005 (2020) E3S Web of Conferences 166, 02005 (2020) ICSF 2020 The cost of production of the concentrate of the designed mining and processing plant will be represented by the expression: of production should increase. With a constant overburden ratio, increasing profitability is possible only with the increase of the price of products. with the increase of the price of products. Fig. 1. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 65%: 1 – at the rate of profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. Fig. 2. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 66%: 1 – at the rate of the profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. Thus, with the increase of the overburden ratio and Fig. 1. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 65%: 1 – at the rate of profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. С௖ ௗ= а೘ ೏ାа೎೏ା௕೏⋅௡ ఊ೎೏ , UAH/t, (3) (3) where а௠ ௗ – the cost of ore mining without the expense of the overburden in the designed open pit, UAH/t; а௖ ௗ – the cost of the ore conversion to the concentrate at the designed mining and processing plant, UAH/t; ܾௗ – the cost of the overburden in the designed open pit, UAH/m3; n – the overburden ration in the designed open pit, m3/t; ߛ௖ ௗ – concentrate output from one ton of ore to the designed mining and processing plant, parts un. 3 Results On this basis, the expression of determination of the concentrate price of the designed mining and processing plant (2) will take the following form: ܲ௖ ௗ= а೘ ೏ାа೎೏ା௕೏⋅௡ ఊ೎೏ ⋅(1 + ߰), UAH/t. (4) (4) Fig. 1. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 65%: 1 – at the rate of profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. Fig. 1. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 65%: 1 – at the rate of profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. The price of concentrate, which is on the world market, depends on the content of iron. Therefore, the price of the concentrate of the designed mining and processing plant should be brought (in terms of the content of iron) to the price of the base enterprise (the price of concentrate and its iron content in the world market). Then the expression will look like this [1]: g p g p p g overburden ratio when the iron content of the concentrate in the world market 65%: 1 – at the rate of profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. ψ 30; 3 when ψ 50. Fig. 2. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 66%: 1 – at the rate of the profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. ܲ′௖ ௗ= ܲ௖ ௐ⋅ ఉ೏ ఉೈ, (5) (5) where ܲ′௖ ௗ – the present value of the concentrate of the designed mining and processing plant at the price of concentrate and the content of iron in it on the world market, UAH/t; ߚௐ і ߚௗ – iron content in the concentrate in the world market and designed mining and processing plant, accordingly, parts un. 3 Results On this basis, the price of the designed concentrate will be defined as: ܲ′௖ ௗ= ఉ೏⋅൫а೘ ೏ାа೎೏ା௕೏⋅௡൯ ఊ೎೏⋅ఉೈ ⋅(1 + ߰), UAH/t. (6) (6) Fig. 2. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 66%: 1 – at the rate of the profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. Fig. 2. Changes in the price of the concentrate of the Northern mining and processing plant, depending on the current overburden ratio when the iron content of the concentrate in the world market 66%: 1 – at the rate of the profit ψ = 0 %; 2 – when ψ = 30; 3 – when ψ = 50. . 2. Changes in the price of the concentrate of the Northern Based on formula (2), the cost of concentrate on the designed mining and processing plant as an indicator of competitiveness, depends on the planned profit and volume of overburden works. Other indicators are determined by the level of technology and technology of ore and dead rocks mining (а௠ ௗ і ܾௗ), achieved at the enterprise, indicators of technology for ore conversion to concentrate and quality of ore (а௖ ௗ, ߛ௖ ௗ, ߚௗ). Thus, with the increase of the overburden ratio and with the increase of the profitability of the enterprise, the price of production of the designed enterprise will increase. There will come a time when the price of production will rise to the limit value ܲ௖ ௐ. Then, substituting in the condition of competitiveness of marketable product of the designed pit (1), the value of its possible price, we obtain: These indicators in determining the competitiveness of similar companies, can be taken constant, as they do not change for a long time, and if they change, then at almost the same rate. For example, let’s consider the graphs of the change in the price of concentrate of the North mining and processing plant from the current overburden ratio with different iron content in concentrate in the world market (Fig. 1, 2). ఉ೏⋅൫а೘ ೏ାа೎೏ା௕೏⋅௡೎.೚.൯⋅(ଵାట) ఊ೎೏⋅ఉೈ ≤ܲ௖ ௐ. 3 Results (7) (7) Based on this inequality, the cut-off overburden ratio is defined as: From the presented pictures we can see that while maintaining the profitability of the enterprise at the same level, with the increase of the overburden ratio, the price 3 3 E3S Web of Conferences 166, 02005 (2020) ICSF 2020 E3S Web of Conferences 166, 02005 (2020) https://doi.org/10.1051/e3sconf/202016602005 ݊с.о. = ು೎ೈ⋅ം೎೏⋅ഁೈ ഁ೏⋅(భశഗ) ି(а೘ ೏ାа೎೏) ௕೏ . (8) an iron content of 65.5% and a profit rate of 40%, the overburden ratio will be 4.8 m3/t, and the price of concentrate – 1228 UAH/t. (8) In the table 1 we report the present value of the concentrate of the Northern mining and processing plant calculated by the formula 6. Fig. 4. The dependence of the cut off overburden ratio on the profit rate for the working conditions of the Pershotravnevij open pit of the Northern mining and processing plant. Table 1. The price of concentrate on the world market and the price of the concentrate of the Northern mining and processing plant depending from the iron content. Iron content in the concentrate, % The price of the concentrate, UAH/t In the world market Discounted value Northern MPP 62 1086,4 1150,2 63 1127,0 1174,2 64 1167,2 1197,1 65 1208,0 1219,9 66 1248,0 1241,2 Substituting the value of economic and technological indicators of the Northern mining and processing plant in expression (8), we determine the cut-off overburden ratio with different values of profitability (profit rate: 1 – 0%; 2 – 10%; 3 – 20%; 4 – 30 %; 5 – 40%; 6 – 50%). This is shown in Fig. 3. The points in the Fig. 3 correspond to the iron content of the concentrate on the world market. Fig. 4. The dependence of the cut off overburden ratio on the profit rate for the working conditions of the Pershotravnevij open pit of the Northern mining and processing plant. Studies on theoretical open pits and powerful iron ore open pits of Kryvbas have shown that the accuracy in determining of the open pit boundaries that do not take into account market conditions is up to 50%. Thus, it was determined that in order to more accurately determine the open pit boundaries, it is not necessary to take into account the levels of competition, supply and demand. 3 Results However, in light of the fact, that it is difficult to predict economic indicators with a high degree of accuracy, it is necessary to move from difficult-to-predicted economic indicators to more stable technological ones when determining the pit outline. Fig. 3. The dependence of the cut-off overburden ratio on the price of concentrate in the world market: 1 – at the profit rate ψ = 0%; 2 – with ψ = 10%; 3 – with ψ = 20%; 4 – with ψ = 30%; 5 – with ψ = 40%; 6 – with ψ = 50%. Such technological indicators may serve changes of the ore volumes and overburdens in open pits- competitions. In this regard, before designing it is important to carry out analytical work, to make forecasts about the future conditions of competition and demand, as well as changes in price trends and cost of marketable product. At the design stage, it is important to work out the most likely scenarios, and the economic effectiveness of the variants of the final pit boundaries must be determined taking into account the risks of the probably change in the market conditions. Fig. 3. The dependence of the cut-off overburden ratio on the price of concentrate in the world market: 1 – at the profit rate ψ = 0%; 2 – with ψ = 10%; 3 – with ψ = 20%; 4 – with ψ = 30%; 5 – with ψ = 40%; 6 – with ψ = 50%. Fig. 3. The dependence of the cut-off overburden ratio on the price of concentrate in the world market: 1 – at the profit rate ψ = 0%; 2 – with ψ = 10%; 3 – with ψ = 20%; 4 – with ψ = 30%; 5 – with ψ = 40%; 6 – with ψ = 50%. In Fig. 4 we present the dependence of overburden ratio from the profit rate for the working conditions of the Pershotravnevij open pit of the Northern MPP. 4 Conclusions A new method of determination of the cut-off overburden ratio has been developed, that allows the determining while the world market price of iron ore products to determine the open pit boundaries, which ensure the competitiveness of mining and processing plant. According to the developed method, the values of the cut off overburden ratio for the Pershotravnevij open pit of the Northern mining and processing plant were determined, which ensure the competitiveness of iron ore For example, it is necessary to find out what the price of a concentrate with an iron content of 65.5% at a rate of 40% will be. To do this, on the curve with the required norm of the profit (curve 5, Fig. 3) we find a given value of iron content. From the value of the desired iron content (between the group of vertical points 65% and 66%), lower the perpendicular (dashed lines, Fig. 3) on the y- axis and abscissa and determine the values of the overburden ratio and the concentrate price. That is, with 4 E3S Web of Conferences 166, 02005 (2020) E3S Web of Conferences 166, 02005 (2020) ICSF 2020 https://doi.org/10.1051/e3sconf/202016602005 16. A. Selyukov, R. Rybár. Calculation of Boundary Stripping Ratio Errors at the Stage of Quarries Designing. E3S Web of Conferences (2019). doi:10.1051/e3sconf/201910501043 marketable product in the world market at different values of the profit rate. Thus, this principle of defining the boundaries of opencast mining takes into account the dynamics of the external competitive environment of the mining enterprise. 17. V. Kalyuzhin, F. Karavaytsyev, V. Shchukina, Determination of the limits of municipal formations in the inhomogeneous geoinformation space. E3S Web of Conferences (2019). doi:10.1051/e3sconf/201911002117 References 1. Opredelenie perspektivnyh granic Annovskogo karera OAO “SevGOK”, Otchet o NIR (Gosudarstvennoe predpriyatie “Nauchno issledovatelskij gornorudnyj institute”, Krivoy Rog, 2010). 18. S. Moldabayev, B. Rysbaiuly, Zh. Sultanbekova, N. Sarybayev, Methodological approach to creation of the 3D model of an oval-shaped open pit mine. E3S Web of Conferences (2019). doi:10.1051/e3sconf/201912301049 2. Opredelenie perspektivnyh granic i pro- izvoditelnosti Pervomajskogo karera PAO “SevGOK”, Otchet o NIR (Akademiya gornyh nauk Ukrainy, Krivoy Rog, 2014) 19. V. Mikhalchenko, The Concept of Resource Use Efficiency as a Theoretical Basis for Promising Coal Mining Technologies. E3S Web of Conferences (2017). doi:10.1051/e3sconf/20172104007 3. Opredelenie perspektivnyh granic karera, obespechivayushih konkurentosposobnost zhele- zorudnoj produkcii Poltavskogo GOKa, Otchet o NIR (Gosudarstvennoe vysshee uchebnoe zavedenie “Krivorozhskij nacionalnyj universitet”, Krivoy Rog, 2014). 20. S.A. Lutsenko, V.G. Bliznyukov, Quality – Access to Success 18 (S1), 226–230 (2017) 4. A.I. Arsentev, Opredelenie proizvoditelnosti i granic karerov, 2nd edn. (Nedra, Moscow, 1970), pр. 175– 198. 5. A.I. Arsentev, A.K. Polishuk, Razvitie metodov opredeleniya granic karerov (Nauka, Leningrad, 1967), pр. 68–82 6. V.V. Rzhevskij, Otkrytye gornye raboty (Nedra, Moscow, 1985), pр. 256–263 7. B.P. Yumatov, Gornyy zhurnal 2, 45–53 (1962) 8. V.S. Hohryakov, Proektirovanie karerov (Nedra, Moscow, 1980), рp. 115–128 9. A.I. Arsentev, Konechnyie granitsyi karerov (Publishing Sankt-Peterburgskiy Gornyiy institute, Sankt-Peterburg, 1995), pр. 29–41 10. V.G. Bliznyukov, Opredelenie glavnyh parametrov karera s uchetom kachestva rudy (Nedra, Moscow, 1978), pр. 105–118 11. V.G. Bliznyukov, I.V. Baranov, Visnik Kryvorizkoho tekhnichnoho universytetu 18, 7–11 (2007) 12. Yu.I. Anistratov, K.Yu. Anistratov, Proektirovanie karerov (NPK “Gemos Limited”, Moscow, 2002), pр. 111–127 13. Opredelenie perspektivnyh granic i proizvoditelnosti karera Inguleckogo gorno-obogatitelnogo kombinata, Otchet o NIR (Gosudarstvennoe predpriyatie “Nauchno issledovatelskij gornorudnyj institute”, Krivoy Rog, 2007) 14. V.G. Bliznyukov, S.A. Lucenko, Scientific bulletin of National Mining University 1(157), 44-49 (2017) 15. V.G. Bliznyukov, I.V. Baranov, A.V. Savickij, Visnik Krivorizkogo nacionalnogo universitetu 31, 3–6 (2012) 5 5
https://openalex.org/W2125471673	https://europepmc.org/articles/pmc3907302?pdf=render	English	null	Crossover Patterning by the Beam-Film Model: Analysis and Implications	PLOS genetics	2,014	cc-by	24,949	Introduction Crossover (CO) recombination interactions occur stochastically at different positions in different meiotic nuclei. Nonetheless, along a given chromosome, COs tend to be evenly spaced. This interesting phenomenon implies the existence of communication along chromosomes, the nature of which is not understood. CO patterning, commonly known as ‘‘CO interference’’, was originally detected from genetic studies in Drosophila [1,2]. It was found that the frequency of meiotic gametes exhibiting two crossovers close together along the same chromosome (‘‘double COs’’) was lower than that expected for their independent occurrence. The implication was that occurrence of one CO (or more correctly one CO-designation) ‘‘interferes’’ with the occurrence of another CO (CO-designation) nearby. CO-designated interactions then undergo multiple additional biochemical steps to finally become mature CO products [5]. Precursors that do not undergo CO-designation mature to other fates, predominantly inter-homolog non-crossovers (NCOs). We previously proposed a model for CO patterning in which macroscopic mechanical properties of chromosomes play gov- erning roles via accumulation, relief and redistribution of stress (Figure 1A) [3,4]. In that model, a chromosome with an array of precursor interactions comes under mechanical stress along its length. Eventually, a first interaction ‘‘goes critical’’, undergoing a stress-promoted molecular change which designates it to eventually mature as a CO. By its intrinsic nature, this change results in local relief of stress. That local relaxation then redistributes outward in the immediate vicinity of its nucleation point, in both directions, dissipating with distance. A new stress distribution is thereby produced, with the stress level reduced in the vicinity of the CO-designation site, to a decreasing extent with increasing distance from its nucleation point. This effect disfavors occurrence of additional (stress-promoted) CO designa- tions in the affected region. The spreading inhibitory signal CO patterning by the above stress-and-stress relief mechanism can be modeled quantitatively by analogy with a known physical system that exhibits analogous behavior, giving the beam-film (BF) model [3]. We note that BF model simulations can be applied to any mechanism whose effects are described by the same mathematical expressions as the beam-film case. In such a more general formulation (Figure 1B), there is again an array of precursor interactions. That array would be acted upon by a ‘‘Designation Driving Force’’ (DDF). Event-designations would occur sequen- tially (or nearly so). Liangran Zhang1, Zhangyi Liang1, John Hutchinson2, Nancy Kleckner1* ar Biology, Harvard University, Cambridge, Massachusetts, United States of America, 2 School of Engineering and Applied Sciences, achusetts, United States of America 1 Department of Molecular and Cellular Biology, Harvard University, Cambridge, Massachusetts, United States of America, 2 Sch Harvard University, Cambridge, Massachusetts, United States of America 1 Department of Molecular and Cellular Biology, Harvard University, Cambridge, Massachusetts, United States of America, 2 School of Engineering and Applied Sciences Harvard University, Cambridge, Massachusetts, United States of America Abstract Crossing-over is a central feature of meiosis. Meiotic crossover (CO) sites are spatially patterned along chromosomes. CO- designation at one position disfavors subsequent CO-designation(s) nearby, as described by the classical phenomenon of CO interference. If multiple designations occur, COs tend to be evenly spaced. We have previously proposed a mechanical model by which CO patterning could occur. The central feature of a mechanical mechanism is that communication along the chromosomes, as required for CO interference, can occur by redistribution of mechanical stress. Here we further explore the nature of the beam-film model, its ability to quantitatively explain CO patterns in detail in several organisms, and its implications for three important patterning-related phenomena: CO homeostasis, the fact that the level of zero-CO bivalents can be low (the ‘‘obligatory CO’’), and the occurrence of non-interfering COs. Relationships to other models are discussed. Editor: R. Scott Hawley, Stowers Institute for Medical Research, United States of America Editor: R. Scott Hawley, Stowers Institute for Medical Research, United States of America Received May 10, 2013; Accepted November 5, 2013; Published January 30, 2014 Received May 10, 2013; Accepted November 5, 2013; Published January 30, 2014 Received May 10, 2013; Accepted November 5, 2013; Published January 30, 2014 Copyright: 2014 Zhang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Funding to LZ, ZL and NK provided by a grant to NK from the National Institutes of Health: NIH/NIGMS R01 GM044794 (http://nih.gov). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: kleckner@fas.harvard.edu comprises ‘‘CO interference’’. More such CO-designations may then occur, sequentially, each accompanied by spreading interference. Each subsequent event will tend to occur in a region where the stress level remains higher, which will necessarily tend to be regions far away from prior CO-designated sites. Thus, as more and more designation events occur, they tend to fill in the holes between prior events, ultimately producing an evenly-spaced array. The most attractive feature of this proposed mechanism is the fact that redistribution of stress is an intrinsic feature of any mechanical system, thus comprising a built-in communication network as required for spreading CO interference. Patterning by the Beam-Film Model: Analysis cations Zhangyi Liang1, John Hutchinson2, Nancy Kleckner1* d Cellular Biology, Harvard University, Cambridge, Massachusetts, United States of America, 2 School of Engineering and Applied Sciences, e, Massachusetts, United States of America central feature of meiosis. Meiotic crossover (CO) sites are spatially patterned along chromosomes. CO- position disfavors subsequent CO-designation(s) nearby, as described by the classical phenomenon of multiple designations occur, COs tend to be evenly spaced. We have previously proposed a mechanical O patterning could occur. The central feature of a mechanical mechanism is that communication along as required for CO interference, can occur by redistribution of mechanical stress. Here we further of the beam-film model, its ability to quantitatively explain CO patterns in detail in several organisms, s for three important patterning-related phenomena: CO homeostasis, the fact that the level of zero-CO w (the ‘‘obligatory CO’’), and the occurrence of non-interfering COs. Relationships to other models are , Hutchinson J, Kleckner N (2014) Crossover Patterning by the Beam-Film Model: Analysis and Implications. PLoS Genet 10(1): e1004042. 004042 towers Institute for Medical Research, United States of America Accepted November 5, 2013; Published January 30, 2014 g et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits on, and reproduction in any medium, provided the original author and source are credited. L and NK provided by a grant to NK from the National Institutes of Health: NIH/NIGMS R01 GM044794 (http://nih.gov). The funders had ata collection and analysis, decision to publish, or preparation of the manuscript. he authors have declared that no competing interests exist. ard.edu Crossover Patterning by the Beam-Film Model: Analysis and Implications Liangran Zhang1, Zhangyi Liang1, John Hutchinson2, Nancy Kleckner1* January 2014 \| Volume 10 \| Issue 1 \| e1004042 Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) CO data sets, whether experimental or from BF simulations, comprise descriptions of the positions of individual COs along the lengths of each of a large number of different chromosomes (‘‘bivalents’’). Each bivalent represents the outcome of CO- designation in a single meiotic nucleus; the entire data set comprises the outcomes of CO patterning for a particular chromosome in many nuclei. multiple designation/interference events occur, they would produce an evenly-spaced array. Maturation of CO-designated and not-CO-designated interactions ensues. CoC relationships. The classical description of CO inter- ference relationships is Coefficient of Coincidence (CoC) analysis [1,2]. For this purpose (Figure 2A), the chromosome of interest is divided into a number of intervals and for each interval the total frequency of COs in the many chromosomes examined is observed. Intervals are then considered in pairs, in all possible pairwise combinations. For each pair, the observed frequency of The present study adds several new features to the BF simulation program and explores in further detail the predictions and implications of the BF model (whether mechanical or general). We evaluate the ability of the model to quantitatively explain experimental CO pattern data sets in budding yeast, tomato, Figure 1. The beam-film model. (A) Beam-film model [3]. CO-designation is promoted by stress. Each stress-promoted CO-designation reduces the stress level to zero at the designation point. That effect redistributes in the vicinity, decreasing exponentially with distance, correspondingly reducing the probability of subsequent designation(s) in the immediate vicinity. Subsequent CO-designations will tend to occur in regions with higher remaining stress levels, thus giving an even distribution. More specifically: with a film attached to a beam, if the beam expands relative to the film, stress arises along the film, causing it to crack at the positions of flaws. A crack at one position will release the stress nearby (with a distance L that is characteristic of the materials) thus reducing the probability that another crack occurs nearby. Assuming a maximal possible stress level of s0, if a crack occurs at an isolated position that is unaffected by any prior cracks, then the stress level at any distance ‘‘x’’ to either side is s = s0 (12e2x/L). If two cracks occur near one another, the stress levels at positions between them is the sum of their individual effects, with additional considerations also coming into play at the ends of the beam as described [3]. Results Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) Author Summary Spatial patterning is a common feature of biological systems at all length scales, from molecular to multi- organismic. Meiosis is the specialized cellular program in which a diploid cell gives rise to haploid gametes for sexual reproduction. Crossing-over between homologous maternal and paternal chromosomes (homologs) is a central feature of this program, playing a role not only for increasing genetic diversity but also for ensuring regular segregation of homologs at the first meiotic division. The distribution of crossovers (COs) along meiotic chromosomes is a paradigmatic example of spatial patterning. Crossovers occur at different positions in different meiotic nuclei but, nonetheless, tend to be evenly spaced along the chromosomes. We previously- described a mechanical ‘‘stress and stress relief’’ model for CO patterning with an accompanying mathematical description (the ‘‘beam-film model’’). In this paper we explore the roles of mathematical parameters in this model; show that it can very accurately describe experi- mental data sets from several organisms, in considerably quantitative depth; and discuss implications of the model for several phenomena that are directly related to crossover patterning, including the features which can ensure that every chromosome always acquires at least one crossover. Introduction Each designation would set up a spreading inhibitory effect that spreads outward in both directions, decreasing in strength with increasing distance, thereby decreasing the ability of the affected precursors to respond to the DDF. When January 2014 \| Volume 10 \| Issue 1 \| e1004042 1 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns grasshopper and Drosophila. Our results show that the logic and mathematics of the BF model are remarkably robust in explaining experimental data. New information of biological interest also emerges. We then present detailed considerations of three phenomena of interest, the so-called ‘‘obligatory CO’’ and ‘‘CO homeostasis’’, and the nature of ‘‘non-interfering COs’’. We discuss how these phenomena are explained by the BF model and show that BF predictions can very accurately explain experimental data pertaining to these effects. Overall, the presented results show that BF simulation analysis is a useful approach for exploring experimental CO patterns. Other applications of this analysis are presented elsewhere. The current study has also provided new criteria for characterization of CO patterns using Coefficient of Coincidence analysis and illustrates both short-comings and useful applications of gamma distribution analysis. Relationships of the BF model to other models are discussed. grasshopper and Drosophila. Our results show that the logic and mathematics of the BF model are remarkably robust in explaining experimental data. New information of biological interest also emerges. We then present detailed considerations of three phenomena of interest, the so-called ‘‘obligatory CO’’ and ‘‘CO homeostasis’’, and the nature of ‘‘non-interfering COs’’. We discuss how these phenomena are explained by the BF model and show that BF predictions can very accurately explain experimental data pertaining to these effects. Overall, the presented results show that BF simulation analysis is a useful approach for exploring experimental CO patterns. Other applications of this analysis are presented elsewhere. The current study has also provided new criteria for characterization of CO patterns using Coefficient of Coincidence analysis and illustrates both short-comings and useful applications of gamma distribution analysis. Relationships of the BF model to other models are discussed. Author Summary Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) Fraction of bivalents exhibiting different numbers of COs and average values. d i 0 3 /j l 00 0 2 002 BF simulations specify a parameter for interference distance, denoted ‘‘L’’ ([3]; below). Figure 2B illustrates CoC curves for simulations at varying values of L. For any actual CoC curve, whether experimental or simulated, a useful parameter for describing the strength of CO interference is the inter-interval distance at which the CoC = 0.5. We define this parameter as LCoC (Figure 2B; vertical arrows). Where appropriate, the value of the BF-specified parameter ‘‘L’’ is denoted alternatively as LBF to distinguish it from LCoC (Figure 2B). Interestingly, the values of LCoC and LBF are always quite similar (e.g. Figure 2B). Figure 2. Descriptors of CO patterns: Coefficient of Coinci- dence (CoC) and Event Distribution (ED). (A) Determination of CoC. Interval sizes can be identical or different (Figure S1 for more details). (B,C) Data sets were generated by BF simulations at the indicated varying values of parameter (L), also called LBF. Other specified parameter values for the simulations are: Smax = 3.5, A = 1, cL = cR = 0.85, N = 13, B = 1, E = 0, M = 1. (B) CoC curves. Inter-interval distances given as fractions of total physical chromosome length in mm. The inter-interval distance at which CoC = 0.5 (vertical arrows) is defined as LCoC. LCoC and LBF are always quite similar in magnitude. (C) EDs. Fraction of bivalents exhibiting different numbers of COs and average values. CoC analysis provides a very accurate and reproducible description of CO patterns for experimental data sets as long as two requirements are met (Figure S1, Protocol S1). First, chromosomes must be divided into a large enough number of intervals that double COs within an interval are rare. If this requirement is not met, closely-spaced double COs will be missed. In general, interval size should be less than ,1/4 the average distance between adjacent COs. Second, the data set must be large enough to give significant numbers of double COs. As a practical matter, where possible (e.g. for cytological markers of CO positions), interval size should be decreased progressively until the CoC curve no longer changes. doi:10.1371/journal.pgen.1004042.g002 ‘‘Event Distribution’’ or ‘‘ED’’. As the ‘‘interference distance’’ increases, the distribution of COs per bivalent shifts to lower numbers with a corresponding decrease in the average number of COs per bivalent (Figure 2C). Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) (B) A generalized version of the beam-film model involving sequential rounds of event designation and spreading interference as described by the mathematical expressions of the BF model. doi:10.1371/journal.pgen.1004042.g001 Figure 1. The beam-film model. (A) Beam-film model [3]. CO-designation is promoted by stress. Each stress-promoted CO-designation reduces the stress level to zero at the designation point. That effect redistributes in the vicinity, decreasing exponentially with distance, correspondingly reducing the probability of subsequent designation(s) in the immediate vicinity. Subsequent CO-designations will tend to occur in regions with higher remaining stress levels, thus giving an even distribution. More specifically: with a film attached to a beam, if the beam expands relative to the film, stress arises along the film, causing it to crack at the positions of flaws. A crack at one position will release the stress nearby (with a distance L that is characteristic of the materials) thus reducing the probability that another crack occurs nearby. Assuming a maximal possible stress level of s0, if a crack occurs at an isolated position that is unaffected by any prior cracks, then the stress level at any distance ‘‘x’’ to either side is s = s0 (12e2x/L). If two cracks occur near one another, the stress levels at positions between them is the sum of their individual effects, with additional considerations also coming into play at the ends of the beam as described [3]. (B) A generalized version of the beam-film model involving sequential rounds of event designation and spreading interference as described by the mathematical expressions of the BF model. doi:10.1371/journal.pgen.1004042.g001 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 2 Logic and Modeling of Meiotic Crossover Patterns Figure 2. Descriptors of CO patterns: Coefficient of Coinci- dence (CoC) and Event Distribution (ED). (A) Determination of CoC. Interval sizes can be identical or different (Figure S1 for more details). (B,C) Data sets were generated by BF simulations at the indicated varying values of parameter (L), also called LBF. Other specified parameter values for the simulations are: Smax = 3.5, A = 1, cL = cR = 0.85, N = 13, B = 1, E = 0, M = 1. (B) CoC curves. Inter-interval distances given as fractions of total physical chromosome length in mm. The inter-interval distance at which CoC = 0.5 (vertical arrows) is defined as LCoC. LCoC and LBF are always quite similar in magnitude. Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) We further note that the appropriate metric for CO interference is physical distance along the chromosome. This has been shown to be the case for mouse and Arabidopsis [6–8]; for tomato (as described below); and for budding yeast (L.Z., unpublished). Accordingly, disruption of chromosome continuity abolishes the transmission of interference in C.elegans [9]. Experimentally, ‘‘interference distance’’ is defined in units of mm pachytene (synaptonemal complex; SC) length. In reality, SC length is often (or always) a proxy for chromosome length at the preceding stage (leptotene): in yeast, Sordaria and likely other organisms, CO patterning occurs at the leptotene stage and nucleates SC formation (e.g., [4]). Part I. Coefficient of Coincidence (CoC) Relationships and the Event Distribution (ED) (C) EDs. Fraction of bivalents exhibiting different numbers of COs and average values. doi:10.1371/journal.pgen.1004042.g002 bivalents exhibiting a CO in both intervals (‘‘double COs’’) is compared with the frequency expected if events occurred independently in the two intervals. The latter frequency is given by the product of the total frequencies of events in the two intervals, each considered individually. For any pair of intervals, the ratio of the frequency of observed double COs to the frequency of expected double COs (Observed/Expected) is the Coefficient of Coincidence (CoC). If events occur independently in the two intervals, the CoC for that pair of intervals is one. If (positive) CO interference is present for the two intervals, the CoC is less than one (some expected COs have been inhibited). CoC values for all interval pairs are then plotted as a function of the distance between the corresponding intervals (defined as the distance between the centers of the two intervals). The classical resulting CO interference CoC pattern is illustrated by an appropriate set of BF simulations (Figure 2B). When intervals are close together (short inter-interval distances), the frequency of observed COs is much less than that expected from independent occurrence (CoC,,1), reflecting ‘‘interfer- ence’’. The CoC increases with increasing inter-interval distance to a value of ,1. Additionally, because COs tend to be evenly spaced, the CoC value rises above ,1 specifically at the average distance between adjacent COs (or multiples thereof): at these particular spacings, the probability of a double CO is higher than that predicted by random occurrence. This tendency is increas- ingly pronounced as interference extends over longer and longer fractions of total chromosome length. Figure 2. Descriptors of CO patterns: Coefficient of Coinci- dence (CoC) and Event Distribution (ED). (A) Determination of CoC. Interval sizes can be identical or different (Figure S1 for more details). (B,C) Data sets were generated by BF simulations at the indicated varying values of parameter (L), also called LBF. Other specified parameter values for the simulations are: Smax = 3.5, A = 1, cL = cR = 0.85, N = 13, B = 1, E = 0, M = 1. (B) CoC curves. Inter-interval distances given as fractions of total physical chromosome length in mm. The inter-interval distance at which CoC = 0.5 (vertical arrows) is defined as LCoC. LCoC and LBF are always quite similar in magnitude. (C) EDs. Part II. Parameters of the BF Model and Their Roles for CO Patterns 0 = random; 1 = even; 0, = E, = 1 Precursor distribution among chromosomes 0 = Poisson; 1 = constant; 0, = B, = 1 Distribution of precursor sensitivities to DDF (currently random) Precursor density can be varied along the chromosome as desired Efficiency with which a CO-designated interaction matures to a detectable CO doi:10.1371/journal.pgen.1004042.t001 parameters describe the nature of this array. (N) specifies the average number of precursors per bivalent. (E) specifies the extent to which the precursors along a given bivalent are evenly or randomly spaced. (B) specifies the extent to which precursors occur at a constant value along a given bivalent in different meiotic nuclei or are randomly (Poisson) distributed among different nuclei. Also, different precursors will naturally exhibit a range of intrinsic sensitivities to the DDF. The parameter(s) specifies the distribution of those sensitivities as specified by the Matlab function ‘‘rand’’. clearly true in some organisms (e.g. [12]). It is not so clear in other organisms, where different regions of the genome can be at significantly different stages within a single nucleus (e.g. [27]). However, BF simulations will still pertain in the latter case if the effects of CO-designation at earlier-evolved positions can be ‘‘stored’’ within the chromosomes and exert their effects upon nearby positions when the appropriate precursors do finally evolve. Finally, it has sometimes been considered that CO patterns evolve in two stages (e.g. [11,15]). In such a case, one round of event-designation is imposed on total DSB-mediated interactions, giving a set of intermediate designated sites. That intermediate set then undergoes a second round of designation. BF simulations can directly model this situation. DSB-mediated interactions are used as a first set of precursors for a first simulation to give the intermediate array of events. That intermediate array is then used as a second set of precursors for a second simulation. The predicted outcomes from one- and two-round scenarios for recombination-related markers in Sordaria meiosis are presented elsewhere. A useful feature in distinguishing between the two scenarios is whether closely-spaced COs ever do, or do not, occur at the specific spacing characteristic of the first precursor array. If COs arise in a single step, closely-spaced events can occur at the positions of adjacent precursors. If COs arise in two steps, then this will not occur; instead, closely-spaced COs can only occur at the spacing of adjacent precursors in the intermediate array (Figure S3). Part II. Parameters of the BF Model and Their Roles for CO Patterns BF simulations require specification of three types of parameters (Table 1). One set describes the nature of the precursor array upon which CO-designation acts; a second set describes features of the patterning process per se; and a third precursor specifies the efficiency with which a designated event matures into a detectable CO or CO-correlated signal. The ED. CO patterns are also reflected in the average number of COs per bivalent and the fractions of chromosomes exhibiting different numbers of COs, which we refer to as the The precursor array. The precursors for CO patterning are generally assumed to be the total array of double strand break (DSB)-initiated interactions between homologs. Several BF The ED. CO patterns are also reflected in the average number of COs per bivalent and the fractions of chromosomes exhibiting different numbers of COs, which we refer to as the January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org January 2014 \| Volume 10 \| Issue 1 \| e1004042 3 Table 1. BF parameters. Precursor parameters N Average number of precursors per bivalent E Precursor distribution along chromosomes. 0 = random; 1 = even; 0, = E, = 1 B Precursor distribution among chromosomes 0 = Poisson; 1 = constant; 0, = B, = 1 s Distribution of precursor sensitivities to DDF (currently random) Density Precursor density can be varied along the chromosome as desired Patterning parameters S (Smax) Designation driving force; increased in simulations to maximum level; usually S.1 L = LBF Distance over which interference signal spreads (total for both directions) A Reactivity of precursor to local value of (s); (A 1 = s2; A2 = s; A3 = 5/s; A4 = 1/s) c End effects on interference (cL and cR) 0 = unclamped; $1 = clamped Other parameters M Efficiency with which a CO-designated interaction matures to a detectable CO doi:10.1371/journal.pgen.1004042.t001 Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns Table 1. BF parameters. Average number of precursors per bivalent Precursor distribution along chromosomes. PLOS Genetics \| www.plosgenetics.org Part II. Parameters of the BF Model and Their Roles for CO Patterns The original BF model included (N) and (s) and assumed a given chromosome has the same number of precursors in different nuclei but assumed that precursors are distributed randomly along a given chromosome (B = 1, E = 0). The latter assumption is likely not the case in vivo. Experimental evidence in several organisms shows that precursors tend to be evenly spaced, sometimes dramatically (e.g. [10–14]). And anecdotal evidence further suggests that the number of precursors tends to be quite constant for a given bivalent in different nuclei (e.g. [10,12,15,16]). We further note that evenly-spaced precursors have not been taken into account in any previous quantitative model for CO patterning (e.g. [17–20]). Variations in the nature of the precursor array can affect CoC relationships, with interactive effects, particularly at low values of (N) (Figure 3ABC; Figure S2). The BF simulation program now also includes a feature which permits the density of precursors to be varied along the chromosome in a desired pattern. This feature is useful for modeling effects such as the paucity of DSBs in centromeric regions, or other regional and domainal variations in DSB levels along chromosomes (e.g. [5,6,10,21–26]). Application to grass- hopper CO patterns is described below. CO patterning parameters: DDF (S), interference distance (L), precursor reactivity (A) and end effects (cL, cR). All of these patterning parameters are present in the original BF model. Detailed explanations of their significance are as follows: CO patterning parameters: DDF (S), interference distance (L), precursor reactivity (A) and end effects (cL, cR). All of these patterning parameters are present in the original BF model. Detailed explanations of their significance are as follows: The BF model assumes that precursors do not turn over, i.e. that a precursor either develops into a CO or into some other type of product, without being recycled to serve again as a precursor in another position. This assumption has not been directly tested. However, it seems reasonable because precursors are known to be highly evolved multi-protein complexes whose numbers can be constant over long periods of time (e.g. [12]). (S,Smax) versus (L): The outcome of the patterning process is determined primarily by two basic parameters: the strength of the (CO)-designation driving force (global stress or the DDF), as given by parameter (S); and the distance over which interference spreads, given by parameter (L). January 2014 \| Volume 10 \| Issue 1 \| e1004042 Part II. Parameters of the BF Model and Their Roles for CO Patterns For simulation purposes, the value of (S) is progressively increased to a specified maximum (Smax). A first, most reactive, precursor goes critical to give a CO-designation with The BF model also assumes that the entire precursor array is established prior to CO-designation (or essentially so). This is January 2014 \| Volume 10 \| Issue 1 \| e1004042 4 Logic and Modeling of Meiotic Crossover Patterns Figure 3. Variations in the precursor array can alter CO patterns. BF simulations were carried out under a set of ‘‘standard conditions’’ except for variations in the parameter of interest as indicated. Panels (A), (B) and (C) illustrate the effects of variations in N, E and B. Standard simulation parameter values are the same as the best-fit values for yeast Chromosome XV (Table 2). Parameter (E) is the standard deviation of the average inter- precursor distance and the corresponding evenness level is also shown by the shape (n) of gamma distribution in the program output. For E = 0, 0.6 and 0.7, the corresponding n = 1, 2.4 and 4.2, respectively. Parameter (B) is set by using the binomial distribution, in which with an average number of precursors on each bivalent (a constant mean), the distribution of the number of precursors among bivalents (probability of success for each trial) can be adjusted by changing the number of the trials. doi:10.1371/journal.pgen.1004042.g003 Figure 3. Variations in the precursor array can alter CO patterns. BF simulations were carried out under a set of ‘‘standard conditions’’ except for variations in the parameter of interest as indicated. Panels (A), (B) and (C) illustrate the effects of variations in N, E and B. Standard simulation parameter values are the same as the best-fit values for yeast Chromosome XV (Table 2). Parameter (E) is the standard deviation of the average inter- precursor distance and the corresponding evenness level is also shown by the shape (n) of gamma distribution in the program output. For E = 0, 0.6 and 0.7, the corresponding n = 1, 2.4 and 4.2, respectively. Parameter (B) is set by using the binomial distribution, in which with an average number of precursors on each bivalent (a constant mean), the distribution of the number of precursors among bivalents (probability of success for each trial) can be adjusted by changing the number of the trials. doi:10.1371/journal.pgen.1004042.g003 Figure 3. Variations in the precursor array can alter CO patterns. Part II. Parameters of the BF Model and Their Roles for CO Patterns BF simulations were carried out under a set of ‘‘standard conditions’’ except for variations in the parameter of interest as indicated. Panels (A), (B) and (C) illustrate the effects of variations in N, E and B. Standard simulation parameter values are the same as the best-fit values for yeast Chromosome XV (Table 2). Parameter (E) is the standard deviation of the average inter- precursor distance and the corresponding evenness level is also shown by the shape (n) of gamma distribution in the program output. For E = 0, 0.6 and 0.7, the corresponding n = 1, 2.4 and 4.2, respectively. Parameter (B) is set by using the binomial distribution, in which with an average number of precursors on each bivalent (a constant mean), the distribution of the number of precursors among bivalents (probability of success for each trial) can be adjusted by changing the number of the trials. doi:10.1371/journal.pgen.1004042.g003 whereas variations in (Smax) primarily affect the ED (Figure 4CDEF). The practical implication for best-fit BF simulations is that the values of these two parameters can be specified independently (Figure S4). accompanying interference; the level of (S) is then further increased, giving a next designation at the next most reactive position. This process is increased up to a final desired level. This procedure gives sequential CO-designations. The higher the final maximum value, (Smax), the more CO-designations. Interference arises instantaneously after each designation, reducing the probability that affected precursors can respond to the driving force over the specified distance (L). We note that, in vivo, (S) and (Smax) could take a variety of forms. (i) The value of Smax could potentially be defined by the time available for CO-designation. Interestingly, in Drosophila, the presence of a structural chromosome heterozygosity (deletion or inversion) results in a delay in meiotic progression and an increase in the number of COs without loss of CO interference [28,29]. This constellation of phenotypes could be modeled by an increase in Smax. (ii) CO designation could occur sequentially without any progressive increase in the DDF, simply because different precursor complexes will tend to undergo designations sooner or later in relation to their intrinsic reactivities, up to the maximum number specified by Smax. The final overall pattern of COs reflects the balance between the CO-designation driving force (DDF) and the interference distance, i.e. the values of (Smax) and (L). January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org Part II. Parameters of the BF Model and Their Roles for CO Patterns Correspondingly, a change in the value of either parameter can confer a similar alteration in CoC relationships and the ED (Figure 4AB). Higher (Smax) or lower (L) permits more COs to occur at shorter inter- interval distances, thus shifting the CoC curve to the left. Concomitantly, the overall level of COs increases. Lower (Smax) or higher (L) has the opposite effects. The ED changes commensurately. Also: in the BF model, the strength of the interference signal decays exponentially with distance from the CO-designation site (Figure 1A) in accord with the way in which stress redistributes in the beam-film system upon which the mathematical expressions are based [3,4] (Figure 1A legend). This decay relationship can be altered in the simulation program. However, we have found no need to do so thus far (e.g. below). To a considerable extent, opposing variations in the two parameters can compensate for one another (Figure 4CD). Nonetheless, in most cases, the effects of variations in (L) and (Smax) can be distinguished. The primary target of (L) is inter- CO communication, with the number of COs affected as a secondary consequence. The primary target of the DDF (Smax) is the number of COs, not inter-CO communication, with inter- CO relationships affected as a secondary consequence. Corre- spondingly, variations in (L) primarily affect CoC relationships Finally, the value of (Smax) actually incorporates the combined effects of the driving force and the sensitivity of precursors to that force. Similarly, the value of (L) incorporates the combined effects of the strength of the interference signal (as it dissipates with January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 5 Logic and Modeling of Meiotic Crossover Patterns Figure 4. Variations in the L and/or Smax can alter CO patterns. All panels: BF simulations as in Figure 3. Panels (A) and (B) illustrate the effects of variations in L or Smax, respectively. Panels (C) and (D) illustrate the fact that very similar CoC and ED relationships can be achieved by appropriate different combinations of L and Smax, but with a differential response of CoC relationships to changes in L (C) and of ED relationships to changes in Smax (D). These differential responses are further documented in Panels (E) and (F). doi:10.1371/journal.pgen.1004042.g004 Figure 4. Variations in the L and/or Smax can alter CO patterns. All panels: BF simulations as in Figure 3. Part II. Parameters of the BF Model and Their Roles for CO Patterns Panels (A) and (B) illustrate the effects of variations in L or Smax, respectively. Panels (C) and (D) illustrate the fact that very similar CoC and ED relationships can be achieved by appropriate different combinations of L and Smax, but with a differential response of CoC relationships to changes in L (C) and of ED relationships to changes in Smax (D). These differential responses are further documented in Panels (E) and (F). doi:10.1371/journal.pgen.1004042.g004 chromosome to some object. In BF simulations, clamping is defined by parameter (c), which can be specified individually at each chromosome end (cL and cR). The default case, fully clamped, is c = 1. distance) and the sensitivity of precursors to that signal. Put another way: any difference that can be modeled by a change in (Smax), or a change in (L) could, potentially, reflect a change either an actual change in the altered feature or a change in the ability of precursors to sense that feature. Other information must be brought to bear to distinguish the two types of effects. In a mechanical model, the chromosome end could alternatively be free in space, i.e. would be ‘‘unclamped’’. Since such a free end cannot support stress, it would behave as if it already had experienced a CO, i.e. with an interference signal having spread inward. The result would be a decreased probability that COs will occur near that end. In this case, c = 0. Intermediate situations can also occur. Thus, c(L) and c(R) can take any value between 0 and 1. (A): A third patterning parameter, (A), describes the dose/ response relationship between precursor sensitivity (s) and the local stress/DDF level at the corresponding position (i.e. the value of s as modified by the effects of any interference signals that have emanated across that position). Parameter (A) can have one of four possible values. In two cases (A = 1, 2), reactivity varies directly with (s); in the other two cases (A = 3, 4), reactivity varies inversely with (s) (Table 1). Variations in (A) can affect CO patterns (Figure 5A). As a practical matter, specification of (cL) and (cR) permits more accurate modeling of in vivo patterns where end effects are prominent. For example, many organisms exhibit a tendency for COs to occur near the ends of chromosomes whereas others do not. Such effects tend to emerge when chromosome ends are clamped (e.g. Part II. Parameters of the BF Model and Their Roles for CO Patterns Figure 5B middle, bottom). [Notably, however, genetic variations that result in paucities or excesses of DSBs near ends (e.g. [6,10,21–26,30]) should be modeled by variations in the precursor density (above) rather than as effects on interference]. Clamping: c(L) and c(R): Special considerations apply to interference at chromosome ends. These effects are incorporated into BF simulations by ‘‘clamping’’ parameters (cL and cR). In the absence of any other consideration, a terminal region will behave the same way as any other region of the chromosome with respect to its response to the DDF (Smax), the interference signal (L) and precursor reactivity (A). The interesting consequence of this effect, not regularly appreciated, is that there will automatically be intrinsic tendency for the ends to exhibit higher frequencies of COs, because these regions will be subjected to interference signals emanating in from only one direction (i.e. from internal regions of the chromosomes and not from regions ‘‘beyond’’ the end of the chromosome) (Figure 5B top). In a mechanical model, this ‘‘default’’ situation is achieved by ‘‘clamping’’ the end of the Variations in (cL) and (cR) primarily affect the distribution of COs along the length of a bivalent but also have secondary effect on overall CoC and ED relationships (Figure 5BC). Maturation efficiency (M). A precursor that undergoes designation may not mature efficiently into the signal used to define designations experimentally. This situation, defined by the value of the parameter (M), occurs in diverse mutant situations. If maturation is less than 100% efficient, the initial array of CO- Maturation efficiency (M). A precursor that undergoes designation may not mature efficiently into the signal used to define designations experimentally. This situation, defined by the value of the parameter (M), occurs in diverse mutant situations. If maturation is less than 100% efficient, the initial array of CO- January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 6 Logic and Modeling of Meiotic Crossover Patterns Figure 5. Variations in A, c or M can alter CO patterns. All panels: BF simulations as in Figure 3. (A) Effects of variations in precursor reactivity relationships. (B, C) Effects of variations in end clamping status on the distribution of COs along the chromosome (Panel B) and on CoC and ED relationships (Panel C). (D) Variations in maturation efficiency (M) do not affect CoC relationships but do affect the ED. doi:10.1371/journal.pgen.1004042.g005 Figure 5. Part II. Parameters of the BF Model and Their Roles for CO Patterns Variations in A, c or M can alter CO patterns. All panels: BF simulations as in Figure 3. (A) Effects of variations in precursor reactivity relationships. (B, C) Effects of variations in end clamping status on the distribution of COs along the chromosome (Panel B) and on CoC and ED relationships (Panel C). (D) Variations in maturation efficiency (M) do not affect CoC relationships but do affect the ED. doi:10.1371/journal.pgen.1004042.g005 Budding yeast. Yeast provides a favorable system for analysis of CO patterning in general, and for application and evaluations of BF modeling in particular, for several reasons. First, in this organism, the sites of patterned (‘‘interfering’’) COs are marked by foci of ZMM proteins Zip2 or Zip3 along pachytene chromosomes ([31]; Materials and Methods; Figure 6A). Zip2/3 foci mark CO sites very soon after they are designated (and independent of the two immediate downstream consequences of CO-designation, i.e. formation of the first known CO-specific DNA intermediate and nucleation of SC formation). Thus, effects of CO maturation defects on CO patterns are minimized. Second, the positions of these foci, and thus of CO-designations, can be determined along any specifically-marked chromosome to the resolution of fluores- cence microscopy (e.g. Figure 6B). Also, Zip2/3 focus positions can be determined even in the absence of SC [31]. Third, the average number of Zip3 foci (COs) per bivalent varies over a significant range (e.g. Table 2). The lowest value described thus far, ,2 for Chromosome III, is close to the number of COs seen in some other organisms (e.g. mouse or human or grasshopper) and thus provide appropriate models for such cases. At the highest values, ,7, multiple foci (COs) occur quite evenly along the length of the chromosome, which is very useful for revealing general patterns. Fourth, analysis of chromosomes in hundreds of nuclei is readily achievable, thus readily providing sufficiently large data sets for both wild-type and diverse mutant cases. Correspondingly, CoC and ED relationships can be determined extremely accurately. For a given data set, CoC values at each inter-interval distance can vary over a significant range (Figure 6C). This designated events undergoes random subtraction such that the final array of detected events reflects only a subset of the original designation array. Variations in (M) do not affect CO patterns. Part II. Parameters of the BF Model and Their Roles for CO Patterns Since maturation efficiency only affects CO status after patterning is established, a decrease in (M) only shifts ED relationships to lower CO numbers, with no/little effect on CoC relationships (Figure 5D). Part III. BF Simulations Accurately Describe Experimental Data Sets PLOS Genetics \| www.plosgenetics.org Part III. BF Simulations Accurately Describe Experimental Data Sets Application of the BF model to an experimental data set permits the identification of a set of parameter values for which simulated CO patterns most closely match those observed experimentally (general strategy described and illustrated in Figure S4). Best-fit simulation analysis for data sets from yeast, Drosophila, tomato and grasshopper demonstrates that the logic and mathematics of the BF model can describe experimental CO patterns with a high degree of quantitative accuracy. This conclusion is evident in descriptions of CoC and ED patterns as described in this section (III). Additional evidence is provided by applications and extensions of BF simulation analysis to CO homeostasis, the obligatory CO and non-interfering COs as described in sections IV–VI. Inspection of experimental CoC relationships has also provided new information regarding the metric of CO interference in tomato and the fact that interference spreads across centromere regions (in grasshopper, as previously described, and also in tomato and yeast). January 2014 \| Volume 10 \| Issue 1 \| e1004042 7 PLOS Genetics \| www.plosgenetics.org Logic and Modeling of Meiotic Crossover Patterns Table 2. Experimental characteristics and best-fit simulation parameters. Experimental data Best-fit simulation values Organisms\ Chromosomes Chrom. length# CO/Biv % Zero- CO Biv LCoC LBF Smax A cL/cR N@ E B M Mbp mm Fraction mm S.cerevisiae Chr III 0.32 1.2 1.8 1 0.3 0.25 0.3 3.5 1 0.85/0.85 6 0.6 1 1 Chr XV 1.05 3.2 4.7 0 0.3 0.1 0.3 3.5 1 0.85/0.85 13 0.6 1 1 Chr IV 1.53 4.8 6.8 0 0.3 0.075 0.3 3.5 1 0.85/0.85 19 0.6 1 1 Chr XIV (BR) 0.79 2.1 3.7 0 0.3 0.15 0.3 3.5 1 0.85/0.85 11 0.6 1 1 Chorthippus L3 1800 94 2.2 0 28 0.3 28 2.3 2 1.1/0.96 14 0 1 1 D.Melanogaster X 22 15 1.4 5 6 0.2 6 2.8 1 0.65/0.5 7 0.4 0.5 1 S. lycopersicum Chr 2–4 70 22 1.4 0 11 0.65 14 1.05 3 1.7/1.1 25 0.6 1 1 70 17 1.2 0 11 0.8 14 1.05 3 1.7/1.1 20 0.6 1 1 by the same set of optimal parameter values with the exception of the predicted number of precursors, which increases with chromosome length as could be expected (Table 2). variation is largely due to sampling variation because it is significantly reduced in BF simulation data sets that involve 5000 chromosomes rather than the 300 that usually comprises a typical data set. The genomic chromosome length and SC length are from the following sources. (1) S.cerevisiae: Saccharomyces Genomic Database (http://www.yeastgenome.org) and SC length are from this study and [31]. (2) Chorthippus: its C value = ,10 pg (http://www.genomesize.com/results.php?page=1), and 1 pg = 978 Mb [70]; thus its genome size is ,9780 Mb, chromosome L3 SC = 94 mm (0.189 of total SC length [71]), thus the predicted L3 is ,1800 Mb assuming the SC length is proportional to genome size in this organism. (3) D. Melanogaster: Fly Database (http://flybase.org/). SC length is based on [72,73]. (4) S. lycopersicum: Sol Genomics Network (http://solgenomics.net). SC length is from [38]. @The number of precursors is based on the following. (1) S. cerevisiae: Spo11 oligos and also microarray data [21,52]. (2) Chorthippus: from [3] (3) D. Melanogaster: based on [3,34,35] (4) S. lycopersicum: based on [38]. doi:10.1371/journal.pgen.1004042.t002 Part III. BF Simulations Accurately Describe Experimental Data Sets Despite this variation, which is present in all experimental data sets (below), the average CoC curve obtained from such an experiment is highly reproducible, as illustrated by the results of four independent experiments (Figure 6D–G). (2) The thus-defined best-fit value of (L), i.e. the distance over which the inhibitory interference signal spreads, is ,300 nm (LBF = ,0.3 mm) for all analyzed chromosomes (Table 2). This value of (LBF) turns out to correspond closely to ‘‘interference distance’’ as defined experimentally by CoC analysis in all cases (LCOC = ,0.3 mm; Table 2). Experimental CO patterns in wild-type SK1: We have defined Zip3 focus patterns along WT chromosomes III, IV and XV in the SK1 strain background. These chromosomes range in size from 330 to 1530 kb (Table 2). For CoC analysis, each chromosome was divided into 100 nm interval, a size that provides maximal accuracy (Figure S1). Inter-interval distances for CoC analysis are expressed in units of mm SC length (rationale above). The three analyzed chromosomes exhibit virtually overlapping CoC curves, with LCOC = ,0.3 mm (Figure 6F; Table 2). Along a pachytene chromosome of SK1, this distance corresponds to ,100 kb. The CoC curve remains less than 1 up to inter-interval distances corresponding to ,150 kb, in accord with the maximal distance over which interference is detected by genetic analyses (e.g. [32]). (3) For each chromosome, the number of precursors used for the best-fit simulation corresponds well to that described exper- imentally by analysis of DSBs and is approximately propor- tional to chromosome length [21]. (4) An optimal match between best-fit CoC curves and experimental data requires that precursors be relatively evenly spaced along the chromosomes (E = 0.6; (Figure 6IJ; Figure S5A–E), thus confirming and extending experimental evi- dence that yeast DSBs are evenly spaced (discussion in [13,14]; see also below). (5) Notably, also, for the shortest chromosome (III), an optimal match is obtained only if precursors are also assumed to occur in a relatively constant number along a given chromosome in different nuclei (B = 1; Figure S5F). At lower values of B, the frequency of zero-Zip3 focus chromosomes is higher than that observed experimentally because a significant fraction of chromosomes fail to acquire enough precursors to give at least one focus. This finding suggests that a given chromosome usually acquires the same (or nearly the same) number of precursors in every meiotic nucleus. #The genomic chromosome length and SC length are from the following sources. Part III. BF Simulations Accurately Describe Experimental Data Sets For longer chromosomes, the value of B is not very important (e.g. Figure 6K; further discussion below). (5) Notably, also, for the shortest chromosome (III), an optimal match is obtained only if precursors are also assumed to occur in a relatively constant number along a given chromosome in different nuclei (B = 1; Figure S5F). At lower values of B, the frequency of zero-Zip3 focus chromosomes is higher than that observed experimentally because a significant fraction of chromosomes fail to acquire enough precursors to give at least one focus. This finding suggests that a given chromosome usually acquires the same (or nearly the same) number of precursors in every meiotic nucleus. For longer chromosomes, the value of B is not very important (e.g. Figure 6K; further discussion below). The value of Zip3 foci as a marker for CO patterns is further confirmed by analysis of an mlh1D mutant. Since Zip3 foci mark CO sites shortly after they are designated, and long before the late step at which Mlh1 is thought to act (above), the mlh1D mutation should not reduce total Zip3 focus levels and should have no effect on CoC relationships for Zip3 foci. This expectation is confirmed (Figure 6H). BF analysis: CoC and ED data for all three analyzed chromosomes can be very closely matched by BF best-fit simulations (e.g. Figure 6I; best-fit simulations for SK1 Chromo- some IV and III and for chromosome XIV in BR are shown in Figure S5 A–D; parameter values in Table 2). Points of note include: (6) (6) Zip3 focus analysis reveals that the shortest yeast chromosome (III) has a significant number of zero-event chromosomes (1%) whereas the two longer chromosomes have much lower (1) Despite differences in absolute chromosome length, and numbers of Zip3 foci, all three chromosomes are described (1) Despite differences in absolute chromosome length, and numbers of Zip3 foci, all three chromosomes are described January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 8 Logic and Modeling of Meiotic Crossover Patterns Figure 6. Experimental and BF analysis of CO patterns in budding yeast. Panels (A, B): Experimental System. (A) Spread yeast pachytene chromosomes fluorescently labeled for SC component Zip1, CO-correlated foci of ZMM protein Zip3, and terminally labeled at the end of Chromosome XV by a lacO/LacI-GFP array. (B) Positions of Zip3 foci along a single Chromosome XV bivalent were defined as shown. Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns inter-interval distance. (G) Shows the EDs for four independent experiments in (D–F) and the average (in red) with standard error. (H) Compares the average CoC curve and ED for an mlh1D mutant (blue) with those for WT (black; average of averages from panels (F) and (G)). Both ED relationships and CoC relationships in the mutant are as WT since Mlh1 acts very late (text). Panels (I–K): BF simulations of CO patterns for Chromosome XV (data from average of averages in panels (F) and (G). (I) Best-fit simulation (red) versus experimental data (black). Best-fit simulation specifies relatively even spacing of precursors (E = 0.6) and a constant number of precursors along the chromosome in all nuclei (B = 1). Other parameter values are in Table 1. (J, K). Experimental data and best-fit simulation data (black and red, from panel (I)) are compared with simulation using the same parameter values as the best-fit simulation except that precursors are either randomly spaced (Panel J; E = 0; green) or Poisson distributed among chromosomes in different nuclei (Panel K; B = 0; blue). Best-fit simulations for Chromosome IV and III data and for chromosome XIV in BR are shown in Figure S5 A–D; parameter values in Table 1. Importantly, even spacing is important for the best fit in all cases. In contrast, constant and Poisson distributions give very similar matches to experimental data except for the case of Chromosome III, where constant distribution must be required to ensure a sufficiently low number of zero-CO chromosomes (Figure S5 EF). doi:10.1371/journal.pgen.1004042.g006 COs/bivalent and no detected zero-CO chromosomes (,0.07%; Figure 8A). numbers (no zero-focus chromosome has been detected among .1000 chromosomes analyzed) (e.g. Figure 6G). These values are recapitulated by BF simulations (Figure 6I; Figure S5EF; further discussion below). A prominent feature of the L3 bivalent is a severe paucity of COs in the centromere region (Figure 8B left panel). This feature presumably reflects a defect in occurrence of precursors (DSBs) in centromeric heterochromatin. Thus, for BF best-fit simulations, the precursor array was adjusted accordingly, to give a strong paucity of precursors in the centromere region (a ‘‘black hole’’; Figure 8B left; C). The ability of BF simulations to accurately describe yeast data is further supported by analysis of CO homeostasis and the ‘‘obligatory CO’’ as described below. Drosophila melanogaster. Part III. BF Simulations Accurately Describe Experimental Data Sets Panels (C–H): Experimental CO patterns for Chromosome XV (C) CoC and ED relationships for a single representative Chromosome XV data set reflecting CO Logic and Modeling of Meiotic Crossover Patterns Figure 6. Experimental and BF analysis of CO patterns in budding yeast. Panels (A, B): Experimental System. (A) Spread yeast pachytene chromosomes fluorescently labeled for SC component Zip1, CO-correlated foci of ZMM protein Zip3, and terminally labeled at the end of Chromosome XV by a lacO/LacI-GFP array. (B) Positions of Zip3 foci along a single Chromosome XV bivalent were defined as shown. Panels (C–H): Experimental CO patterns for Chromosome XV. (C) CoC and ED relationships for a single representative Chromosome XV data set reflecting CO positions defined along .300 bivalents (as in (B)). Average CoC curve (black line) shows LCOC = 0.3 mm. (D–G) CoC curves and EDs for four independent experiments like that in (C). (D) shows the four individual average CoC curves; data set from panel (C) in black. (E) shows the four curves from four independent data sets and their average (in red). (F) shows the average of the four average CoC curves with the standard error at each Figure 6. Experimental and BF analysis of CO patterns in budding yeast. Panels (A, B): Experimental System. (A) Spread yeast pachytene chromosomes fluorescently labeled for SC component Zip1, CO-correlated foci of ZMM protein Zip3, and terminally labeled at the end of Chromosome XV by a lacO/LacI-GFP array. (B) Positions of Zip3 foci along a single Chromosome XV bivalent were defined as shown. Panels (C–H): Experimental CO patterns for Chromosome XV. (C) CoC and ED relationships for a single representative Chromosome XV data set reflecting CO positions defined along .300 bivalents (as in (B)). Average CoC curve (black line) shows LCOC = 0.3 mm. (D–G) CoC curves and EDs for four independent experiments like that in (C). (D) shows the four individual average CoC curves; data set from panel (C) in black. (E) shows the four curves from four independent data sets and their average (in red). (F) shows the average of the four average CoC curves with the standard error at each January 2014 \| Volume 10 \| Issue 1 \| e1004042 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 9 Logic and Modeling of Meiotic Crossover Patterns CO interference was first discov- ered, and CoC analysis developed, by genetic analysis of Drosophila X chromosome (Introduction). These classical genetic data do provide for a quantitatively accurate CoC curve with LCOC = ,6 mm (Figure 7A), because the interval sizes are small enough and the data set is large enough [33]. The average number of COs per bivalent defined by experimental analysis is 1.44 with an unusually high level of zero-CO bivalents (5%) (Figure 7B). Recent studies of c-His2Av foci have defined a total of ,24 DSBs for the entire genome in Drosophila female meiosis [34]. Assuming that DSBs are proportional to genome length, this implies 6 DSBs for the X-chromosome (N = 6). A BF simulation with N = 6 can quantitatively match the classical Drosophila X chromosome data for all descriptors (Figure 7AB). Best-fit simulation requires differential clamping at the two ends, with less clamping at the centromeric end (Table 2). This feature provides for the experimentally-observed tendency for CO distributions to be shifted away from that end (illustrated in [35]). N = ,14 (Figure 8H). Solanum lycopersicum (tomato). CO patterns in tomato have been defined by analysis of Mlh1 foci ([38]; Lhuissier F.G. personal communication). Chromosomes in this organism exhibit a range of different pachytene bivalent lengths [38]. Bivalents comprise two groups, chromosomes 2–4 and 5–11, on the basis of longer and shorter SC lengths respectively (Figure S6A). We find that experimental CoC curves are significantly offset for the two groups when the metric of inter-interval distance is Mb (Figure 9A); in contrast, CoC curves for the two groups are superimposable when inter-interval distance is mm SC length (physical distance), with LCoC = 11 mm (Figure 9B). The ability of BF simulations to accurately describe Drosophila data is further supported by analysis of CO homeostasis and the ‘‘obligatory CO’’ as described below. Chorthippus bruneus (grasshopper). CO sites along the L3 bivalent of grasshopper have been defined by analysis of chiasmata in 1466 diplotene nuclei [36,37]. These data yield a CoC curve with LCoC = 28 mm and an ED with an average of 2.2 Chorthippus bruneus (grasshopper). CO sites along the L3 bivalent of grasshopper have been defined by analysis of chiasmata in 1466 diplotene nuclei [36,37]. Logic and Modeling of Meiotic Crossover Patterns (H e perimentall (black) ere compared b those gi en b BF Figure 8. Experimental and BF analysis of CO patterns in grasshopper (Chorthippus bruneus). (A, B) Experimental data. (A) CoC and ED relationships; (B) Distribution of COs along the bivalent for total COs (left) and for bivalents with either two or three COs with different colors for first, second (and third) COs from the left end of the bivalent. The centromeric region is labeled by a red bar. (C–H) BF simulation analyses. (C) Precursor density (frequency of precursors pre bivalent per interval specified for simulations, where the number of intervals = 17 as for CoC analysis) for BF simulations that used either an even distribution along the chromosome (blue; 2 Black hole) or a distribution where precursor levels decrease to zero over a region corresponding the paucity of COs in the centromeric region (red; + Black hole; centromeric region defined in Panel B). (D) BF simulations with best-fit parameter values (Table 2) using the two precursor distributions defined in (C), i.e. with or without the centromere region black hole. CoC and ED relationships are the same in both cases (D and E); the distribution of COs along the chromosomes are well-fit when the black hole is included (F) as compared to when it is not (G). (H) BF simulations were used to estimate the likely value of N. CoC relationships seen experimentally (black) were compared by those given by BF simulations that use all best-fit parameter values except that the value of N (precursors per bivalent), which was varied from N = 7 to N = 40 (colors). Left panel: CoC relationships match the experimental curve for any N$14 (i.e. all curves except gold and green which are N = 7 and N = 10). Right panel: ED relationships are best fit by N = 14 (compare red and black), with less good fits at lower and higher values (right side). doi:10.1371/journal.pgen.1004042.g008 Figure 8. Experimental and BF analysis of CO patterns in grasshopper (Chorthippus bruneus). (A, B) Experimental data. (A) CoC and ED relationships; (B) Distribution of COs along the bivalent for total COs (left) and for bivalents with either two or three COs with different colors for first, second (and third) COs from the left end of the bivalent. The centromeric region is labeled by a red bar. (C–H) BF simulation analyses. Logic and Modeling of Meiotic Crossover Patterns These data yield a CoC curve with LCoC = 28 mm and an ED with an average of 2.2 Physical distance has shown to be the appropriate metric for CO interference in mouse and Arabidopsis (above). The experimental results for tomato described above (Figure 9 AB) imply that this is also true in tomato. In addition, these data imply that the ratio of Mb to mm SC length is higher for shorter chromosomes than for longer chromosomes. This difference is explained, quantitatively, by two facts that: (i) heterochromatic DNA is much more densely packed along the SC than euchromatin and (ii) shorter chromosomes have a higher proportion of heterochromatin than euchromatin (and thus shorter SC lengths) (Figure S6). Figure 7. Experimental and BF analysis of CO patterns in Drosophila. (A, B). Experimental data (black) and BF simulations (red) for the D. melanogaster X chromosome. doi:10.1371/journal.pgen.1004042.g007 CoC and ED patterns for both groups of chromosomes are well- described by BF simulations (Figure 9CD). All parameters have the same values in both cases, including the interference distance (L) when expressed in mm SC length, except that precursor number varies with CO number/SC length (Figure 9CD legend). Interference spans centromeres in grasshopper, yeast and tomato. Previous analyses of interference have shown that CO interference is transmitted across centromeric regions [6,37– 42]. Correspondingly, CoC values for interval pairs that span centromeres are almost indistinguishable from those for pairs separated by the same distances that do not span centromeres Figure 7. Experimental and BF analysis of CO patterns in Drosophila. (A, B). Experimental data (black) and BF simulations (red) for the D. melanogaster X chromosome. doi:10.1371/journal.pgen.1004042.g007 PLOS Genetics \| www.plosgenetics.org January 2014 \| Volume 10 \| Issue 1 \| e1004042 10 Logic and Modeling of Meiotic Crossover Patterns Figure 8. Experimental and BF analysis of CO patter relationships; (B) Distribution of COs along the bivalent for to second (and third) COs from the left end of the bivalent. Th density (frequency of precursors pre bivalent per interval sp simulations that used either an even distribution along the ch over a region corresponding the paucity of COs in the ce simulations with best-fit parameter values (Table 2) using th black hole. CoC and ED relationships are the same in both ca hole is included (F) as compared to when it is not (G). Logic and Modeling of Meiotic Crossover Patterns Figure 9. Experimental and BF analysis of CO patterns in tomato. (A, B) Experimental CoC data for chromosomes with shorter or longer SC lengths (2–4 and 5–11, respectively). Curves for the two groups are offset when inter-interval distance is expressed in units of Mb (genomic distance) (Panel A) and are overlapping when distance is expressed in units of mm SC (physical distance) (Panel B). (C, D) BF best- fit simulations for both groups of chromosomes using the physical distance metric (mm SC length) and adjusting precursor number as required to give ,20 precursors per CO for each type of chromosome. This relationship is based on observation of ,280 Rad51 foci [69] and 15 Mlh1 foci genome wide [38]. Interestingly, this relationship further implies that N is proportional to SC length. doi:10.1371/journal.pgen.1004042.g009 accompanied by corresponding variations in the number of COs. When DSB levels are either reduced or increased, CO levels are not reduced or increased commensurately [15,34,43–46]. This phenomenon is referred to as CO homeostasis [43]. According to the BF model, CO homeostasis is dependent upon, and in fact is a direct consequence of, CO interference (Figure 11A), as proposed [43,46]. In the absence of interference, the probability that a precursor will give rise to a CO is a function only of its own intrinsic properties, independent of the presence/ absence of other precursors nearby. Thus, as the number of precursors decreases, the number of COs will decrease propor- tionately. In contrast, if interference is present, each individual precursor is subject to interference that emanates across its position from CO-designation events at neighboring positions. The lower the number of precursors, the less this effect will be. Thus, assuming a fixed level of CO interference, the frequency of COs per precursor will increase as the number of precursors decrease. Put another way: as the density of precursors decreases, the ratio of COs to precursors increases, even though there is no change in CO interference. Importantly, since CO homeostasis requires CO interference, its magnitude will also depend on the strength of CO interference as discussed below. In the BF model, CO homeostasis involves interplay between N and patterning parameters, e.g. L and Smax. Logic and Modeling of Meiotic Crossover Patterns This is basically because interference precludes CO-designation at all other precursor sites, thus rendering variations in CO density irrelevant. This latter situation has recently been documented for C.elegans [44,45]. Analogous effects can be seen for variations in Smax (not shown), in accord with the fact that both L and Smax play important roles for LCoC (above). Figure 9. Experimental and BF analysis of CO patterns in tomato. (A, B) Experimental CoC data for chromosomes with shorter or longer SC lengths (2–4 and 5–11, respectively). Curves for the two groups are offset when inter-interval distance is expressed in units of Mb (genomic distance) (Panel A) and are overlapping when distance is expressed in units of mm SC (physical distance) (Panel B). (C, D) BF best- fit simulations for both groups of chromosomes using the physical distance metric (mm SC length) and adjusting precursor number as required to give ,20 precursors per CO for each type of chromosome. This relationship is based on observation of ,280 Rad51 foci [69] and 15 Mlh1 foci genome wide [38]. Interestingly, this relationship further implies that N is proportional to SC length. doi:10.1371/journal.pgen.1004042.g009 BF simulations of CO homeostasis in experimental data sets can be used to evaluate the validity of the BF model. BF best-fit simulations for wild type experimental data sets provide specific predicted values of all BF parameters, including N, L and Smax (above). Given this starting point, BF simulations can then specifically predict how CO levels will vary if the level of DSBs (precursors) is decreased or increased. If the BF model accurately describes CO patterning, the best-fit simulation will accurately describe the experimental data. Application of this approach to budding yeast and Drosophila shows that the BF (Figure 10B). This is remarkable given the manifestly different structure in these regions. We further find that the same is true for budding yeast, based on Zip3 focus analysis (Figure 10A) and for tomato, from Mlh1 focus analysis (Figure 10C). Logic and Modeling of Meiotic Crossover Patterns (C) Precursor density (frequency of precursors pre bivalent per interval specified for simulations, where the number of intervals = 17 as for CoC analysis) for BF simulations that used either an even distribution along the chromosome (blue; 2 Black hole) or a distribution where precursor levels decrease to zero over a region corresponding the paucity of COs in the centromeric region (red; + Black hole; centromeric region defined in Panel B). (D) BF simulations with best-fit parameter values (Table 2) using the two precursor distributions defined in (C), i.e. with or without the centromere region black hole. CoC and ED relationships are the same in both cases (D and E); the distribution of COs along the chromosomes are well-fit when the black hole is included (F) as compared to when it is not (G). (H) BF simulations were used to estimate the likely value of N. CoC relationships seen experimentally (black) were compared by those given by BF simulations that use all best-fit parameter values except that the value of N (precursors per bivalent), which was varied from N = 7 to N = 40 (colors). Left panel: CoC relationships match the experimental curve for any N$14 (i.e. all curves except gold and green which are N = 7 and N = 10). Right panel: ED relationships are best fit by N = 14 (compare red and black), with less good fits at lower and higher values (right side). doi:10.1371/journal.pgen.1004042.g008 January 2014 \| Volume 10 \| Issue 1 \| e1004042 11 PLOS Genetics \| www.plosgenetics.org Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns CO homeostasis for a given condition can be defined quantitatively by BF simulations in which the value of N (which is the number of precursors per bivalent and thus corresponds to precursor density) is varied, with all other parameter values remaining constant. Homeostasis can be described by plotting, as function of (N), either the frequency of COs per precursor (CO/N) or the total number of COs per bivalent (which corresponds to CO density) (Figure 11BC). Alternatively, CO homeostasis can be seen from the perspective of a starting wild-type situation with the number of COs per bivalent at the wild type precursor level taken as the point of reference and variations in the number of COs and precursors expressed relative to those reference values (Figure 11D). The magnitude of CO homeostasis, i.e. the extent to which CO levels fail to respond to changes in precursor levels, will vary with the level of CO interference (above). This relationship can be described quantitatively by carrying out simulations for different values of (N) at different values of patterning parameter(s), illustrated here for variations in the interference distance (L) (Figures 11B–D). In the absence of interference, CO levels vary directly with precursor levels; the greater the interference distance, the less the change in CO levels with precursor levels. Notably, at very long interference distances, CO levels do not change at all with precursor levels. This is basically because interference precludes CO-designation at all other precursor sites, thus rendering variations in CO density irrelevant. This latter situation has recently been documented for C.elegans [44,45]. Analogous effects can be seen for variations in Smax (not shown), in accord with the fact that both L and Smax play important roles for LCoC (above). The magnitude of CO homeostasis, i.e. the extent to which CO levels fail to respond to changes in precursor levels, will vary with the level of CO interference (above). This relationship can be described quantitatively by carrying out simulations for different values of (N) at different values of patterning parameter(s), illustrated here for variations in the interference distance (L) (Figures 11B–D). In the absence of interference, CO levels vary directly with precursor levels; the greater the interference distance, the less the change in CO levels with precursor levels. Notably, at very long interference distances, CO levels do not change at all with precursor levels. Part IV. CO Homeostasis Experimental evidence has revealed that variations in the level of recombination-initiating double-strand breaks (DSBs) are not January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 12 Logic and Modeling of Meiotic Crossover Patterns Figure 10. Interference spreads across centromeres in yeast, grasshopper and tomato. CoC values for interval pairs that span centromere regions, and the corresponding average CoC curves, closely match the average CoC curves for all interval pairs for yeast (Panel A), grasshopper (Panel B) and tomato, both bivalent groups as described in Figure 9B (Panel C). doi:10.1371/journal.pgen.1004042.g010 Figure 10. Interference spreads across centromeres in yeast, grasshopper and tomato. CoC values for interval pairs that span centromere regions, and the corresponding average CoC curves, closely match the average CoC curves for all interval pairs for yeast (Panel A), grasshopper (Panel B) and tomato, both bivalent groups as described in Figure 9B (Panel C). doi:10 1371/journal pgen 1004042 g010 Figure 11. CO homeostasis and quantification by BF simula- tions. (A) CO homeostasis is the phenomenon that, as the level of DSBs (precursors) increases or decreases, there is a less than proportional change in the frequency of COs. The basis for this effect is illustrated. At lower (higher) precursor density, a given precursor will be less (more) likely to be subject to interference and thus more (less) likely to give a CO. These relationships further imply that the extent of CO homeostasis at a given precursor density will also vary with interference distance and the strength of the DDF (L and Smax) and other patterning features as reflected in LCoC (text). (B, C, D) CO homeostasis was modeled by BF simulations using standard parameter values (Figure 3) with the indicated variations in the number of precursors (N), and varying values for the interference distance (L). CO homeostasis can be viewed as a function of N and L, as an effect on the probability that a single precursor will become a CO (B); or as an effect on the total number of COs along a chromosome (C). Also, the values of N and of COs can be defined as 1 for the wild-type reference situation and the effects of varying N (as a percentage of the reference value) can be seen directly as effects on CO levels (also as a percentage of the reference value). Part IV. CO Homeostasis BF simulations should (and do) accurately predict CO homeostasis in interference-defective mutants: CO homeostasis relationships should be altered in mutants where interference is defective, in a predictable way according to the magnitude of the reduction (Figure 12ABC, grey lines above the curves describing the wild-type relationships). We will present elsewhere data showing that BF best-fit simulations for a yeast mutant specifically defective in interference accurately predict CO homeostasis relationships in that mutant, thus further supporting the validity of BF simulation analysis. Figure 11. CO homeostasis and quantification by BF simula- tions. (A) CO homeostasis is the phenomenon that, as the level of DSBs (precursors) increases or decreases, there is a less than proportional change in the frequency of COs. The basis for this effect is illustrated. At lower (higher) precursor density, a given precursor will be less (more) likely to be subject to interference and thus more (less) likely to give a CO. These relationships further imply that the extent of CO homeostasis at a given precursor density will also vary with interference distance and the strength of the DDF (L and Smax) and other patterning features as reflected in LCoC (text). (B, C, D) CO homeostasis was modeled by BF simulations using standard parameter values (Figure 3) with the indicated variations in the number of precursors (N), and varying values for the interference distance (L). CO homeostasis can be viewed as a function of N and L, as an effect on the probability that a single precursor will become a CO (B); or as an effect on the total number of COs along a chromosome (C). Also, the values of N and of COs can be defined as 1 for the wild-type reference situation and the effects of varying N (as a percentage of the reference value) can be seen directly as effects on CO levels (also as a percentage of the reference value). This approach considers the ‘‘densities’’ of precursors and COs rather than the absolute levels. The BF model quantitatively explains CO homeostasis in Drosophila: We analogously evaluated whether BF simulations accurately predict CO homeostasis relationships in Drosophila, which were defined experimentally by analysis of a fragment of chromosome 3 [34]. In Drosophila, as in yeast, CO-versus-DSB experimental data exhibit a very good match to the CO-versus-(N) relationships predicted using BF best-fit parameter values, uniquely and specifically at the best-fit value of (L) (Figure 12C). Part IV. CO Homeostasis This approach considers the ‘‘densities’’ of precursors and COs rather than the absolute levels. model can quantitatively predict experimental CO homeostasis patterns in both organisms. The BF model quantitatively describes CO homeostasis in yeast: We asked whether the BF model could quantitatively explain CO homeostasis along chromosomes XV and III as defined by Zip3 focus analysis (above). We determined experimentally the number of Zip3 foci that occur along the two test chromosomes in a series of mutants that are known to exhibit particular, defined decreases or increases in DSB levels. Reductions in DSB levels were provided by the hypomorphic alleles of DSB transesterase Spo11 used to originally define CO homeostasis [43]. An increase in DSB levels was provided by a tel1D mutation, which increases DSB levels without significantly altering CO interference ([13]; unpub- lished). The observed experimental relationships are described by the BF-predicted relationships (Figures 12AB, L = 0.3 mm). The robustness of BF simulations is further supported by analysis of CoC and ED relationships for Zip3 foci in each of the DSB mutant strains. Best-fit simulations of these data sets should occur at exactly the same values of all parameters except for the number of precursors (N), which should match that defined by experimental analysis of DSBs. Both of these predictions are fulfilled (Figure S7). These comparisons also reveal some interesting subtleties to DSB formation in the mutants (Figure S7). The robustness of BF simulations is further supported by analysis of CoC and ED relationships for Zip3 foci in each of the DSB mutant strains. Best-fit simulations of these data sets should occur at exactly the same values of all parameters except for the number of precursors (N), which should match that defined by experimental analysis of DSBs. Both of these predictions are fulfilled (Figure S7). These comparisons also reveal some interesting subtleties to DSB formation in the mutants (Figure S7). The BF model quantitatively explains CO homeostasis in Drosophila: We analogously evaluated whether BF simulations accurately predict CO homeostasis relationships in Drosophila, which were defined experimentally by analysis of a fragment of chromosome 3 [34]. In Drosophila, as in yeast, CO-versus-DSB experimental data exhibit a very good match to the CO-versus-(N) relationships predicted using BF best-fit parameter values, uniquely and specifically at the best-fit value of (L) (Figure 12C). Part IV. CO Homeostasis doi:10.1371/journal.pgen.1004042.g011 doi:10.1371/journal.pgen.1004042.g011 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 13 Logic and Modeling of Meiotic Crossover Patterns Figure 12. CO homeostasis in yeast, Drosophila, tomato and grasshopper. (A, B) CO homeostasis relationships were determined experimentally for yeast chromosomes XV and III (text; Figure S7); plotted (filled circles); and compared with the values predicted from BF simulations based on best-fit parameter values for the two chromosomes (L = 0.3 mm, black lines). Predictions for other values of L are shown for comparison (grey lines). Experimental data precisely match BF simulation predictions. (C) CO homeostasis relationships determined experimentally for Drosophila chromosome 3 were compared with the predictions of the corresponding BF simulation (L = 6 mm, black line). Predictions for other values of L are shown for comparison (grey lines). Experimental data precisely match BF simulation predictions. [Note: Drosophila analysis was carried out as follows. Variation in CO number as a function of DSB level was determined experimentally for a fragment of chromosome 3 [34]. We first defined the theoretical CO homeostasis curve (black line) for full length chromosome 3 by BF simulations using the same set of parameters defined for chromosome X except that the number of precursors was adjusted in proportion to relative chromosome length (chromosome 3 is 1.56the length of chromosome X). The CO levels observed experimentally on the chromosome 3 fragment at the different DSB levels were then adjusted to those expected for the full length chromosome under the assumption that CO frequency is proportional to chromosome length, as in all other cases (above). The resulting experimental values were then compared with the theoretical curve and its relatives constructed at varying values of L (grey lines).] doi:10.1371/journal.pgen.1004042.g012 Figure 12. CO homeostasis in yeast, Drosophila, tomato and grasshopper. (A, B) CO homeostasis relationships were determined experimentally for yeast chromosomes XV and III (text; Figure S7); plotted (filled circles); and compared with the values predicted from BF simulations based on best-fit parameter values for the two chromosomes (L = 0.3 mm, black lines). Predictions for other values of L are shown for comparison (grey lines). Experimental data precisely match BF simulation predictions. (C) CO homeostasis relationships determined experimentally for Drosophila chromosome 3 were compared with the predictions of the corresponding BF simulation (L = 6 mm, black line). Predictions for other values of L are shown for comparison (grey lines). Part IV. CO Homeostasis Experimental data precisely match BF simulation predictions. [Note: Drosophila analysis was carried out as follows. Variation in CO number as a function of DSB level was determined experimentally for a fragment of chromosome 3 [34]. We first defined the theoretical CO homeostasis curve (black line) for full length chromosome 3 by BF simulations using the same set of parameters defined for chromosome X except that the number of precursors was adjusted in proportion to relative chromosome length (chromosome 3 is 1.56the length of chromosome X). The CO levels observed experimentally on the chromosome 3 fragment at the different DSB levels were then adjusted to those expected for the full length chromosome under the assumption that CO frequency is proportional to chromosome length, as in all other cases (above). The resulting experimental values were then compared with the theoretical curve and its relatives constructed at varying values of L (grey lines).] doi:10.1371/journal.pgen.1004042.g012 The strength of CO homeostasis reflects the ratio of inter- precursor distance and interference distance (LCoC). It would be convenient to have a standard way of comparing different situations (e.g. different chromosomes, mutants or organisms) with respect to the ‘‘strength’’ of CO homeostasis. In principle, the strength of homeostasis should vary according to the ratio between the CO interference distance (LCoC) and the distance between adjacent precursors (roughly given by the number of precursors divided by mm SC length). If this ratio is higher homeostasis will be stronger because a greater fraction of precursors are within the interference distance and thus can be eliminated without effect; and in the limit, reduction of precursor density will have no effect whatsoever. Oppositely, if this ratio is lower; homeostasis will be weaker because a greater fraction of precursors are outside of the interference distance and thus, when eliminated, will directly reduce CO levels; and in the limit, when the interference distance is zero, CO homeostasis is absent. By this criterion, i.e. [LCOC/average inter-precursor distance from the data in Table 2], the strength of CO homeostasis is the same for all yeast chromosomes (0.3 mm/ 0.3 mm = 1); are essentially the same for yeast chromosomes as for the Drosophila X chromosome (6 mm/5 mm = 1.1); is significantly greater for grasshopper (28 mm/2.6 mm = 10.8); and is even greater for tomato chromosomes, both groups (11 mm/0.888 mm = 12.5). Systematic exploration of such relationships by BF simulations remains for future studies. Part IV. CO Homeostasis meiosis in all organisms, in at least one sex and usually both, the requisite physical connection is provided by the combined effects of a crossover between non-sister chromatids of homologs and connections between sister chromatids along the chromosome arms. Correspondingly, in such organisms, in wild-type meiosis, every bivalent almost always acquires at least one CO [47]. This first CO that is essential for homolog segregation is often referred to as the ‘‘obligatory CO’’. In fact, the obligatory CO is simply a biological imperative: the level of zero-CO chromosomes should be low. The CO patterning process, by whatever mechanism, must somehow explain this feature. In most situations, the frequency of zero-CO bivalents is extremely low (,1023), but higher frequencies also occur in certain wild-type situations as well as in certain mutants (below). In some models for CO patterning, the obligatory CO is ensured by a specific ‘‘added’’ feature of the patterning process (e.g. the King and Mortimer model; Discussion). In contrast, in the beam-film model, the requirement for one CO per bivalent is satisfied as an intrinsic consequence of the basic functioning of the process, as follows. In the BF model, the ‘‘obligatory CO’’ is independent of (L) and (E) and requires an appropriate combination of values for (Smax), (N), (B) and (M). In the BF model, the obligatory CO is ensured as an intrinsic consequence of all of the features that ensure occurrence of a first event; features that act later in the process are not relevant (Figure 13A). L and E: Variations in (L) and (E) have no effect on the level of zero-CO chromosomes because: (i) spreading inhibition of CO- designation (‘‘interference’’) only affects the number of COs after In the BF model, the ‘‘obligatory CO’’ is independent of (L) and (E) and requires an appropriate combination of values for (Smax), (N), (B) and (M). In the BF model, the obligatory CO is ensured as an intrinsic consequence of all of the features that ensure occurrence of a first event; features that act later in the process are not relevant (Figure 13A). Part V. The ‘‘Obligatory CO’’ L and E: Variations in (L) and (E) have no effect on the level of zero-CO chromosomes because: (i) spreading inhibition of CO- designation (‘‘interference’’) only affects the number of COs after L and E: Variations in (L) and (E) have no effect on the level of zero-CO chromosomes because: (i) spreading inhibition of CO- designation (‘‘interference’’) only affects the number of COs after Regular segregation of homologs to opposite poles at the first meiotic division requires that they be physically connected. During January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 14 Logic and Modeling of Meiotic Crossover Patterns Figure 13. How is a low level of zero-CO bivalents ensured by the beam-film model? (A) BF simulations were carried out under a set of ‘‘standard conditions’’ (as described in Figure 3), except that Smax = 5, N = 8, while the value of one parameter was systematically varied as indicated in each panel. The frequency of zero-CO chromosomes as a function of the value of the varied parameter is plotted. Variations in L and E have no effect; variations in Smax, N, B and M all have effects. (B,C). A given frequency of zero-CO chromosomes can be achieved by diverse constellations of parameter values that play off against one another. This situation is illustrated by BF simulations under above conditions where the frequency of zero-CO chromosomes is determined over a range of combinations of values of two parameters. This interplay is illustrated for combinations of precursor number (N) and either parameter Smax (Panel B) or parameter B (Panel C). The lower the number of precursors, (N), the higher the Smax needed to ensure that at least one will be sensitive enough to undergo CO designation and the more important it will be for precursors to occur at the same average number along the bivalent in every nucleus, thus minimizing the probability that a bivalent will have too-few precursors (higher B) (text). doi:10 1371/journal pgen 1004042 g013 Figure 13. How is a low level of zero-CO bivalents ensured by the beam-film model? (A) BF simulations were carried out under a set of ‘‘standard conditions’’ (as described in Figure 3), except that Smax = 5, N = 8, while the value of one parameter was systematically varied as indicated in each panel. Part V. The ‘‘Obligatory CO’’ The frequency of zero-CO chromosomes as a function of the value of the varied parameter is plotted. Variations in L and E have no effect; variations in Smax, N, B and M all have effects. (B,C). A given frequency of zero-CO chromosomes can be achieved by diverse constellations of parameter values that play off against one another. This situation is illustrated by BF simulations under above conditions where the frequency of zero-CO chromosomes is determined over a range of combinations of values of two parameters. This interplay is illustrated for combinations of precursor number (N) and either parameter Smax (Panel B) or parameter B (Panel C). The lower the number of precursors, (N), the higher the Smax needed to ensure that at least one will be sensitive enough to undergo CO designation and the more important it will be for precursors to occur at the same average number along the bivalent in every nucleus, thus minimizing the probability that a bivalent will have too-few precursors (higher B) (text). d i 10 1371/j l 1004042 013 (A): The more likely a precursor is to give a CO-designation in response to a particular local level of interference, the lower will be the frequency of zero-CO chromosomes. The frequency of zero- CO chromosomes is lowest for A = 4 and increases progressively for A = 3, 2 and 1 (not shown). the first CO designation; and (ii) the distribution of multiple events along a chromosome also comes into play only after the first designation event has occurred. In contrast, essentially all other basic BF parameters are important in ensuring a low level of zero- CO chromosomes: Smax, N and B: The frequency of zero-CO chromosomes will be minimized if every chromosome has at least one precursor that is adequately sensitive to the DDF. This effect can be favored by either (i) higher Smax; (ii) higher N; or (iii) higher B. Higher Smax means that a higher fraction of sites in a particular precursor array will be adequately sensitive. Higher N means that there will be more sites and there is a higher chance that an adequately- sensitive precursor will be present. Higher B means a reduced probability that a bivalent will have a lower-than-average number of precursors and thus a lower-than-average chance for an adequately-sensitive precursor to be present. Evolutionary implications. Logic and Modeling of Meiotic Crossover Patterns Notably, also, in both Drosophila and yeast, additional mechanisms exist which complement the patterned CO system to ensure regular homolog segregation. Drosophila exhibits robust CO-independent segregation [48] and, in budding yeast, the significant level of non-interfering COs also ensure disjunction. For example: 1% of chromosome III’s exhibit zero Zip3 foci whereas only 0.1% of chromosome IIIs exhibit no COs as defined genetically (Hunter N and Bishop D, personal communication). Notably, also, in both Drosophila and yeast, additional mechanisms exist which complement the patterned CO system to ensure regular homolog segregation. Drosophila exhibits robust CO-independent segregation [48] and, in budding yeast, the significant level of non-interfering COs also ensure disjunction. For example: 1% of chromosome III’s exhibit zero Zip3 foci whereas only 0.1% of chromosome IIIs exhibit no COs as defined genetically (Hunter N and Bishop D, personal communication). Application to experimental data: When the obligatory CO appears to be ‘‘missing’’, in wild-type meiosis. In most organisms, for most chromosomes, in wild-type meiosis the observed frequency of zero-CO chromosomes is ,0.1% (e.g. Table 1). However, zero-CO chromosomes occur at significantly higher levels on the Drosophila X chromosome (5%) and on yeast chromosome III (1% by Zip3 foci). Yeast: the importance of N: BF analysis suggests that, in yeast, the feature responsible for the high level of zero-Zip3 focus chromosomes along chromosome III is simply a paucity of precursors and thus is a function of its diminutive size, per se. In this organism, patterns on all analyzed chromosomes can be described by the same set of BF parameters with the exception of (N), which varies roughly in proportion to chromosome length as defined in mm SC (above). The high level of zero-Zip3 focus chromosomes along chromosome III, relative to other chromosomes, is thus solely a reflection of the fact that it is much shorter than other chromosomes (N = 6 versus N$13; Table 2). This conclusion is directly and strongly supported by experimental analysis of Zip3 foci along chromosome III in mutants where precursor levels are gradually decreased or increased (by alterations in DSB levels; described below). In such mutants, the frequencies of zero-Zip3 focus chromosomes are commensurately increased or decreased. In mutants with relative DSB levels of 1.5, 1 (WT), 0.8 and 0.7, the frequencies of zero-CO chromosomes are, respectively, ,0.003; 0.01; 0.02; and 0.08 (Figure S7B). Logic and Modeling of Meiotic Crossover Patterns Corre- spondingly, there are multiple indications that Mlh1 acts very late in recombination, for maturation of dHJs to COs [50,51]. Analysis of yeast chromosome III further shows that the best-fit simulation requires that precursors occur in a constant number per bivalent in all nuclei (B = 1; frequency of zero-Zip3 focus bivalents = 0.01). If precursors are Poisson distributed among bivalents in different nuclei (B = 0), the frequency of zero-Zip3 focus bivalents increases to 0.04 (Figure S5). This comparison not only suggests that DSBs/precursors always occur at the same number along a different chromosome but provide a rationale for the existence of this feature which, in the general case, is essential to minimize the frequency of zero-CO bivalents along short chromosomes. This phenotype could also be conferred by genetic shortening of a chromosome. A reduction in the Mb length of the chromosome will decrease the precursor number (N) without alternation of any other properties and thus could push N below the minimum necessary threshold. Such an effect could explain the high level of zero-CO chromosomes seen when plant chromosomes are shortened by centric fission [47] and when yeast chromosome III is reduced in length (Hunter N., and Bishop, D.K., personal communication). Drosophila X chromosome: The high level of zero-CO bivalents for the Drosophila X chromosome is recapitulated in BF simulations without adding any unusual features (Figure 7B), suggesting that there is nothing remarkable about this chromo- some. Further, this chromosome the same number of precursors as yeast chromosome III (N = 6), suggesting that here, too, the fact that the chromosome is ‘‘too short’’ could be an important factor for the high level of zero-CO chromosomes. In accord with this possibility, Drosophila chromosome 3 is 50% longer than chromosome X; and BF simulation analysis shows that if X chromosome length is increased by 50%, with a proportional increase in precursors (to N = 9), and without any change in any other parameter, the frequency of zero-CO would decrease to 0.008. IF2 OC2: This phenotype could be conferred in several ways. One example would be a defect early in the recombination process that eliminate the ability of recombination precursors to both generate COs and generate/respond to the CO interference signal. Implications: The existence of IF+ OC2 mutants has some- times been cited as evidence for the existence of a specific feature, separate from interference, that ‘‘ensures the obligatory CO’’. Logic and Modeling of Meiotic Crossover Patterns Application to experimental data: When the obligatory CO appears to be ‘‘missing’’ in mutant meiosis. Mutant phenotypes that affect interference and/or the ‘‘obligatory CO’’ could fall into three different categories [47]: IF2 OC+: The BF model predicts that mutants with decreased CO interference, as defined specifically by decreased L, will show no defect in formation of the first (obligatory) CO, i.e. will show the same level of zero-CO chromosomes as that seen in WT meiosis. This phenotype, ‘‘IF2 OC+’’, not previously reported, is now described by observations in yeast to be presented elsewhere. IF+ OC2: In certain mutants, the level of zero-CO bivalents is increased but interference is unaltered. In the context of the BF model, this phenotype could arise from several types of defects. For example: this phenotype is expected for mutants with altered recombination biochemistry such that CO designation is normal but CO maturation is inefficient, i.e. M,1 (Figure 5A). For example, mutants lacking Mlh1 exhibit reduced levels of COs and significant levels of zero-CO chromosomes but relatively normal CO interference, as shown by genetic/chiasma analysis for several organisms [49–51]. On the other hand, by Zip3 focus analysis, both CoC and ED in a yeast mlh1D are essentially the same as in WT (Figure 6H), implying that CO-designation is normal. Thus the obligatory CO defect seen by genetic/chiasma analysis is specifically attributable to a maturation defect (M,1). Corre- spondingly, there are multiple indications that Mlh1 acts very late in recombination, for maturation of dHJs to COs [50,51]. IF+ OC2: In certain mutants, the level of zero-CO bivalents is increased but interference is unaltered. In the context of the BF model, this phenotype could arise from several types of defects. For example: this phenotype is expected for mutants with altered recombination biochemistry such that CO designation is normal but CO maturation is inefficient, i.e. M,1 (Figure 5A). For example, mutants lacking Mlh1 exhibit reduced levels of COs and significant levels of zero-CO chromosomes but relatively normal CO interference, as shown by genetic/chiasma analysis for several organisms [49–51]. On the other hand, by Zip3 focus analysis, both CoC and ED in a yeast mlh1D are essentially the same as in WT (Figure 6H), implying that CO-designation is normal. Thus the obligatory CO defect seen by genetic/chiasma analysis is specifically attributable to a maturation defect (M,1). Logic and Modeling of Meiotic Crossover Patterns The existence of IF2 OC2 mutants has sometimes been cited as evidence that interference is required for the ‘‘obligatory CO’’. Both of these phenotypes have alternative explanations in the context of the beam-film model. In contrast, the existence of IF2 OC+ mutants is specifically predicted by the beam-film model. However: the number of precursors cannot be the only relevant feature, because the level of zero-CO chromosomes in Drosophila is higher than that for yeast chromosome III even though they both have N = 6. This difference could be attributable in part to less regular distribution of precursors among chromosomes in Drosophila (B = 0.5 versus B = 1 in yeast) and to a lower DDF level (Smax = 2.8 versus 3.5 in yeast [Table 2]). Part V. The ‘‘Obligatory CO’’ The above considerations imply that, according to the BF model, the ‘‘obligatory CO require- ment’’ will be met in any given organism because the relevant features have been coordinately tuned by evolution into a combination that ensures a low level of zero-CO chromosomes. That is: a suitably low level of zero-CO chromosomes can be achieved by a variety of combinations, with more and less favorable values of different parameters in different cases. Interplay between pairs of parameters is illustrated for various combinations of (Smax) and (N) (Figure 13B) and for various combinations of (N) and (B) (Figure 13C). The lower the number of precursors, (N), the higher the Smax needed to ensure that at least one will be sensitive enough to undergo CO designation and the more important it will be for precursors to occur in a constant level along the bivalent in every nucleus (higher B). (M): Even if the CO patterning process ensures the occurrence of at least one CO per chromosome, a defect in maturation of CO- designated interactions to detectable COs will tend to counteract that effect, converting chromosomes with one (or a few) COs to chromosomes with zero-COs. January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 15 PLOS Genetics \| www.plosgenetics.org Part VII. Variations in Diverse BF Parameters Alter the Value of the Gamma Shape Parameter Many studies of CO interference characterize CO patterns by defining a gamma distribution that best describes an experimen- tally observed distribution of the distances between adjacent COs, often with the assumption (implicit or explicit) that a higher value of the gamma shape parameter (n) corresponds to ‘‘stronger’’ CO interference (e.g. [11]). We have examined the way in which (n) varies as a function of changes in the values of several BF parameters. Variations in L or Smax increase or decrease the value of (n) in correlation with increased or decreased LCOC and in opposition to the average number of COs per bivalent (Figure 15AB, compare green line and blue/pink distributions with red and black lines). This is the pattern expected for a change in the ‘‘strength of interference’’. In contrast, the value of (n) is also altered by variations in M or N, which have little or no effect on LCOC; moreover, the change in (n) co-varies with the change in the average number of COs per bivalent (Figure 15CD, compare green line and blue/pink distributions with red and black lines). The BF model thus implies that a change in the value of (n), e.g. in a mutant as compared to wild type, may or may not imply a change in the patterning process per se. However, comparison of the variation in (n) with the variation in average COs per bivalent can distinguish between the two possibilities, with opposing variation implying a patterning difference and co-variation implying a difference in some other feature. To model Scenario 2, in which non-patterned COs arise from an unrelated set of precursors, a standard CO-designation BF simulation is performed to define interfering COs. Then a second, independent simulation is performed using a specified number of precursors that are unrelated to the first set and random selection of COs from among that precursor set. COs generated by the two types of simulations are then again combined and analyzed. CoC relationships for total COs (interfering plus non-interfer- ing) will depend significantly on whether the precursors that give rise to the ‘‘non-interfering’’ COs are evenly or randomly spaced along the chromosomes. CoC curves for total COs reflect the combined inputs of CoC relationships for interfering COs and non-interfering COs. CoC curves for interfering COs are affected only modestly by even-versus-random spacing due to the overriding effects of CO interference (above; e.g. Figure 14B left). Discussion The presented analysis has provided new information on CO patterning from several different perspectives. BF Simulations Accurately, Quantitatively Describe Experimental Data Sets Given this framework, we defined CoC curves for total COs along yeast chromosomes IV and XV as defined by microarray analysis (Figure 14D left panel). The general shapes of these experimental curves correspond qualitatively to those predicted for emergence of non-interfering COs from an evenly-spaced precursor array, with a closer correspondence to those predicted for Scenario 1 than to those predicted for Scenario 2 (compare Figure 14D left panel with Figure 14C). This is true not only with respect to CoC and ED relationships but with respect to more detailed effects such as CO homeostasis and the obligatory CO. These matches, and the information that emerges there-from, support the notion that the basic logic of the BF model provides a robust and useful way of thinking about CO patterning. These matches are also specifically supportive of the proposed mechanical stress-and-stress relief mechanism. This impression is further supported by BF simulations. To model Scenario 1, we began with the set of best-fit parameters defined for interfering COs (Zip3 foci) above (Figure 6I) and generated predicted total CoC curves, assuming that non-interfering COs comprise 30% of the total (above), for each Part VII. Variations in Diverse BF Parameters Alter the Value of the Gamma Shape Parameter However, non-interfering CO relationships are a direct reflection of precursor relationships, which differ dramatically in the two cases. For precursors, CoC = 1 for random spacing and significant ‘‘interference’’ for even spacing; Figure 14B second from left). CoC relationships for non-interfering COs alone exhibit the same features (Figure 14B, rightmost two panels). These differences are directly visible in CoC curves for total COs, with greater or lesser prominence according to the relative abundance of non-interfering COs versus interfering COs (Figure 14C). Notably, CoC relationships for Scenario 1, where precursors exhibit the even spacing defined by BF best-fit simulations (E = 0.6), show a qualitatively different shape than CoC relationships under Scenario 2. Logic and Modeling of Meiotic Crossover Patterns total COs from genetic and microarray analyses) (e.g. [22,31,52,53]; below). of the three possible case of non-interfering COs: Scenario 1 (where precursors are assumed to be evenly spaced as for interfering COs); and Scenario 2 with precursors assumed to be either evenly or randomly spaced (Figure 14D, second panel from left). The CoC curve for the first of these three cases has the same shape as the experimental CoC curves for total COs (compare Figure 14D left and second from left panels) and direct comparison shows that it gives a quite good quantitative match with the experimental curves (Figure 14D third panel from left). Scenario 2 with evenly-spaced precursors is a less good match (Figure 14D, right panel). Scenario 2 with randomly-spaced precursors (Figure 14D, second panel from left, red) is a quite poor match (not shown). The origin of non-interfering COs is unknown. One possibility is that they arise from the majority subset of interactions that do not undergo CO-designation [5]. By this model (‘‘Scenario 1’’; Figure 14A left), not-CO-designated interactions would mostly mature to NCOs but sometimes would mature to COs, analogously to the situation in mitotic DSB-initiated recombina- tional repair [54]. Alternatively, such COs might arise from some other set of DSBs that arise outside of the normal process, e.g. because they occur later in prophase after CO-designation is completed or earlier in prophase before patterning conditions are established (‘‘Scenario 2’’; Figure 14A right). Both scenarios can be examined using the BF simulation program. To simulate the outcome of Scenario 1, where non- patterned COs arise from non-designated interactions left over after patterning, a standard CO-designation BF simulation is performed to define the interfering COs; the precursors that have not undergone CO-designation are then used as the starting array of precursors for a second round of CO-designation. In this second round, COs are randomly selected from among the precursors remaining after the first round of designation. The COs resulting from the two simulations are then combined and the total pattern is analyzed. These analyses suggest that, in yeast, non-interfering COs arise from the not-CO-designated precursors as a minority outcome of the ‘‘NCO’’ default pathway (Figure 14A, Scenario 1). Part VII. Variations in Diverse BF Parameters Alter the Value of the Gamma Shape Parameter Part VI. Non-interfering COs Implications: The general implication of these considerations is that CO patterning features have evolved to give a very low level of zero-CO bivalents, in accord with the biological imperative. However, in certain organisms, the constellation of features may be tuned to just such a level that the number of precursors along shorter chromosomes is just at the limit of the necessary threshold. In some organisms, a significant fraction of COs arises outside of the patterning process. The existence of these ‘‘non-interfer- ing’’ COs is most rigorously documented for budding yeast, where the number of ‘‘non-interfering’’ COs is ,30% among total COs (by compassion the number of patterned COs defined by analysis of CO-correlated Zip2/Zip3 foci with the number of January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 16 Logic and Modeling of Meiotic Crossover Patterns New Information about CO Patterning in Several Organisms In budding yeast: (i) CO patterning has the same basic features for shorter and longer chromosomes; (ii) Mlh1 is required January 2014 \| Volume 10 \| Issue 1 \| e1004042 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 17 Logic and Modeling of Meiotic Crossover Patterns Figure 14. Modeling of non-interfering COs with the example of budding yeast. (A) Non-interfering COs could arise from the same array of precursors as designated COs by low probability of CO formation at remaining not-CO-designated sites (left); alternatively, non-interfering COs could arise from a different set of precursors (right). (B) BF simulations. (Far left) Simulations under standard parameter conditions (Figure 3) illustrate the facts that CoC relationships for interfering COs vary relatively little according to whether precursors are randomly or evenly spaced; see also Figure 3 and Figure S2). (Second from left) An array of precursors was defined under standard parameter conditions with E = 0.6 (relatively even spacing) or random spacing (E = 0) and subjected to a simulation at very high Smax such that all precursors were manifested as ‘‘events’’, for which CoC curves were then constructed. Precursor CoC relationships are dramatically different for random spacing (CoC = 1 for all inter-interval distances) or even spacing (significant interference at small inter-interval distances). (Right two panels) CoC curves for non-interfering COs generated by Scenario 1 or Scenario 2 with even or random precursor spacing (E = 0.6 or E = 0) directly match the CoC curves for the precursors from which they arose. (C) BF simulations for mixtures of interfering and non-interfering COs, at the different indicated proportions of non-interfering COs. Interfering COs were generated under standard parameter conditions which correspond to yeast chromosome XV (Figure 3; E = 0.6). Non-interfering COs were generated according to Scenario 1 (left) or by Scenario 2 from either even or random precursors (E = 0.6; E = 0) (middle; right). (D) Left: Experimental CoC curves for CO data from microarray analysis of yeast Chromosomes IV and XV [22,52]. Data from the two studies were combined and the positions of COs Figure 14. Modeling of non-interfering COs with the example of budding yeast. (A) Non-interfering COs could arise from the same array of precursors as designated COs by low probability of CO formation at remaining not-CO-designated sites (left); alternatively, non-interfering COs could arise from a different set of precursors (right). Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns defined under the assumption, as made in both studies, that the CO position is at the site of the exchanged polymorphism. CoC analysis of these COs was then performed as for Zip3 foci. Second from left: CoC curves for yeast Chromosome XV corresponding to the combination of interfering COs as described by the best-fit simulation (Figure 6I; E = 0.6) plus non-interfering COs at a level comprising 30% of total COs as described by best-fit simulations under the indicated Scenario and precursor array conditions. Two right-most panels: Experimental CoC curves for Chromosomes IV and XV (from left panel) plus best-fit simulation CoC curves for the indicated mixtures of interfering and non-interfering COs (from the second from left panel). The best match is provided if non-interfering COs arise from precursors left over after CO-designation (Scenario 1 with relatively evenly-spaced precursors; E = 0.6) as shown in second panel from right. doi:10.1371/journal.pgen.1004042.g014 beam-film model are significantly different, in three respects. First, in the King and Mortimer model, the final array of COs reflects the relative rates of CO-designation and polymerization. Thus it is the kinetics of the system that governs its outcome. In the beam- film model, where interference arises immediately after each CO- designation, kinetics does not play a role. Second, in the King and Mortimer model, the interference signal continues to spread until it runs into another signal approaching from the opposite direction. In the beam-film model, the interference signal is nucleated and spreads for an intrinsically limited distance, with an intrinsic tendency to dissipate with distance from its nucleation site. Third, the King and Mortimer model envisioned that precursors were Poisson distributed among chromosomes. As a result, significant numbers of chromosomes would initially acquire no precursors if the average number of precursors is low and thus would never give a CO, thereby giving an unacceptably high level of zero-CO chromosomes. To compensate for this effect, the model proposed that the effect of interference was to release encountered precursors, which then rebound in regions that were not yet affected by interference (and thus on chromosomes with no precursors). This precursor turnover would ensure that all chromosomes achieved a precursor that could ultimately give a CO. Because of this feature, the King and Mortimer model envisions that interference is required to ensure a low level of zero- CO chromosomes (i.e. Models and Mechanisms of CO Interference The central issue for CO patterning is how information is communicated along the chromosomes. Three general types of mechanisms have been envisioned. (1) A molecular signal spreads along the chromosomes, e.g. as in the polymerization model of King and Mortimer [55] or the ‘‘counting model’’ of Stahl and colleagues [20,35]. (2) A biochemical reaction/diffusion process surfs along the chromosomes [56], as recently described in detail for bacterial systems [57,58]. (3) Communication occurs via redistribution of mechanical stress, as in the beam-film model [3,4] or via other mechanical mechanisms (e.g. [59]). New Insights into CO Homeostasis and the ‘‘Obligatory CO’’ With respect to CO homeostasis, the importance of CO interference as a determinant in the strength of homeostasis is emphasized and BF simulations are shown to permit accurate quantitative descriptions of homeostasis. Also, the strength of homeostasis can be seen to reflect the ratio of interference distance (LCoC) to the distance between adjacent precursors. With respect to the obligatory CO, the general logic of the BF model (Figure 1) suggests that occurrence of a low level of zero-CO chromosomes is independent of CO interference (and precursor spacing) and is achieved by an appropriate evolved constellation of all other parameters. Explanations can also be provided for several known cases where the level of zero-CO chromosomes is unusually high, but interference is robust, and potential explanations for other mutant phenotypes are suggested. Importantly, the logic of the beam-film model predicts the existence of mutants that lack interference but still exhibit the obligatory CO, evidence for which will be presented elsewhere. Logic and Modeling of Meiotic Crossover Patterns to ensure the ‘‘obligatory CO’’). By the beam-film model, instead, precursors do not turn over and interference is not required to ensure a low level of zero-CO chromosomes, which results instead from an appropriate constel- lation of other features, as described above. The beam-film model predicts the existence of mutants that are defective in interference but do not exhibit an increase in the frequency of zero-CO chromosomes. specifically for CO maturation not for CO patterning; and (iii) Precursors are evenly spaced, as shown by both CoC analysis and analysis of total (interfering-plus-non-interfering) COs. In tomato (and, to be described elsewhere, in budding yeast), the metric of CO interference is physical chromosome length (mm) not genomic length (Mb). In the case of tomato, differences in CoC relationships expressed in the two different metrics is attributable to differential packaging of heterochromatin versus euchromatin along the chromosome plus differential proportions of hetero- chromatic versus euchromatic regions among different chromo- somes. In tomato and yeast, as previously described for grasshopper, human and several other organisms, crossover interference spreads across centromeres with the same metric as along chromosome arms. In budding yeast, non-interfering COs arise from evenly-spaced precursors, most probably by occasional resolution of NCO-fated precursors to the CO fate. Yeast Strains Yeasts SK1 strains (Figure 6 and S7) are described in Table S1. In all strains, ZIP3 carries a MYC epitope tag; a construct expressing LacI-GFP and is integrated at either LEU2 or URA3, and a lacO array [61] is inserted at HMR (chromosome III), Scp1 (Chromosome XV) or Chromosome IV telomere (SGD1522198) to specifically label each chromosomes by binding of LacI-GFP. New Information about CO Patterning in Several Organisms (B) BF simulations. (Far left) Simulations under standard parameter conditions (Figure 3) illustrate the facts that CoC relationships for interfering COs vary relatively little according to whether precursors are randomly or evenly spaced; see also Figure 3 and Figure S2). (Second from left) An array of precursors was defined under standard parameter conditions with E = 0.6 (relatively even spacing) or random spacing (E = 0) and subjected to a simulation at very high Smax such that all precursors were manifested as ‘‘events’’, for which CoC curves were then constructed. Precursor CoC relationships are dramatically different for random spacing (CoC = 1 for all inter-interval distances) or even spacing (significant interference at small inter-interval distances). (Right two panels) CoC curves for non-interfering COs generated by Scenario 1 or Scenario 2 with even or random precursor spacing (E = 0.6 or E = 0) directly match the CoC curves for the precursors from which they arose. (C) BF simulations for mixtures of interfering and non-interfering COs, at the different indicated proportions of non-interfering COs. Interfering COs were generated under standard parameter conditions which correspond to yeast chromosome XV (Figure 3; E = 0.6). Non-interfering COs were generated according to Scenario 1 (left) or by Scenario 2 from either even or random precursors (E = 0.6; E = 0) (middle; right). (D) Left: Experimental CoC curves for CO data from microarray analysis of yeast Chromosomes IV and XV [22,52]. Data from the two studies were combined and the positions of COs January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 18 PLOS Genetics \| www.plosgenetics.org Zip2/Zip3 Foci on Yeast Pachytene Chromosomes Correspond to Programmed (‘‘Interfering’’) COs Pachytene chromosomes exhibit ,65 foci of Zip2, Zip3 and Msh4, with strong colocalization of Zip3 and Msh4 foci ([31,62]; this work). Zip2 foci [63] exhibit interference as defined by CoC relationships for random adjacent pairs of intervals [31]. We further show here that Zip2 and Zip3 foci exhibit interference as defined by full CoC relationships along specific individual chromosomes (Figure 6 and Figure S5). Zip2 and Zip3 foci also both occur specifically on association sites of zip1D chromosomes [31,64]. The total number of COs per yeast nucleus as defined by microarray and genetic analysis is ,90 [22,52,65] implying that The counting model can provide good explanations of experimental data; however, the underlying mechanism is contradicted by experimental findings ([43]; but see [60]). No specific reaction/diffusion mechanism has been suggested thus far for CO interference. The King and Mortimer model and the January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 19 Figure 15. Application of the gamma distribution to CO patterns: Risks and rewards. The value of the gamma distribution shape parameter (n) varies not only with variations in interference distance (L) but with variations in other parameters. Notably, the value of n and the value of the average number of COs per bivalent vary inversely with changes in patterning parameters L and Smax but vary directly, or very little, with changes in precursor and maturation parameters. This is shown by BF simulations performed under standard conditions (Figure 3) except that individual parameters were systematically varied. The values of each of the four indicated descriptors (average CO/bivalent; LCoC; average distance between adjacent COs (‘‘inter-CO distance’’) and the gamma distribution shape parameter n) were plotted as a function of the value of the varied parameter (left). Also, the distribution of distances between adjacent COs for representative values of the varied parameter are shown (right). (A–D) Effects of systematic variation of parameters L, Smax, N and M. (E, F) Effects of variation of E and B at two different values of N. doi:10.1371/journal.pgen.1004042.g015 Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns Figure 15. Application of the gamma distribution to CO patterns: Risks and rewards. The value of the gamma distribution shape parameter (n) varies not only with variations in interference distance (L) but with variations in other parameters. Cytological Mapping Zip3 Foci in Yeast y g pp g p Synchronous meiotic cultures (SPS sporulation procedure from [66]) were prepared and harvested at a time when pachytene nuclei are most abundant (,4–5 hours). Cells were spheroplasted and chromosomes spread on glass slides according to Loidl et al. and Kim et al. [67,68]. Primary antibodies were mouse monoclonal anti-myc, goat polyclonal anti-Zip1 (Santa Cruz) and rabbit polyclonal anti-GFP (Molecular Probes). Each was diluted appropriately in the above BSA/TBS blocking buffer. Secondary antibodies were donkey anti-mouse, donkey anti-goat, and donkey anti-rabbit IgG labeled with Alexa488, Alexa645 or 594 and Alexa555 (Molecular Probes), respectively. Stained slides were mounted in Slow Fade Light or Prolong Gold Antifade (Molecular Probes). Spread chromosomes were visualized on an Axioplan IEmot microscope (Zeiss) with appropriate filters. Images were collected using Metamorph (Molecular Devices) image acquisition and analysis software. Acquired images were then analyzed with Image J software (NIH), with total SC length and positions of Zip3 foci for the specifically labeled bivalent were measured from the lacO/LacIGFP-labeled end to the other end (Figure 6B bottom). For each type of chromosome analyzed (III, IV and XV) in each experiment, measurements were made for .300 bivalents, one from each of a corresponding number of spread nuclei. Resulting data were transferred into an EXCEL worksheet for further analysis. (2) Do a single BF simulation with a particular set of specified parameter values. This option gives the same outputs as for analysis of an existing CO data set. It can help the user to understand how the BF model works (e.g. by extension of examples presented in the Results). It also enables a skilled user to do a single round of BF simulation at some single particular parameter condition. (3) Scan a range of parameters to get a BF best-fit simulation of an experimental data set. This option automatically scans all parameter combinations (over specified ranges of each parameter) and outputs the results in rank order according to the goodness of fit levels calculated based on PLS (Projected Likelihood Score) as defined by Falque et al. [18,19]. However, the rank order defined by PLS is not a maximum likelihood method. As a result, the best fit judged by PLS is not always the actual best fit. To overcome this drawback, the software outputs the results for all parameter combinations scanned, which the user can further evaluate to select the actual best fit. Logic and Modeling of Meiotic Crossover Patterns Zip2/Zip3/Msh4 foci represent 65/90 = 70% of the total. Correspondingly, mutant analysis suggests that ‘‘non-interfering’’ COs comprise ,30% of total COs (e.g. [50]). Additionally, BF analysis accurately explains CoC relationships for total COs on the assumption of 70% patterned COs and 30% ‘‘non-interfering’’ COs (Figure 14, Results). Zip2/Zip3/Msh4 foci represent 65/90 = 70% of the total. Correspondingly, mutant analysis suggests that ‘‘non-interfering’’ COs comprise ,30% of total COs (e.g. [50]). Additionally, BF analysis accurately explains CoC relationships for total COs on the assumption of 70% patterned COs and 30% ‘‘non-interfering’’ COs (Figure 14, Results). program structure and application are provided in the Protocol S1 section. However, briefly, there are three options in the software that serve three different purposes: (1) Analyze existing CO data. This option allows the user to process an experimental CO data set. Outputs include a variety of different CO distribution descriptors including CoC curves, ED relationships including the average number of COs/bivalent, average inter-CO distance and evaluation of the gamma distribution shape parameter (n). Other Data Used in This Study The Chorthippus L3 chiasmata data were generously provided by Gareth Jones (University of Birmingham, UK). The Drosophila X- chromosome crossover data are from [33]. The tomato (S. lycopersicum) Mlh1 foci date are from [38] (generously provided by F. Lhuissier). Zip2 data in the S.cerevisiae BR background are from [31] (generously provided by J. Fung). Cytological Mapping Zip3 Foci in Yeast We normally choose the best fit by comparing experimental and simulated data sets with respect to the CoC curve, the average number of COs per bivalent and the distribution of CO number per bivalents (the ED distribution). (3) Scan a range of parameters to get a BF best-fit simulation of an experimental data set. This option automatically scans all parameter combinations (over specified ranges of each parameter) and outputs the results in rank order according to the goodness of fit levels calculated based on PLS (Projected Likelihood Score) as defined by Falque et al. [18,19]. However, the rank order defined by PLS is not a maximum likelihood method. As a result, the best fit judged by PLS is not always the actual best fit. To overcome this drawback, the software outputs the results for all parameter combinations scanned, which the user can further evaluate to select the actual best fit. We normally choose the best fit by comparing experimental and simulated data sets with respect to the CoC curve, the average number of COs per bivalent and the distribution of CO number per bivalents (the ED distribution). CoC Calculations for Yeast Zip3 Foci p Coefficient of coincidence (CoC) curves were generated from SC length and Zip3 focus positions determined as described above. Each analyzed bivalent was divided into a series of intervals of 0.1 mm in length (corresponding to the resolution with which adjacent Zip3 foci can be resolved). Chromosome III, IV and XV were thus usually divided into 9, 42 and 30 intervals with equal size, respectively. Each chromosome length was normalized to 100% and each Zip3 focus position was also normalized correspondingly. Each Zip3 focus was then assigned to a specific interval according to its coordinate. The total frequency of bivalents having a Zip3 focus in each interval was calculated. For each pair of intervals, the frequency of bivalents having a Zip3 focus in both intervals was determined to give the ‘‘observed’’ frequency of double COs. For each pair of intervals, the total CO frequencies for the two intervals were multiplied to give the frequency of double COs ‘‘expected’’ on the hypothesis of independent occurrence. The ratio of these two values is the CoC. Thus in each pair of intervals, CoC = (Obs DCO)/(Pred DCO). CoC values for all pairs of intervals can be plotted as a function of the distance between the midpoints of the two involved intervals (‘‘inter-interval distance’’). However, for all of the data shown here, the CoC values from all pairs of intervals having same inter-interval distance were averaged and this average CoC was plotted as a function of inter-interval distance (e.g. Figure 6C and others). Supporting Information Figure S1 Determination of interval sizes required for accurate CoC curves. Bivalents must be divided into a sufficiently large number of intervals that few if any closely-spaced COs are missed. A general rule is that the interval size should be less than 1/4 the average distance between COs. Operationally, where possible, interval size should be progressively decreased until the position of the CoC curve no longer changes. The simulations presented were performed under standard parameter conditions (text Figure 3) except that the number of intervals (and thus the inter-interval distance) was progressively increased. For this particular case, the CoC curves do not change significantly once the number of intervals is at or above 20 (interval size = 5% total chromosome length in mm). Figure S2 Interplay among precursor parameter values at low precursor number (N). Variations in the distributions of precursors along or between bivalents (parameters E and B) have more significant effects at lower average precursor numbers (N). Panels (A) and (B): Effects of variations in B and E are illustrated by Zip2/Zip3 Foci on Yeast Pachytene Chromosomes Correspond to Programmed (‘‘Interfering’’) COs Notably, the value of n and the value of the average number of COs per bivalent vary inversely with changes in patterning parameters L and Smax but vary directly, or very little, with changes in precursor and maturation parameters. This is shown by BF simulations performed under standard conditions (Figure 3) except that individual parameters were systematically varied. The values of each of the four indicated descriptors (average CO/bivalent; LCoC; average distance between adjacent COs (‘‘inter-CO distance’’) and the gamma distribution shape parameter n) were plotted as a function of the value of the varied parameter (left). Also, the distribution of distances between adjacent COs for representative values of the varied parameter are shown (right). (A–D) Effects of systematic variation of parameters L, Smax, N and M. (E, F) Effects of variation of E and B at two different values of N. doi:10.1371/journal.pgen.1004042.g015 Figure 15. Application of the gamma distribution to CO patterns: Risks and rewards. The value of the gamma distribution shape parameter (n) varies not only with variations in interference distance (L) but with variations in other parameters. Notably, the value of n and the value of the average number of COs per bivalent vary inversely with changes in patterning parameters L and Smax but vary directly, or very little, with changes in precursor and maturation parameters. This is shown by BF simulations performed under standard conditions (Figure 3) except that individual parameters were systematically varied. The values of each of the four indicated descriptors (average CO/bivalent; LCoC; average distance between adjacent COs (‘‘inter-CO distance’’) and the gamma distribution shape parameter n) were plotted as a function of the value of the varied parameter (left). Also, the distribution of distances between adjacent COs for representative values of the varied parameter are shown (right). (A–D) Effects of systematic variation of parameters L, Smax, N and M. (E, F) Effects of variation of E and B at two different values of N. doi:10.1371/journal.pgen.1004042.g015 January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 20 Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns simulations using the same parameter values as for text Figure 3 except that N = 7. and B = 1, which give relatively even spacing and a constant number of COs per bivalent; Table 2) (red) are compared with BF simulations using the same parameter values except that precursors were considered to be randomly spaced along chromosomes (E = 0; Panel E, green) or Poisson-distributed among chromosomes III in different nuclei (B = 0; Panel F, blue). Even- versus-random spacing affects CoC relationships, confirming that precursors are evenly spaced (text), but does not affect ED relationships. Oppositely, constant-versus-Poisson distribution among chromosomes does not affect CoC relationships but significantly alters ED relationships, with a decrease in the average number of COs per bivalent overall but, more importantly, a significant increase in the frequency of zero-CO bivalents, from 1% to 4%. This effect strongly suggests that a given chromosome always acquires the same/similar number of precursors in every meiotic nucleus. (TIF) simulations using the same parameter values as for text Figure 3 except that N = 7. (TIF) Figure S3 CoC curves can have signatures that reflect inter- precursor spacing. BF simulations show that if precursors are very evenly spaced (E$0.8, the corresponding n.10), and if the interference distance is relatively long as compared to the average distance between precursors (e.g. L = 0.15 versus 0.06–0.17), CoC curves can exhibit ‘‘humps’’ corresponding to the average inter- precursor distance. These humps reflect the fact that closely- spaced double COs will tend to occur specifically at adjacent precursors, and when those precursors are evenly spaced, there is an elevated probability of double CO occurrence at that particular inter-interval distance. This feature is not apparent in standard simulation conditions (text Figures 3–5) because, in those conditions, precursors are less evenly spaced (E = 0.6, the corresponding n = 2.4). Other BF parameter values for the simulations show in this figure are: L = 0.15, Smax = 1.8, A = 2, cL = cR = 1, E = 0.8, B = 0.9. (TIF) Figure S6 Heterochromatin/euchromatin ratios underlie differ- ences in SC length in tomato. CoC analysis suggests that the metric for CO interference is physical chromosome length (manifested as SC length) and that the density of DNA, i.e. Mb per mm SC length, is greater for short chromosomes than for longer chromosomes (text). Logic and Modeling of Meiotic Crossover Patterns The latter feature reflects the combination of two effects: (i) short chromosomes have a higher percentage of their DNA in heterochromatin and (ii) there is more DNA per mm SC length in heterochromatin versus euchromatin. This conclusion requires knowing: (i) the ‘‘packing ratio’’ of Mb/ mm SC length for heterochromatin versus euchromatin; and (ii) the fraction of each pachytene SC length that underlies heterochro- matin versus euchromatin. This information is available from published data (Panel A; http://solgenomics.net; [38,74,75]. Panel B: (i) Genome-wide, total DNA in heterochromatin and euchromatin are 693 Mb and 203 Mb, respectively and total SC length in heterochromatin and euchromatin are 78 mm and 152 mm, respectively. Thus, Mb/mm SC is 8.9 for heterochroma- tin and 1.3 for euchromatin, respectively. (ii) For shorter and longer chromosomes, the fractions of SC length in euchromatin and heterochromatin in the two groups are 66% and 34% respectively. This information implies that the group of shorter chromosomes will comprise 70 Mb, with an average Mb/mm SC length of 3.35, while the group of longer chromosomes will comprise 73 Mb, with an average Mb/mm SC length of 4.12. This difference quantitatively explains the offset in CoC curves when inter-interval distance is expressed in Mb versus mm SC length. LCoC (and LBF) are 14 mm SC length, which thus corresponds to 47 Mb and 58 Mb, respectively, which is a difference of 11 Mb. The offset in the experimental CoC curves for the two groups when inter-interval distance is expressed in Mb exactly matches this difference: LCoC occurs at inter-interval distances of 32 Mb and 43 Mb respectively, which is a difference of 11 Mb. Notably, these considerations also show that, despite their differences in SC lengths, all of the chromosomes in tomato contain about the same total amount of DNA. Figure S4 How to obtain the best-fit BF simulation for an experimental data set. For each data set, the constellation of BF parameters that provides the best fit to the data set was obtained in three stages, as illustrated for data from yeast Chromosome XV (text). (1) A simulation was carried out at probable approximate values of N and L (Panel A). The range of sensible values of (N) is suggested by total DSB levels, total levels of inter-homolog events (COs+NCOs), numbers of inter-axis bridges, immunofluorescent foci and/or EM-defined SC-associated recombination nodules, all of which approximately reflect total precursor interactions. BF Software and Simulations BF Software and Simulations The previous Beam Film program [3] was rewritten in MATLAB (R2010a) for easy use and modified to include more features as described in the text. Extensive details regarding PLOS Genetics \| www.plosgenetics.org January 2014 \| Volume 10 \| Issue 1 \| e1004042 January 2014 \| Volume 10 \| Issue 1 \| e1004042 21 Logic and Modeling of Meiotic Crossover Patterns The initial value of L (LBF) was generally set at LCOC With respect to other parameters: values of (cL) and (cR) were selected based on the distribution of COs along the chromosome; the value of M was assumed to be 100% for wild-type meiosis; the value of A = 1 was selected as a reasonable first approximation. (2) The value of Smax was then refined so as to optimize the fit between experimental and predicted ED arrays with respect to both the average number of COs per bivalent, including the probability of zero-CO chromosomes (Panel B). (3) The values of all parameters were then further refined by empirical trial-and-error, guided by knowledge as to the general effects of changes in each parameter on CoC and ED outputs as described above. Panel C describes initial refinements with respect to L and Smax; Panel D describes subsequent refinements of these two parameters. The final selected best-fit simulation parameters are those described in Panel D, L = 0.1; Smax = 3.5 (other parameter values in text Table 2). (TIF) Figure S5 Experimental data and BF simulations for yeast chromosomes. Panels A–C: CoC and ED relationships, and best- fit BF simulations, for SK1 Chromosomes XV (from text Figure 6I) and analogously analyzed Chromosomes IV and III. Panel D: CoC and ED relationships and BF simulation for Chromosome XIV in the BR background ([31]; J. Fung, personal communica- tion). Parameter values for simulations in A–D in text Table 2. Note that all chromosomes, in both strain backgrounds, have the same CoC relationships when inter-interval distance is expressed in mm SC length. Further, pachytene SC length is ,10% less in BR than in SK1. This comparison, along with other comparisons (L.Z., unpublished), shows that the metric for interference is physical distance in yeast as in other analyzed organisms (text). Panels E and F: the average CoC curve and the ED relationships for Chromosome III observed experimentally (black) and BF best- fit simulations using optimal parameter values (including E = 0.6 References 25. Blat Y, Protacio RU, Hunter N, Kleckner N (2002) Physical and functional interactions among basic chromosome organizational features govern early steps of meiotic chiasma formation. Cell 111: 791–802. 1. Muller HJ (1916) The Mechanism of Crossing Over. Am Nat 50: 193–434. 1. Muller HJ (1916) The Mechanism of Crossing Over. Am Nat 50: 193–434. 2. Sturtevant AH (1915) The behavior of the chromosomes as studied through linkage. Z indukt Abstamm-u VererbLehre 13: 234–287. linkage. Z indukt Abstamm-u VererbLehre 13: 234–287. 26. Buhler C, Borde V, Lichten M (2007) Mapping meiotic single-strand DNA reveals a new landscape of DNA double-strand breaks in Saccharomyces cerevisiae. PLoS Biol 5: e324. 3. Kleckner N, Zickler D, Jones GH, Dekker J, Padmore R, et al. (2004) A mechanical basis for chromosome function. Proc Natl Acad Sci U S A 101: 12592–12597. 27. Kauppi L, Barchi M, Lange J, Baudat F, Jasin M, et al. (2013) Numerical constraints and feedback control of double-strand breaks in mouse meiosis. Genes Dev 27: 873–886 4. Borner GV, Kleckner N, Hunter N (2004) Crossover/noncrossover differenti- ation, synaptonemal complex formation, and regulatory surveillance at the leptotene/zygotene transition of meiosis. Cell 117: 29–45. p yg 5. Hunter N (2006) Meiotic recombination. In: Aguilera A, Rothstein R editors. Molecular Genetics of Recombination. pp 381–442. 28. Joyce EF, McKim KS (2010) Chromosome axis defects induce a checkpoint- mediated delay and interchromosomal effect on crossing over during Drosophila meiosis. PLoS Genet 6: e1001059. Molecular Genetics of Recombination. pp 381–442 6. Drouaud J, Mercier R, Chelysheva L, Berard A, Falque M, et al. (2007) Sex- specific crossover distributions and variations in interference level along Arabidopsis thaliana chromosome 4. PLoS Genet 3: e106. 29. Joyce EF, McKim KS (2011) Meiotic checkpoints and the interchromosomal effect on crossing over in Drosophila females. Fly (Austin) 5: 134–140. 7. Petkov PM, Broman KW, Szatkiewicz JP, Paigen K (2007) Crossover interference underlies sex differences in recombination rates. Trends Genet 23: 539–542. 30. Higgins JD, Perry RM, Barakate A, Ramsay L, Waugh R, et al. (2012) Spatiotemporal asymmetry of the meiotic program underlies the predominantly distal distribution of meiotic crossovers in barley. Plant Cell 24: 4096–4109. 31. Fung JC, Rockmill B, Odell M, Roeder GS (2004) Imposition of crossover interference through the nonrandom distribution of synapsis initiation complexes. Cell 116: 795–802. 8. Billings T, Sargent EE, Szatkiewicz JP, Leahy N, Kwak IY, et al. References Saintenac C, Falque M, Martin OC, Paux E, Feuillet C, et al. (2009) Detailed recombination studies along chromosome 3B provide new insights on crossover distribution in wheat (Triticum aestivum L.). Genetics 181: 393–403. 16. Mets DG, Meyer BJ (2009) Condensins regulate meiotic DNA break distribution, thus crossover frequency, by controlling chromosome structure. Cell 139: 73–86. 41. Lian J, Yin Y, Oliver-Bonet M, Liehr T, Ko E, et al. (2008) Variation in crossover interference levels on individual chromosomes from human males. Hum Mol Genet 17: 2583–2594. 17. McPeek S, Speed TP, Mor H (1995) Modeling Interference in Genetic Recombination. Genetics 1044: 1031–1044. 42. Broman KW, Rowe LB, Churchill GA, Paigen K (2002) Crossover interference in the mouse. Genetics 160: 1123–1131. 18. Falque M, Anderson LK, Stack SM, Gauthier F, Martin OC (2009) Two types of meiotic crossovers coexist in maize. Plant Cell 21: 3915–3925. 43. Martini E, Diaz RL, Hunter N, Keeney S (2006) Crossover homeostasis in yeast meiosis. Cell 126: 285–295. 19. Gauthier F, Martin OC, Falque M (2011) CODA (crossover distribution analyzer): quantitative characterization of crossover position patterns along chromosomes. BMC Bioinformatics 12: 27. 44. Rosu S, Libuda DE, Villeneuve AM (2011) Robust crossover assurance and regulated interhomolog access maintain meiotic crossover number. Science 334: 1286–1289. 20. Foss E, Lande R, Stahl FW, Steinberg CM (1993) Chiasma interference as a function of genetic distance. Genetics 133: 681–691. 45. Yokoo R, Zawadzki KA, Nabeshima K, Drake M, Arur S, et al. (2012) COSA-1 reveals robust homeostasis and separable licensing and reinforcement steps governing meiotic crossovers. Cell 149: 75–87. 21. Pan J, Sasaki M, Kniewel R, Murakami H, Blitzblau HG, et al. (2011) A hierarchical combination of factors shapes the genome-wide topography of yeast meiotic recombination initiation. Cell 144: 719–731. 22. Chen SY, Tsubouchi T, Rockmill B, Sandler JS, Richards DR, et al. (2008) Global analysis of the meiotic crossover landscape. Dev Cell 15: 401–415. 46. Globus ST, Keeney S (2012) The joy of six: how to control your crossovers. Cell 149: 11–12. 47. Jones GH, Franklin FC (2006) Meiotic crossing-over: obligation and interfer- ence. Cell 126: 246–248. 23. Borde V, Lin W, Novikov E, Petrini JH, Lichten M, et al. (2004) Association of Mre11p with double-strand break sites during yeast meiosis. Mol Cell 13: 389– 401. 48. Hawley RS, Theurkauf WE. (1993) Requiem for distributive segregation: achiasmate segregation in Drosophila females. Trends Genet 9: 310–317. 24. References (2010) Patterns of recombination activity on mouse chromosome 11 revealed by high resolution mapping. PLoS One 5: e15340. 32. Malkova A, Swanson J, German M, McCusker JH, Housworth EA, et al. (2004) Gene conversion and crossing over along the 405-kb left arm of Saccharomyces cerevisiae chromosome VII. Genetics 168: 49–63. 9. Hillers KJ, Villeneuve AM (2003) Chromosome-wide control of meiotic crossing over in C. elegans. Curr Biol 13: 1641–1647. 10. Oliver-Bonet M, Campillo M, Turek PJ, Ko E, Martin RH (2007) Analysis of replication protein A (RPA) in human spermatogenesis. Mol Human Reprod 13: 837–844. 33. Charles DR (1938) The spatial distribution of cross-overs in X-chromosome tetrads of Drosophila melanogaster. J Genetics 36: 103–126. 34. Mehrotra S, McKim KS (2006) Temporal analysis of meiotic DNA double- strand break formation and repair in Drosophila females. PLoS Genet 2: e200. 11. de Boer E, Stam P, Dietrich AJ, Pastink A, Heyting C (2006) Two levels of interference in mouse meiotic recombination. Proc Natl Acad Sci U S A 103: 9607–9612. 35. Lande R, Stahl FW (1993) Chiasma interference and the distribution of exchanges in Drosophila melanogaster. Cold Spring Harb Symp Quant Biol 58: 543–552. 12. Storlazzi A, Gargano S, Ruprich-Robert G, Falque M, David M, et al. (2010) Recombination proteins mediate meiotic spatial chromosome organization and pairing. Cell 141: 94–106. 36. Laurie DA, Jones GH (1981) Inter-individual variation in chiasma distribution in Chorthippus brunneus (Orthoptera: Acrididae). Heredity 47: 409–416. p g 13. Zhang L, Kim KP, Kleckner NE, Storlazzi A (2011) Meiotic double-strand breaks occur once per pair of (sister) chromatids and, via Mec1/ATR and Tel1/ ATM, once per quartet of chromatids. Proc Natl Acad Sci U S A 108: 20036– 41. 37. Jones GH (1984) The control of chiasma distribution. Symp Soc Exp Biol 38: 293–320. 38. Lhuissier FG, Offenberg HH, Wittich PE, Vischer NO, Heyting C (2007) The mismatch repair protein MLH1 marks a subset of strongly interfering crossovers in tomato. Plant Cell 19: 862–876. 14. Berchowitz LE, Copenhaver GP (2010) Genetic interference: don’t stand so close to me. Curr Genomics 11: 91–102. 39. Colombo PC, Jones GH (1997) Chiasma interference is blind to centromeres. Heredity (Edinb) 79 (Pt 2): 214–227. 15. Cole F, Kauppi L, Lange J, Roig I, Wang R, et al. (2012) Homeostatic control of recombination is implemented progressively in mouse meiosis. Nat Cell Biol 14: 424–430. 40. (TIF) Figure S7 BF simulations of CO patterns in yeast mutants with altered DSB levels. CoC and ED relationships for yeast mutants with changed DSB levels and BF best-fit simulations (Panels A, B). Also indicated are the number of DSBs predicted from experimental analysis (relative DSBs levels from pulse-field gels along chromosome III, VII and VIII in a rad50S strain background; [43]) and the number of precursors required to give a best fit simulation. Predicted values of (N) and observed levels of DSBs match very precisely for most of the mutants; however, January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 22 Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns predicted values are slightly but significantly higher than experimental values at the very lowest DSB levels. This could mean that rad50S DSB levels are underestimated at low DSB levels; that DSB levels are auto-catalytic such that occurrence of DSBs above a threshold level tends to promote the formation of additional DSBs; or that best-fit simulations do not give precisely the correct values at low DSB levels. (TIF) Acknowledgments We thank Gareth Jones (University of Birmingham, UK) for grasshopper chiasmata data, Jennifer Fung (UCSF) for Zip2 data in the S.cerevisiae BR background and Franck G.P. Lhuissier (Netherlands) for tomato Mlh1 foci data. We thank Scott Hawley for provoking deeper thinking about the obligatory CO; Aurora Storlazzi and Denise Zickler for comments; and Guillaume Witz and Frederick Chang for help with mathematical analyses. Author Contributions Protocol S1 Instructions for the BF program. (DOCX) Conceived and designed the experiments: LZ JH NK. Performed the experiments: LZ. Analyzed the data: LZ NK. Contributed reagents/ materials/analysis tools: LZ ZL JH. Wrote the paper: LZ NK. Table S1 Strains used in this study. (DOCX) Table S1 Strains used in this study. (DOCX) Table S1 Strains used in this study. Logic and Modeling of Meiotic Crossover Patterns for Mlh1p in meiotic crossing over and in vegetative and meiotic mismatch repair. Mol Cell Biol 23: 873–886. 63. Agarwal S, Roeder GS (2000) Zip3 provides a link between recombination enzymes and synaptonemal complex proteins. Cell 102: 245–255. p 50. Argueso JL, Wanat J, Gemici Z, Alani E (2004) Competing crossover pathways act during meiosis in Saccharomyces cerevisiae. Genetics 168: 1805–1816. 64. Chua PR, Roeder GS (1998) Zip2, a meiosis-specific protein required for the initiation of chromosome synapsis. Cell 93: 349–359. g y 51. Zakharyevich K, Tang S, Ma Y, Hunter N (2012) Delineation of joint molecule resolution pathways in meiosis identifies a crossover-specific resolvase. Cell 149: 334–347. y p 65. Cherry JM, Ball C, Weng S, Juvik G, Schmidt R, et al. (1997) Genetic and physical maps of Saccharomyces cerevisiae. Nature 387: 67–73. physical maps of Saccharomyces cerevisiae. Nature 387: 67–73. 66. Koszul R, Kleckner N (2009) Dynamic chromosome movements during meiosis: a way to eliminate unwanted connections? Trends Cell Biol 19: 716–724. 52. Mancera E, Bourgon R, Brozzi A, Huber W, Steinmetz LM (2008) High- resolution mapping of meiotic crossovers and non-crossovers in yeast. Nature 454: 479–485. a way to eliminate unwanted connections? Trends Cell Biol 19: y 67. Loidl J, Klein F, Engebrecht J (1998) Genetic and morphological approaches for the analysis of meiotic chromosomes in yeast. Methods Cell Biol 53: 257–285. 53. Franklin WS, Foss HM (2010) A two-pathway analysis of meiotic crossing over and gene conversion in Saccharomyces cerevisiae. Genetics 186: 515–536. 68. Kim PK, Weiner BM, Zhang L, Jordan A, Dekker J, et al. (2010) Sister cohesion and structural axis components mediate homolog bias of meiotic recombination. Cell 143: 924–937. 54. Bzymek M, Thayer NH, Oh SD, Kleckner N, Hunter N (2010) Double holliday junctions are intermediates of DNA break repair. Nature 464: 937–941. nctions are intermediates of DNA break repair. Nature 464: 937–941 69. Qiao H, Lohmiller LD, Anderson LK (2011) Cohesin proteins load sequentially during prophase I in tomato primary microsporocytes. Chromosome Res 19: 193–207. 55. King JS, Mortimer RK (1990) A polymerization model of chiasma interference and corresponding computer simulation. Genetics 126: 1127–1138. and corresponding computer simulation. Genetics 126: 1127–1138 56. Fujitani Y, Mori S, Kobayashi I (2002) A reaction-diffusion model for interference in meiotic crossing over. Genetics 161: 365–372. 70. Dolezel J, Bartos J, Voglmayr H, Greilhuber J (2003) Nuclear DNA content and genome size of trout and human. References Gerton JL, Derisi J, Shroff R, Lichten M, Brown PO, et al. (2000) Global mapping of meiotic recombination hotspots and coldspots in the yeast Saccharomyces cerevisiae. Proc Natl Acad Sci U S A 97: 11383–11390. 49. Argueso JL, Kijas AW, Sarin S, Heck J, Waase M, et al. (2003) Systematic mutagenesis of the Saccharomyces cerevisiae MLH1 gene reveals distinct roles PLOS Genetics \| www.plosgenetics.org 23 January 2014 \| Volume 10 \| Issue 1 \| e1004042 January 2014 \| Volume 10 \| Issue 1 \| e1004042 Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns Logic and Modeling of Meiotic Crossover Patterns Cytometry A 51: 127–128. g 57. Han YW, Mizuuchi K (2010) Phage Mu transposition immunity: protein pattern formation along DNA by a diffusion-ratchet mechanism. Mol Cell 39: 48–58. 71. Santos J, del Cerro A, Dı´ez M (1993) Spreading synaptonemal complexes from the grasshopper Chorthippus jacobsi: pachytene and zygotene observations. Hereditas 118: 235–241. Hereditas 118: 235–241. g y 58. Vecchiarelli AG, Hwang LC, Mizuuchi K (2013) Cell-free study of F plasmid partition provides evidence for cargo transport by a diffusion-ratchet mechanism. Proc Natl Acad Sci U S A 110: E1390–1397. 72. Page SL, Hawley RS (2001) c(3)G encodes a Drosophila synaptonemal complex protein. Genes Dev 15: 3130–3143. 59. Hulten MA (2011) On the origin of crossover interference: A chromosome oscillatory movement (COM) model. Mol Cytogenet 4: 10. p 73. Carpenter AT (1979) Recombination nodules and synaptonemal complex in recombination-defective females of Drosophila melanogaster. Chromosoma 75: 259–292. 60. Getz TJ, Banse SA, Young LS, Banse AV, Swanson J, et al. (2008) Reduced mismatch repair of heteroduplexes reveals ‘‘non’’-interfering crossing over in wild-type Saccharomyces cerevisiae. Genetics 178: 1251–1269. 74. Peterson DG, Stack SM, Price JH, Johnston JS. DNA content of heterochro- matin and euchromatin in tomato (Lycopersicon esculentum) pachytene chromosomes. Genome 39: 77–82. yp y 61. Straight AF, Belmont A S, Robinett CC, Murray AW (1996) GFP tagging of budding yeast chromosomes reveals that protein-protein interactions can mediate sister chromatid cohesion. Curr Biol 6: 1599–1608. 75. Sherman JD, Stack SM (1995) Two-dimensional spreads of synaptonemal complexes from solanaceous plants. VI. High-resolution recombination nodule map for tomato (Lycopersicon esculentum). Genetics 141: 683–708. 62. Hollingsworth NM, Ponte L, Halsey C (1995) MSH5, a novel MutS homolog, facilitates meiotic reciprocal recombination between homologs in Saccharomy- ces cerevisiae but not mismatch repair. Genes Dev 9: 1728–1739. January 2014 \| Volume 10 \| Issue 1 \| e1004042 PLOS Genetics \| www.plosgenetics.org 24
https://openalex.org/W2977389905	https://europepmc.org/articles/pmc6777196?pdf=render	English	null	Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry	EGEMS	2,019	cc-by	7,023	Pratt, J, et al. 2019 Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry. eGEMs (Generating Evidence & Methods to improve patient outcomes), 7(1): 51, pp. 1–10. DOI: https://doi.org/10.5334/egems.262 Pratt, J, et al. 2019 Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry. eGEMs (Generating Evidence & Methods to improve patient outcomes), 7(1): 51, pp. 1–10. DOI: https://doi.org/10.5334/egems.262 EMPIRICAL RESEARCH EMPIRICAL RESEARCH Introduction There is growing interest in the potential for clinical registries that can simultaneously support clinical care, quality improvement (QI), and research. This multi-purpose model is consistent with the Institute of Medicine’s (IOM’s) vision of a Learning Health System which “draws research closer to clinical practice by building knowledge development and application into each stage of the health care delivery process” [1]. Gliklich and Dreyer [2] define a registry as “an organized system that uses observational study methods to collect uniform data (clinical and other) to evaluate speci fied outcomes for a population defined by a particular disease, condition, or exposure, and that serves one or more predetermined scientific, clinical, or policy purposes.” Most pediatric chronic illnesses meet the NIH definition for rare disease [3] and, as such, multi-center registries are especially important to study and improve care for children with chronic diseases. Some multi-center networks are beginning to adopt principles of open science, or network-based production [4], to foster collaborative improvement, research, data sharing, and innovation. In this setting, the registry functions not only to provide access to condition-specific information in a uniform way to support clinical care but also to support QI and research to improve patient outcomes. The challenges and opportunities in managing data from multi-purpose clinical registries that are used for care, QI, and research are distinct from those that arise in the management of data collected specifically for study purposes, particularly clinical trials. This is largely due to the differences in the purpose of and resources available for data collec tion. In clinical trials, data collection involves a limited and pre-specified number of participants (based on a sample size determination). Data collection occurs at pre-specified time intervals (i.e. study visits) for a defined period of time. In addition, the trial data collection system is closed at the end of the study. In contrast, registries are designed to support real time care, quality improvement, and knowledge development. They involve data collection as part of routine care and must embed the process of data collection into the clinical workflow. The data reflect actual practice and patient care. Challenges in this setting may include data collection at every patient visit over an extended period of time, unstandardized visit schedules, and large numbers of data elements needed to support chronic care activities such as population management [5] and pre-visit planning [6] for an entire patient population. Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Jesse Pratt1, Daniel Jeffers2, Eileen C. King3, Michael D. Kappelman4, Jennifer Collins5, Peter Margolis3, Howard Baron6, Julie A. Bass7, Mikelle D. Bassett8, Genie L. Beasley9, Keith J. Benkov10, Jeffrey A. Bornstein11, José M. Cabrera12, Wallace Crandall13, Liz D. Dancel14, Monica P. Garin-Laflam15, John E. Grunow16, Barry Z. Hirsch17, Edward Hoffenberg18, Esther Israel19, Traci W. Jester20, Fevronia Kiparissi21, Arathi Lakhole22, Sameer P. Lapsia23, Phillip Minar3, Fernando A. Navarro24, Haley Neef25, KT Park26, Dinesh S. Pashankar27, Ashish S. Patel28, Victor M. Pineiro29, Charles M. Samson30, Kelly C. Sandberg31, Steven J. Steiner32, Jennifer A. Strople33, Boris Sudel34, Jillian S. Sullivan35, David L. Suskind36, Vikas Uppal37 and Prateek D. Wali38 1 Pharmaceutical Product Development, US 1 Pharmaceutical Product Development, US 2 Total Quality Logistics, US 3 Cincinnati Children’s Hospital Medical Center, University of Cincinnati, US 3 Cincinnati Children’s Hospital Medical Center, University 4 University of North Carolina at Chapel Hill, US 5 Cincinnati Children’s Hospital Medical Center, US 6 Pediatric Gastroenterology & Nutrition Associates, US 8 OHSU Doernbecher Children’s Hospital, US 9 UF Health Pediatric Gastroenterology, Hepatology and Nutrition, US 10 Kravis Children’s Hospital at Mount Sinai, US 14 Greenville Health System, Children’s Hospital, US 22 UCSF Benioff Children’s Hospital Oakland, US 23 Children’s Hospital of the King’s Daughters, US 24 Children’s Memorial Hermann Hospital – UT Houston, U 27 Yale-New Haven Children’s Hospital, US 27 Yale-New Haven Children’s Hospital, US 28 UT Southwestern/Children’s Health, US 28 UT Southwestern/Children’s Health, US 29 Levine Children’s Hospital, US 30 St. Louis Children’s Hospital – Washington University, US 30 St. Louis Children’s Hospital – Washington University, US 31 Dayton Children’s Hospital, US 31 Dayton Children’s Hospital, US 32 Riley Hospital for Children, US 33 Ann and Robert H. Lurie Children’s Hospital of Chicago, US 33 Ann and Robert H. Lurie Children’s Hospital of Chicago, US 34 University of Minnesota, US 35 The University of Vermont Children’s Hospital, US 35 The University of Vermont Children’s Hospital, US Art. 51, page 2 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry 36 Seattle Children’s Hospital, US 37 Nemours Children’s Health System – Wilmington, US 38 Upstate Golisano Children’s Hospital, US Corresponding author: Eileen King (eileen.king@cchmc.org) Objective: To implement a quality improvement based system to measure and improve data observational clinical registry to support a Learning Healthcare System. Keywords: Quality Improvement; Data Quality; Registry Keywords: Quality Improvement; Data Quality; Registry Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Data Source: ImproveCareNow Network registry, which as of September 2019 contained data from 314,250 visits of 43,305 pediatric Inflammatory Bowel Disease (IBD) patients at 109 participating care centers. Study Design: The impact of data quality improvement support to care centers was evaluated using statistical process control methodology. Data quality measures were defined, performance feedback of those measures using statistical process control charts was implemented, and reports that identified data items not following data quality checks were developed to enable centers to monitor and improve the quality of their data. q y Principal Findings: There was a pattern of improvement across measures of data quality. The proportion of visits with complete critical data increased from 72 percent to 82 percent. The percent of registered patients improved from 59 percent to 83 percent. Of three additional measures of data consistency and timeliness, one improved performance from 42 percent to 63 percent. Performance declined on one measure due to changes in network documentation practices and maturation. There was variation among care centers in data quality. q y Conclusions: A quality improvement based approach to data quality monitoring and improvement is feasible and effective. Setting and Centers The ImproveCareNow (ICN) Network (www.improvecarenow.org) is a multi-center international research and quality improvement network whose purpose is to transform the health, care, and costs for all children and adolescents with inflammatory bowel disease (IBD), specifically Crohn’s disease and ulcerative colitis. The network seeks to enable patients, families, clinicians and researchers to work together to accelerate innovation, discovery and the application of new knowledge. All 74 participating care centers entering data in the ICN registry from June 2010 through June 2016 were included in this study, representing data from 162,626 visits from 24,309 patients. y p g p The design of the Network has been described in detail previously [11, 12]. Briefly, ICN care centers include a mix of large and small academic medical centers and private practices over diverse geographic regions (urban/rural) and include approximately 60 percent of all pediatric gastroenterologists in the United States. Providers at each center receive instruction and ongoing coaching in QI methods, use of tools, and performance reports. Monthly webinars and semi-annual Community Conferences are held to provide ongoing training. Participating clinicians have devel oped care guidelines, tools, and processes to reduce variation in care. Centers within the network collect standardized data elements at the time patients are enrolled into the registry, at all follow-up visits, in the event of a hospitaliza tion, and when the patient discontinues participation in the registry. These elements include patient demograph ics, specific disease characteristics, level of disease activity, test results, treatments, and clinical outcomes. Registry data are used to support chronic care management reports that enable pre-visit planning, population management and patient-tracking, and comparative performance measurement (monthly charts displaying clinical, process, and outcome measures). Registry data can also be used to conduct various types of research, including comparative effectiveness studies. Centers provide their own resources to support data capture as part of their participation in the network. Development of the Quality Improvement Intervention to Enhance Data Quality The ICN Data Management Committee was formed as part of the network’s data coordinating center at Cincinnati Children’s Hospital Medical Center to inform the design, development, and testing the data quality QI process. Representatives include staff from the data coordinating center for the network as well as indi viduals from participating ICN clinical care centers. The following disciplines are included on the committee: clinical medicine, research coordination, data management, biostatistics, clinical epidemiology, project management, and QI. Introduction In addition, care centers par ticipating in multi-purpose registries participate voluntarily. Many members of the clinical care team are involved, and resources for data capture and cleaning, such as clinical auditing and source document verification, are substantially less compared with clinical trials. The same staff responsible for transcribing data from the medical record and entering into the electronic case report forms may also be responsible for completing source document verification, in addition to other administrative and/or clinical responsibilities. Such systems cannot support the data cleaning efforts typical of clinical trials that involve large numbers of queries sent to care centers for response. A key challenge to using data from registries for research is that the quality may not match that of data collected using other, more rigorous and expensive, Art. 51, page 3 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry study support [7]. To date, studies of data quality in registries have focused on retrospective assessments of the “fit for use” model which indicates that the data quality is appropriate for the intended use [8, 9]. study support [7]. To date, studies of data quality in registries have focused on retrospective assessments of the “fit for use” model which indicates that the data quality is appropriate for the intended use [8, 9]. Multi-center registries have used quality improvement methodology to improve patient care and outcomes. These same methods may be extended to interventions that enable teams to improve data quality. The impact of a data qual ity improvement project based on good clinical data management practices [10] was evaluated within a multi-center registry for clinical care, QI, and research. Setting and Centers The data coordinating center of the network reports to the network’s executive leadership and is an ongoing network function. The DCC and the committee are accountable for the ongoing maintenance and improvement of data quality. Its work is funded as part of network operations. Figure 1: ImproveCareNow’s data quality key driver diagram. reports, and pre-visit planning reports. In addition, the reporting system was modified so that the results were updated on a nightly basis to align with daily data collection updates. This enabled centers to view the charts at any time, whereas previously changes were not reflected until the next reporting month. Components of the Data Quality Intervention A full population registry contains complete data for all visits for all patients in the registerable (target) population. Determining if all potential patients were included in the registry and if all visits had been entered could not be done solely by querying the registry but required additional work for the centers to identify and report their eligible popula tion and evaluate which visits should be entered into the registry. To obtain an accurate and timely population total, centers must query clinical records to identify the number of eligible patients followed in the practice. Early efforts focused on developing center-specific procedures to identify their eligible population and subsequently providing this total to ICN on a quarterly basis. Centers with early success were asked to share the details of their process with all cent ers at network-wide meetings and webinars. This total population is now entered by each center into the registry on a quarterly basis. If they do not update the number, it carries forward. Using center reported denominators, the percent age of eligible patients with data in the registry is calculated. When a patient no longer participates in the registry, the center changes the status for that patient to “deactivated”. This can occur for numerous reasons including transferring to an adult gastroenterologist, moving, etc. To ensure that the patient population at each center is accurately repre sented, the percentage of patients with a visit in the past 13 months is calculated and monitored. This measure helps centers identify and remove patients no longer receiving care at the center. If a patient has not been seen, this could be an indication that they should be deactivated and should not be included as part of the practice population or the registry moving forward. g y g To estimate the percentage of visits captured in the registry each month, centers are asked to provide the visit dates and patient IDs for any registered patient having a visit during the first full week of each month. The percentage of visits captured in the registry for that week is then calculated and used to estimate the percentage of visits captured monthly. Hospitalization is a serious outcome for any patient, and data quality is monitored using two mechanisms. The first is a measure assessing whether centers entered at least one hospitalization in the past 90 days. Development of the Quality Improvement Intervention to Enhance Data Quality p p A structured QI process was used to design a data quality system with the following aim: “To assist ICN care centers in creating a reliable, accurate and consistent data collection system and full population registry (all patients, all visits, all data items) so patients can rely on the recommendations generated from the reg istry and physicians and researchers can use data to make decisions with confidence about the care and outcomes of children with IBD.” A logic model known as a key driver diagram was developed to identify changes and interventions that support the drivers of good data quality (Figure 1). These key drivers included having consistent and timely information about the population, having reliable data collection practices, and the application of QI methodology to improve data quality. As illustrated in the key driver diagram, specific interventions in several processes were identified: creating and maintain ing a full population registry, developing a reliable process for data collection, and monitoring data quality. The data quality improvement effort was built upon ongoing QI training and support processes that were already established for centers. The team initially defined nine measures that measured completeness, consistency, and timeli ness of data entered into the ICN registry (see Appendix A). These measures were implemented in April 2011 and the consistency measures were bundled together in late 2013 due to high performance on each measure and similarity of content. The measures still covered the same broad categories (completeness, consistency, and timeliness). Two addi tional measures for monitoring hospitalization data were added in 2014. Only current measures only are presented. The original data quality reports were developed to look like the reports for the clinical process and outcomes measures already utilized by centers and distributed on a monthly basis. By July 2013, the data quality charts were made available electronically in conjunction with the other QI tools in the system including clinical measures, population management Art. 51, page 4 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Figure 1: ImproveCareNow’s data quality key driver diagram. Monitoring Data Quality In September 2011, network centers were asked to conduct a failure mode and effect analysis [13] of their data collec tion and capture process based on the data quality and exceptions reports. This exercise led centers toward the appli cation of high reliability processes for data collection and capture including documenting the process of data capture with process flow diagrams. These tools were subsequently incorporated into training for centers joining the network. Ideas to improve data collection continue to be posted on a website that enables centers to share knowledge learned. Center-level and aggregate statistical process control charts were developed to show performance on each of the measures of data quality. A sample data quality report that ICN centers see in the registry is displayed in Figure 2. Measures are designed to assess completeness, timeliness, and consistency of the data. Four measures of completeness for population, outpatient visits, and hospitalizations were described above. An additional measure of completeness examines the data elements considered critical that are entered for each outpatient visit. These data elements are critical because they are required to determine disease activity and track medication use, key components of patient outcome data. Specifically, the set of critical variables include height, weight, medication data, Physician’s Global Assessment (PGA), and the individual components needed to score the short Pediatric Crohn’s Disease Activity Index (sPCDAI) [14] and Pediatric Ulcerative Colitis Activity Index (PUCAI) [15]. Two additional measures are used to assess timeliness: visits and hospitalizations entered within 30 days. Finally, consistency was evaluated using four measures. The first required age adjusted height, weight, and BMI z-scores to be between –4 and 4; the second required physician global assess ment and PUCAI/sPCDAI to differ by no more than one classification category; the third required height to be no more than 1 centimeter less than it was on the prior visit; the fourth required changes in disease extent and phenotype to be clinically acceptable. Due to high performance on each of four measures, they were bundled, or combined into a single measure of intra- and inter-visit consistency, thereby requiring each of the four previous measures to be satis fied to pass the bundle measure. Data quality improvement training began in March 2011, and centers received their first data quality reports in May 2011. Components of the Data Quality Intervention 51, page 5 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Components of the Data Quality Intervention As the network has become more diverse, it has become apparent that the assumption that every center will have a hospitalization to enter every 90 days is not necessarily valid. To assess the timeliness of data entry for hospitalizations, a second measure was implemented, tracking the percentage of hospitalizations entered into the registry within 30 days of discharge. More recently, centers have been asked to start entering the total number of hospitalizations that occur each month to com pare with the number entered, a process similar to how visits are tracked. g y g To estimate the percentage of visits captured in the registry each month, centers are asked to provide the visit dates and patient IDs for any registered patient having a visit during the first full week of each month. The percentage of visits captured in the registry for that week is then calculated and used to estimate the percentage of visits captured monthly. Hospitalization is a serious outcome for any patient, and data quality is monitored using two mechanisms. The first is a measure assessing whether centers entered at least one hospitalization in the past 90 days As the network has To estimate the percentage of visits captured in the registry each month, centers are asked to provide the visit dates and patient IDs for any registered patient having a visit during the first full week of each month. The percentage of visits captured in the registry for that week is then calculated and used to estimate the percentage of visits captured monthly. Hospitalization is a serious outcome for any patient, and data quality is monitored using two mechanisms. The first is a measure assessing whether centers entered at least one hospitalization in the past 90 days. As the network has become more diverse, it has become apparent that the assumption that every center will have a hospitalization to enter every 90 days is not necessarily valid. To assess the timeliness of data entry for hospitalizations, a second measure was implemented, tracking the percentage of hospitalizations entered into the registry within 30 days of discharge. More recently, centers have been asked to start entering the total number of hospitalizations that occur each month to com pare with the number entered, a process similar to how visits are tracked. Art. Monitoring Data Quality Center personnel were trained to monitor the data quality reports for trends to evaluate improvement efforts or to detect problems so that process failures could be identified and improvements implemented. New centers are trained in a similar manner during their onboarding process. To enable centers to easily identify data quality failures, a data quality “exceptions report” was developed, similar in appearance to the population management reports already in use by centers. These reports allow users easy access to specific information (a list of data elements that may represent an error) about each data quality failure. One example of a center’s exception report for the complete critical data measure is displayed in Figure 3. Figure 2: Sample center level data quality report as received by participating centers monthly. Figure 2: Sample center level data quality report as received by participating centers monthly. Figure 2: Sample center level data quality report as received by participating centers monthly. Art. 51, page 6 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry page 6 of 10 3: Sample exceptions report demonstrating center’s view of missing data elements. Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry age 6 of 10 : Sample exceptions report demonstrating center’s view of missing data elements. Figure 3: Sample exceptions report demonstrating center’s view of missing data elements. Figure 3: Sample exceptions report demonstrating center’s view of missing data elements. Figure 3: Sample exceptions report demonstrating center’s view of missing data elements. Beginning in January 2015, the ICN Data Management Committee began monitoring data quality charts monthly for the network as a whole and on a quarterly basis for each center. Follow-up is conducted with centers to identify oppor tunities for shared learning or improvement. Centers are also reminded about data quality through routine calls with their quality improvement consultants and on network-wide webinars. Statistical Analysis i i l Statistical process control (SPC) methods were used to monitor changes in performance for each data quality measure. Control charts were used to analyze process improvement efforts and identify and classify causes of variation into com mon cause variation (inherent in the system) and special cause variation (variations not normally part of the system that arise due to specific events) [16]. Center lines were estimated and control limits were calculated as the center line pro portion ±3 standard deviations. Each control chart was monitored for evidence of special cause changes using standard SPC rules [17–19]. Center-lines were shifted if a sustainable change had taken place based upon review by the ICN Data Management Committee taking into account SPC rules and subject matter expertise. Results By June 2016, 74 centers were participating in the ICN registry and had enrolled 24,309 patients and entered data from 162,626 visits. There was a pattern of improvement across 8 measures of data quality. The percentage of registered patients increased from 59 percent to 83 percent, with 51 percent of centers having registered at least 90 percent of their IBD population. Overall, 92 percent of visits of registered patients were recorded. In October 2011, there was evi dence of a significant improvement in the percent of visits with all critical variables recorded based on the SPC criterion of one point outside of the control limits (Figure 4). By November 2014, there was evidence of additional improvement. Based on this special cause and subject matter expertise indicating purposeful changes to the system of data quality, the center line and associated control limits were shifted from 72 percent for the period June 2010 to September 2011 to 78 percent for the period October 2011 to October 2014, and again to 82 percent for the period November 2014 to June 2016. Figure 5 is a small multiple display of the complete critical data measure for participating individual cent ers. This display highlights the variability among centers for this measure. Some of the centers had high data quality across all time points. Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Art. 51, Figure 4: Annotated control chart displaying change in aggregate data element completeness. i d l f d d li i d i i h i i Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Art. 51, page 7 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Art. 51, page 7 of 10 Figure 4: Annotated control chart displaying change in aggregate data element completeness. Figure 4: Annotated control chart displaying change in aggregate data element completeness. Figure 4: Annotated control chart displaying change in aggregate data element completeness. Figure 4: Annotated control chart displaying change in aggregate data element completenes Figure 5: Random sample of 25 centers exposed to data quality reporting and training, showing intra-vis completeness over time. Discussion We found that efforts to improve data quality that included training in quality improvement methods, and tools includ ing control charts, exception reports and failure mode and effect analysis resulted in improvement in a range of data quality measures. These measures have continued to improve since 2016. This improvement was augmented by the development of electronically available control charts and reports that were updated daily and available to the centers on demand. We also observed variation among centers in data quality that could be due to the variation in use of these tools by a center. Measures of completeness of critical data and the percentage of patients registered increased. Meas ures of consistency and the percentage of visits recorded in the registry remained at their high baseline levels. The success of the ICN registry rests on the QI training participating centers had already received to standardize clinical care processes to improve outcomes. Compared to comprehensive data cleaning that is typically used for clini cal trials, this approach is much less labor intensive and requires fewer resources. Our approach is conceivably more sustainable given the focus on changing processes to improve data quality at the point of data entry instead of correct ing errors after entry takes place. g y p Several authors have discussed the importance of high data quality in data registries [20–26]. Previous studies have emphasized methods to assess the quality of data in registries, but there has been little work on models to improve registry data quality that go beyond centralized data quality support. Kahn et al [24] proposed several data quality assessment methods to gauge single-site data quality that include rules for evaluating single-item completeness and validity (e.g. toddler’s height cannot be 7 ft.) and cross-item temporal (e.g. surgery date cannot be after death date) and relational consistency (e.g. males cannot have positive pregnancy test). The authors also discussed analyzing temporal data to evaluate logical consistency relative to the evolution of a process or set of states over time and the examina tion of conditional dependencies based on knowledge of a clinical scenario (e.g. gestational period for human cannot be 48 months). Brown et al [22] discusses best practices and provides recommendations for data quality checking in distributed data networks. In a distributed data network, a common data model is developed to standardize the content and format of observational data that facilitates cross-site analyses. Discussion Data are checked to ensure they conform to the common data model (e.g. valid values for SEX “M”, “F”, “U”). In addition to evaluating conformity with the data model, Brown states that additional checks should be done by evaluating individual item completeness and validity and cross- item temporal and logical consistency [22]. Despite differences in methodology, both aforementioned groups recom mend making information about data quality of a database available to users of the database. To our knowledge, this is the first report detailing an effort to improve data quality using quality improvement methods and additional tools described in this manuscript. A limitation of this study is that centers were not randomly assigned to the data quality QI support intervention. Although it is possible that improvement could have taken place on its own, prior to the intervention, the measures were stable over a long period of time, and we did not observe a special cause signal until the implementation of the data quality program. This observation is limited by the inability to track the use of the data quality tools. However, this is no longer a problem as the reporting system has gone online and now has usage tracking capabilities. The ICN Data Management Committee provides a structure to enable the network’s focus on improving data quality to evolve over time. Ongoing monitoring of data quality is part of the network’s monthly dashboard of measures. The auto mated assessment of data quality measures allows the network to monitor data quality over time and to detect changes in data quality that need further exploration as well as to inform the ongoing improvement of the measures themselves. q y p g g p Complete and accurate registry data are essential for managing chronically ill patients, guiding improvement efforts, and research. As the emphasis on multi-site registries and networks increases, so will the importance of their data qual ity. Future work should focus on designing, developing, and testing additional interventions to further increase the quality of registry data. By raising awareness of the importance of high data quality and supporting member centers’ use of QI training and tools, the completeness and consistency of data within the ICN registry has increased. In conclusion, our work demonstrates that QI approaches to improving data quality are effective and feasible to implement in research and improvement networks. Results The measure of patients active in the registry with a visit in the past 13 months increased over time from 89 percent to 91 percent. By June 2016, the percentage of centers having at least one hospital discharge in the past 90 days was 82 percent and the percentage of hospitalizations entered into the registry within 30 days of discharge improved from 42 percent to 63 percent. By June 2016, the percentage of centers having at least one hospital discharge in the past 90 days was 82 percent and the percentage of hospitalizations entered into the registry within 30 days of discharge improved from 42 percent to 63 percent. Results For each center, the percentage of visits each month with complete critical data elem indicated on the y-axis. The x-axis indicates June 2010 through June 2016. Figure 5: Random sample of 25 centers exposed to data quality reporting and training, showing intra-visit data completeness over time. For each center, the percentage of visits each month with complete critical data elements is indicated on the y-axis. The x-axis indicates June 2010 through June 2016. Art. 51, page 8 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Improvements in data quality over time can be seen in other measures. From June 2010 to June 2016, the completion rate for PUCAI increased from 73 percent to 91 percent and that for sPCDAI increased from 60 percent to 68 percent. The bundled measure of intra- and inter-visit consistency of the data (Appendix A) demonstrated common cause varia tion. During this time there was a change in the definition of extent of disease from the Montreal classification to the Paris classification. The percentage of visits entered within 30 days of visit date varied widely over time but demonstrated improvement as well. New chronic care management reports (i.e. population management and pre-visit planning) were implemented in 2012 that required timely data entry for their effectiveness, leading to improvements in this measure. The measure of patients active in the registry with a visit in the past 13 months increased over time from 89 percent to 91 percent. By June 2016, the percentage of centers having at least one hospital discharge in the past 90 days was 82 percent and the percentage of hospitalizations entered into the registry within 30 days of discharge improved from 42 percent to 63 percent. The percentage of visits entered within 30 days of visit date varied widely over time but demonstrated improvement as well. New chronic care management reports (i.e. population management and pre-visit planning) were implemented in 2012 that required timely data entry for their effectiveness, leading to improvements in this measure. The measure of patients active in the registry with a visit in the past 13 months increased over time from 89 percent to 91 percent. in 2012 that required timely data entry for their effectiveness, leading to improvements in this measure. Acknowledgements The research reported in this article was conducted in the ImproveCareNow Network, a Learning Health System for children and youth with Inflammatory Bowel Disease. The authors acknowledge all of the patients, clinicians and staff who make together make up the network. They particularly acknowledge Melissa Disorda who served as a research coor dinator and facilitated the work described here. Pamela J. Schoettker assisted with the preparation of the manuscript. This work was partially supported by the ImproveCareNow Network, NIH TRO1 5R01DK085719-05, AHRQ Enhanced Registries 1R01HS022974-01, PCORI PPRN-1306-01754. Discussion Furthermore, this results in a higher level of confidence when accessing the data for various purposes, including clinical decision making. Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Art. 51, page 9 of 10 Competing Interests The authors have no competing interests to declare. References 1. Olsen, L, Aisner, D and McGinnis, J. The Learning Healthcare System. Washington, DC: National Academy Press; 2007. 1. Olsen, L, Aisner, D and McGinnis, J. The Learning Healthcare System. Washington, DC: National Academy Press; 2007. 2. Gliklich, R and Dreyer, N. Registries for Evaluating Patient Outcomes: A User’s Guide. (Prepared by Outcome DEcIDE Center [Outcome Sciences, Inc. dba Outcome] under Contract No. HHSA290200500351 T01). Rockville, MD: Agency for Healthcare Research and Quality, April 2007 Contract No.: AHRQ Publication No. 07-EHC001-1. 2. Gliklich, R and Dreyer, N. Registries for Evaluating Patient Outcomes: A User’s Guide. (Prepared by Outcome DEcIDE Center [Outcome Sciences, Inc. dba Outcome] under Contract No. HHSA290200500351 T01). Rockville, MD: Agency for Healthcare Research and Quality, April 2007 Contract No.: AHRQ Publication No. 07-EHC001-1. Agency for Healthcare Research and Quality, April 2007 Contract No.: AHRQ Publication No. 07-EHC001- g p 3. U.S. Department of Health & Human Services, National Institutes of Health. National Center for Advancing Translational Sciences. Genetic and Rare Disease Information Center. [November 2, 2017]. Available from: https:// rarediseases.info.nih.gov/. 3. U.S. Department of Health & Human Services, National Institutes of Health. National Center for Advancing Translational Sciences. Genetic and Rare Disease Information Center. [November 2, 2017]. Available from: https:// rarediseases.info.nih.gov/. 4. Benkler, Y. Intellectual property. Commons-based strategies and the problems of patients. Science. 2004; 305(5687): 1110–1. DOI: https://doi.org/10.1126/science.1100526 4. Benkler, Y. Intellectual property. Commons-based strategies and the problems of patients. Science. 2004; 305(5687): 1110–1. DOI: https://doi.org/10.1126/science.1100526 5. Backus, L, Gavrilov, S, Loomis, T, Halloran, J, Phillips, B, Belperio, P, et al. Clinical Case Registries: simulta- neous local and national disease registries for population quality management. J Am Med Inform Assoc. 2009; 16(6): 775–83. DOI: https://doi.org/10.1197/jamia.M3203 5. Backus, L, Gavrilov, S, Loomis, T, Halloran, J, Phillips, B, Belperio, P, et al. Clinical Case Registries: simulta- neous local and national disease registries for population quality management. J Am Med Inform Assoc. 2009; 16(6): 775–83. DOI: https://doi.org/10.1197/jamia.M3203 6. Wagner, E, Austin, B, Davis, C, Hindmarsh, M, Schaefer, J and Bonami, A. Improving chronic illness care: translating evidence into action. Health Aff (Millwood). 2001; 20(6): 64–78. DOI: https://doi.org/10.1377/ hlthaff.20.6.64 7. Botsis, T, Hartvigsen, G, Chen, F and Weng, C. Secondary use of EHR data quality issues and informatics opportunities. AMIA Jt Summits Transl Sci Proc. 2010; 2010: 1–5. 8. Gryna, F, Chua, R, Defeo, J and Juran, J. Juran’s Quality Planning and Analysis for Enterprise Quality. Boston: McGraw-Hill; 2006. 9. References Juran, J and Gryna, F. Quality Planning and Analysis: From Product Development Through Use. New York: McGraw-Hill; 1980. 10. Society for Clinical Data Management. Good Clinical Data Management Practices (GCDMP). [November 2, 2017]. Available from: http://scdm.org/publications/gcdmp/. 11. Crandall, W, Boyle, B, Colletti, R, Margolis, P and Kappelman, M. Development of process and outcome measures for improvement: lessons learned in a quality improvement collaborative for pediatric bowel disease. Inflamm Bowel Dis. 2011; 17(10): 2184–91. DOI: https://doi.org/10.1002/ibd.21702 12. Crandall, W, Kappelman, M, Colletti, R, Leibowitz, I, Grunow, J, Ali, S, et al. ImproveCareNow: The development of a pediatric inflammatory bowel disease improvement network. Inflamm Bowel Dis. 2010; 17(1): 450–7. DOI: https://doi.org/10.1002/ibd.21394 p g 13. Rausand, M and Hoyland, A. System Reliability Theory: Models, Statisticial Methods, and Applications. Hoboken, NJ: Wiley-Interscience; 2004. 14. Kappelman, M, Crandall, W, Colletti, R, Goudie, A, Leibowitz, I, Duffy, L, et al. Short pediatric Crohn’s disease activity index for quality improvement and observational research. Inflamm Bowel Dis. 2011; 17(1): 112–7. DOI: https://doi.org/10.1002/ibd.21452 15. Turner, D, Otley, A, Mack, D, Hyams, J, de Bruijne, J, Uusoue, K, et al. Development, validation, and evaluation of a pediatric ulcerative colitis activity index: a prospective multicenter study. Gastroenterology. 2007; 133(2): 423–32. DOI: https://doi.org/10.1053/j.gastro.2007.05.029 p g j g 16. Moen, R, Nolan, T and Provost, L. Quality Improvement through Planned Experimentation. NY: 1999. 17. Amin, S. Control charts 101: a guide to health care applications. Qual Manag Health Care. 2001; 9: 1–27. DOI: https://doi.org/10.1097/00019514-200109030-00003 p g 18. Benneyan, J, Lloyd, R and Plsek, P. Statistical process control as a tool for research and healthcare improve- ment. Qual Saf Health Care. 2003; 12(6): 458–64. DOI: https://doi.org/10.1136/qhc.12.6.458 19. Provost, L and Murray, S. The Health Care Data Guide: Learning from Data for Improvement. San Francisco, CA: Jossey-Bass; 2011. 20. Aronsky, D and Haug, P. Assessing the quality of clinical data in a computer-based record for calculating the pneumonia severity index. J Am Med Inform Assoc. 2000; 7(1): 55–65. DOI: https://doi.org/10.1136/ jamia.2000.0070055 Art. 51, page 10 of 10 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry 21. Arts, D, De Keizer, N and Scheffer, G. Defining and improving data quality in medical registries: a literature review, case study, and generic framework. References J Am Med Inform Assoc. 2002; 9(6): 600–11. DOI: https://doi. org/10.1197/jamia.M1087 j J, Kahn, M and Toh, S. Data quality assessment for comparative effectiveness research in distributed data ks. Med Care. 2013; 51(8 Suppl 3): S22–S9. DOI: https://doi.org/10.1097/MLR.0b013e31829b1e2c 22. Brown, J, Kahn, M and Toh, S. Data quality assessment for comparative effectiveness research in distributed data networks. Med Care. 2013; 51(8 Suppl 3): S22–S9. DOI: https://doi.org/10.1097/MLR.0b013e31829b1e2c 23. Hlaing, T, Hollister, L and Aaland, M. Trauma registry data validation: essential for quality trauma care. J Trauma. 2006; 61(6): 1400–7. DOI: https://doi.org/10.1097/01.ta.0000195732.64475.87 networks. Med Care. 2013; 51(8 Suppl 3): S22 S9. DOI: https://doi.org/10.1097/MLR.0b013e31829b1e2c 23. Hlaing, T, Hollister, L and Aaland, M. Trauma registry data validation: essential for quality trauma care. J Trauma. 2006; 61(6): 1400–7. DOI: https://doi.org/10.1097/01.ta.0000195732.64475.87 ( ) p g 24. Kahn, M, Raebel, M, Glanz, J, Riedlinger, K and Steiner, J. A pragmatic framework for single-site and multisite data quality assessment in electronic health record-based clinical research. Med Care. 2012; 50(Suppl): S21–S9. DOI: https://doi.org/10.1097/MLR.0b013e318257dd67 25. Messenger, J, Ho, K, Young, C, Slattery, L, Draoui, J, Curtis, J, et al. The National Cardiovascular Data Registry (NCDR) Data Quality Brief: the NCDR Data Quality Program in 2012. J Am Coll Cardiol. 2012; 60(16): 1484–8. DOI: https://doi.org/10.1016/j.jacc.2012.07.020 p g j j 26. Weiner, M and Embi, P. Toward reuse of clinical data for research and quality improvement: the end of the beginning? Ann Intern Med. 2009; 151(5): 359–60. DOI: https://doi. org/10.7326/0003-4819-151-5-200909010-00141 How to cite this article: Pratt, J, Jeffers, D, King, EC, Kappelman, MD, Collins, J, Margolis, P, Baron, H, Bass, JA, Bassett, MD, Beasley, GL, Benkov, KJ, Bornstein, JA, Cabrera, JM, Crandall, W, Dancel, LD, Garin-Laflam, MP, Grunow, JE, Hirsch, BZ, Hoffenberg, E, Israel, E, Jester, TW, Kiparissi, F, Lakhole, A, Lapsia, SP, Minar, P, Navarro, FA, Neef, H, Park, KT, Pashankar, DS, Patel, AS, Pineiro, VM, Samson, CM, Sandberg, KC, Steiner, SJ, Strople, JA, Sudel, B, Sullivan, JS, Suskind, DL, Uppal, V and Wali, PD 2019 Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry. eGEMs (Generating Evidence & Methods to improve patient outcomes), 7(1): 51, pp. 1–10. DOI: https://doi.org/10.5334/ egems.262 Published: 30 September 2019 Submitted: 19 March 2018 Pratt et al: Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry Submitted: 19 March 2018 Copyright: © 2019 The Author(s). This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC-BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See http://creativecommons.org/licenses/by/4.0/. eGEMs (Generating Evidence & Methods to improve patient outcomes) is a peer-reviewed open access journal published by Ubiquity Press. OPEN ACCESS
https://openalex.org/W4312070625	https://www.researchsquare.com/article/rs-2389077/latest.pdf	English	null	ALL ROAD USER CASUALTIES (KILLED) IN GREAT BRITAIN FROM 1926. LINEAR AND NONLINEAR TRENDS WITH PERSISTENT DATA	Research Square (Research Square)	2,022	cc-by	4,402	ALL ROAD USER CASUALTIES (KILLED) IN GREAT BRITAIN FROM 1926. LINEAR AND NONLINEAR TRENDS WITH PERSISTENT DATA Luis Alberiko Gil-Alana (  alana@unav.es ) University of Navarra Research Article Keywords: road casualties, Great Britain, fractional integration, persistence, breaks. ALL ROAD USER CASUALTIES (KILLED) IN GREAT BRITAIN FROM 1926. LINEAR AND NONLINEAR TRENDS WITH PERSISTENT DATA Luis Alberiko Gil-Alana University of Navarra, Pamplona, Spain and Universidad Francisco de Vitoria, Madrid, Spain Abstract This paper deals with the analysis of road casualties in Great Britain, using annual data since 1926. Based on the persistent nature of the data, fractional integration methods are used that include linear and non-linear (structural breaks) models. The results indicate that when the whole data set is employed the series is nonstationary I(1) implying permanency of shocks. However, considering data starting in 1964 we observe a significant negative time trend along with a lower degree of integration that implies transitory shocks. In order to avoid the abrupt change produced by the break, a nonlinear deterministic trend model based on Chebyshev polynomials in time is also considered with the whole sample, and though the order of integration is much lower than 1, the unit root null hypothesis cannot yet be rejected. Keywords: road casualties; Great Britain; fractional integration; persistence; breaks. JEL Classification: C22; C13; C54 JEL Classification: C22; C13; C54 Prof. Luis Alberiko Gil-Alana University of Navarra Faculty of Economics, Edificio Amigos E-31009 Pamplona, Spain email: alana@unav.es Prof. Luis Alberiko Gil-Alana University of Navarra Faculty of Economics, Edificio Amigos E-31009 Pamplona, Spain email: alana@unav.es DOI: https://doi.org/10.21203/rs.3.rs-2389077/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Additional Declarations: No competing interests reported. Additional Declarations: No competing interests reported. Corresponding author: * The three authors acknowledge support from an Internal Project of the Universidad Francisco de Vitoria. Luis A. Gil-Alana gratefully acknowledges financial support from the MINEIC-AEI-FEDER PID2020- 113691RB-I00 project from ‘Ministerio de Economía, Industria y Competitividad’ (MINEIC), `Agencia Estatal de Investigación' (AEI) Spain and `Fondo Europeo de Desarrollo Regional' (FEDER). 1 1 1. Introduction Stationarity is a minimal requirement in time series to make statistical inference. Until the 1980s, the standard approach for removing the nonstationary component was to make regressions on time, under the assumption that the estimated errors were then stationary. Later on, and especially after the seminal paper by Nelson and Plosser (1982), taking first (or second) differences in the series became standard practice. This came after the development of various unit root tests, starting with Fuller (1976) and Dickey and Fuller (1979), and followed by other such as Phillips and Perron (1988), Kwiatkowski et al. (1992), Elliot et al., (1996), etc. Nevertheless, all these methods consider only two alternatives for the series of interest: stationarity I(0) or nonstationarity I(1). In other words, you have to determine whether to work with the series itself, in levels, if it is I(0), or in first differences if it is I(1) (in some special cases the series might require twice differentiation if it is I(2)) . During the 90s, however, a more flexible approach became very popular that allows the number of differences to be adopted in the series to be a fractional value. Thus, it may be 0 or 1, but also a value below 0, in the interval (0, 1) or even above 1. Granger (1980) proposed this type of model under the observation that the periodogram of many aggregated time series possessed a very large value at the smallest (zero) frequency, which may be consistent with the need for differentiation, but once the series were differentiated, the periodograms displayed a value close to zero at the zero frequency, which was consistent with overdifferentiation. Years later, Gil-Alana and Robinson (1997) examined an updated version of the fourteen macro series studied in Nelson and Plosser (1982). The latter authors, using the tests of Dickey and Fuller (1979) found that all except one of the series were consistent with I(1) nonstationary models. Gil-Alana and Robinson (1997), however, discovered that fractional degrees of 2 2 differentiation were appropriate for all the series examined. Fractional integration then became popular in the forthcoming years (see, e.g., Michelacci and Zaffaroni, 2000; Box- Steffensmeier and Tomlinson, 2000; Candelon and Gil-Alana, 2004; Gil-Alana and Moreno, 2007; Abbritti et al., 2016; etc.) In this paper, various models based on fractional integration are employed in the analysis of road casualties in Great Britain. 1. Introduction Linear and non-linear trends along with breaks are permitted in the data, and the results display different interpretations depending on the model employed. The rest of the paper is structured as followed. Section 2 is devoted to the methodology and the different models used in the manuscript. Section 3 presents the data and the main empirical results, while Section 4 contains the most relevant conclusions. 2. Methodology On the other hand, authors such as Diebold and Rudebusch (1991), Hassler and Wolters (1994), Lee and Schmidt (!996) and others showed that most of the above mentioned approaches have very low power if the true data generating process is I(d) where d is a fractional value. If the differencing parameter d is fractional, equation (1) can be extended as .... ,1 , 0 ) ( ) ( ) 1(  = = − t t u t x L d (3) (3) and, as long as d is positive, x(t) in (3) becomes long memory in the sense that the infinite sum of the autocorrelations is infinite, i.e., and, as long as d is positive, x(t) in (3) becomes long memory in the sense that the infinite sum of the autocorrelations is infinite, i.e., . ) (  =   − = j u  (4) (4) Alternatively, we can describe the two types of processes, long memory and short memory in the frequency domain. Defining the spectral density function, f(λ), as the Fourier transform of the autocovariances, i.e., Alternatively, we can describe the two types of processes, long memory and short memory in the frequency domain. Defining the spectral density function, f(λ), as the Fourier transform of the autocovariances, i.e., ). ( cos ) ( ) ( ) ( u u e u f j u i j         − =  − = = = (6) (6) we say that x(t) is long memory if the spectrum goes to infinity at least at one point in the frequency [0, π), we say that x(t) is long memory if the spectrum goes to infinity at least at one point in the frequency [0, π), ). ,0 [ some for , ) (      → f (6) (6) On the other hand, it is said to be short memory if f(λ) is positive and bounded at all frequencies, On the other hand, it is said to be short memory if f(λ) is positive and bounded at all frequencies,    ) ( 0  f , (7)    ) ( 0  f , (7) (7) as in the white noise case, where f(λ) is constant across λ. 2. Methodology We define L as the lag operator, implying that Lkx(t) =x(t-k). In this context, a series is We define L as the lag operator, implying that Lkx(t) =x(t-k). In this context, a series is said to be integrated of order 1, or I((1) if it can be represented as id to be integrated of order 1, or I((1) if it can be represented as (1 − 𝐿) 𝑥(𝑡) = 𝑢(𝑡), 𝑡= 1, 2, …, (1) (1) where x(t) represents a time series, and u(t) is a covariance stationary process that is short memory or integrated of order 0. Short memory processes are characterized because the infinite sums of its autocovariances, defined as γ(u) = E[(x(t) – μ)(x(t+u) – μ), is finite, i.e., where x(t) represents a time series, and u(t) is a covariance stationary process that is short memory or integrated of order 0. Short memory processes are characterized because the infinite sums of its autocovariances, defined as γ(u) = E[(x(t) – μ)(x(t+u) – μ), is finite, i.e., i.e., . ) (     − = j u  (2) . ) (     − = j u  (2) (2) 3 In this category of short memory processes, we can include the classical AutoRegressive Moving Average (ARMA) processes such that if u(t) is ARMA(p, q), x(t) is said to be ARIMA(p, 1, q). As mentioned in the above section, there are many statistical testing methods to determine if a series is I(1) or I(0) (also named short memory) such as ADF 3 (Dickey and Fuller, 1979), PP (Phillips and Perron, 1988), KPSS (Kwiatkowski et al., 1992), ERS (Elliot et al., 1996), NP (Ng and Perron, 2001), but all them only consider integer orders of integration. On the other hand, authors such as Diebold and Rudebusch (1991), Hassler and Wolters (1994), Lee and Schmidt (!996) and others showed that most of the above mentioned approaches have very low power if the true data generating process is I(d) where d is a fractional value. If the differencing parameter d is fractional, equation (1) can be extended as (Dickey and Fuller, 1979), PP (Phillips and Perron, 1988), KPSS (Kwiatkowski et al., 1992), ERS (Elliot et al., 1996), NP (Ng and Perron, 2001), but all them only consider integer orders of integration. 2. Methodology In the context of (3) it can be as in the white noise case, where f(λ) is constant across λ. In the context of (3) it can be shown that the spectral density function of x(t) is , 1 1 2 ) ( d 2     ie f − = (8) (8) and it tends to infinity as λ → 0+ with d > 0 justifying its long memory property. The fractional integrated process in (3) allows for a great variety of specifications, including among others: The fractional integrated process in (3) allows for a great variety of specifications, including among others: i) anti-persistence, if d < 0, i) anti-persistence, if d < 0, ii) short memory, if d = 0 iii) long memory, though covariance stationary processes, if 0 < d < 0.5, iv) nonstationary mean reverting processes, if 0.5 ≤ d < 1, iv) nonstationary mean reverting processes, if 0.5 ≤ d < 1, ) it t if d 1 v) unit roots if d = 1, vi) and explosive processes if d > 1. vi) and explosive processes if d > 1. In addition, the model in (1) can be extended to allow for linear (or even non- linear) trends. In the most classical way, and following the standard unit roots literature (Bhargava, 1986, Schmidt and Phillips, 1992), we consider a linear time trend model of the following form, 𝑦(𝑡) = 𝛼 + 𝛽 𝑡 + 𝑥(𝑡), 𝑡= 1, 2, …, (9) where α and β are unknown coefficients to be estimated, specifically a constant and a time trend, and where xt is given by equation (3). 𝑦(𝑡) = 𝛼 + 𝛽 𝑡 + 𝑥(𝑡), 𝑡= 1, 2, …, (9) where α and β are unknown coefficients to be estimated, specifically a constant and a time trend, and where xt is given by equation (3). 𝑦(𝑡) = 𝛼 + 𝛽 𝑡 + 𝑥(𝑡), 𝑡= 1, 2, …, (9) (9) In the empirical application conducted in the following section, we first estimate a model given by the equations (9) and (3), assuming that the error term u(t) adopts first a white noise process, then an AR(1) structure, and finally the exponential spectral approach of Bloomfield (1973), which is a non-parametric technique that approximates AR structures by means of an approximation of the logged spectral density function. 2. Methodology The presence of structural breaks is also taken into account by using Bai and Perron’s (2003) and Gil-Alana’s (2008) approaches, the latter being an extension of the 5 former to the fractional case. This is used noting that various authors have argued that fractional integration might be an artificial artifact generated by the presence of breaks in the data that have not been taken into account (Diebold and Inoue, 2001; Granger and Hyung, 2004; etc.). Finally, in order to avoid the abrupt change produced by the use of structural breaks, a non-linear deterministic model, based on the Chebyshevs polynomials in time, and used in Gil-Alana and Cuestas (2016) in the context of I(d) models is also employed in the paper. Using this model, we replace equation (9) by , ... , 2 ,1 ,) ( ) ( ) ( 0 = + =  = t t x t P t y m i iT i (10) (10) with m indicating the order of the Chebyshev polynomial Pi,T(t) defined as: ,1 ) ( ,0 = t P T ( ) ... , 2 ,1 ; , ... , 2 ,1 , /) 5.0 ( cos 2 ) ( , = = − = i T t T t i t P T i  . (11) (11) (11) See Hamming (1973) and Smyth (1998) for a detailed description of these polynomials. Authors such as Bierens (1997) and Tomasevic and Stanivuk (2009) argued that it is possible to approximate highly non-linear trends with rather low degree polynomials. If m = 0 the model contains an intercept, and if m > 0, it becomes non-linear - the higher m is the less linear the approximated deterministic component becomes. TABLES 1 AND 2 ABOUT HERE The first thing we observe in Table 1 is that the time trend coefficient is not required in any single case, while the intercept is significant in the three cases. The estimates of the differencing parameter are 1.05, 0.71 and 0.86 respectively for white noise, AR and Bloomfield disturbances, and the unit root null hypothesis cannot be rejected in any single case, implying permanency of shocks according to this simple model. Table 2 displays the estimated coefficients of the selected model for each type of disturbances. Performing the methods for testing the presence of structural breaks we find a single break at 1963. See Table 3. Tables 4 and 5 re-estimates the model given by Equation (12) for each subsample, using both uncorrelated (white noise) and Bloomfield disturbances. 3. Data and empirical results We use annual data corresponding to the all road used casualties (killed) in Great Britain from 1926 until 2021. They have been obtained from the UK Department for Transport (https://en.wikipedia.org/wiki/Department_for_Transport). The time series plot of the data is displayed in Figure 1. It can be observed that the highest death rate corresponds to 1941 during World War II, while 1966 is the year with the highest value 6 at peacetime. Since then, the reported deaths have been generally decreasing and the lowest number corresponds to 2020 (1460). Several national casualty reduction targets were proposed at years 1987, 1999 and 2010, being successful in all three cases. The first model examined is the one given by Equations (9) and (3), i.e., , ... , 2 ,1 , ) 1( ;) ( ) ( = = − + + = t u x L t x t t y t t d   (12) (12) and the results are displayed across Tables 1 and 2. In Table 1 we report the estimates of d under the three classical scenarios in the unit root literature of i) no deterministic terms, i.e., imposing that both unknown coefficients, α and β are equal to zero a priori in (12); ii) with an intercept (β = 0) and iii) with an intercept and a linear time trend. The values in bold in the table refer to the selected specification for each type of residuals, which are 1) white noise, 2) an AR(1) process, and 3) the exponential spectral model of Bloomfield (1973). TABLES 3 – 6 ABOUT HERE 7 We observe that if u(t) is a white noise process, the estimates of d are 1.01 for the first subsample, and 0.85 for the second one, and the unit root null cannot be rejected in either of the two subsamples. However, the time trend coefficients is now significantly negative in the second subsample. Allowing for autocorrelation throughout the model of Bloomfield (1973) the estimates of d are now much lower, 0.51 for the first subsample and 0.11 for the second one, and the confidence intervals are wider. In fact, for the first subsample, we cannot reject either the I(0) or the I(1) hypothesis. In the second one, however, the I(1) hypothesis is rejected in favour of mean reversion and thus transitory shocks. The time trend coefficient is also negative in this case. Thus, according to this specification, if there is an exogenous shock affecting the death rates, its impact will not be permanent, since the series will return by itself to its trend. As a final approach, the non-linear deterministic trend model in (10) is employed along with (3), and the results are displayed across Table 6. We see that if u(t) is white noise, all the Chebyshev coefficients are found to be statistically insignificant. However, allowing for autocorrelation, a non-linear structure is observed, with an estimated value of d equal to 0.60, and where the unit root null cannot be rejected. 4. Conclusions We have examined in this paper the total number of all road user casualties (killed) in Great Britain for the time period from 1926 to 2021. Using a variety of linear and non- linear models, all based on fractionally integration, our results indicate that if the sample is considered as a whole, the data are very persistent, supporting the existence of unit roots and implying permanency of shocks. However, if segmented trends are permitted, a structural break is found at 1963 and a different pattern is observed before and after the break. (see Figure 1 with the estimated trend for the model with autocorrelated errors). 8 Thus, for the sample ending at 1963 no time trend is observed and the estimated order of integration is 0. 51, though the unit root null cannot be rejected due to the large confidence interval for the value of d; for the second subsample, however, the estimated value of d is 0.11 and the hypothesis of d is decisively rejected in favour of d < 1 implying transitory shocks. Moreover, a significant time trend is also observed for this subperiod. A clear implication of this result is that mean reversion seems to take place in the data under examination and thus, there is no need for strong actions in the event of exogenous shocks since the series will return to its original long term projection. Alternative methods based for instance on time varying differencing parameters or alternative non-linear structures such as those based on Fourier functions of time (Gil- Alana and Yaya, 2021) or neural networks (Yaya et al., 2021) are being investigated at present. References Abbritti, M., L.A. Gil-Alana, Y. Lovcha and A. Moreno (2016), Term structure persistence, Journal of Financial Econometrics 14, 2, 331-352. Bai, J. and P. Perron (2003), Computation and analysis of multiple structural change models, Journal of Applied Econometrics 18, 1, 1-22. Bhargava, A. (1986), On the Theory of Testing for Unit Roots in Observed Time Series, Review of Economic Studies 53, 3, 369-384. Bierens, H. J. (1997). Testing the unit root with drift hypothesis against nonlinear trend stationarity with an application to the US price level and interest rate. Journal of Econometrics 81, 29-64. Bloomfield, P. (1973). An exponential model in the spectrum of a scalar time series, Biometrika 60, 217-226. Box-Steffensmeier, J.M. and A.R. Tomlinson, (2000), Fractional integration methods in political science, Electoral Studies 19, 63-76. Candelon, B. and L.A. Gil-Alana (2004), Fractional integration and business cycle features, Empirical Economics 29, 343-359. 9 Dickey, D. A and Fuller, W. A. (1979). Distributions of the estimators for autoregressive time series with a unit root, Journal of American Statistical Association, 74 (366), 427- 481. Diebold, F.X. and A. Inoue (2001), Long memory and regime switching, Journal of Econometrics 105, 1,131-159. Diebold, F. X. and Rudebush, G. D. (1991). On the power of Dickey‐ Fuller tests against fractional alternatives. Economics Letters, 35, 155-160. Elliot, G., T.J. Rothenberg and J.H. Stock (1996). Efficient tests for an autoregressive unit root, Econometrica 64, 813-836. Fuller, W. A. (1976). Introduction to Statistical Time Series, New York: JohnWiley. Fuller Introduction to Statistical Time Series 1976. Gil-Alana, L.A. (2008) Fractional integration and structural breaks at unknown periods of time, Journal of Time Series Analysis 29, 1, 163-185. Gil-Alana, L.A, & Cuestas, J. (2016). A nonlinear approach with long range sependence based on Chebyshev polynomials. Studies in Nonlinear Dynamics and Econometrics 16, 5, 445-468. Gil-Alana, L.A. and A. Moreno (2007), Uncovering the US term premium, Journal of Banking and Finance 36, 4, Gil-Alana, L.A. and P.M. Robinson (1997), Testing of unit roots and other nonstationary hypotheses, Journal of Econometrics 80, 2, 241-268. Gil-Alana, L.A. and O. Yaya (2021), Testing fractional unit roots with non-linear smooth break approximations using Fourier functions, Journal of Applied Statistics 48, 13-15, 2542-2559. Granger, C.W.J. (1980), Long memory relationships and the aggregation of dynamic models, Journal of Econometrics 14, 2, 227-238. Granger, C.W.J. and N. ---: the coefficients were found to be statistically insignificant. ---: the coefficients were found to be statistically insignificant. References Hyung (2004), Occasional structural breaks and long memory with an application to the S&P500 absolute stock returns, Journal of Empirical Finance 11, 3, 399-421. Hamming, R. W. (1973). Numerical Methods for Scientists and Engineers. Dover. Hassler, U., and Wolters, J. (1994). On the power of unit root tests against fractional alternatives. Economics Letters, 45(1), 1-5. Kwiatkowski, D., Phillips, P. C., Schmidt, P. and Shin, Y. (1992). Testing the null hypothesis of stationarity against the alternative of a unit root. Journal of Econometrics, 54(1-3), 159-178. 10 Lee, D., and Schmidt, P. (1996). On the power of the KPSS test of stationarity against fractionally-integrated alternatives. Journal of Econometrics, 73(1), 285-302. Michelacci, C. and P. Zaffaroni (2000) (Fractional) beta convergence, Journal of Monetary Economics 45, 1, 129-153. Nelson, C.R. and C.I. Plosser (1982) Trends and random walks in macroeconomic time series: Some evidence and implications, Journal of Monetary Economics 10:2, 139-162. Phillips, P.C.B. and P. Perron, (1988). Testing for a unit root in time series regression, Biometrika 75, 335-346. Schmidt, P. and P.C.B. Phillips (1992), LM tests for a unit root in the presence of deterministic trends, Oxford Bulletin of Economics and Statistics 54, 257-287. Smyth, G.K. (1998). Polynomial Approximation. John Wiley & Sons, Ltd, Chichester. Tomasevic, N. M. and Stanivuk, T. (2009). Regression Analysis and approximation by means of Chebyshev Polynomial. Informatologia 42, 3, 166-172. Yaya, O., Ogbonna, A.E., Gil-Alana, L.A. and Furuoka, F. (2021). A new unit root analysis for testing hysteresis in unemployment, Oxford Bulletin of Economics and Statistics 83(4), 960-981. 11 Table 1: Estimates of the differencing parameter d Type of errors No terms With an intercept With an intercept and a linear time trend White noise 1.09 (0.96, 1.27) 1.05 (0.91, 1.27) 1.05 (0.90, 1.27) AR(1) 1.09 (0.89, 1.59) 0.71 (0.40, 1.14) 0.72 (0.38, 1.14) Bloomfield (1973) 1.00 (0.80, 1.31) 0.86 (0.68, 1.14) 0.84 (0.61, 1.15) Values in bold refer to the selected specification. In parenthesis the 95% confidence band of the values of d. Table 1: Estimates of the differencing parameter d Table 2: Estimated coefficients of the selected specifications in Table 1 Type of errors d (95% conf. References band) Intercept (t-value) Time trend (t-value) White noise 1.05 (0.91, 1.27) 4.83991 (10.373) --- AR(1) 0.71 (0.40, 1.14) 5.25604 (11.483) --- Bloomfield (1973) 0.86 (0.68, 1.14) 5.05277 (11.112) --- th ffi i t f d t b t ti ti ll i i ifi t Table 2: Estimated coefficients of the selected specifications in Table 1 12 Table 3: Structural breaks Time series Break date All road user casualties in Great Britain since 1926 1963 Table 3: Structural breaks Table 4: Estimates of d No terms With an intercept With an intercept and a linear time trend White noise 1926 – 1963 1.06 (0.87, 1.35) 1.01 (0.75, 1.37) 1.01 (0.77, 1.36) 1964 - 2022 0.98 (0.82, 1.23) 0.91 (0.77, 1.27) 0.85 (0.56, 1.27) Bloomfield 1926 – 1963 0.95 (0.56, 1.43) 0.51 (-0.04, 1.25) 0.64 (-0.71, 1.27) 1964 - 2022 0.84 (0.58, 1.20) 0.69 (0.57, 0.88) 0.11 (-0.50, 0.76) Values in bold refer to the selected specification. In parenthesis the 95% confidence band of the values of d. Values in bold refer to the selected specification. In parenthesis the 95% confidence band of the values of d. Values in bold refer to the selected specification. In parenthesis the 95% confidence band of the values of d. Table 5: Estimated coefficients No terms With an intercept With an intercept and a linear time trend White noise 1926 – 1963 1.01 (0.75, 1.37) 4.87543 (7.544) ---- 1964 - 2022 0.85 (0.56, 1.27) 7.96145 (31.866) -0.1130 (-5.823) Bloomfield 1926 – 1963 0.51 (-0.04, 1.25) 5.87920 (13.098) ---- 1964 - 2022 0.11 (-0.50, 0.76) 8.10967 (26.837) -0.12199 (-14.100) ---: the coefficients were found to be statistically insignificant. Table 6: Non-linear deterministic terms Type of errors d (95% bad) White noise 0.99 (0.82, 1.24) 3.6673 (1.51) 1.4747 (1.02) -0.9155 (-1.26) 0.3026 (0.62) Bloomfield 0.60 (0.19, 1.04) 4.8171 (8.39) 1.3174 (4.00) -0.9884 (-4.01) 0.2365 (1.20) Table 5: Estimated coefficients No terms With an intercept With an intercept and a linear time trend White noise 1926 – 1963 1.01 (0.75, 1.37) 4.87543 (7.544) ---- 1964 - 2022 0.85 (0.56, 1.27) 7.96145 (31.866) -0.1130 (-5.823) Bloomfield 1926 – 1963 0.51 (-0.04, 1.25) 5.87920 (13.098) ---- 1964 - 2022 0.11 (-0.50, 0.76) 8.10967 (26.837) -0.12199 (-14.100) ---: the coefficients were found to be statistically insignificant. References Table 6: Non-linear deterministic terms Type of errors d (95% bad) White noise 0.99 (0.82, 1.24) 3.6673 (1.51) 1.4747 (1.02) -0.9155 (-1.26) 0.3026 (0.62) Bloomfield 0.60 (0.19, 1.04) 4.8171 (8.39) 1.3174 (4.00) -0.9884 (-4.01) 0.2365 (1.20) 13 Figure 1: Time series data and estimated segmented trend Figure 1: Time series data and estimated segmented trend 0 1 2 3 4 5 6 7 8 9 10 1926 2021 1964 Figure 1: Time series data and estimated segmented trend 0 1 2 3 4 5 6 7 8 9 10 1926 2021 1964 14
https://openalex.org/W2124999939	https://figshare.utas.edu.au/articles/journal_contribution/Peer_Review_The_Importance_of_Education_for_Best_Practice/22915856/1/files/40650293.pdf	English	null	Peer Review: The Importance Of Education For Best Practice	Journal of college teaching & learning/Journal of college teaching and learning	2,010	cc-by	270	Peer Review: The Importance Of Education For Best Practice Goldberg, Lynette R;Parham, Douglas F;Coufal, Kathy L;Maeda, Masako;Scudder, Rosalind R;Sechtem, ... Journal of College Teaching and Learning; Feb 2010; 7, 2; ProQuest pg. 71 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ed with permission of the copyright owner. Further reproduction prohibited without permission Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. duced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. h permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited with h permission of the copyright owner. Further reproduction prohibited without permission Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
https://openalex.org/W4224244167	https://publikationen.bibliothek.kit.edu/1000147680/148863188	English	null	What Is Worth Knowing in Interventional Practices about Medical Staff Radiation Exposure Monitoring: A Review of Recent Outcomes of EURADOS Working Group 12	Environments	2,022	cc-by	9,937	10 Faculty of Medicine, University of Osijek, 31000 Osijek, Croatia; dariofaj@mefos.hr 11 Radiation, Chemical and Environmental Hazards (RCE), UKHSA, Oxford OX11 0RQ, UK; jan.jansen@phe.gov.uk j j p g 12 Faculty of Science, University of Kragujevac, 34000 Kragujevac, Serbia; z.jovanovic@kg.ac.rs (Z.J.); Ž 12 Faculty of Science, University of Kragujevac, 34000 Kragujevac, Serbia; z.jovanovic@kg.ac.rs (Z.J dragana@kg.ac.rs (D.K.); milena.zivkovic@pmf.kg.ac.rs (M.Ž.) g g p g 13 RBI, Ru ¯der Boškovi´c Institute, 10000 Zagreb, Croatia; zeljka.knezevic@irb.hr RBI, Ruder Boškovic Institute, 10000 Zagreb, Croatia; zeljka.knezevic@irb.hr 14 CHUV, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland; marta.san 14 CHUV, Centre Hospitalier Universitaire Vaudois, 1011 Lausanne, Switzerland; marta.sansmerce@hcuge.ch 15 Departamento de Física e Astronomia, Faculdade de Ciências da Universidade do Porto, 4169-007 Porto, Portugal; ppteles@fc.up.pt 15 Departamento de Física e Astronomia, Faculdade de Ciências da Universidade do Porto, 4169 007 P t P t l t l @f t * Correspondence: paolo.ferrari@enea.it Abstract: EURADOS (European Radiation Dosimetry Group) Working Group 12 (WG12) SG1 activi- ties are aimed at occupational radiation protection and individual monitoring in X-ray and nuclear medicine practices. In recent years, many studies have been carried out in these ﬁelds, especially for interventional radiology and cardiology workplaces (IC/IR). The complexity of the exposure conditions of the medical staff during interventional practices makes the radiation protection and monitoring of the exposed workers a challenging task. The scope of the present work is to review some of the main results obtained within WG12 activities about scattered ﬁeld characterization and personal dosimetry that could be very useful in increasing the quality of radiation protection of the personnel, safety, and awareness of radiation risk. Two papers on Monte Carlo modelling of inter- ventional theater and three papers on active personal dosimeters (APDs) for personnel monitoring were considered in the review. More speciﬁcally, Monte Carlo simulation was used as the main tool to characterize the levels of exposure of the medical staff, allowing to determine how beam energy and direction can have an impact on the doses received by the operators. Indeed, the simulations provided information about the exposure of the operator’s head, and the study concluded with the determination of an eye-lens protection factor when protection goggles and a ceiling shielding are used. Moreover, the review included the results of studies on active personal dosimeters, their use in IC/IR workplaces, and how they respond to calibration ﬁelds, with X-ray standard and pulsed beams. environments environments environments environments What Is Worth Knowing in Interventional Practices about Medical Staff Radiation Exposure Monitoring: A Review of Recent Outcomes of EURADOS Working Group 12 Paolo Ferrari 1,* , Mercè Ginjaume 2 , Oliver Hupe 3, Una O’Connor 4, Filip Vanhavere 5, Elena Bakhanova 6, Frank Becker 7, Lorenzo Campani 1, Eleftheria Carinou 8, Isabelle Clairand 9, Dario Faj 10, Jan Jansen 11, Zoran Jovanovi´c 12, Željka Kneževi´c 13 , Dragana Krsti´c 12, Francesca Mariotti 1, Marta Sans-Merce 14, Pedro Teles 15 and Milena Živkovi´c 12 1 ENEA—Radiation Protection Institute, 40129 Bologna, Italy; lorenzo.campani@enea.it (L.C.); mariotti.francesca@enea.it (F.M.) 1 ENEA—Radiation Protection Institute, 40129 Bologna, Italy; lorenzo.campani@enea.it (L.C.); mariotti.francesca@enea.it (F.M.) 2 Institut de Tècniques Energètiques, Universitat Politècnica de Catalunya, 08034 Barcelona, Spain; merce.ginjaume@upc.edu 4 St. James’s Hospital, D08 NHY1 Dublin, Ireland; uoconnor@stjames.ie 5 SCK-CEN, Belgian Nuclear Research Centre, 2400 Mol, Belgium; ﬁlip.vanhavere@sckcen.be 6 National Research Center for Radiation Medicine, 04050 Kyiv, Ukraine; elena.bakhanova@gmail.com 7 KIT, Institute for Nuclear Waste Disposal, 76344 Eggenstein-Leopoldshafen, Germany; frank.becker@kit.edu 8 EEAE, Greek Atomic Energy Commission, 15341 Athens, Greece; eleftheria.carinou@eeae.gr 9 IRSN, Institute Radiological Protection and Nuclear Safety, 92260 Fontenay-Aux-Roses, France; isabelle.clairand@irsn.fr 5 SCK-CEN, Belgian Nuclear Research Centre, 2400 Mol, Belgium; ﬁlip.vanhavere@sckcen.be 6 National Research Center for Radiation Medicine, 04050 Kyiv, Ukraine; elena.bakhanova@gmail.com 7 KIT, Institute for Nuclear Waste Disposal, 76344 Eggenstein-Leopoldshafen, Germany; frank.becker@kit.edu Citation: Ferrari, P.; Ginjaume, M.; Hupe, O.; O’Connor, U.; Vanhavere, F.; Bakhanova, E.; Becker, F.; Campani, L.; Carinou, E.; Clairand, I.; et al. What Is Worth Knowing in Interventional Practices about Medical Staff Radiation Exposure Monitoring: A Review of Recent Outcomes of EURADOS Working Group 12. Environments 2022, 9, 53. https://doi.org/10.3390/ environments9040053 Academic Editors: Gaetano Licitra, Mauro Magnoni and Gian Marco Contessa Received: 16 March 2022 Accepted: 12 April 2022 Published: 18 April 2022 1. Introduction The complexity of the exposure conditions of the medical staff during interventional practices makes radiation protection and monitoring of the involved workers a challenging task [1–3]. The scattered radiation field produced by the patient from the primary X-ray beam and propagating towards the operators can be mitigated by wearable protective means (aprons and collars) and a properly positioned shielding [4]. Notwithstanding that, because of the increasing numbers of procedures performed in a year, due to their diagnostic and therapeutic effectiveness [5], the annual doses in personnel performing this kind of practice are among the highest registered for workers in the medical field [1,2,6,7]. For instance, increasing knowledge about exposure conditions and optimizing radiation protection are fundamental to fulfill the implementation of the ALARA (as low as reasonably achievable) principle for those workers. If it is possible to assess dose in the hand and ﬁnger employing ring or bracelet dosimeters [8,9], effective dose estimation, evaluated through a personal equivalent dose, Hp(10) measurements, is even more complex. A protective apron and collar only partially protect the operator’s body from the scattered radiation. Indeed, the operator’s exposure is dependent on the X-ray tube projection, which is changed during practice, according to diagnostic needs. For these reasons, a single dosimeter put under the apron or on the apron (in various positions, at the waist or breast or at clavicle height) cannot mimic the exact exposure situation of the operator. Different correction factors are available in the literature, to be applied to a single dosimeter reading, or algorithms that can be employed when two dosimeters (one on and one under the apron) are worn simultaneously [10,11]. The re-evaluation of the tissue reaction and the radiosensitivity of the eye lens [12,13] and the subsequent annual dose limit reduction introduced in the EU 2013/59/Euratom directive [14] involved an update on the methodology of the eye lens dose estimation in interventional practices [15,16]. Lastly, the introduction of electronic dosimeters and online devices for operator dose monitoring [17–20], on the one hand, opens new perspectives for individual dosimetry, and on the other hand, it requires better knowledge of their behavior in interventional ﬁelds and a thorough check of their reliability. It was shown that APDs are insensitive to backscatter radiation, but some of them could not respond correctly to the very intense pulsed ﬁelds (as those next to the patient in interventional practices). The measurements during interventional procedures showed the potential capability of the employment of APDs in hospitals. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). https://www.mdpi.com/journal/environments Environments 2022, 9, 53. https://doi.org/10.3390/environments9040053 Environments 2022, 9, 53 2 of 13 Keywords: interventional radiology; interventional cardiology; medical staff radiation dose moni- toring; Monte Carlo simulations; dosimetry; health physics; electronic personal dosimeters; medical staff ionizing radiation exposure 1. Introduction g y For all these reasons, a great part of the EURADOS (European Radiation Dosimetry Group) Working Group 12 Dosimetry in medical imaging (WG12) activities has been devoted to performing a series of studies on the exposure of medical staff performing interventional practices. The scope of the present work is presenting a review of the main results obtained in WG12 and published in the literature aimed at characterizing the irradiation ﬁeld and the exposure of the operator’s head and studying the response of an electronic dosimeter in interventional practices. Some new outcomes, not yet published, are reported to complete the discussion of the results. 2. Simulation of the Scattered Radiation Field and the Operator Doses 2.1. Interventional Theater Modelling Monte Carlo (MC) simulations, mainly obtained by the MCNP (Monte Carlo N- Particle) code family [21,22], supplied the primary data for the scattered ﬁeld study. A modiﬁed version of a MIRD (Medical Internal Radiation Dose Committee) type model [23] was used to mimic the patient and two operators. The two operators were covered with a 0.5 mm Pb layer, properly shaped, representing the protective apron and collar, and their arms were bent to reproduce a “realistic” position. The kerma area product transmission chamber (KAP), normally employed to monitor the X-ray emission during the clinical exam, Environments 2022, 9, 53 3 of 13 ct g 3 of 13 ct g was reproduced as a simple air parallelepiped put at a given distance from the source. The KAP calculated values were used to normalize the dose-calculated values. The dosimeters, put in selected positions on the operators, were simulated as simple air-ﬁlled spheres, with a 1 cm radius. The corresponding dose equivalent was calculated by folding the photon ﬂuence reaching the sphere with the ICRU (International Commission on Radiation Units and Measurements) conversion coefﬁcients [24]. ues. The dosimeters, put in selected positions on the operators, were simulated as ple air-filled spheres, with a 1 cm radius. The corresponding dose equivalent was ulated by folding the photon fluence reaching the sphere with the ICRU (International mmission on Radiation Units and Measurements) conversion coefficients [24]. Initially, the geometry included only a single operator and the undercouch shielding; n a second operator was added on the right of the first operator (Figure 1). The two Initially, the geometry included only a single operator and the undercouch shielding; then a second operator was added on the right of the ﬁrst operator (Figure 1). The two operators’ positions were changed, as their “direction” (rotating the models on their vertical axis toward the patient head or feet), and a 0.5 or 1 mm rectangular lead ceiling shielding was opportunely positioned between the operators and scattering source when needed. a seco d ope ato was added o t e ig t o t e i st ope ato ( igu e ). e two rators’ positions were changed, as their “direction” (rotating the models on their tical axis toward the patient head or feet), and a 0.5 or 1 mm rectangular lead ceiling lding was opportunely positioned between the operators and scattering source when ded. ure 1. 2. Simulation of the Scattered Radiation Field and the Operator Doses 2.1. Interventional Theater Modelling Monte Carlo simulated scenarios with two operators and the patient (PA X-ray beam ection). Interventional cardiology (IR) on the left (first operator next to the image device), rventional cardiology (IC) on the right (first operator near the right patient’s leg) Figure 1. Monte Carlo simulated scenarios with two operators and the patient (PA X-ray beam projection). Interventional cardiology (IR) on the left (ﬁrst operator next to the image device), interventional cardiology (IC) on the right (ﬁrst operator near the right patient’s leg). re 1. Monte Carlo simulated scenarios with two operators and the patient (PA X-ray beam ction). Interventional cardiology (IR) on the left (first operator next to the image device), ventional cardiology (IC) on the right (first operator near the right patient’s leg) Figure 1. Monte Carlo simulated scenarios with two operators and the patient (PA X-ray beam projection). Interventional cardiology (IR) on the left (ﬁrst operator next to the image device), interventional cardiology (IC) on the right (ﬁrst operator near the right patient’s leg). The X-ray tube beam qualities were simulated according to the IPEM (Institute of sics and Engineering in Medicine) spectral data [25]. Different beam qualities anging kV and copper and aluminum filtrations) were employed, and, besides teroanterior (PA), different beam projections were considered (left anterior oblique O), right anterior oblique (RAO), cranial (CRA), caudal (CAU)) at various impinging l (25° 30° 45° 60° 90°) The X-ray tube beam qualities were simulated according to the IPEM (Institute of Physics and Engineering in Medicine) spectral data [25]. Different beam qualities (changing kV and copper and aluminum ﬁltrations) were employed, and, besides posteroanterior (PA), different beam projections were considered (left anterior oblique (LAO), right anterior oblique (RAO), cranial (CRA), caudal (CAU)) at various impinging angles (25◦, 30◦, 45◦, 60◦, 90◦). les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure Scattering Field Evaluation and Operator’s Whole Body Exposure During the European ORAMED (Optimization of Radiation Protection of Medical f) [26] project, the effect of the LAO and CRA projections in increasing the exposure he operator with respect to the RAO projections was demonstrated [4]. These effects e confirmed in more recent works, and in particular, it was seen how the image device, nterventional radiology (IR) practice, can have a “protective” effect on the first During the European ORAMED (Optimization of Radiation Protection of Medical Staff) [26] project, the effect of the LAO and CRA projections in increasing the exposure of the operator with respect to the RAO projections was demonstrated [4]. These ef- fects were conﬁrmed in more recent works, and in particular, it was seen how the image device, in interventional radiology (IR) practice, can have a “protective” effect on the ﬁrst operator [27]. gy ( ) p , p rator [27]. The personal dose equivalent, Hp(10), was calculated through dosimeters placed at ee heights (waist, breast, and neck) on the apron, along the sagittal median plane of the hropomorphic model, for sake of simplicity. The study [27] allowed investigating the p The personal dose equivalent, Hp(10), was calculated through dosimeters placed at three heights (waist, breast, and neck) on the apron, along the sagittal median plane of the anthropomorphic model, for sake of simplicity. The study [27] allowed investigating the differences in the scattering ﬁeld reaching the two operators during IR practice. A detailed analysis of the scattering can be found in the cited paper here, as can be seen in Figure 2, for PA projection (the most common projection employed in those procedures). The highest dose is registered at the level of the waist, and it decreases moving toward the neck. Because of the distances with the patient and beam, the second operator’s doses were lower than the ﬁrst with the only exception of the dosimeter put on the neck. Here, the values were more similar due to the reduced amount of radiation coming directly from the patient and because the possible increase of the contribution derived from the Environments 2022, 9, 53 4 of 13 irectly the mu multiple scattering. Indeed, as said above, the image device can play a role in reducing the exposure of the ﬁrst operator upper parts. les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure An analogous exposure condition was recently investigated, even in the presence of a ceiling shielding, by the group of Nowak et al. [28], who published the spectra of the scattered photons in selected positions in an interventional theater. The complexity of the situation of the scatter reaching the operators in IR can be guessed also considering what happens to the LAO 25◦(left anterior oblique 25◦) projection (Figure 3). In that case, the image device does not offer an additional protection to the ﬁrst operator, and the doses registered at the neck (these data have not been published previously) depend more directly on the distances from the beam (Figure 4). In that case, the doses to the ﬁrst operator are higher. As can be seen, indeed, the doses for the LAO projections are higher than those evaluated for the PA projection due to the direction of the X-ray beam impinging on the patient, which increases the backscatter component directed towards the operator, as already demonstrated in a previous work [4]. ure of the first operator upper parts. An analogous exposure condition was recentl estigated, even in the presence of a ceiling shielding, by the group of Nowak et al. who published the spectra of the scattered photons in selected positions in an inter ional theater. The complexity of the situation of the scatter reaching the operators an be guessed also considering what happens to the LAO 25° (left anterior oblique rojection (Figure 3). In that case, the image device does not offer an additional prote o the first operator, and the doses registered at the neck (these data have not been shed previously) depend more directly on the distances from the beam (Figure 4). In ase, the doses to the first operator are higher. As can be seen, indeed, the doses fo LAO projections are higher than those evaluated for the PA projection due to the dire f the X-ray beam impinging on the patient, which increases the backscatter compo directed towards the operator, as already demonstrated in a previous work [4]. Figure 2. The response of a dosimeter put on the operator’s apron at three different heights in ventional radiology procedures (MC simulations for PA projection, data taken from [27]). Figure 2. The response of a dosimeter put on the operator’s apron at three different heights in interventional radiology procedures (MC simulations for PA projection, data taken from [27]). les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure The calculated “male effective dose” for the first operator ranges from 0.03 μSv/Gy·cm2 to 0.4 μSv/Gy·cm2, with the X-ray mean energy varying from 40 keV to 70 keV, and from 0.015 Because the anthropomorphic models were based on a MIRD male phantom [23], it was possible only to evaluate a sort of “male effective dose” (the effective dose per se would require the average of the doses evaluated in the female and male models). The calculated “male effective dose” for the ﬁrst operator ranges from 0.03 µSv/Gy·cm2 to 0.4 µSv/Gy·cm2, with the X-ray mean energy varying from 40 keV to 70 keV, and from 0.015 µSv/Gy·cm2 to 0.1 µSv/Gy·cm2 for the second operator [27]. Such estimates are in agreement with the data published by Kim et al. [30]. μSv/Gy·cm2 to 0.1 μSv/Gy·cm2 for the second operator [27]. Such estimates are in agree- ment with the data published by Kim et al. [30]. Some algorithms available in the literature employ Hp(10), evaluated by only one do- imeter put on the protective collar, to estimate the effective dose [11]. In particular, the NCRP report suggests dividing Hp(10) by 21 [31], and McEwan suggests multiplying it by Some algorithms available in the literature employ Hp(10), evaluated by only one dosimeter put on the protective collar, to estimate the effective dose [11]. In particular, the NCRP report suggests dividing Hp(10) by 21 [31], and McEwan suggests multiplying it by 0.08 [32] (i.e., dividing it by 12.5). These two factors can be compared with those that can be calculated from WG12 studies in IR. .08 [32] (i.e., dividing it by 12.5). These two factors can be compared with those that can e calculated from WG12 studies in IR. In Figure 5, the ratios between Hp(10), evaluated with the neck (unshielded) dosime- er, and the “male effective dose” versus beam energy are reported (these data have not een published previously). They vary from 3 to 25 for the first operator and from 15 to In Figure 5, the ratios between Hp(10), evaluated with the neck (unshielded) dosimeter, and the “male effective dose” versus beam energy are reported (these data have not been published previously). They vary from 3 to 25 for the ﬁrst operator and from 15 to 45 for the second operator, and their mean values are 10 and 27, respectively. les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure REVIEW Figure 2. The response of a dosimeter put on the operator’s apron at three different heights in ventional radiology procedures (MC simulations for PA projection, data taken from [27]). Figure 2. The response of a dosimeter put on the operator’s apron at three different heights in interventional radiology procedures (MC simulations for PA projection, data taken from [27]). Figure 3. Monte Carlo simulated irradiation in interventional radiology for the LAO Figure 3. Monte Carlo simulated irradiation in interventional radiology for the LAO 25◦projection. Figure 3. Monte Carlo simulated irradiation in interventional radiology for the LA Figure 3. Monte Carlo simulated irradiation in interventional radiology for the LAO 25◦projection. 5 of 13 Environments 2022, 9, 53 Figure 4. The response of a dosimeter put on the operator’s apron MC simulations for the LAO 25° Figure 4. The response of a dosimeter put on the operator’s apron. MC simulations for the LAO 25◦projection). igure 4 The response of a dosimeter put on the operator’s apron MC simulations for the LAO 25° Figure 4. The response of a dosimeter put on the operator’s apron. MC simulations for the LAO 25◦projection). projection). In the case of IR, the exposure of the first operator, registered by a dosimeter put on he breast, on an apron, is of the order of 10 μSv/Gy·cm2, a value in agreement with those eported by Leyton et al. [29] and Martin [2], and that can be reduced by about two orders f it d h dditi l hi ldi i d i l ti d t In the case of IR, the exposure of the ﬁrst operator, registered by a dosimeter put on the breast, on an apron, is of the order of 10 µSv/Gy·cm2, a value in agreement with those reported by Leyton et al. [29] and Martin [2], and that can be reduced by about two orders of magnitude when an additional shielding is used, as simulation and measurement demonstrated [27]. of magnitude when an additional shielding is used, as simulation and measurement demonstrated [27]. Because the anthropomorphic models were based on a MIRD male phantom [23], it was possible only to evaluate a sort of “male effective dose” (the effective dose per se would require the average of the doses evaluated in the female and male models). les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure The factors are lower for the ﬁrst operator because the doses evaluated in the collar are comparable between the two operators, but the effective dose for the ﬁrst operator is higher; thus the “correction factor” is lower. It is worth recalling that, due to the geometries and distances from the patient, the position of the second operator (near the patient’s right leg) can mimic the position of the ﬁrst operator during interventional cardiology (IC) procedures (through femoral access; see also Figure 1). Thus, the mean values obtained from the simu- lations can justify the discrepancy of the correction factors in the two papers cited above, considering that they could have been evaluated in different conditions (cardiology vs. radiology, beam energy, distances, etc.) and that they are plausible, taking into account the variability of the exposure due to the complex scattered ﬁeld reaching the operators during interventional practices. 6 of 13 tors Environments 2022, 9, 53 igure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male ffe ti e do e” e u X ay ea e e y (MC i ulatio fo PA oje tio ) Figure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). igure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male ff d l f A Figure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). gure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male ective dose” versus X ray mean energy (MC simulations for PA projection) Figure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). gure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male fecti e dose” ersus X ray mean energy (MC simulations for PA projection) Figure 5. Ratio between Hp(10) evaluated with a neck unshielded dosimeter and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). les (25 , 30 , 45 , 60 , 90 ). 2.2. Scattering Field Evaluation and Operator’s Whole Body Exposure y gy ( p j ) In Figure 6, the ratios between Hp(10), evaluated with a dosimeter put on the breast nshielded), and the “male effective dose” are reported, and they vary from 25 to 125 for e first operator and from 25 to 190 for the second operator. The large variability of those ctors for the effective dose evaluation with a single dosimeter is coherent (for X-ray b h h h 4 k V) h h f l d b J In Figure 6, the ratios between Hp(10), evaluated with a dosimeter put on the breast (unshielded), and the “male effective dose” are reported, and they vary from 25 to 125 for the ﬁrst operator and from 25 to 190 for the second operator. The large variability of those factors for the effective dose evaluation with a single dosimeter is coherent (for X-ray mean beam energies higher than 45 keV) with the range of values reported by Jarvinen et al. [33]. EVIEW 7 of 13 ean beam energies higher than 45 keV) with the range of values reported by Jarvinen et . [33]. mean beam energies higher than 45 keV) with the range of values reported by Jarvinen et l. [33]. Figure 6. Ratio between Hp(10) evaluated with unshielded dosimeter put on the apron at the breast level and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA pro- jection) Figure 6. Ratio between Hp(10) evaluated with unshielded dosimeter put on the apron at the breast level and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). Figure 6. Ratio between Hp(10) evaluated with unshielded dosimeter put on the apron at the breast evel and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA pro- ection) Figure 6. Ratio between Hp(10) evaluated with unshielded dosimeter put on the apron at the breast level and calculated “male effective dose” versus X-ray mean energy (MC simulations for PA projection). jection). 2 3 E f th O t ’ H d d E L 2.3. Exposure of the Operator’s Head and Eye Lens jection). 2 3 E f th O t ’ H d d E L 2.3. Exposure of the Operator’s Head and Eye Lens j ) 2 3 E f th O t ’ H d d E L 2.3. Exposure of the Operator’s Head and Eye Lens 2.3. Exposure of the Operator’s Head and Eye Lens In studying the scattering radiation impinging on an IR operator in WG12, a partic- ular attention was addressed to the head and eye lens exposure. In an early work [23], considering only one operator and femoral access position in IC procedures, the exposure of the eye lens was investigated, taking into account different X-ray qualities, projections, d t ti ( l d ili hi ldi ) Th i l ti d t t d h In studying the scattering radiation impinging on an IR operator in WG12, a partic- ular attention was addressed to the head and eye lens exposure. In an early work [23], considering only one operator and femoral access position in IC procedures, the exposure of the eye lens was investigated, taking into account different X-ray qualities, projections, and protective means (goggles and ceiling shielding). The simulation demonstrated how Environments 2022, 9, 53 7 of 13 7 of 13 the efﬁcacy of the protective goggles was very dependent on their shape and ability to attenuate the radiation coming from below, in agreement with experimental data produced by Domienik and Brodecki [34]. In order to determine the direction of incidence of the scattered radiation towards the eye lens, the geometry was simpliﬁed, modelling the eye as a simple sphere of the eye-bulb size, divided in spherical sectors by median axial, sagittal, and coronal planes (more details can be found in the cited paper [23]). It was shown that properly shaped goggles can reduce the absorbed dose up to nearly a factor of 10 (but this is highly dependent on their shape; see also [35]), and that is particularly valid for the left eye lens, which usually receives twice the amount of the radiation reaching the right eye lens. The latter was in fact reached by the radiation scattered inside the head of the operator, an effect suggested by Martin [36] and proven in the cited work [23]. Protection of the eye lens is achievable, in a less straightforward way, through the ceiling shielding, provided that it is correctly positioned with respect to the beam and the scattering source. j ) 2 3 E f th O t ’ H d d E L 2.3. Exposure of the Operator’s Head and Eye Lens In the same study, it was shown how in case of a LAO 90◦projection, with the X-ray tube on the left side of the ﬁrst operator, the protective efﬁciency of the shielding is reduced by a factor of the order of 10 when the shielding is not correctly placed (i.e., shifted in a position that does not protect the operator from the scattered radiation [23]). Similar considerations have been drawn for the other projections tested and for radial access and can be stressed to improve the compliance of operators to radioprotection strategies [37,38]. Analogous remarks on the effectiveness of the shielding can be made for the protection of the head (brain) of the operator involved in interventional practices. Recent studies have put the attention on the possible increase in the malignant late effect of radiation in IR and cardiology personnel due to the left part of the brain exposure [39–42]. In order to study that exposure, the brain of the numerical models was properly segmented to roughly reproduce human brain lobes (frontal, temporal, parietal, and occipital), a cerebellum, and a stem [43]. The absorbed doses in these brain sectors, normalized to the KAP values, were calculated for the two operators of IR, in the absence of a ceiling shielding (a condition not so infrequent in practice) for different X-ray beam qualities and projections (PA, RAO 25◦, LAO 25◦, CAR 25◦). Even in the approximations introduced by the rigidity of the models (considering the variability of the real scenario), in interventional radiology, because of the position with respect to the X-ray beam, the left and right brain parts of the ﬁrst operator are affected by nearly the same amount of radiation, with a maximum corresponding to the frontal lobes (as expected from a previous study on the incidence direction of the scattered radiation on the head [23]). On the other hand, for the second operator, placed on the right side of the ﬁrst operator, in a position that can resemble the position of the ﬁrst operator in IC, the left part of the brain receives twice the amount of the radiation received by the right part (in agreement with what happens with the eye lens in the same situation). The highest values were obtained for the LAO and PA projections, then for the CRA and RAO projections. j ) 2 3 E f th O t ’ H d d E L 2.3. Exposure of the Operator’s Head and Eye Lens p j p j Depending on the beam quality, projection, and operator position, in the absence of a ceiling shielding, the range of the absorbed doses to the brain varies between 0.1 Gy·cm2 and 0.7 µGy/Gy·cm2 for the ﬁrst operator in IR and between 0.1 Gy·cm2 and 1.1 µGy/Gy·cm2 for the second operator. It is worth recalling here that, because of the peculiar exposure condition in IR, the ﬁrst operator is nearer to the patient but partially shielded by the image device in the PA projection; thus the absorbed dose to the brain of the ﬁrst operator can be lower than the absorbed dose to the brain of the second operator. Considering a typical KAP value ranging from 20 Gy·cm2 to 40 Gy·cm2 [9], one gets about 2 µGy to 28 µGy for the ﬁrst and 4 µGy to 44 µGy for the second operator per procedure, respectively. These values are consistent with the (48 +/−28) µGy per procedure reported by Watson et al. [44] and with those recently evaluated using plastic phantoms and dosimeters by Hattori et al. [45]. Environments 2022, 9, 53 8 of 13 8 of 13 Taking into account 200 to 500 procedures in a year [5], the annual absorbed dose can vary between 0.4 mGy and 22 mGy. The latter value is in agreement with the exposure considered by Reeves in “The Brain Study” [41]. 3.1. APD Response in Continuous and Pulsed X-ray Field Commercially available APDs were tested in a calibration laboratory at KIT (Karlsruhe Institute of Technology), NIOM (Nofer Institute of Occupational Medicine), and PTB (Physikalisch-Technische Bundesanstalt) with 60Co, 137Cs sources, and selected RQR X-ray beams (in continuous and pulsed emission) [47]. All the measurements were performed on an ISO water slab phantom in ISO reference conditions [46]. The behavior of the dosimeters against an increasing dose rate was checked, showing that some of them tended to underestimate doses because they were not able to correct for the electronic pile-up at a certain dose rate [48]. The IEC (International Electrotechnical Commission) is preparing a technical report describing a general procedure for testing the suitability of a dosimeter for pulsed radiation ﬁelds, IEC/DTS 63050 [49]. Due to the pulsation of the radiation ﬁeld, the response will not change by more than ±20%. Additionally, the smallest acceptable value for the maximum measurable dose rate of a dosimeter will be at least 1 Sv/h. Both criteria were only fulﬁlled by 3 of 10 tested APD types (see the cited paper [48] for more details). 3. Testing Active Personal Dosimeters in Interventional Practices A reliable evaluation of the IR and IC medical staff exposure is of primary concern, and this can be pursued even taking advantage of new technologies, such as electronic personal devices and online measuring systems [18–20] possibly coupled with fast Monte Carlo codes [17]. For that reason, WG12 performed a series of studies addressed to analyze the behavior of the APDs (active personal dosimeters) in interventional practices. The study was divided into three lines of investigation: (a) the response of some commercially available APDs was tested in primary standard dosimetry laboratories, em- ploying continuous and pulsed X-ray ﬁeld (like those used in interventional practices); (b) the responses of both APDs and passive-type dosimeters (thermoluminescence (TLD), radio-photoluminescence (RPL), optically stimulated luminescence (OSL), and ﬁlm dosime- ter) were studied in secondary standard dosimetry laboratories, employing the standard ISO (International Organization for Standardization) water slab phantom [46] covered by a “typical” interventional radiology operator’s protective apron, determining the effect of such protective means on the dosimeter response; (c) a measuring campaign in an in- terventional theater was performed, with operators wearing a device that housed both an RPL dosimeter and an APD, to guarantee that they were exposed simultaneously in the analogous conditions. 3 3 Comparison of the Responses of Active and Passive Dosimeters in the Hospital 3.3. Comparison of the Responses of Active and Passive Dosimeters in the Hospital 3.3. Comparison of the Responses of Active and Passive Dosimeters in the Hospital With the aim of testing simultaneously passive (RPL) and active dosimeters in the hospital, a simple device housing the two dosimeters (Figure 8) and to be worn on an apron was distributed among physicians performing interventional procedures. A meas- urement campaign was performed comparing the doses received by the two types of do- simeters. Because direct reading was possible in the active dosimeter, after a certain level of cumulated dose, that was registered, the passive dosimeter was removed and substi- tuted, and the active dosimeter reset for a new measurement. The comparison of the read- ings of the passive and active dosimeters showed a certain spread in the results, particu- larly at lower doses, and, in general, an underestimation of the doses in the case of the active ones. In particular, it was observed that the three tested APD, EPD Mk2.3, DMC- With the aim of testing simultaneously passive (RPL) and active dosimeters in the hospital, a simple device housing the two dosimeters (Figure 8) and to be worn on an apron was distributed among physicians performing interventional procedures. A measurement campaign was performed comparing the doses received by the two types of dosimeters. Because direct reading was possible in the active dosimeter, after a certain level of cu- mulated dose, that was registered, the passive dosimeter was removed and substituted, and the active dosimeter reset for a new measurement. The comparison of the readings of the passive and active dosimeters showed a certain spread in the results, particularly at lower doses, and, in general, an underestimation of the doses in the case of the active ones. In particular, it was observed that the three tested APD, EPD Mk2.3, DMC-3000, and RaySafe-i2 had, on average, a 30%, 10%, and 10%, respectively, lower response than the reference RPL passive dosimeter. Due to the fact they were worn in a random variety of standard interventional procedures, with a random range of energies, angles, and pulse ﬁeld characteristics, the standard deviations ranged from 40% to 50%, but the tendency is clear for the three devices (more details can be found in the published paper [51]). EVIEW 10 of 13 3000, and RaySafe-i2 had, on average, a 30%, 10%, and 10%, respectively, lower response than the reference RPL passive dosimeter. 3.2. Effect of the Protective Apron on a Dosimeter’s Response 3 3 Comparison of the Responses of Active and Passive Dosimeters in the Hospital 3.3. Comparison of the Responses of Active and Passive Dosimeters in the Hospital 3.2. Effect of the Protective Apron on a Dosimeter’s Response The response of passive (TLD, OSL, RPL, and ﬁlms) and active dosimeters was tested in secondary standard dosimetry laboratories with 137Cs and selected RQR X-ray beam qualities (continuous emission). The dosimeter was placed on an ISO water slab phantom, and the measurement was repeated, putting an apron or an equivalent thickness of lead on the phantom between the slab and the dosimeter, mimicking the wearing condition of the dosimeter put on the apron in interventional procedures. The measurements were coupled with Monte Carlo simulations. In the presence of a protective apron on the slab, the study showed a reduction of the order of 30% for the dose reading in the case of passive dosimeters. The dose readings were less affected by the apron for electronic dosimeters due to the reduced sensitivity to backscatter of the latter device [50]. This behavior is clearly visible in Figure 7, where the ratios of the responses of the tested dosimeters placed on the calibration phantom, with and without the lead layer, were reported. 9 of 13 vior is placed Environments 2022, 9, 53 Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qual- ities (data taken from [50]). Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qualities (data taken from [50]). Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qual- ities (data taken from [50]). Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qualities (data taken from [50]). Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qual- ities (data taken from [50]). Figure 7. Ratios of Hp(10) with lead apron and Hp(10) without lead apron of different types of active and passive dosimeters placed on ISO calibration phantoms for different beam and radiation qualities (data taken from [50]). 4. Conclusions EURADOS WG12 activities are driven to radiation protection in the medical ﬁeld. The works reviewed in the present paper aimed at a better knowledge of personnel radiation exposure in interventional practices. These studies were performed through MC simula- tions and measurements, in calibration laboratories and hospitals, and brought results that can be used to increase the quality of radiation protection of personnel and the awareness of radiation risk. Two studies of the scattering around the operator’s head [23,43] provided information about the exposure of the eye lens and the operator’s brain, offering strategies to optimize radiation protection. They stated that the exposure of the eye lens can be reduced by a factor of 2 up to nearly a factor of 10 employing proper goggles (i.e., goggles adherent to the operator face, capable of absorbing the radiation coming from the bottom (see also [34],) but goggles with an unsuited shape offer a lower protection level [35,36]). A reduction of the order of 10 can be obtained also from a ceiling shielding (see also [53]), but it is important to underline that a non-optimal placement of the shielding, with respect the main source of the scatter, can heavily reduce its protection capability. Thus to avoid exposure of the operator’s eyes and head, a correct shielding positioning should be stressed [54], particularly when goggles are not used, to improve radiation safety and awareness [37]. The same ceiling shielding can offer a protection also for the operator’s brain that during an interventional procedure can receive exposure ranging from 0.1 µGy/Gy·cm2 and 1 µGy/Gy·cm2 in the absence of a proper shielding, implying absorbed doses that are coherent with those evaluated for long-term effects in medical staff cohorts [40]. A study on operator exposure in interventional radiology [27] allowed a better knowl- edge of the scattering ﬁeld in relation to the Hp(10) evaluated with dosimeters put on an apron. During interventional procedures, a dosimeter put on an apron can receive a dose, expressed in terms of Hp(10), of the order of 10 µSv/Gy·cm2 per procedure, but the same studies also showed the sensitivity of such kind of evaluation due to the dosimeter’s position coupled with the complexity of the scattering scenario. Indeed, three studies on APDs [48,50,51] considered the opportunity of using these devices. 3 3 Comparison of the Responses of Active and Passive Dosimeters in the Hospital 3.3. Comparison of the Responses of Active and Passive Dosimeters in the Hospital Due to the fact they were worn in a random variety of standard interventional procedures, with a random range of energies, angles, and pulse field characteristics, the standard deviations ranged from 40% to 50%, but the tendency is clear for the three devices (more details can be found in the published paper [51]). Figure 8. The simple device employed the measurements in the hospital during the interventional procedures. The APD was attached to the upper section of the device, while the lower section hosted the RPL. Figure 8. The simple device employed the measurements in the hospital during the interventional procedures. The APD was attached to the upper section of the device, while the lower section hosted the RPL. Figure 8. The simple device employed the measurements in the hospital during the interventional procedures. The APD was attached to the upper section of the device, while the lower section hosted the RPL. Figure 8. The simple device employed the measurements in the hospital during the interventional procedures. The APD was attached to the upper section of the device, while the lower section hosted the RPL. 10 of 13 Environments 2022, 9, 53 10 of 13 The causes of that behavior can be associated with the pulsed ﬁeld and the slightly different heights of the two dosimeters that can play a role in such discrepancy. A series of recommendations for APD employment in the medical ﬁeld have been published [52]. Author Contributions: Conceptualization and methodology M.G., E.C., O.H., I.C., U.O., M.S.-M., Ž.K., F.V. and P.F., simulations P.F., E.B., F.B., J.J., Z.J., D.K., P.T. and M.Ž., data acquisition and measurements O.H., F.V., L.C., E.C., I.C., D.F., P.F., F.M., M.G., U.O. and Ž.K.; writing—review and editing, P.F., M.G., E.C., E.B., D.F., I.C., J.J., P.T., Ž.K. and F.V. All authors have read and agreed to the published version of the manuscript. References 1. Covens, P.; Berus, D.; Buls, N.; Clerinx, P.; Vanhavere, F. Personal dose monitoring in hospitals: Global assessment, critical applications and future needs. Radiat. Prot. Dosim. 2007, 124, 250–259. [CrossRef] [PubMed] 2. Martin, C.J. A review of radiology staff doses and dose monitoring requirements. Radiat. Prot. Dosim. 2009, 136, 140–157. [CrossRef] [PubMed] 1. Covens, P.; Berus, D.; Buls, N.; Clerinx, P.; Vanhavere, F. Personal dose monitoring in hospitals: Global assessment, critical applications and future needs. Radiat. Prot. Dosim. 2007, 124, 250–259. [CrossRef] [PubMed] pp 2. Martin, C.J. A review of radiology staff doses and dose monitoring requirements. Radiat. Prot. Dosim. 2009, 136, 140–157. [CrossRef] [PubMed] 2. Martin, C.J. A review of radiology staff doses and dose monitoring requirements. Radiat. Prot. Dosim. 2009, 136, 140 157. [CrossRef] [PubMed] 3 Padovani R ; Rodella C A Staff dosimetry in interventional cardiology Rad Prot Dosim 2001 94 99 103 [CrossRef] [CrossRef] [PubMed] 3. Padovani, R.; Rodella, C.A. Staff dosimetry in interventional cardiology. Rad. Prot. Dosim. 2001, 94, 99–103. [CrossRef] ] ; Rodella, C.A. Staff dosimetry in interventional cardiology. Rad. Prot. Dosim. 2001, 94, 99–103. [CrossRef] y gy 4. Koukorava, C.; Carinou, E.; Ferrari, P.; Krim, S.; Struelens, L. Study of the parameters affecting operator doses in interventional radiology using Monte Carlo simulations. Rad. Meas. 2011, 46, 1216–1222. [CrossRef] gy g 5. ESR. European Society of Radiology (ESR) and Cardiovascular and Interventional Radiological Society of Europe (CIRSE). Interventional Radiology in European Radiology Departments: A Joint Survey from the European Society of Radiology (ESR) and the Cardiovascular and Interventional Radiological Society of Europe (CIRSE). Insights into Imaging; 10:16. 2019. Available online: https://doi.org/10.1186/s13244-019-0698-6 (accessed on 2 February 2022). p g y 6. On behalf of the ICRP; Cousins, C.; Miller, D.L.; Bernardi, G.; Rehani, M.M.; Schofield, P.; Vañó, E.; Einstein, A.J.; Geiger, B.; Heintz, P.; et al. International Commission on Radiological Protection: Radiological protection in cardiology. Ann. ICRP 2013, 42, 1–125. [CrossRef] [PubMed] 7. Vañó, E.; González, L.; Guibelalde, E.; Fernández, J.M.; Ten, J.I. Radiation exposure to medical staff radiology. Brit. J. Radiol. 1998, 71, 954–960. [CrossRef] [PubMed] 8. Donadille, L.; Carinou, E.; Brodecki, M.; Domienik, K.; Jankowskic, J.; Koukorava, C.; Krim, S.; Nikodemova, D.; Ruiz-Lopez, N.; Sans-Merce, M.; et al. Staff eye lens and extremity exposure in interventional cardiology: Results of the ORAMED project. Rad. Meas. 2011, 46, 1203–1209. [CrossRef] 9. 4. Conclusions In particular, they showed that during a calibration procedure on a phantom, APDs are insensitive to the backscatter radiation, but some of them could not respond correctly in case of a very intense pulsed X-ray beam. That means that some aspects related to APDs probably deserve supplementary investigations. Based on the work performed on these devices, a series of recommendations were proposed as a practical guide for hospital users, medical physicists, and radiation protection experts [52]. Future works will be addressed to better assess APDs’ response in these pulsed radiation ﬁelds and to study the inﬂuence of the position of the dosimeter on the operator on the evaluated doses. A comprehensive document summarizing all the works performed in this field by WG12 is under preparation, and it is expected to be issued soon as a EURADOS technical report. Author Contributions: Conceptualization and methodology M.G., E.C., O.H., I.C., U.O., M.S.-M., Ž.K., F.V. and P.F., simulations P.F., E.B., F.B., J.J., Z.J., D.K., P.T. and M.Ž., data acquisition and measurements O.H., F.V., L.C., E.C., I.C., D.F., P.F., F.M., M.G., U.O. and Ž.K.; writing—review and editing, P.F., M.G., E.C., E.B., D.F., I.C., J.J., P.T., Ž.K. and F.V. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Funding: This research received no external funding. Data Availability Statement: All the data refer to published study cited throughout the text. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 11 of 13 Environments 2022, 9, 53 References Nikodemová, D.; Brodecki, M.; Carinou, E.; Domienik, K.; Donadille, L.; Koukorava, C.; Krim, S.; Ruiz-López, N.; Sans-Merce, M.; Struelens, L.; et al. Staff extremity doses in interventional radiology. Results of the ORAMED measurement campaign. Rad. Meas. 2011, 46, 1210–1215. [CrossRef] 10. Clerinx, P.; Buls, N.; Bosmans, H.; de Mey, J. Double dosimetry algorithm for workers in interventional radiology. Rad. Prot. Dosim. 2008, 129, 321–327. [CrossRef] 11. Padovani, R.; Foti, C.; Malisan, M.R. Staff dosimetry protocols in interventional radiology. Rad. Prot. Dosim 2001, 94, 193–196. [CrossRef] [PubMed] ] mmission on Radiological Protection. The 2007 recommendations of the International Commission on Radiol nn. ICRP 2007, 103, 2–4. 12. International Commission on Radiological Protection. The 2007 recommendations of the Internation cal Protection. Ann. ICRP 2007, 103, 2–4. 13. International Commission on Radiological Protection. ICRP Statement on tissue reactions/early and late effects of radiation in normal tissues and organs threshold doses for tissue reactions in a radiation protection context. Ann. ICRP 2012, 41, 1–322. [CrossRef] [PubMed] [ ] [ ] 14. Council Directive 2013/59/Euratom of 5 December 2013 Laying down Basic Safety Standards for Protection against the Dangers Arising from Exposure to Ionising Radiation, and Repealing Directives. 89/618/Euratom, 90/641/Euratom, 96/29/Euratom, 97/43/Euratom and 2003/122/Euratom. Available online: https://eur-lex.europa.eu/eli/dir/2013/59/oj (accessed on 14 April 2022). p p j p 15. Carinou, E.; Ferrari, P.; Ciraj Bjelac, O.; Ginjaume, M.; Sans Merce., M.; O’Connor, U. Eye lens monitoring for interventional radiology personnel: Dosemeters, calibration and practical aspects of Hp(3) monitoring. A 2015 review. J. Radiol. Prot. 2015, 35, R17–R34. [CrossRef] 16. Ciraj Bjelac, O.; Carinou, E.; Ferrari, P.; Ginjaume, M.; Sans Merce, M.; O’Connor, U. Occupational Exposure of the Eye Lens in Interventional Procedures: How to Assess and Manage Radiation Dose. J. Am. Coll. Radiol. 2016, 3, 1347–1353. [CrossRef] [PubMed] 17. Abdelrahman, M.; Lombardo, P.; Vanhavere, F.; Seret, A.; Phillips, C.; Covens, P. First steps towards online personal dosimetry using computational methods in interventional radiology: Operator’s position tracking and simulation input generation. Radiat. Phys. Chem. 2020, 171, 108702. [CrossRef] y 18. Inaba, Y.; Chida, K.; Kobayashi, R.; Kaga, R.; Zuguchi, M. Fundamental study of a real-time occupatio interventional radiology staff. J. Radiol. Prot. 2014, 34, N65–N71. [CrossRef] [PubMed] 18. Inaba, Y.; Chida, K.; Kobayashi, R.; Kaga, R.; Zuguchi, M. Fundamental study of a real-time occupational dosimetry system for interventional radiology staff. J. Radiol. Prot. 2014, 34, N65–N71. [CrossRef] [PubMed] 19. Vañó, E.; Fernandez, J.M.; Sanchez, R. g y EM. Catalogue of Diagnostic X ray and Other Spectra; Scientiﬁc Report Series No.78; IPEM: York, UK, 1997. g p D.B. (Ed.) MCNPX User’s Manual; LA-CP-05-0369; Los Alamos Laboratoty: Los Alamos, NM, USA, 2005. References Occupational dosimetry in real time. Beneﬁts for interventional radiology Rad. Meas. 2011, 46, 1262–1265. 19. Vañó, E.; Fernandez, J.M.; Sanchez, R. Occupational dosimetry in real time. Beneﬁts for interventional radiology Rad. Meas. 2011, 46, 1262–1265. 20. Vañó, E.; Sanchez., R.; Fernandez, J.M. Strategies to optimise occupational radiation protection in interventional cardiology using simultaneous registration of patient and staff doses. J. Radiol. Prot. 2018, 38, 1077–1088. [CrossRef] [PubMed] g p 21. Pelowitz, D.B. (Ed.) MCNPX User’s Manual; LA-CP-05-0369; Los Alamos Laboratoty: Los Alamos, NM 22. Pelowitz, D.B. (Ed.) MCNP6 User’s Manual; Version 1.0, s LA-CP-13-00634; Los Alamos Laboratoty: Los Alamos, NM, USA, 2013. y 23. Ferrari, P.; Becker, F.; Carinou, E.; Chumak, V.; Farah, J.; Jovanovic, Z.; Krstic, D.; Morgun, A.; Principi, S.; Teles, P. Monte Carlo study of the scattered radiation ﬁeld near the eyes of the operator in interventional procedures. J. Radiol. Prot. 2016, 36, 902–921. [CrossRef] 24. International Commission on Radiation Units and Measurements. ICRU Report 57 Conversion Coefﬁcients for Use in Radiological Protection against External Radiation; Oxford University Press: Bethesda, MD, USA, 1996. g y 25. IPEM. Catalogue of Diagnostic X ray and Other Spectra; Scientiﬁc Report Series No.78; IPEM: York, UK, 1 12 of 13 12 of 13 Environments 2022, 9, 53 26. Ginjaume, M.; Vanhavere, F.; Carinou, E.; Gualdrini, G.; Clairand, I.; Sans-Merce, M. (Eds.) International Workshop on Optimization of Radiation Protection of Medical Staff. ORAMED Rad. Meas. 2011, 46, 11. p 27. Ferrari, P.; Becker, F.; Jovanovic, Z.; Khan, S.; Bakhanova, E.; Principi, S.; Krstic, D.; Pierotti, L.; Mariotti, F.; Faj, D. Simulation of Hp(10) and effective dose received by the medical staff in interventional radiology procedures. J. Radiol. Prot. 2019, 39, 809–824. [CrossRef] [PubMed] [ ] [ ] 28. Nowak, M.; Carbonez., P.; Krauss, M.; Verdun, F.R.; Damet, J. Characterisation and mapping of scattered radiation ﬁelds in interventional radiology theatres. Sci. Rep. 2020, 30, 18754. [CrossRef] [PubMed] gy p 29. Leyton, F.; Nogueira, M.S.; Gubolino, L.A.; Pivetta, M.R.; Ubeda, C. Correlation between scatter radiation dose at height of operator’s eye and dose to patient for different angiographic projections. Appl. Radiat. Isot. 2016, 117, 100–105. [CrossRef] [PubMed] [ ] 30. Kim, K.P.; Miller, D.L.; Berrington de Gonzalez, A.; Balter, S.; Kleinerman, R.A.; Ostroumova, W.; Simon, S.L.; Linet, M.S. Occupational radiation doses to operator performing ﬂuoroscopically-guided procedures. Health Phys. 2012, 103, 80–99. [CrossRef] [ ] 31. References National Council on Radiation Protection and Measurements Use of Personal Monitors to Estimate Effective Dose Equivalent to Workers for External Exposure to Low-LET Radiation. Bethesda, MDNCRP NCRP Report No. 122. 1995. Available online: https://ncrponline.org/shop/reports/report-no-122-use-of-personal-monitors-to-estimate-effective-dose-equivalent- and effective dose to workers for external exposure to low let radiation 1995/ (accessed on 14 April 2022) p ( p ) 32. McEwan, A.C. Assessment of occupational exposure in New Zealand from personal monitoring records. Radiat. Prot. Australas. 2000, 17, 60–66. 33. Jarvinen, H.; Buls, N.; Clerinx, P.; Jansen, J.; Miljanic, S.; Nikodemova, D.; Ranogajec-Komor, M.; d’Errico, F. Overview of double dosimetry procedures for the determination of the effective dose to the interventional radiology staff. Radiat. Prot. Dosim. 2008, 129, 333–339. [CrossRef] 34. Domienik, J.; Brodecki, M. The effectiveness of lead glasses in reducing the doses to eye lenses during cardiac implantation procedures performed using x-ray tubes above the patient table. J. Radiol. Prot. 2016, 36, N19–N25. [CrossRef] 35. Van Roojien, B.D.; de Haan, M.W.; Das, M.; Arnoldusse, C.W.K.P.; de Graaf, R.; van Zwam, W.H.; Backes, W.H.; Jeuken, C.R.L.P.N. Efﬁcacy of Radiation Safety Glasses in Interventional Radiology Cardio. Vasc. Interv. Radiol. 2014, 37, 1149–1155. 36. Martin, C.J.; Magee, J.S.; Sandblom, V.; Almén, A.; Lundh, C. Eye dosimetry and protective eyewear for interventional clinicians. Radiat. Prot. Dosim. 2015, 165, 284–288. [CrossRef] 37. Dominiek, J.; Brodecki, M.; Carinou, E.; Donadille, L.; Jankowski, J.; Koukorava, C.; Krim, S.; Nikodemova, D.; Ruiz-Lopez, N.; Sans-Merce, M.; et al. Extremity and eye lens doses in interventional radiology and cardiology procedures: First results of the ORAMED project. Radiat Prot Dosim. 2011, 144, 442–447. [CrossRef] 38. Sailer, A.M.; Vergoossen, L.; Paulis, L.; van Zwam, W.H.; Das, M.; Wildberger, J.E.; Jeukens, C.R. Personalized Feedback on Staff Dose in Fluoroscopy-Guided Interventions: A New Era in Radiation Dose Monitoring. Cardiovasc. Intervent. Radiol. 2017, 40, 1756–1762. [CrossRef] 39. Picano, E.; Vano, E.; Domenici, L.; Bottai, M.; Thierry-Chef, I. Cancer and non-cancer brain and eye effects of chronic low-dose ionizing radiation exposure. BMC Cancer 2012, 12, 157. [CrossRef] [PubMed] 40. Roguin, A.; Goldstein, J.; Bar, O.; Goldestein, J.A. Brain and neck tumours among physicians performing interventional procedures. Am. J. Cardiol. 2013, 111, 1368–1372. [CrossRef] [PubMed] 41. Reeves, R.R.; Ang, L.; Bahadorani, J.; Naghi, J.; Dominguez, A.; Palakodeti, V.; Tsimikas, S.; Patel, M.P.; Mahmud, E. Invasive Cardiologists are exposed to greater left sided cranial radiation: The Brain study. JACC Cardiovasc. Interv. 2015, 8, 1197–1206. [CrossRef] [PubMed] 42. 47. IEC 61267:2005; Medical Diagnostic X-ray Equipment—Radiation Conditions for use in the Determination of Characteristics. IEC: Geneva, Switzerland. Available online: https://webstore.iec.ch/publication/5079 (accessed on 14 April 2022). References Kithara, C.M.; Linet, M.S.; Balter, S.; Miller, D.L.; Rajaraman, P.; Cahoon, E.K.; Velazquez-Kronen, R.; Simon, S.L.; Little, M.P.; Doody, M.M.; et al. Occupational radiation exposure and deaths from malignant intracranial neoplasms of the brain and CNS in U.S. radiologic technologists, 1983–2012. Am. J. Roentgenol. 2017, 208, 1278–1284. [CrossRef] 43. Ferrari, P.; Jovanovic, Z.; Bakhanova, E.; Becker, F.; Krstic, D.; Jansen, J.; Principi, S.; Teles, P.; Clairand, I.; Knezevic, Ž. Absorbed dose in the operator’s brain in interventional radiology practices: Evaluation through KAP value conversion factors. Phys. Med. 2020, 76, 177–181. [CrossRef] 44. Watson, L.E.; Riggs, M.W.; Bourland, P.D. Radiation exposure during cardiology fellowship training. Health Phys. 1997, 73, 690–693. [CrossRef] 45. Hattori, S.; Monzen, H.; Tamura, M.; Kosaka, H.; Nakamura, Y.; Nishimura, Y. Estimating radiation exposure of the brain of a physician with a protective ﬂap in interventional radiology: A phantom study. J. Appl. Clin. Med. Phys. 2022, 23, e13532. [CrossRef] 46. UNE EN ISO 4037-3; 2021 Radiological Protection–X and Gamma Reference Radiation for Calibrating Dosemeters and Doserate Meters and for Determining Their Response as a Function of Photon Energy–Part 3: Calibration of Area and Personal Dosemeters and the Measurement of Their Response as a Function of Energy and Angle of Incidence (ISO 4037-3:2019). ISO: Geneva, Switzerland. Available online: https://www.iso.org/standard/66874.html (accessed on 14 April 2022). p g p 47. IEC 61267:2005; Medical Diagnostic X-ray Equipment—Radiation Conditions for use in the Determination of Characteristics. IEC: Geneva, Switzerland. Available online: https://webstore.iec.ch/publication/5079 (accessed on 14 April 2022). 13 of 13 13 of 13 Environments 2022, 9, 53 48. Hupe, O.; Friedrich, S.; Vanhavere, F.; Marcin, B. Determining the dose rate dependence of different active personal dosemeters in standardized pulsed and continuous radiation ﬁelds. Rad. Prot. Dosim. 2019, 187, 345–352. [CrossRef] p 49. International Electrotechnical Commission. Radiation Protection Instrumentation–Dosemeters for Pulsed Fields of Ionizing Radiation. International Draft Technical Speciﬁcation IEC/DTS 63050, Community Draft. Available online: https://webstore.iec. ch/preview/info_iects63050%7Bed1.0%7Den.pdf (accessed on 14 April 2022). p p p 50. Ginjaume, M.; Carinou, E.; Marcin, B.; Clairand, I.; Domienik-Andrzejewska, J.; Exner, L.; Ferrari, P.; Jovanovi´c, Z.; Krstic, D.; Principi, S.; et al. Effect of the radiation protective apron on the response of active and passive personal dosemeters used in interventional radiology and cardiology. J. Radiol. Prot. 2019, 39, 97–112. [CrossRef] gy gy 51. Vanhavere, F.; Carinou, E.; Clairand, I.; Ciraj-Bjela, O.; De Monte, F.; Domienik-Andrzejewska, J.; Ferrari, P.; Ginjaume, M.; Hrsak, H.; Hupe, O.; et al. 54. Martin, C.J. Eye lens dosimetry for ﬂuoroscopically guided clinical procedures: Practical approaches to protection and dose monitoring. Radiat. Prot. Dosim. 2016, 169, 286–291. [CrossRef] [PubMed] References The use of active personal dosemeters in interventional workplaces in hospitals: Comparison between active and passive dosemeters worn simultaneously by medical staff. Rad. Prot. Dosim. 2020, 188, 22–29. [CrossRef] [PubMed] p y y 52. O’Connor, U.; Carinou, E.; Clairand, I.; Ciraj-Bjelac, O.; De Monte, F.; Domienik-Andrzejewska, J.; Ferrari, P.; Ginjaume, M.; Hršak, H.; Hupe, O.; et al. Recommendations for the use of active personal dosemeters (APDs) in interventional workplaces in hospitals. Phys. Med. 2021, 87, 131–135. [CrossRef] [PubMed] 53. Maeder, M.; Brunner-La Rocca, H.P.; Wolber, T.; Ammann, P.; Roelli, H.; Rohner, F.; Rickli, H. Impact of a Lead Glass Screen on Scatter Radiation to Eyes and Hands in Interventional Cardiologists. Catheter. Cardiovasc. Interv. 2006, 67, 18–23. [CrossRef] [PubMed] 54. Martin, C.J. Eye lens dosimetry for ﬂuoroscopically guided clinical procedures: Practical approaches to protection and dose monitoring. Radiat. Prot. Dosim. 2016, 169, 286–291. [CrossRef] [PubMed] 54. Martin, C.J. Eye lens dosimetry for ﬂuoroscopically guided clinical procedures: Practical approaches to protection and dose monitoring. Radiat. Prot. Dosim. 2016, 169, 286–291. [CrossRef] [PubMed]
https://openalex.org/W3019285824	https://ieeexplore.ieee.org/ielx7/6287639/8948470/09076081.pdf	English	null	Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers Without Duty-Cycle Optimization	IEEE access	2,020	cc-by	9,159	Received March 30, 2020, accepted April 16, 2020, date of publication April 22, 2020, date of current version May 7, 2020. Received March 30, 2020, accepted April 16, 2020, date of publication April 22, 2020, date of current version May 7, 2020. Digital Object Identifier 10.1109/ACCESS.2020.2989505 XIAOLONG SHI 1, (Student Member, IEEE), JIANGUO ZHU 2, (Senior Member, IEEE), LI LI 1, (Member, IEEE), AND DYLAN DAH-CHUAN LU 1, (Senior Member, IEEE) 1School of Electrical and Data Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia 2School of Electrical and Information Engineering, The University of Sydney, Sydney, NSW 2006, Australia Corresponding author: Xiaolong Shi (sydney3168@gmail.com) XIAOLONG SHI 1, (Student Member, IEEE), JIANGUO ZHU 2, (Senior Member, IEEE), LI LI 1, (Member, IEEE), AND DYLAN DAH-CHUAN LU 1, (Senior Member, IEEE) 1School of Electrical and Data Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia 2School of Electrical and Information Engineering, The University of Sydney, Sydney, NSW 2006, Australia Corresponding author: Xiaolong Shi (sydney3168@gmail.com) ABSTRACT The conventional model-predictive-based direct power control (MPDPC) of the three-phase full-bridge AC/DC converters chooses the best single voltage vector for the following control period, which results in variable switching frequency and power distortion, and thus a relatively higher sampling frequency is needed to achieve acceptable results. This paper proposes a simpliﬁed dual-vector-based predictive direct duty-cycle-control (SPDDC) with an additional zero vector implemented in contrast to the MPDPC. With the same best vector selection method, the proposed strategy has retained the control simplicity with just one more step added and much better control performance as well as a ﬁxed switching frequency in comparison to the MPDPC. On the other hand, the duty-cycle optimization procedure is eliminated while the negative duration issue is essentially resolved compared with the conventional dual-vector-based model predictive duty-cycle-control (MPDCC). Comprehensive comparisons of various control methods by numerical simulation and experimental testing show that the SPDDC can achieve better steady state and dynamic performance than the MPDPC and simpler algorithms than the MPDCC. INDEX TERMS AC/DC conversion, cost function, duty-cycle-control, duty-cycle optimization, model predictive control. INDEX TERMS AC/DC conversion, cost function, duty-cycle-control, duty-cycle optimization, model predictive control. table [6]–[9]. However, the STDPC bears with variable switching frequency and irregular power ripples because of the use of hysteresis comparators and switching tables, result- ing in broadband harmonic spectrum range, which requires high sampling frequency for the acceptable performance. is work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ I. INTRODUCTION Various control methods have been proposed in the liter- ature to further enhance the control performance and realize the ﬁxed switching frequency by implementing two or more vectors in one control section. In [23]–[31], the concept of duty-cycle-control and SVM-based methods have been introduced and studied. The SVM-based method obtains the desired vector by nullifying the error of the control variable at the end of the following control period, which uses the SVM to generate the gating pulses [12], [27]. The three- vector-based duty-cycle-control in [23] and [24] obtains the best adjacent non-zero vector pair and a zero vector by sector information, namely the grid-voltage vector location. How- ever, the sector information-based vector selection algorithm might select the non-best vectors [25], this further results in negative duration value and power control deterioration, such as signiﬁcant power notches and current spikes. To solve this issue, in [25] and [26], a complementary vector sequence table is added with an additional vector sequence table, but the negative duration issue cannot be fully solved after recalcula- tion. The authors in [27] proposed an improved method with- out selecting the voltage vectors, but it needs a procedure of equivalent reconstruction of switching signals. Therefore, for multi-vector-based duty-cycle-control, using sector informa- tion for the best non-zero vector selection is not a good choice since the negative duration issue could be serious. Recently, the popular and effective solution is to use the model predic- tive control method for the best vector selection, it has been applied in several multi-vector-based approaches [28]–[33], while the procedure of determining these vectors become more complex [10], [33]. Taking advantage of the feature that the power variation rates of reversible vector pair are symmet- rical for that of zero vector, the proposed three-vector-based method with reversible vector selection in [33] cunningly resolved the negative duration issue. However, the control complexity is also increased in comparison to the MPDPC and dual-vector-based methods. To overcome the critical issues such as complicated vec- tor selection approach and high computational burden of durations, this paper proposes a simpliﬁed predictive direct duty-cycle (SPDDC) approach. The SPDDC can solve the negative duration issue essentially and simplify the control scheme by eliminating the duty-cycle optimization. Firstly, the cost function of MPDPC is employed to choose the best active voltage vector to avoid selecting the non-optimal vectors. I. INTRODUCTION The three-phase full-bridge AC/DC converter is a popular device widely applied in various applications, such as inte- gration of renewable energy resources, electric drives, volt- age source converter transmission, and so on. It has several merits such as four-quadrants power control, ﬂexible DC voltage control, and low DC capacitance with high-quality DC voltage [1]–[4]. To improve the performance of STDPC, various kinds of methods have been combined with the direct power control (DPC), such as the fuzzy logic, space vector modulation (SVM), sliding mode, virtual ﬂux, and model predictive con- trol [10]–[15]. The model-predictive-based direct power con- trol (MPDPC) is a quite popular control method in the area of power converters and motor drives along with control unit development, since it is a promising algorithm with advan- tages like rapid instantaneous response, no need of modula- tor, and ﬂexibility to add various constraints [13], [16]–[22]. Though the power ripples can be reduced compared with the STDPC, the MPDPC still cannot achieve satisfactory steady state performance since only one single switching vector selected in each control period and limited number of voltage vectors. Besides, its switching frequency is not The classical voltage-oriented control (VOC) method con- trols the input power by regulating the decoupled AC cur- rents [5]. Although it has good dynamic and steady state performance, the main drawback is that VOC highly depends on the inner current controller and coordinate transformation accuracy. Similar to the direct torque control (DTC) of elec- trical machine drive, the conventional switching-table-based direct power control (STDPC) regulates the power directly by choosing one voltage vector from a predeﬁned switching The associate editor coordinating the review of this manuscript and approving it for publication was Dušan Grujić . 77049 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers constant, leading to the spread spectrum nature of harmonics and complicated ﬁlter design. functions for the best vector pair selection and only selects the adjacent non-zero vector pair with or without the zero vector. The performance is not evaluated compared with other multi-vector-based methods. Some new insights on reduc- ing the calculation burden of vector selection and duration have been provided in [34]–[36]. The algorithms for vector selection and duration calculation are reconstructed from the SVM-based method in [34] and from different control objectives in [35], [36], which are still quite complex and not intuitive. I. INTRODUCTION Then, the duty-cycle is derived directly by using the cost function value obtained by applying the selected vectors without compensation and optimization. It can thus eliminate the calculation of power slopes and reduce the computational burden. While the negative duration is essentially avoided, the merits, such as good steady state and dynamic perfor- mance, of the duty-cycle-control are retained. The control algorithm is more intuitive with better performance compared with the conventional methods. Numerical simulation and experiments have been conducted and the results discussed to conﬁrm the theoretical analyses and the advantages of the novel method. III. MODEL-PREDICTIVE-BASED DUTY-CYCLE- CONTROL III. MODEL-PREDICTIVE-BASED DUTY-CYCLE- CONTROL A typical dual-vector-based model predictive duty-cycle- control (MPDCC) method is presented in [30]. The power differential equation can be derived from (4) as ( Pk+1 = Pk + δpntn + δpztz Qk+1 = Qk + δqntn + δqztz (12) (12) d dt P Q = 3 2 iα d dt eα eβ + diα dt eα eβ + iβ d dt eβ −eα + diβ dt eβ −eα (5) where δpn and δpz are the P slopes of the non-zero voltage vectors, Vn, and zero voltage vector, Vz; δqn and δqz the Q slopes; tn and tz the durations of Vn and Vz, respectively, and tn + tz = Ts. (5) In the complex form, the sinusoidal and balanced three- phase emf can be expressed in the αβ-coordinate system as The errors between the predicted and reference power values can be derived as e = eα + jeβ = \|e\|ejωt (6) (6)    Perr = Pref − Pk + δpntn + δpztz Qerr = Qref − Qk + δqntn + δqztz (13) (13) where ω is the grid angular frequency. Taking derivatives on the both sides of (6) yields d dt eα eβ = ω · −eβ eα (7) where Perr and Qerr are the power errors, and Pref and Qref are the reference power, respectively. The optimal durations can then be calculated by using the least-square optimization method [28]–[31] to minimize the cost function of power errors deﬁned as (7) Substituting (3), (4) and (7) into (5), one obtains Substituting (3), (4) and (7) into (5), one obtains d dt Pi Q i = −R L Pi Q i + ω −Qi Pi + 3 2L \|e\|2 −Re eV∗ i −Im eV∗ i (8    J = P2 err + Q2 err ∂J ∂tn = 0 (14) (14) (8) Finally, the optimal durations can be derived as where Vi is the voltage vector, and i = 0,1,2...6. II. MODELING OF THREE-PHASE GRID-CONNECTED CONVERTER Fig.1 shows a topology of the three-phase full-bridge grid-connected AC/DC converter, which is connected to the main AC power supply via three inductors, L, and resistors, R, where ea, eb, and ec are the electromotive forces (emf) of three-phase AC power source; va, vb, and vc the three-phase terminal voltages; and ia, ib, and ic the three-phase currents. At the DC side, a DC load is connected to the DC bus in parallel to a capacitor C. Apart from the efforts on the improvement of best vec- tor or vector pair selection, the signiﬁcantly increased computational burden on hardware from the vector dura- tion calculation should also be considered, since the least-square optimization method is generally employed in the above two or three-vector-based duty-cycle-control methods [25]–[31], [33]. The duration calculation approach is computationally intensive since the active and reac- tive power slope calculations of the selected vectors are needed [30]. Meanwhile, the negative duration of zero vec- tor cannot be avoided especially during the dynamic instant with large instantaneous power errors, and thus the com- pensation measure of negative duration issue is also needed, which further increases the control complexity. By emulat- ing the SVM strategy, the modulated model predictive con- trol (MMPC) in [32] uses a different method to derive the durations simply. However, it employs two complicated cost FIGURE 1. Topology of the AfC/DC three-phase converter. FIGURE 1. Topology of the AfC/DC three-phase converter. In the αβ-coordinate system, the AC power source emf and current vectors, eαβ and iαβ, can be derived from the 77050 77050 VOLUME 8, 2020 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers three-phase emf and currents, respectively, as 2 " 1 1 2 1 2 # ea  predicted active and reactive powers of the converter switch- ing state can be derived as: three-phase emf and currents, respectively, as predicted active and reactive powers of the converter switch- ing state can be derived as: predicted active and reactive powers of the converter switch- ing state can be derived as: three-phase emf and currents, respectively, as eαβ = eα eβ = 2 3 " 1 −1 2 −1 2 0 √ 3 . 2 − √ 3 . II. MODELING OF THREE-PHASE GRID-CONNECTED CONVERTER 2 #   ea eb ec   (1) iαβ = iα iβ = 2 3 " 1 −1 2 −1 2 0 √ 3 . 2 − √ 3 . 2 #   ia ib ic   (2) predicted active and reactive powers of the converter switch- ing state can be derived as: Pk+1 i Qk+1 i = Ts −R L Pk i Qk i + ω −Qk i Pk i + 3 2L \|e\|2 −Re eV∗ i −Im eV∗ i + Pk i Qk i (10) where T is the sampling period For the conventional dual eαβ = eα eβ = 2 3 " 1 −1 2 −1 2 0 √ 3 . 2 − √ 3 . 2 #   ea eb ec   (1) iαβ = iα iβ = 2 3 " 1 −1 2 −1 2 0 √ 3 . 2 − √ 3 . 2 #   ia ib ic   (2) ing state Pk+1 i Qk+1 i where T (10) (2) where Ts is the sampling period. For the conventional dual- vector-based MPDCC, the P and Q differences between the references, Pref and Qref , and predictive values, Pk+1 i and Qk+1 i , of the i-th switching state with non-zero voltage vector, Vi, at the (k + 1)-th sampling instant are evaluated by a predeﬁned cost function The line currents can be calculated by solving eαβ = L diaβ dt + Riaβ + vαβ (3) (3) where vαβ is the terminal voltage vector in the αβ-coordinate system. The exchange between the active power P and the reactive power Q can be derived as ji = (Pref −Pk+1 i )2 + (Qref −Qk+1 i )2 (11) (11) P Q = 3 2 eα eβ eβ −eα iα iβ (4) and the non-zero voltage vector is pick as the optimal one if it yields a minimum cost function value. (4) Assuming the power slopes are kept constant during quite a small sampling period, one can predict the active and reactive powers at the end of a control period by VOLUME 8, 2020 III. MODEL-PREDICTIVE-BASED DUTY-CYCLE- CONTROL The voltage vector corresponding to the i-th switching state, [Viα, Viβ], can be derived as    tn = (Pref −Pk)(δpn −δpz) + (Qref −Qk)(δqn −δqz) (δpn −δpz)2 + (δqn −δqz)2 + Ts(δ2 pz + δ2 qz −δpnδpz −δqnδqz) (δpn −δpz)2 + (δqn −δqz)2 tz = Ts −tn (15) Viα Viβ = 2 3Vdc   Sia −1 2(Sib + Sic) √ 3 2 (Sib −Sic)   (9) (15) (9) where Sia, Sib and Sic are the converter switching states, and Vdc is the DC-bus voltage. Once tn and tz are calculated, the pulse width modulation (PWM) switching signals can be generated at a ﬁxed switch- ing frequency. However, the duration based on (15) could be negative or over Ts, especially at a dynamic instant with large instantaneous active or reactive power error. The measure is to force the action time to zero whenever the negative duration Assume the tracking error of the DC-bus voltage to be constant during two sampling periods. The instant power at the beginning of the following (k + 1)-th sampling instant can be evaluated by linear extrapolation, and the corresponding 77051 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers where dn and dz are the duty cycles of non-zero vector Vn and zero vector Vz. If Ji/λJ0 decreases, dn would increase, mean- ing the best active vector constitutes an increased propor- tion of the control period. Otherwise, the zero vector would constitute an increased proportion of the control period. The balanced modulation scheme can adjust the duty cycle automatically. exists and saturate the other duration to Ts, which may lead to steady state performance deterioration or dynamic power overshoot. The hardware with high computing capability is required for the calculation of power slopes and optimal durations. A. PRINCIPLE OF THE PROPOSED SPDDC A. PRINCIPLE OF THE PROPOSED SPDDC Based on the cost function Based on the cost function Based on the cost function Ji = (P∗−Pk+1 i )2 + (Q∗−Qk+1 i )2 (16) (16) the best non-zero vector, Vi, yielding the minimum Ji is selected. The cost function can also be deﬁned in the abstract form as IV. SIMPLIFIED PREDICTIVE DIRECT DUTY-CYCLE-CONTROL Different to the MPDCC, the durations are simply calcu- lated by (18) rather than (15) and directly allocated to the corresponding vectors. The calculations of power slopes for each vector are eliminated, which reduces the calculation burden and is intuitive for implementation. Since tn and tz are undoubtedly within the range of 0 to Ts in inverse pro- portion to Ji and λJ0, respectively, it can eliminate the need of compensation, and the negative duration issue is essen- tially solved. Table 1 compares the complexities of MPDCC, SPDDC, and MPDPC. To solve the issues mentioned above, this paper proposes the SPDDC, which differs from the MPDPC by just adding one more step of the duty-cycle calculation. By emulating the implementation of SVM, the proposed method assigns a por- tion of the control period directly in reciprocal proportional with the cost function value of corresponding optimal dual vectors, which means the cost function is not only used for the best vector selection but also applied for duration calculation. To solve the issues mentioned above, this paper proposes the SPDDC, which differs from the MPDPC by just adding one more step of the duty-cycle calculation. By emulating the implementation of SVM, the proposed method assigns a por- tion of the control period directly in reciprocal proportional with the cost function value of corresponding optimal dual vectors, which means the cost function is not only used for the best vector selection but also applied for duration calculation. Thus, it can eliminate the power slope calculation of each vector as in (12), which is more intuitive compared with the MPDCC while retaining the ﬁxed-switching frequency feature. Besides, the negative duration issue of MPDCC is essentially resolved. Merits of both the MPDPC and MPDCC are retained. TABLE 1. Complexity comparisons of each control scheme. TABLE 1. Complexity comparisons of each control scheme. B. MATHEMATICAL ANALYSIS OF THE PROPOSED SPDDC Ji = Pref −Pk+1 i + Qref −Qk+1 i (17) (17) Using complex apparent power, (8) can be rewritten as Using complex apparent power, (8) can be rewritten as Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers The predicted value (−S∗)k+1 can be derived from (21) as (−S∗)k+1 = (−S∗)k + Ts L (R + jωL) · (S∗)k −3 2E2 \| {z } (−S∗)k+1 0 + 3ETs 2L vdq (23) The predicted value (−S∗)k+1 can be derived from (21) as The predicted value (−S∗)k+1 can be derived from (21) as Therefore, the SPDCC with (19) can be rearranged as Therefore, the SPDCC with (19) can be rearranged as dn = λG0 λG0 + G1 = \|a\| λ \|a\| + \|b\| = \|a\| \|a\| + \|b\| 1 λ (27) (−S∗)k+1 = (−S∗)k + Ts L (R + jωL) · (S∗)k −3 2E2 \| {z } (−S∗)k+1 0 (27) By replacing cos α/ cos β with λ, the calculation complex- ity can be signiﬁcantly reduced. + 3ETs 2L vdq (23) (23) In the steady state, generally, both the zero vector and best non-zero vector are selected for implementation, which means 0 ≤dn−opt ≤1 and 0 ◦≤β ≤90 ◦. Since dn−opt has a large chance to be in the middle range between 0 and 1, we can assume that 0 ◦≤β ≤60 ◦in most cases. Also, it is reasonable to suppose that generally the active vector error b has a lower value than the zero vector error a, namely α < β, since the active vector is selected instead of the zero vector in most cases with single-vector-based methods, such as STDPC and MPDPC. Based on the assumption 0 ◦≤β ≤ 60 ◦and α < β, as 0 ◦≤α ≤30 ◦and λ = cos α/ cos β, it can be deduced that 1 ≤λ ≤2 in general. This has also been veriﬁed through trial and error in simulations and experiments. According to (23), the cost function in (22) can be rear- ranged as G = 1(−S∗)k+1 0 −3ETs 2L vdq (24) (24) where 1(−S∗)k+1 0 = (−S∗)ref −(−S∗)k+1 0 represents the error vector caused by the zero voltage vector. where 1(−S∗)k+1 0 = (−S∗)ref −(−S∗)k+1 0 represents the error vector caused by the zero voltage vector. FIGURE 2. Principle of calculating duty cycle dn−opt of optimized non-zero vector in a synchronous frame, where a = 1(−S∗)k+1 0 , b = 1(−S∗)k+1 0 −mv1, and m = 1.5ETs/2L. Using complex apparent power, (8) can be rewritten as p Since the complex mathematical calculation is no longer needed and λ theoretically equals to cos α/ cos β, it is impos- sible to give the exact range of the λ. The above analy- ses about λ are just based on a reasonable assumption and derivation. In some special cases, λ could be out the range mentioned above. Through the numerical simulation and experimental tests presented in sections V and VI, it has been veriﬁed that the control method works well if λ is in the range of 1 and 2 in the steady state, though the dynamic performance is inﬂuenced a lot by the choice of λ. At the beginning of design, the value of λ can be optimized for good dynamic performance by simulation in MATLAB/Simulink in the range of 1 and 2. It can then be adjusted slightly and applied in experiments. FIGURE 2. Principle of calculating duty cycle dn−opt of optimized non-zero vector in a synchronous frame, where a = 1(−S∗)k+1 0 , b = 1(−S∗)k+1 0 −mv1, and m = 1.5ETs/2L. As shown in Fig. 2, the cost function error G0 caused by the zero vector, namely a = 1(−S∗)k+1 0 , is supposed to be located in sector S1. The non-zero vector closest to a is the best one that can minimize the cost function in (24). Thus, v1 would be selected and 0 ◦≤α ≤30 ◦. Based on (24), the error G1 caused by v1 can be illustrated as vector b in Fig. 2, where m v1 indicates the second term of (23), m = 1.5ETs/2L. The combined error vector c caused by the zero vector and v1 with non-zero vector duty cycle dn can then be derived as Using complex apparent power, (8) can be rewritten as though they do not have noticeable differences [16]. dS dt = 1 L −(R −jωL) · S + 3 2 \|¯e\|2 −¯V ∗¯e (20) The duration can then be directly derived from the inverse proportion to the cost function results of the selected dual vectors, which further takes full use of the cost function values, as (20) where S is the complex power from the power grid. Thus, the slope of the negative conjugate of complex appar- ent power in the synchronous dq frame, denoted as −S∗, can be derived as    k Ji + k λJ0 = Ts tn = k 1 Ji tz = k 1 λJ0 (18) (18) d(−S∗) dt = 1 L (R + jωL) · S∗−3 2E2 + 3 2Evdq (21) (21) where k is the gain for duration calculation and can be solved easily, J0 the cost function value of the zero vector, λ an added parameter to balance the ratio of J0 and Ji, which will be veriﬁed in the following subsection, and tn and tz are the duration of the non-zero vector and zero vector, respectively, which are all positive. Usually, λ can be selected in the range between 1 and 2 to achieve better steady state and dynamic performance simultaneously. where E and vdq are the amplitudes of grid emf vector e and terminal voltage vector V. Since the complex apparent power is irrelevant to the transformation frame, the subscript ‘‘dq’’ of −S∗is neglected. In terms of the apparent power, the cost function (17) can be rewritten as According to (18), the corresponding duty cycles for dual vectors can be calculated by G = Sref −Sk+1 = (−S∗)ref −(−S∗)k+1 (22) (22)    dn = λJ0 Ji + λJ0 tn = dnTs dz = Ji Ji + λJ0 tz = dzTs (19) (19) where the superscript ‘‘ref’’ stands for reference, k + 1 denotes the predictive value for the next control period. 77052 VOLUME 8, 2020 VOLUME 8, 2020 X. A. STEADY STATE PERFORMANCE COMPARISON V = Vndn + Vzdz (28) (28) For the steady state performance comparison, the AC three- phase voltage, current, and instantaneous active and reactive powers of the system under each of the above three control methods are depicted, where the reactive power reference is set as 200 VAR and the active power reference as 400 W. As shown in Fig. 4, both the active and reactive powers can track the reference values with high accuracy. Substituting (28) into (10), one can obtain the predicted values of Pk+1 and Qk+1 at the (k + 1)-th instant with V. Based on Pk+1 and Qk+1, the one-step-ahead prediction Pk+2 i and Qk+2 i of each active vector can be calculated for the next best vector selection by Fig. 4(a) shows that under the MPDPC method, the system exhibits high P and Q ripples of 11.05 W and 12.32 VAR, respectively, and a current total harmonic distortion (THD) of 3.77%. On the other hand, the dual-vector-based methods can achieve much better results than the MPDPC method. For instance, under the control of MPDCC method, as shown in Fig. 4(b), the system exhibits a current THD of 1.49%, and P and Q ripples of 4.87 W and 4.27 VAR, respectively, which are less than half of those under the MPDPC method. Pk+2 i Qk+2 i = Ts −R L Pk+1 Qk+1 + ω −Qk+1 Pk+1 + 3 2L \|e\|2 −Re eV∗ i −Im eV∗ i + Pk+1 Qk+1 (29) (29) The cost function (17) can be revised as The cost function (17) can be revised as Ji = P∗−Pk+2 i + Q∗−Qk+2 i (30) (30) Under the control of proposed SPDDC, while the steady state performance is very close to that of MPDCC method, there exists some small difference. For instance, compared with MPDCC, the Q ripple is further reduced to 3.98 VAR, and the current THD 1.43%, but the P ripple is 5.12 W when λ = 1, which is slightly higher than that of MPDCC. When λ = 1.5, the steady state performance is not far away from that when λ = 1 with slightly increased Q ripple and THD. A. STEADY STATE PERFORMANCE COMPARISON Table 3 compares quantitatively the performance indicators of the above three control methods, such as the current THD and active and reactive power ripples in different states (P = 400 W, Q = 0 VAR). As shown, the MPDPC method yields the highest current THD with a wide harmonic spectrum. On the other hand, the proposed SPDDC method yields a slightly better current THD with much lower low- order harmonic contents than those of the MPDCC. The current harmonics of the MPDCC and SPDDC methods con- centrate at 20 kHz, which is a common feature of dual- vector-based control and can result in much simpler ﬁlter design. In conclusion, the proposed SPDDC can achieve a steady state performance much better than that of the MPDPC method and a steady state performance comparable to that of the MPDCC method. The duration for each vector can still be calculated by (19). Fig.3 illustrates schematically the block diagram of the pro- posed SPDDC. FIGURE 3. Block diagram of SPDDC for grid-connected AC/DC converter. FIGURE 3. Block diagram of SPDDC for grid-connected AC/DC converter. D. DESIGN OF ONE-STEP-DELAY COMPENSATION The one-step-delay between the commanding voltage vector and the applied voltage vector caused by the discrete-time digital implementation can signiﬁcantly deteriorate the con- trol performance, such as power control performance and power prediction error [8], [19]. Thus, the one-step-ahead prediction is required during implementation of SPDDC. Firstly, the combined voltage V is built with currently implemented vectors and durations as C. VECTOR SEQUENCE FOR SWITCHING FREQUENCY OPTIMIZATION In regard to the vector sequence to realize the minimal commutation between vectors in an application, two aspects about vector sequence should comply in order. Firstly, if the vector sequence during the previous cycle is with the zero vector at the end, the same zero vector should be selected and applied ﬁrst in the next cycle to reduce the switching commutation. Otherwise, the sequence of non-zero vector and zero vector should be changed in order to achieve the minimal commutation between the current and the previous vector sequences. Meanwhile, the proper zero vector that requires the least commutation of the current vector sequence should be selected. c = dnb + (1 −dn)a = dn(a −mv1) + (1 −dn)a = a −dnmv1 (25) (25) To minimize the combined error, vector c should be per- pendicular to m v1, as illustrated in Fig. 2. Finally, the opti- mized duty cycle dn−opt can be obtained by To minimize the combined error, vector c should be per- pendicular to m v1, as illustrated in Fig. 2. Finally, the opti- mized duty cycle dn−opt can be obtained by dn−opt = \|a\| cos α m \|v1\| = \|a\| cos α \|a\| cos α + \|b\| cos β = \|a\| \|a\| + \|b\| cos β cos α (26) For example, if active vector ‘‘100’’ is chosen to implement and the last vector of the previous sequence is ‘‘010,’’ the proper zero vector will be ‘‘000’’ instead of ‘‘111,’’ and the vector applied ﬁrstly should be ‘‘000’’ rather than ‘‘100,’’ as it has only two commutations rather than three. By realizing the ﬁxed switching frequency, the switching frequency can be reduced to a certain degree. (26) According to (26), dn−opt can be greater than 1 if β > 90 ◦, especially at a dynamic instant with large instant power error, as illustrated in Fig. 2(b). In a real control system, 0 ≤dn−opt ≤1 should be forced. 77053 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers TABLE 2. Electrical parameter of the power circuit. TABLE 2. Electrical parameter of the power circuit. V. NUMERICAL SIMULATION The performance of proposed SPDDC has been simulated in MATLAB/Simulink, and compared with the performances of the conventional single-vector-based MPDPC and the dual- vector-based MPDCC. The one-step-delay compensation is applied to the model-predictive-based control methods during the procedure of the best vector selection by using equations similar to (29) and (30). The power ﬂow from the AC side to the DC load is deﬁned as positive. Table 2 lists the power circuit parameters, and the sampling frequency is 20 kHz. A. COMPARISON OF STEADY STATE PERFORMANCE For each of the three control methods, the current THD, P and Q ripples at P = 200 W and Q = 400 VAR have been recorded at the sampling frequency of 10 kHz except where specialized. The MPDPC with a sampling fre- quency of 20 kHz and three-vector-based predictive duty- cycle-control (TPDCC) with an additional vector sequence in [25] are also implemented for comparison. Among vari- ous three-vector predictive control methods, TPDCC is one of the most popular and known methods. For simplicity, the MPDPC control with 10 kHz and 20 kHz sampling fre- quency are indicated as MPDPC-10kHz and MPDPC-20kHz, and the SPDDC method with λ = 1 and λ = 1.5 as SPDDC-1 and SPDDC-1.5, respectively. One-step-delay compensation is applied to each of the control methods. Among various predictive control methods, these above men- tioned are the most popular and known methods. While for comparisons, due to the page limit, the comparisons of eight different control methods including the single-vector- based, dual-vector-based, and three-vector-based methods are presented in [37]. FIGURE 4. Steady state performance at P = 200 W, Q = 400 VAR. Top to bottom: ic, P, Q, and harmonic spectrum of ic. (a). MPDPC. (b). MPDCC. (c). SPDDC-1 (λ = 1). (d). SPDDC-1.5 (λ = 1.5). the Q reference drops to −300 VAR. Then, at 0.04 s, the active power reference steps up to 800 W, and at 0.06 s, the reactive power reference drops to −400 VAR. Finally, at 0.11 s, the P reference is reduced to 200 W. As shown, the dynamic response of every control method can track the reference values of P and Q with high accuracy and stability. Under the control of either MPDPC or MPDCC method, an overshoot of P appears at the instant of Q step change. Under the control of proposed SPDDC, when λ = 1, the dynamic response to a step change of P reference at 0.02 s is much slower than those of the MPDPC and MPDCC. However, when λ = 1.5, as shown in Fig. 5(d), the overshoot almost disappeared, indicating a good dynamic performance with rapid dynamic response and accurate reference tracking ability. This is also shown in the quantitative comparison of instantaneous states presented in Table 3. Fig. B. DYNAMIC PERFORMANCE COMPARISON The dynamic performances of these control methods under the same step power change references are compared. As illustrated in Fig. 5, the active and reactive power refer- ences are set as 400 W and 0 VAR, respectively, from 0 s. At 0.02 s, the P reference drops to 100 W, and at 0.03 s, 77054 77054 VOLUME 8, 2020 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers FIGURE 4. Steady state performance at P = 200 W, Q = 400 VAR. Top to bottom: ic, P, Q, and harmonic spectrum of ic. (a). MPDPC. (b). MPDCC. (c). SPDDC-1 (λ = 1). (d). SPDDC-1.5 (λ = 1.5). TABLE 3. Quantitative comparison of simulation results. TABLE 3. Quantitative comparison of simulation results. TABLE 4. Electrical parameter of prototype. TABLE 3. Quantitative comparison of simulation results. TABLE 4. Electrical parameter of prototype. TABLE 4. Electrical parameter of prototype. The parameters are listed in Table 4. The power circuit was controlled by a TMS320F28335 ﬂoating-point digital signal processor (DSP) based on Texas Instruments (TI) C2000 tar- get board for A/D sampling, PWM signal generation and D/A output. The variables such as the reference value changes are controlled by the real-time data exchange communication between the DSP and PC. A. COMPARISON OF STEADY STATE PERFORMANCE 7(a) shows the PWM driving signal of the upper switch of phase A, input phase to phase voltage Vab, and input current ia and ib of the MPDPC method with 10 kHz sampling frequency. The PWM signals are input signals, while Vab, ia and ib shown on the oscilloscope are output signals. Also, with the experimental data acquired from the oscilloscope to PC, the instantaneous P and Q, harmonic spectrum analyses of ia are presented. The current THD is 4.95% as shown by the harmonic spectrum of ia. The active and reactive power tracks the references very well. The P and Q ripples are 19.55 W and 22.88 VAR, respectively. In conclusion, the proposed SPDDC method is an effective and intuitive method with better steady state and dynamic performance than those of the MPDPC method, and it can be simply transformed from the MPDPC method. VI. EXPERIMENTAL TESTS To verify the effectiveness of the proposed SPDDC method, a scaled-down prototype is built as shown in Fig. 6. 77055 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers URE 5. Dynamic-state performance with bi-directional power flow to bottom: Va and ia, three-phase current, P and Q. (a). MPDPC. MPDCC. (c). SPDDC-1 (λ = 1). (d). SPDDC-1.5 (λ = 1.5). comparison, the performance is enhanced a lot w PDPC-20kHz as shown in Fig. 7(b), since the s FIGURE 6. Experimental setup of three-phase AC/DC converter, (1) Three-phase isolated transformer, (2) Main circuit of the converter, (3) Inductors, (4) TI C2000 target and interface board (5) DC resistive load, and (6) PC. FIGURE 6. Experimental setup of three-phase AC/DC converter, (1) Three-phase isolated transformer, (2) Main circuit of the converter, (3) Inductors, (4) TI C2000 target and interface board (5) DC resistive load, and (6) PC. FIGURE 6. Experimental setup of three-phase AC/DC converter, (1) Three-phase isolated transformer, (2) Main circuit of the converter, (3) Inductors, (4) TI C2000 target and interface board (5) DC resistive load, and (6) PC. TABLE 5. Quantitative comparison of steady states. TABLE 5. Quantitative comparison of steady states. TABLE 5. Quantitative comparison of steady states. burden on hardware and increased switching frequency. With MPDPC-20kHz, the current THD is reduced to 3.54%, the active and reactive power ripples are also decreased obviously as presented in Table 5. As shown in Fig. 7(c), the steady state performance of TPDCC with 10 kHz sampling frequency is close to that of MPDPC-20kHz, showing that the multi-vector-based control method can perform much better than the single-vector-based control method. The steady state performance of the MPDCC in Fig. 7(d) shows a slight deterioration in comparison to that of TPDCC due to the application of dual-vectors, but it is still better than that of the single-vector-based MPDPC-10kHz method. The performances of the proposed SPDDC-1 and SPDDC-1.5 methods are presented in Fig. 7(e) and Fig. 7(f), respectively. As shown, they both can achieve slightly lower current THD and Q ripples compared with the MPDCC, though the P ripple is increased. It is conﬁrmed that the proposed method is supe- rior to the MPDPC method and has comparable performance to the MPDCC method. As shown in Table 5, the experimen- tal results agree well with the numerical simulation results. VI. EXPERIMENTAL TESTS It should be noted that the deviations between the simulated and experimental results are mainly caused by the difference of sampling frequency. FIGURE 5. Dynamic-state performance with bi-directional power flow. Top to bottom: Va and ia, three-phase current, P and Q. (a). MPDPC. (b). MPDCC. (c). SPDDC-1 (λ = 1). (d). SPDDC-1.5 (λ = 1.5). The harmonic analysis shows that under the control of the proposed SPDDC, a signiﬁcant proportion of the low-order harmonics is reduced in contrast to the MPDPC, which has a high proportion of low-order harmonics and a wide harmonic spectrum. Therefore, using the SPDDC method can reduce the complexity of ﬁlter design. In comparison, the performance is enhanced a lot with MPDPC-20kHz as shown in Fig. 7(b), since the sam- pling frequency is increased with a heavier computational 77056 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers IGURE 7. Steady state performance at P = 200 W, Q = 400 VAR. Top to ottom: Oscilloscope screenshot, P, Q, and harmonic spectrum of ia. a) MPDPC-10kHz. (b) MPDPC-20kHz. (c) TPDCC. (d) MPDCC. (e) SPDDC- ) SPDDC-1.5. FIGURE 7. Steady state performance at P = 200 W, Q = 400 VAR. Top to bottom: Oscilloscope screenshot, P, Q, and harmonic spectrum of ia. (a) MPDPC-10kHz. (b) MPDPC-20kHz. (c) TPDCC. (d) MPDCC. (e) SPDDC-1. (f) SPDDC-1 5 FIGURE 8. Dynamic-state performance when P varying from 200 W to 400 W, Q = 200 VAR. Top: Oscilloscope screenshot. Bottom: ia, P and Q. (a) MPDPC-10kHz, (b) MPDPC-20kHz, (c) TPDCC, (d) MPDCC, (e) SPDDC-1, and (f) SPDDC-1.5. reference remains at 400 VAR. As shown, all control methods can follow the references accurately with similar response times. Fig. 8(f) shows that the SPDDC-1’s dynamic response of reactive power is a bit slower than that of the SPDDC-1.5 method when the P reference has a step change. This agrees with the numerical simulation results. FIGURE 8. Dynamic-state performance when P varying from 200 W to 400 W, Q = 200 VAR. Top: Oscilloscope screenshot. Bottom: ia, P and Q. (a) MPDPC-10kHz, (b) MPDPC-20kHz, (c) TPDCC, (d) MPDCC, (e) SPDDC-1, FIGURE 8. Dynamic-state performance when P varying from 200 W to 400 W, Q = 200 VAR. Top: Oscilloscope screenshot. Bottom: ia, P and Q. (a) MPDPC-10kHz, (b) MPDPC-20kHz, (c) TPDCC, (d) MPDCC, (e) SPDDC-1, FIGURE 8. VI. EXPERIMENTAL TESTS Dynamic-state performance when P varying from 200 W to 400 W, Q = 200 VAR. Top: Oscilloscope screenshot. Bottom: ia, P and Q. (a) MPDPC-10kHz, (b) MPDPC-20kHz, (c) TPDCC, (d) MPDCC, (e) SPDDC-1, and (f) SPDDC-1.5. reference remains at 400 VAR. As shown, all control methods can follow the references accurately with similar response times. Fig. 8(f) shows that the SPDDC-1’s dynamic response of reactive power is a bit slower than that of the SPDDC-1.5 method when the P reference has a step change. This agrees with the numerical simulation results. FIGURE 7. Steady state performance at P = 200 W, Q = 400 VAR. Top to bottom: Oscilloscope screenshot, P, Q, and harmonic spectrum of ia. (a) MPDPC-10kHz. (b) MPDPC-20kHz. (c) TPDCC. (d) MPDCC. (e) SPDDC-1. (f) SPDDC-1.5. C. SYSTEM PARAMETER REDUNDANCY 9(a), when the inductance set in SPDDC is 50% of the actual value (20 mH), the Q ripple increases and a negative DC offset of P appears. As shown in Figs. 9(c) and (d), the increase of inductance value used in the control algorithm to 30 mH and 40 mH has almost no inﬂuence on the Q ripple and positive DC offset of P. As conﬁrmed by the experiments, under the control of proposed SPDDC, the precision of induc- tance, in a range of 50% to 200% of the actual value, has little inﬂuence on the steady state performance and control stability, which is superior to the MPDPC method [27]. [9] J. Alonso-Martínez, J. E. Carrasco, and S. Arnaltes, ‘‘Table-based direct power control: A critical review for microgrid applications,’’ IEEE Trans. Power Electron., vol. 25, no. 12, pp. 2949–2961, Dec. 2010. [10] W. Xiong, Y. Sun, J. Lin, M. Su, H. Dan, M. Rivera, and J. M. Guerrero, ‘‘A cost-effective and low-complexity predictive control for matrix con- verters under unbalanced grid voltage conditions,’’ IEEE Access, vol. 7, pp. 43895–43905, 2019. [11] A. Bouaﬁa, F. Krim, and J.-P. Gaubert, ‘‘Fuzzy-Logic-Based switching state selection for direct power control of three-phase PWM rectiﬁer,’’ IEEE Trans. Ind. Electron., vol. 56, no. 6, pp. 1984–1992, Jun. 2009. [12] A. Bouaﬁa, J.-P. Gaubert, and F. Krim, ‘‘Predictive direct power con- trol of three-phase pulsewidth modulation (PWM) rectiﬁer using space- vector modulation (SVM),’’ IEEE Trans. Power Electron., vol. 25, no. 1, pp. 228–236, Jan. 2010. [13] S. Vazquez, J. Rodriguez, M. Rivera, L. G. Franquelo, and M. Norambuena, ‘‘Model predictive control for power converters and drives: Advances and trends,’’ IEEE Trans. Ind. Electron., vol. 64, no. 2, pp. 935–947, Feb. 2017. VII. CONCLUSION A simpliﬁed dual-vector-based SPDDC strategy for three-phase AC/DC converters is proposed in this paper to eliminate the time consuming procedure of duty-cycle optimization. In SPDCC, the cost function is used for both best vector selection and duration derivation. Compared with MPDPC, since dual vectors are implemented in each control session, SPDDC can achieve much better performance with a ﬁxed switching frequency. Compared with MPDCC, SPDDC is more intuitive. The deduced value of duration is guaranteed to be within a deﬁnite range, and thus the issue of negative duration is essentially resolved. The procedure of duty-cycle optimization can be eliminated by the simpliﬁed calculation. C. SYSTEM PARAMETER REDUNDANCY Fig. 8 compares the experimental dynamic performance when the P reference increases from 200 W to 400 W, while the Q The robustness of the proposed SPDDC is examined through a series of experiments by setting the line inductance in 77057 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers FIGURE 9. Responses of active and reactive power for proposed SPDDC when the actual inductance in control differs from the real value. (a) 10 mH, (b) 20 mH, (c) 30 mH, and (d) 40 mH. FIGURE 9. Responses of active and reactive power for proposed SPDDC when the actual inductance in control differs from the real value. [2] T. Geyer and D. E. Quevedo, ‘‘Multistep ﬁnite control set model predictive control for power electronics,’’ IEEE Trans. Power Electron., vol. 29, no. 12, pp. 6836–6846, Dec. 2014. [3] S. Chakraborty, B. Kramer, and B. Kroposki, ‘‘A review of power elec- tronics interfaces for distributed energy systems towards achieving low- cost modular design,’’ Renew. Sustain. Energy Rev., vol. 13, no. 9, pp. 2323–2335, Dec. 2009. [4] N. Flourentzou, V. G. Agelidis, and G. D. Demetriades, ‘‘VSC-based HVDC power transmission systems: An overview,’’ IEEE Trans. Power Electron., vol. 24, no. 3, pp. 592–602, Mar. 2009. [5] A. Rahoui, A. Bechouche, H. Seddiki, and D. O. Abdeslam, ‘‘Grid voltages estimation for three-phase PWM rectiﬁers control without AC voltage sen- sors,’’ IEEE Trans. Power Electron., vol. 33, no. 1, pp. 859–875, Jan. 2018. [6] Y. Zhang, J. Gao, and C. Qu, ‘‘Relationship between two direct power control methods for PWM rectiﬁers under unbalanced network,’’ IEEE Trans. Power Electron., vol. 32, no. 5, pp. 4084–4094, May 2017. [7] Y. Zhang and C. Qu, ‘‘Table-based direct power control for three-phase AC/DC converters under unbalanced grid voltages,’’ IEEE Trans. Power Electron., vol. 30, no. 12, pp. 7090–7099, Dec. 2015. FIGURE 9. Responses of active and reactive power for proposed SPDDC when the actual inductance in control differs from the real value. (a) 10 mH, (b) 20 mH, (c) 30 mH, and (d) 40 mH. [8] J. Hu, J. Zhu, and D. G. Dorrell, ‘‘In-depth study of direct power con- trol strategies for power converters,’’ IET Power Electron., vol. 7, no. 7, pp. 1810–1820, Jul. 2014. the algorithm different from the actual value. As shown in Fig. C. SYSTEM PARAMETER REDUNDANCY The vector sequence optimization and one-step-delay com- pensation are considered in the proposed method. [14] Y. Cho and K.-B. Lee, ‘‘Virtual-Flux-Based predictive direct power control of three-phase PWM rectiﬁers with fast dynamic response,’’ IEEE Trans. Power Electron., vol. 31, no. 4, pp. 3348–3359, Apr. 2016. [15] J. Liu, S. Vazquez, L. Wu, A. Marquez, H. Gao, and L. G. Franquelo, ‘‘Extended state observer-based sliding-mode control for three-phase power converters,’’ IEEE Trans. Ind. Electron., vol. 64, no. 1, pp. 22–31, Jan. 2017. [16] S. Kouro, P. Cortes, R. Vargas, U. Ammann, and J. Rodriguez, ‘‘Model pre- dictive control—A simple and powerful method to control power convert- ers,’’ IEEE Trans. Ind. Electron., vol. 56, no. 6, pp. 1826–1838, Jun. 2009. [17] S. Vazquez, J. I. Leon, L. G. Franquelo, J. Rodriguez, H. A. Young, A. Marquez, and P. Zanchetta, ‘‘Model predictive control: A review of its applications in power electronics,’’ IEEE Ind. Electron. Mag., vol. 8, no. 1, pp. 16–31, Mar. 2014. [18] B. Stellato, T. Geyer, and P. J. Goulart, ‘‘High-speed ﬁnite control set model predictive control for power electronics,’’ IEEE Trans. Power Electron., vol. 32, no. 5, pp. 4007–4020, May 2017. [19] Y. Zhang and W. Xie, ‘‘Low complexity model predictive Control—Single vector-based approach,’’ IEEE Trans. Power Electron., vol. 29, no. 10, pp. 5532–5541, Oct. 2014. The proposed SPDDC is compared comprehensively with various kinds of typical multi-vector-based control meth- ods through numerical simulations and experimental tests. The correctness and effectiveness of the proposed SPDDC are validated by the results through superior steady state and dynamic performances with ﬁxed switching frequencies, lower current THD, and lower active and reactive power ripples. [20] C. Xia, T. Liu, T. Shi, and Z. Song, ‘‘A simpliﬁed Finite-Control-Set model- predictive control for power converters,’’ IEEE Trans. Ind. Informat., vol. 10, no. 2, pp. 991–1002, May 2014. [21] H. A. Young, M. A. Perez, and J. Rodriguez, ‘‘Analysis of Finite-Control- Set model predictive current control with model parameter mismatch in a three-phase inverter,’’ IEEE Trans. Ind. Electron., vol. 63, no. 5, pp. 3100–3107, May 2016. [22] H. T. Nguyen, J. Kim, and J.-W. Jung, ‘‘Improved model predictive control by robust prediction and stability-constrained ﬁnite states for three-phase inverters with an output LC ﬁlter,’’ IEEE Access, vol. 7, pp. 12673–12685, 2019. [23] Y. Zhang, Y. Bai, and H. Yang, ‘‘A universal Multiple-Vector-Based model predictive control of induction motor drives,’’ IEEE Trans. Power Elec- tron., vol. [1] J. R. Rodriguez, J. W. Dixon, J. R. Espinoza, J. Pontt, and P. Lezana, ‘‘PWM regenerative rectiﬁers: State of the art,’’ IEEE Trans. Ind. Electron., vol. 52, no. 1, pp. 5–22, Feb. 2005. C. SYSTEM PARAMETER REDUNDANCY Zou, and Z. Chen, ‘‘Predictive duty cycle control of three-phase Active-Front-End rectiﬁers,’’ IEEE Trans. Power Electron., vol. 31, no. 1, pp. 698–710, Jan. 2016. [30] Y. Zhang, W. Xie, Z. Li, and Y. Zhang, ‘‘Model predictive direct power control of a PWM rectiﬁer with duty cycle optimization,’’ IEEE Trans. Power Electron., vol. 28, no. 11, pp. 5343–5351, Nov. 2013. [31] Y. Zhang, Y. Peng, and H. Yang, ‘‘Performance improvement of Two- Vectors-Based model predictive control of PWM rectiﬁer,’’ IEEE Trans. Power Electron., vol. 31, no. 8, pp. 6016–6030, Aug. 2016. From 2005 to 2007, he was a Research Asso- ciate with the University of New South Wales, Australian Defence Force Academy (UNSW ADFA), Canberra, ACT, Australia. From 2007 to 2011, he was a Researcher with National ICT Australia, Victoria Research Laboratory, Department of Electrical and Electronics Engineering, The University of Melbourne, Melbourne, VIC, Australia. He held several visiting positions at the Beijing Institute of Technology, Beijing, Tsinghua University, and UNSW ADFA. He joined the University of Technology Sydney, Ultimo, NSW, Australia, in 2011, where he is currently a Senior Lecturer. His current research interests are control theory and power system control. [32] L. Tarisciotti, J. Lei, A. Formentini, A. Trentin, P. Zanchetta, P. Wheeler, and M. Rivera, ‘‘Modulated predictive control for indirect matrix con- verter,’’ IEEE Trans. Ind. Appl., vol. 53, no. 5, pp. 4644–4654, Sep. 2017. [33] X. Shi, J. Zhu, L. Li, D. D.-C. Lu, J. Zhang, and H. Yang, ‘‘Predictive duty cycle control with reversible vector selection for three-phase AC/DC converters,’’ IEEE Trans. Power Electron., vol. 34, no. 5, pp. 4868–4882, May 2019. [34] Y. Zhang, J. Liu, H. Yang, and S. Fan, ‘‘New insights into model predictive control for three-phase power converters,’’ IEEE Trans. Ind. Appl., vol. 55, no. 2, pp. 1973–1982, Mar. 2019. [35] S.-Y. Park and S. Kwak, ‘‘Comparative study of three model predictive current control methods with two vectors for three-phase DC/AC VSIs,’’ IET Electric Power Appl., vol. 11, no. 7, pp. 1284–1297, Aug. 2017. [36] T. Jin, J. Guo, M. A. Mohamed, and M. Wang, ‘‘A novel model predictive control via optimized vector selection method for common-mode voltage reduction of three-phase inverters,’’ IEEE Access, vol. 7, pp. 95351–95363, 2019. DYLAN DAH-CHUAN LU (Senior Member, IEEE) received the Ph.D. degree in electronics and information engineering from The Hong Kong Polytechnic University, Hong Kong, in 2004. C. SYSTEM PARAMETER REDUNDANCY 33, no. 8, pp. 6957–6969, Aug. 2018. 77058 77058 VOLUME 8, 2020 X. Shi et al.: Low-Complexity Dual-Vector-Based Predictive Control of Three-Phase PWM Rectifiers [24] J. Hu and Z. Q. Zhu, ‘‘Investigation on switching patterns of direct power control strategies for grid-connected DC–AC converters based on power variation rates,’’ IEEE Trans. Power Electron., vol. 26, no. 12, pp. 3582–3598, Dec. 2011. JIANGUO ZHU (Senior Member, IEEE) received the B.E. degree from the Jiangsu Institute of Tech- nology, China, in 1982, the M.E. degree from the Shanghai University of Technology, Shanghai, China, in 1987, and the Ph.D. degree from the University of Technology Sydney (UTS), Sydney, Australia, in 1995. [25] J. Hu and Z. Q. Zhu, ‘‘Improved voltage-vector sequences on dead-beat predictive direct power control of reversible three-phase grid-connected voltage-source converters,’’ IEEE Trans. Power Electron., vol. 28, no. 1, pp. 254–267, Jan. 2013. He is currently a Professor of electrical engi- neering and the Head of the School of Electrical and Information Engineering, The University of Sydney. His research interests include electromagnetics, magnetic properties of materials, electrical machines and drives, power electronics, and renew- able energy systems. He is currently a Professor of electrical engi- neering and the Head of the School of Electrical and Information Engineering, The University of Sydney. His research interests include electromagnetics, magnetic properties of materials, electrical machines and drives, power electronics, and renew- able energy systems. [26] S. Vazquez, A. Marquez, R. Aguilera, D. Quevedo, J. I. Leon, and L. G. Franquelo, ‘‘Predictive optimal switching sequence direct power control for grid-connected power converters,’’ IEEE Trans. Ind. Electron., vol. 62, no. 4, pp. 2010–2020, Apr. 2015. vol. 62, no. 4, pp. 2010–2020, Apr. 2015. [27] Z. Song, W. Chen, and C. Xia, ‘‘Predictive direct power control for three-phase grid-connected converters without sector information and volt- age vector selection,’’ IEEE Trans. Power Electron., vol. 29, no. 10, pp. 5518–5531, Oct. 2014. [28] Y. Zhang, W. Xie, Z. Li, and Y. Zhang, ‘‘Low-complexity model predictive power control: Double-Vector-Based approach,’’ IEEE Trans. Ind. Elec- tron., vol. 61, no. 11, pp. 5871–5880, Nov. 2014. LI LI (Member, IEEE) received the B.S. degree from the Huazhong University of Science and Technology, Wuhan, China, in 1996, the M.S. degree from Tsinghua University, Beijing, China, in 1999, and the Ph.D. degree from the University of California at Los Angeles, Los Angeles, CA, USA, in 2005. [29] Z. Song, Y. Tian, W. Chen, Z. C. SYSTEM PARAMETER REDUNDANCY In 2003, he joined PowerELab Ltd., as a Senior Design Engineer, where he was responsible for industrial switching power supply projects. He was a full-time Faculty Member of The University of Sydney, from 2006 to 2016, where he currently holds an honorary position. Since July 2016, he has been an Associate Professor with the School of Electrical and Data Engi- neering, University of Technology Sydney, Australia. He has been heading the discipline of electrical power and energy systems, since December 2018. He has authored or coauthored more than 100 international journals and held two patents in power electronics. His current research interests include efﬁcient and reliable power conversion for renewable sources, energy storage systems, and microgrids. He is currently serving as a Chair of the IEEE NSW Joint Chapter IAS/IES/PELS and an Associate Editor for the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS. [37] X. Shi, ‘‘Advanced control of three-phase full-bridge converter in micro- grids,’’ Ph.D. dissertation, Dept. Elect. Data Eng., Univ. Technol. Sydney, Ultimo, Australia, 2017. XIAOLONG SHI (Student Member, IEEE) received the B.E. and M.E. degrees in electri- cal engineering from Beijing Jiaotong Univer- sity, China, in 2011 and 2014, respectively, and the Ph.D. degree from the University of Tech- nology Sydney (UTS), Sydney, NSW, Australia, in 2017. Since 2013, he has been with the School of Data and Electrical Engineering, UTS. His research interests include design, optimization, advanced control algorithms of power converters and renewable energy systems, and advanced digital control with real-time implementation. 77059 VOLUME 8, 2020
https://openalex.org/W2394578779	https://www.nature.com/articles/am201663.pdf	English	null	Gadolinium polytungstate nanoclusters: a new theranostic with ultrasmall size and versatile properties for dual-modal MR/CT imaging and photothermal therapy/radiotherapy of cancer	NPG Asia materials	2,016	cc-by	10,452	1CAS Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety, Institute of High Energy Physics and National Center for Nanoscience and Technology of China, Chinese Academy of Sciences, Beijing, People’s Republic of China and 2Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing, People’s Republic of China Correspondence: Professor Z Gu or Professor Y Zhao, CAS Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety, Institute of High Energy Physics and National Center for Nanoscience and Technology of China, Chinese Academy of Sciences, Yuquan Road 19b Shijingshan District, 100049 Beijing, People’s Republic of China. E-mail: zjgu@ihep.ac.cn or zhaoyuliang@ihep.ac.cn or Professor G Zhang, Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Zhongguancun South Road No. 12 Haidian District, 100190 Beijing, People’s Republic of China. E-mail: zhanggj@ipe.ac.cn 3These authors contributed equally to this work. Received 16 November 2015; revised 28 January 2016; accepted 15 March 2016 OPEN OPEN NPG Asia Materials (2016) 8, e273; doi:10.1038/am.2016.63 www.nature.com/am NPG Asia Materials (2016) 8, e273; doi:10.1038/am.2016.63 ORIGINAL ARTICLE Gadolinium polytungstate nanoclusters: a new theranostic with ultrasmall size and versatile properties for dual-modal MR/CT imaging and photothermal therapy/radiotherapy of cancer Yuan Yong1,3, Liangjun Zhou1,3, Shuangshuang Zhang2, Liang Yan1, Zhanjun Gu1, Guangjin Zhang2 and Yuliang Zhao1 ngjun Zhou1,3, Shuangshuang Zhang2, Liang Yan1, Zhanjun Gu1, Guangjin Zhang2 and The development of a new generation of nanoscaled theranostics with simple compositions and versatile properties to realize enhanced diagnoses and treatment outcomes and to avoid side effects is highly desirable but remains a great challenge. Here we report a new ultrasmall theranostic based on bovine serum albumin-coated GdW10O36 nanoclusters (GdW10 NCs) as multifunctional theranostics for multifunctional bio-imaging, highly effective in vitro photothermal ablation of cancer cells and in vivo radiotherapy of tumors. The as-synthesized GdW10 NCs afford signiﬁcantly enhanced computed tomography/magnetic resonance imaging signals and remarkable photothermal therapy (PTT)/radiotherapy therapy (RT) therapeutic effects for tumor treatment. Moreover, toxicity studies conﬁrmed their low toxicity and efﬁcient renal clearance, suggesting their potential for practical applications. These results indicated that the as-prepared GdW10 nanoclusters are promising as multifunctional nanotheranostics for multimodal imaging-guided PTT/RT of tumors. These results also encourage the further exploration of other polyoxometalate-based multifunctional nanotheranostics for cancer diagnoses and therapy. NPG Asia Materials (2016) 8, e273; doi:10.1038/am.2016.63; published online 27 May 2016 INTRODUCTION and Cu, to impart various properties.36–39 These nanoclusters can then be used in biomedical,40–49 analytical,50–54 magnetic,55–60 optical61,62 and catalytic applications.63–67 For example, paramagnetic rare earth elements such as Gd3+ can be easily incorporated into polytungstate (a type of POM) structures to form new molecules.68–70 These Gd-containing polytungstates have been shown to be efﬁcient magnetic resonance (MR) imaging contrast agents and exhibited better performances than those of commercial MR imaging contrast agents due to their large molecular weights and rigid structures.71–76 In addition, due to the strong X-ray attenuation ability of Gd and W atoms in gadolinium polytungstates, these nanoclusters could also be used as efﬁcient contrast agents for enhanced signaling in computed tomography (CT) imaging. Therefore, gadolinium polytungstates can act as dual-functional imaging probes in MR and CT imaging. The combination of MR with CT imaging could offer whole-body imaging capabilities at high spatial resolutions as well as the real-time visualization of nanoparticles in tumor microstructures. In addition to its dual-modal imaging capabilities, gadolinium polytungstates have great potential as radiosensitizers and photothermal absorbing agents for the radiotherapy therapy (RT) and photothermal therapy (PTT) of tumors, respectively. As high Z-materials, gadolinium polytungstates have been shown to accumulate in tumor sites via the enhanced permeability and retention (EPR) effect and are perceived to absorb much more X-ray irradiation than normal tissues can. Thus, gadolinium polytungstates can act as radiosensitizers to improve antitumor radiotherapy therapeutic efﬁciency. Moreover, after electrochemical or photochemical To this end, we synthesized bovine serum albumin (BSA)-coated GdW10O36 (GdW10) nanoclusters (NCs) and, for the ﬁrst time, applied them as multifunctional theranostics for multimodal bio-imaging. We showed that these nanoclusters were highly effective in the in vitro photothermal ablation of cancer cells and in vivo radiotherapy of tumors (Scheme 1). Utilizing their high r1 relaxivity and strong X-ray attenuation abilities, GdW10@BSA NCs can be successfully used as efﬁcient MR/CT imaging contrast agents to complement multimodal imaging. Moreover, the as-prepared NCs produced signiﬁcant heat upon 808 nm laser irradiation. Simultaneously, the NCs generated dose- enhanced sensitization effects in RT, which resulted in remarkable in vitro cell damage and in vivo tumor regression. Toxicity screening conﬁrmed that the GdW10@BSA NCs were biocompatible and had efﬁcient renal clearance. Our work highlights the potential of gadoli- nium polytungstates for cancer imaging and therapy and encourages further in-depth investigations of this novel type of POM for biomedical applications. INTRODUCTION limiting their practical applications. Moreover, most of the previously reported multifunctional nanoprobes have relatively large hydrodynamic diameters (typically 410 nm), which are larger than the kidney ﬁltration thresholds (that is, 5.5 nm).30,31 Consequently, these nanoparticles cannot be metabolized by the kidney and are instead absorbed by the reticuloendothelial system in the liver and spleen, causing potential organ and immune system damage. Therefore, there is a great demand to develop small-sized nanoscale theranostics that have simple compositions with versatile properties to realize enhanced diagnostic and treatment capabilities and to avoid side effects in vivo. New nanomaterials (NMs) for the diagnosis and treatment of cancer have recently received signiﬁcant attention and have become an important ﬁeld in biomedical research.1–10 One major advantage of NMs is their capability to integrate multiple imaging/treatment modalities into one particle, thereby developing a new generation of theranostics.11–16 The synergistic combination of bio-imaging and localized treatment capabilities is unique for nano-sized objects and may offer unprecedented opportunities to achieve better anti-cancer efﬁcacies at tumor sites while mitigating undesired side effects.17–23 Over the past decade, theranostics based on NMs have been extensively explored. Studies have mainly focused on the construction or optimization of hybrid nanocomposites that integrate multiple imaging approaches and therapeutic modalities to realize concurrent diagnostic and therapeutic capabilities.24–29 However, the integration of two or more different components into one nanoparticle is difﬁcult and typically requires complicated synthetic processes, thus greatly Polyoxometalates (POMs) are potential candidates for ultrasmall nanotheranostics due to their unique molecular structure, good water solubility and small particle size.32–35 As a class of important inorganic nanoclusters, POMs are composed of transition metals and/or rare earth elements bridged by oxo ligands. INTRODUCTION Many heteroatoms can be incorporated into a POM framework, including Gd, Mn Gadolinium polytungstate nanoclusters Y Yong et al 2 and Cu, to impart various properties.36–39 These nanoclusters can then be used in biomedical,40–49 analytical,50–54 magnetic,55–60 optical61,62 and catalytic applications.63–67 For example, paramagnetic rare earth elements such as Gd3+ can be easily incorporated into polytungstate (a type of POM) structures to form new molecules.68–70 These Gd-containing polytungstates have been shown to be efﬁcient magnetic resonance (MR) imaging contrast agents and exhibited better performances than those of commercial MR imaging contrast agents due to their large molecular weights and rigid structures.71–76 In addition, due to the strong X-ray attenuation ability of Gd and W atoms in gadolinium polytungstates, these nanoclusters could also be used as efﬁcient contrast agents for enhanced signaling in computed tomography (CT) imaging. Therefore, gadolinium polytungstates can act as dual-functional imaging probes in MR and CT imaging. The combination of MR with CT imaging could offer whole-body imaging capabilities at high spatial resolutions as well as the real-time visualization of nanoparticles in tumor microstructures. In addition to its dual-modal imaging capabilities, gadolinium polytungstates have great potential as radiosensitizers and photothermal absorbing agents for the radiotherapy therapy (RT) and photothermal therapy (PTT) of tumors, respectively. As high Z-materials, gadolinium polytungstates have been shown to accumulate in tumor sites via the enhanced permeability and retention (EPR) effect and are perceived to absorb much more X-ray irradiation than normal tissues can. Thus, gadolinium polytungstates can act as radiosensitizers to improve antitumor radiotherapy therapeutic efﬁciency. Moreover, after electrochemical or photochemical reduction, reduced gadolinium polytungstates with dark blue colors can efﬁciently absorb and convert NIR light into heat77 for use as PTT agents in the thermal ablation of cancer cells. Most importantly, the ultrasmall size of gadolinium polytungstates enables a rapid renal clearance after treatment and greatly reduces their potential side effects. These unique features could not be replicated by most other hybrid nanocomposite-based theranostic agents. These nanoclusters could be used in the simultaneous multimodal imaging and RT/PTT of cancerous tumors. However, no studies have addressed or considered these potential roles. Cytotoxicity and cellular of GdW10@BSA NCs 16HBE cells (bronchial epithelioid cells) and HeLa cells (human cervical cancer cells) were grown in 96-well culture plates with high-glucose DMEM supplemented with 10% FBS(GIBCO) at 37 °C under a humidiﬁed 5% CO2 atmosphere. Cell viabilities were evaluated using the Cell Counting Kit-8 (CCK-8, Dojindo Laboratories, Shanghai, China) assay. The evaluation was performed as follows. Cells were cultured in a 96-well plate at a density of MATERIALS AND METHODS Materials All reagents and solvents were of commercial grade and were used without any further puriﬁcation. Na2WO4·2H2O and GdCl6·6H2O were provided by Alfa Aesar Reagent Company (China, Shanghai). Acetic acid was obtained from Beijing Chemical Works (Beijing, China). BSA was adopted from Amresco Company (Shanghai, China). Fetal bovine serum (FBS) and Dulbecco’s modiﬁed Eagle’s medium (DMEM) were adopted from Gibco Company. Calcein-AM (CA), propidium iodide (PI) stain and Cell Counting Kit-8 (CCK-8) were all provided from Dojindo Laboratories. Puriﬁed water was used throughout. Synthesis of GdW10 nanoclusters and GdW10@BSA NCs The GdW10 NCs were prepared according to a previously described procedure.78 In the synthesis, 8.3 g (25 mmol) of Na2WO4·2H2O was ﬁrst dissolved in 20 ml of puriﬁed water to form a homogeneous solution. The homogeneous solution was adjusted to pH 7.4–7.5 with acetic acid under magnetic stirring at room temperature. Subsequently, 2 ml of GdCl6·6H2O (477.95 g mol−1, 2.5 mmol) aqueous solution was added dropwise to the aforementioned solution under continuous stirring and heated to 85 °C. After cooling to room temperature, crude crystals were isolated. The mixture was dialyzed against puriﬁed water for 1 day using a dialysis tube (2.0 kDa molecular cutoff) to remove impurities. To enhance the biocompatibility of the synthesized GdW10 nanoclusters, BSA was functionalized onto the surfaces of the GdW10 nanoclusters. Brieﬂy, 40 mg BSA was mixed with 2 ml GdW10 (2 mg ml−1) aqueous solution and then stirred for 6 h at room temperature. The resulting GdW10@BSA NCs were stored at room temperature for further experiments.79–81 GdW10-enhanced radiotherapy in vivo GdW10 enhanced radiotherapy in vivo All animals were purchased, maintained and killed the under protocols approved by the Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety (Institute of High Energy Physics, CAS). The BEL-7402 tumors were generated by the subcutaneous implantation of 1.0 × 106 BEL-7402 cells suspended in 100 μl phosphate-buffered saline (PBS) into the right rear legs of 16 BALB/c female mice. The mice were assigned to four groups based on treatment: (a) PBS injection; (b) GdW10@BSA NCs i.t. injection; (c) X-ray only; and (d) GdW10@BSA NCs+X-ray when the tumor volume reached ~ 75 mm3. For groups b and d, the BEL-7402 tumor-bearing BALB/c nude mice were intratumorally injected with 20 μl GdW10@BSA NCs (2 mg ml−1). As the control group, 20 μl of phosphate buffer saline was intratumorally injected into each mouse. The mice in groups c and d were irradiated by X-ray (6 Gy) to receive radiotherapy. After the treatments, the mice weights and tumor volumes were recorded every 2 or 3 days using a vernier caliper. At the end of the experiment, all BEL-7402 tumor-bearing BALB/c nude mice were killed under protocols approved by the Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety (Institute of High Energy Physics, CAS). The main organs and tumors of each mice were collected and stained with hematoxylin & eosin (H&E) for tissue lesion analyses. CT/MR imaging in vitro To test the linearity of the CT signal as a function of the GdW10@BSA NC concentration, various concentrations (0, 6.25, 12.5, 25, 50 and 100 mM) of the GdW10@BSA NC solution and iopromide (Ultravist300, Bayer, Shanghai, China) were separately dispersed in 0.5% agarose gel. Samples were then aliquoted into 1.5-ml centrifuge tubes for in vitro CT signal detection. CT images were acquired on the Gamma Medica-Ideas. A Triumph X-O CT system was employed to obtain the Hounsﬁeld unit (HU) values and CT images in vitro. The imaging parameters were as follows: effective pixel size, 50 μm; 80 kV, 270 μA; and ﬁeld of view, 1024 pixels × 1024 pixels. To detect the MRI signals, the GdW10@BSA NCs and Gd-DTPA at various concentra- tions (0, 0.05, 0.1, 0.2, 0.5, 1 mM) were dispersed in 0.5 % agarose gel. Samples were then aliquoted into 1.5-ml NMR tubes for MRI signal detection. A 4.7-T MR imaging instrument (Biospec; Bruker; Ettlingen; Germany) was adopted to obtain the T1 values and T1-weight images of GdW10@BSA NCs and Gd-DTPA. The parameters used were as follows: matrix size, 128 × 128; ﬁeld of view, 40 × 40 mm; and slice thickness, 1.20 mm. Characterizations of GdW10 nanoclusters and GdW10-BSA Characterizations of GdW10 nanoclusters and GdW10-BSA The size distributions of the GdW10 NCs and the GdW10@BSA NCs were measured on a Nicomp380 ZLS plus ZETADi. The X-ray photoelectron spectra of the GdW10 NCs were obtained on a Kratos AXIS-165 surface analysis system. An energy dispersive spectrometer (FESEM, Hitachi S-4800) was used to obtain the energy-dispersive X-ray spectra of the GdW10 NCs. The Fourier transform infrared spectra of the GdW10 NCs were measured on a micro-Fourier transform infrared spectrophotometer (iN10-IZ10, Thermo Fisher, Shanghai, China). A Nikon D3100 digital camera was used to capture images. Cell viability was recorded using a microplate reader (SpectraMax M2 MDC, Molecular Devices, Sunnyvale, CA, USA) at an optical absorbance of 450 nm. Induced coupled plasma mass spectrometry (Thermo Elemental X7, Thermal Scientiﬁc, Waltham, MA, USA) was employed to determine the concentration of Gd3+ in each organ. NPG Asia Materials MATERIALS AND METHODS intratumorally injected (i.t.) or intravenously injected (i.v.) with the GdW10@BSA NCs. The MDA-MB-231 tumor-bearing BALB/c nude mice were then imaged on the animal X-ray CT instrument (Gamma Medica-Ideas) after the intravenous injection of the GdW10@BSA NCs (75 μmol kg−1 ). In vivo images were carried out using a Triumph X-O CT system. The imaging parameters of the in vivo CT were as follows: effective pixel size, 50 μm; 80 kV, 270 μA; and ﬁeld of view, 1024 pixels× 1024 pixels. For the in vivo MR imaging, the MDA-MB-231 tumor-bearing BALB/c nude mice were (i.v.) injected with the GdW10@BSA NCs and commercially available Gd-DTPA, respectively. After the intravenous injection of the GdW10@BSA NCs (75 μmol kg−1) and Gd-DTPA, the tumor-bearing rat was imaged on a Bruker MR imaging instrument (Bruker; Biospec 70/20; Germany). The MR images were reformed with a microgel system (America). After all experiments were concluded, the animals were euthanized according to the guidelines formulated by the Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety (Institute of High Energy Physics, CAS). In vitro photothermal therapy After electrochemical reduction, the resulting dark blue GdW10@BSA NCs efﬁciently absorbed and converted NIR light to heat. To investigate the photothermal effect of reduced GdW10@BSA NCs, 1 mg ml1 of reduced GdW10@BSA NCs was irradiated with an 808-nm laser at a power density of 1 W cm−2. The NaCl solution was used as a negative control group. An IR thermal camera was used to detect temperature changes in the reduced GdW10@BSA NCs and the NaCl solution at each time point. In addition, to test the photothermal effect of the reduced GdW10@BSA NCs on cancer cells, HeLa cells were incubated in a six-well culture plate at a density of 1 × 106 per well. Then, different concentrations of the reduced GdW10@BSA NCs (0, 62.5, 125, 250, 500 and 1000 μg ml−1) were added to the HeLa cells. After 4 h, the culture medium was replaced with fresh DMEM. The cells incubated with the reduced GdW10@BSA NCs were irradiated by an 808-nm laser lamp at a power density of 1 W cm−2 for 10 min. After incubating for an additional 12 h, the culture medium was washed twice with PBS. Subsequently, the calcein-AM (CA)/propidium iodide (PI) stain applied for 15 min. Finally, the cell samples were washed twice with PBS, and an inverted ﬂuorescence microscope (Olympus X73, Tokyo, Japan) was employed to obtain ﬂuorescent images. INTRODUCTION Scheme 1 Schematic illustration of GdW10@BSA NCs for dual-modal MR/CT imaging-guided photothermal therapy/radiotherapy of cancer. (a) The synthesis process of GdW10@BSA. (b) The application of the as-made GdW10@BSA for bio-imaging and treatment of tumor. Scheme 1 Schematic illustration of GdW10@BSA NCs for dual-modal MR/CT imaging-guided photothermal therapy/radiotherapy of cancer. (a) The synthesis process of GdW10@BSA. (b) The application of the as-made GdW10@BSA for bio-imaging and treatment of tumor. Scheme 1 Schematic illustration of GdW10@BSA NCs for dual-modal MR/CT imaging-guided photothermal therapy/radiotherapy of cancer. (a) The synthesis process of GdW10@BSA. (b) The application of the as-made GdW10@BSA for bio-imaging and treatment of tumor. Scheme 1 Schematic illustration of GdW10@BSA NCs for dual-modal MR/CT imaging-guided photothermal therapy/rad process of GdW10@BSA. (b) The application of the as-made GdW10@BSA for bio-imaging and treatment of tumor. NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 3 CT/MR imaging in vivo In vivo CT images were accomplished on the Gamma Medica-Ideas (North- ridge, CA, USA). MDA-MB-231 tumor-bearing BALB/c nude mice were NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al Gadolinium polytungstate nanoclusters Y Yong et al 2 × 104 cells per well. The cells were washed twice with PBS (pH = 7.4) and incubated with different concentrations of GdW10@BSA NCs for 24 h. Subsequently, 10 μl of fresh CCK-8 was added into the culture medium, and the cells were incubated for an additional 1 h (37 °C under 5% CO2). Finally, the absorbance was measured at 450 nm using a microplate reader (Thermo Scientiﬁc, Multiscan MNK3, Waltham, MA, USA). 2 × 104 cells per well. The cells were washed twice with PBS (pH = 7.4) and incubated with different concentrations of GdW10@BSA NCs for 24 h. Subsequently, 10 μl of fresh CCK-8 was added into the culture medium, and the cells were incubated for an additional 1 h (37 °C under 5% CO2). Finally, the absorbance was measured at 450 nm using a microplate reader (Thermo Scientiﬁc, Multiscan MNK3, Waltham, MA, USA). using protocols approved by the Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety (Institute of High Energy Physics, CAS). The major tissues and organs, including heart, liver, spleen, lungs, kidneys and blood, were weighed and dissolved in 5 ml of concentrated nitric acid solution (HNO3, BV-III). The samples were then heated to 180 °C for 2 h. Subsequently, 2 ml concentrated nitric acid solution was added. When the solution cleared and cooled to room temperature, 1 ml H2O2 solution was added to neutralize the nitric acid solution. The resulting solution was diluted with 2% HNO3 (V/V0) to 5.0 ml and analyzed by induced coupled plasma mass spectrometry (Thermo Elemental X7, USA) to determine the concentration of Gd3+ in each organ or tissue. The BALB/c female mice were killed after the intravenous injection of 200 μl PBS and GdW10@BSA NCs (0.02 M, 75 μmol kg−1) solutions at days 1 and 7. Organs, including heart, liver, spleen, lungs and kidneys, were obtained. These organs were weighed, set and ﬁxed in 4% paraformaldehyde solution, processed in parafﬁn, sectioned and stained with H&E. The representative H&E images were analyzed with an inverted ﬂuorescence microscope (Olympus X73, Japan). using protocols approved by the Key Laboratory for Biomedical Effects of Nanomaterials and Nanosafety (Institute of High Energy Physics, CAS). CT/MR imaging in vivo The major tissues and organs, including heart, liver, spleen, lungs, kidneys and blood, were weighed and dissolved in 5 ml of concentrated nitric acid solution (HNO3, BV-III). The samples were then heated to 180 °C for 2 h. Subsequently, 2 ml concentrated nitric acid solution was added. When the solution cleared and cooled to room temperature, 1 ml H2O2 solution was added to neutralize the nitric acid solution. The resulting solution was diluted with 2% HNO3 (V/V0) to 5.0 ml and analyzed by induced coupled plasma mass spectrometry (Thermo Elemental X7, USA) to determine the concentration of Gd3+ in each organ or tissue. The BALB/c female mice were killed after the intravenous injection of 200 μl PBS and GdW10@BSA NCs (0.02 M, 75 μmol kg−1) solutions at days 1 and 7. Organs, including heart, liver, spleen, lungs and kidneys, were obtained. These organs were weighed, set and ﬁxed in 4% paraformaldehyde solution, processed in parafﬁn, sectioned and stained with H&E. The representative H&E images were analyzed with an inverted ﬂuorescence microscope (Olympus X73, Japan). To observe the cellular absorption of GdW10@BSA NCs, 16HBE cells and HeLa cells were cultured in a six-well plate at a density of 1 × 106 cells per well, respectively. Then, the 16HBE cells and HeLa cells were treated with GdW10@BSA NCs (400 μg ml1) for 24 h. Subsequently, the cells were washed twice with PBS and stained with CA-PI for 15 min. Then, the cells were washed twice with PBS, and ﬂuorescent images were acquired with an inverted ﬂuorescence microscope (Olympus X73, Japan). Blood hematology and biochemistry analysis Blood hematology and biochemistry analysis Blood hematology and biochemistry analysis Blood samples were harvested from the mice fundus artery after the intravenous administration of 200 μl PBS and GdW10@BSA NCs (0.02 M, 75 μmol kg1) solutions, respectively, at 5 h, and at days 1, 3 and 7. One milliliter of blood was obtained from each mouse, 0.5 ml of which was centrifuged at 1500 r.p.m. for 5 min to obtain blood plasma. One-hundred- microliter blood samples in solution were treated with anticoagulant (potassium EDTA) for routine analysis. Blood hematology and biochemistry analyses were accomplished at the Animal Department of Peking University Medical Laboratory. MR imaging of GdW10@BSA NCs in vitro and in vivo Ideal nano-theranostic systems should possess multimodal bio-imaging capabilities for precise spatial and temporal-speciﬁc tracking in vivo. This would help guide and monitor the therapeutic process and response, improve therapeutic outcomes and reduce the associated side effects. To assess the MR contrast performance of the GdW10@BSA NCs in vitro, the contrast effect of the GdW10@BSA NCs at different Gd3+ concentrations (0.05–1.0 mM) was examined with a 4.7-T MR imaging instrument. For increasing Gd3+ concentrations, the T1-weighted MR imaging intensity substantially brightened (Figure 2a). The same concentration gradient of GdW10@BSA NCs was investigated to determine the speciﬁc relaxivity coefﬁcient (r1), which was calculated from the slope of a linear-ﬁtted plot of 1/T1 versus Gd concentration. As shown in Figure 2b, a good linear relationship was observed; the slope, that is, the r1 value for the GdW10@BSA NCs, was calculated to be 9.45 s −1 mM−1, which was much higher than that of conventional Gd-DTPA (3.90 s −1 mM−1) at the same concentration. These results indicated that the GdW10@BSA NCs exhibited a positive signal-enhancement ability and could potentially serve as a T1-weighted MR imaging contrast agent. Encouraged by this performance in vitro, we next evaluated the RESULTS AND DISCUSSION Structure diagram and characterization of GdW10 nanoclusters GdW10 nanoclusters were prepared according to a previously described method.78 As illustrated in Figure 1a, a Gd atom is sandwiched by two lacunary W5O18 caps to form a GdW10 nanocluster. The as-synthesized GdW10 was characterized using Fourier transform infrared (FT-IR) analysis; the resulting spectrum correlated with those of previous reports (Figure 1b).78 This was further conﬁrmed by energy-disperse X-ray spectra, indicating the successful formation of GdW10 nanoclusters (Supplementary Figure S1). Furthermore, X-ray photoelectron spectroscopy analysis was carried out to determine the composition and chemical state of the as-prepared sample. The two strong peaks at 35.4 and 37.6 eV corresponded with the W4f 7/2 and W4f 5/2 states of W6+, respectively (Figure 1c). The peak at 141.4 eV was ascribed to the Gd4d state of Gd3 + (Figure 1d). No other impurities were observed, indicating the high purity of the as-prepared GdW10 clusters. The as-obtained GdW10 nanoclusters were easily dissolved in water. To improve their biocompatibility and solubility in physiological solutions, the GdW10 nanoclusters were further functionalized with BSA to form a hybrid Figure 2 (a) T1-weighted MRI images of commercially available Gd-DTPA and the as-synthesized GdW10@BSA NCs; (b) relaxation rate (1/T1, R1) at various molar concentrations of Gd3+ in GdW10@BSA NCs (0, 0.05, 0.1, 0.2, 0.5, 1 mM of Gd3+), comparing with Gd-DTPA; (c) In vivo T1-weighted MRI images of BALB/c mice bearing MDA-MB-231 tumor before and after i.v. injection of the as-made GdW10@BSA NCs and commercially available Gd-DTPA. (i: Tumor-GdW10@BSA NCs, ii: Tumor-Gd-DTPA, iii: Kidney-GdW10@BSA NCs. dose: 75 μmol kg−1). Figure 2 (a) T1-weighted MRI images of commercially available Gd-DTPA and the as-synthesized GdW10@BSA NCs; (b) relaxation rate (1/T1, R1) at various molar concentrations of Gd3+ in GdW10@BSA NCs (0, 0.05, 0.1, 0.2, 0.5, 1 mM of Gd3+), comparing with Gd-DTPA; (c) In vivo T1-weighted MRI images of BALB/c mice bearing MDA-MB-231 tumor before and after i.v. injection of the as-made GdW10@BSA NCs and commercially available Gd-DTPA. (i: Tumor-GdW10@BSA NCs, ii: Tumor-Gd-DTPA, iii: Kidney-GdW10@BSA NCs. dose: 75 μmol kg−1). Figure 2 (a) T1-weighted MRI images of commercially available Gd-DTPA and the as-synthesized GdW10@BSA NCs; (b) relaxation rate (1/T1, R1) at various molar concentrations of Gd3+ in GdW10@BSA NCs (0, 0.05, 0.1, 0.2, 0.5, 1 mM of Gd3+), comparing with Gd-DTPA; (c) In vivo T1-weighted MRI images of BALB/c mice bearing MDA-MB-231 tumor before and after i.v. Biodistribution and pathology analysis Twenty-one BALB/c female mice were purchased, maintained and killed after the intravenous injection of PBS and GdW10@BSA NCs at different time points Figure 1 (a) Schematic illustration of the structure of GdW10 NCs; (b) FT-IR spectra of as-fabricated GdW10 NCs, FT-IR data in cm−1: 935 (s (m), 706 (m), 586 (w), 545 (m), 490 (w); (c) W 4f and (d) Gd 4d XPS spectra of the synthesized GdW10 NCs. (e) The hydrodynamic d as-made GdW10-BSA nanoclusters. (f) The photograph of GdW10@BSA nanoclusters dissolved in different solutions. Figure 1 (a) Schematic illustration of the structure of GdW10 NCs; (b) FT-IR spectra of as-fabricated GdW10 NCs, FT-IR data in cm−1: 935 (s), 845 (s), 798 (m), 706 (m), 586 (w), 545 (m), 490 (w); (c) W 4f and (d) Gd 4d XPS spectra of the synthesized GdW10 NCs. (e) The hydrodynamic diameter of the as-made GdW10-BSA nanoclusters. (f) The photograph of GdW10@BSA nanoclusters dissolved in different solutions. NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 5 complex through physical adsorption. The hydrodynamic diameter of the formed GdW10-BSA nanoclusters was measured by dynamic light scattering and were ~ 3.5 nm in water (Figure 1e), which is below the commonly recognized threshold for kidney ﬁltration (~5.5 nm). This feature may increase the clearance rates in biosystems and minimize the potential toxic side effects. Moreover, the BSA- modiﬁed GdW10 nanoclusters exhibited good solubility in various physiological solutions (Figure 1f). RESULTS AND DISCUSSION This could be due to the W and Gd atoms having higher X-ray absorption efﬁciencies than I atoms (W, 4.44 cm2 g −1; Gd, 3.11 cm2 g −1; and I, 1. 94 cm2 g −1 at 100 keV).82,83 We next evaluated the feasibility of using the nanoclusters for in vivo CT imaging in MDA-MA-231 tumor- bearing nude mice. After the i.t. injection of GdW10@BSA NCs in physiological saline (0.02 M, 50 μl), the signal at the tumor site was immediately and clearly seen. This indicates that the GdW10@BSA NCs have great potential as efﬁcient contrast agents for tumor CT imaging (Figure 3c). Thus, GdW10@BSA NCs have remarkable potential as dual-mode contrast agents for MR/CT imaging. Photothermal evaluation of the as-fabricated GdW10@BSA NCs in vitro We next pursued the application of GdW10@BSA NCs as therapeutics for cancer therapy. Although many POMs have been used as drugs for the treatment of HIV and other diseases,84,85 the use of GdW10@BSA NCs as photothermal absorbing agents and radiosensitizers for photothermal therapy and tumor radiotherapy has not been reported. After electrochemical reduction, the transparent GdW10@BSA NCs changed to a dark blue color. The UV-Vis absorption spectrum of the reduced GdW10@BSA NCs showed a strong absorbance around 800 nm (Figure 4a), which was attributed to an intervalence charge- transfer transition between W6+ and W5+. The high NIR absorption capability of the reduced GdW10@BSA NCs motivated us to investigate RESULTS AND DISCUSSION injection of the as-made GdW10@BSA NCs and commercially available Gd-DTPA. (i: Tumor-GdW10@BSA NCs, ii: Tumor-Gd-DTPA, iii: Kidney-GdW10@BSA NCs. dose: 75 μmol kg−1). NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 6 application of the nanoclusters as MR imaging contrast agents in vivo. MDA-MB-231 tumor-bearing nude BALB/c mice were intravenously administered with a PBS solution of GdW10@BSA NCs and the T1-weighted MRI images were acquired before and at 5, 30, 60, 90 and 120 min after injection (Figure 2c). Before the intravenous injection of the GdW10@BSA NCs, the tumor appeared dark and homogeneous. After the i.v. administration of the GdW10@BSA NCs, the signal at the tumor site was signiﬁcantly enhanced and the tumor appeared inhomogeneous. No obvious changes were observed in tumors after the i.v. administration of Gd-DTPA due to its smaller molecular size and quick clearance from the tumor. At 2 h after the injection of NCs, the MR signal of the nanocluster injection was two times higher than that of the Gd-DTPA (Supplementary Figure S2), demonstrating that the GdW10@BSA NCs could efﬁciently and passively target the tumor site with a longer residence time compared with the commercial MR imaging contrast agent. In addition, signiﬁcant signal enhancements in the kidney after the i.v. injection of GdW10@BSA NCs were observed (Figure 2c), which indicated that the ultrasmall-sized GdW10@BSA NCs could be cleared via renal excretions, which greatly enhances their clearance rate. This feature of the GdW10@BSA NCs minimizes their potential side effects and increases their suitability for clinical use. application of the nanoclusters as MR imaging contrast agents in vivo. MDA-MB-231 tumor-bearing nude BALB/c mice were intravenously administered with a PBS solution of GdW10@BSA NCs and the T1-weighted MRI images were acquired before and at 5, 30, 60, 90 and 120 min after injection (Figure 2c). Before the intravenous injection of the GdW10@BSA NCs, the tumor appeared dark and homogeneous. After the i.v. administration of the GdW10@BSA NCs, the signal at the tumor site was signiﬁcantly enhanced and the tumor appeared inhomogeneous. No obvious changes were observed in tumors after the i.v. administration of Gd-DTPA due to its smaller molecular size and quick clearance from the tumor. RESULTS AND DISCUSSION At 2 h after the injection of NCs, the MR signal of the nanocluster injection was two times higher than that of the Gd-DTPA (Supplementary Figure S2), demonstrating that the GdW10@BSA NCs could efﬁciently and passively target the tumor site with a longer residence time compared with the commercial MR imaging contrast agent. In addition, signiﬁcant signal enhancements in the kidney after the i.v. injection of GdW10@BSA NCs were observed (Figure 2c), which indicated that the ultrasmall-sized GdW10@BSA NCs could be cleared via renal excretions, which greatly enhances their clearance rate. This feature of the GdW10@BSA NCs minimizes their potential side effects and increases their suitability for clinical use. NCs solution were used to obtain X-ray phantom images in vitro. Figure 3a reveals that the X-ray CT images become gradually brighter with increasing nanocluster concentration. The CT values, that is, HUs, increased linearly with GdW10@BSA NC concentration (Figure 3b). The measured X-ray CT absorption coefﬁcient of the GdW10@BSA NCs was 19.2 HU mM−1, which was three times higher than that of commercial iopromide (6.09 HU mM−1), indicating the good contrast efﬁcacy of the GdW10@BSA NCs for CT imaging. This could be due to the W and Gd atoms having higher X-ray absorption efﬁciencies than I atoms (W, 4.44 cm2 g −1; Gd, 3.11 cm2 g −1; and I, 1. 94 cm2 g −1 at 100 keV).82,83 We next evaluated the feasibility of using the nanoclusters for in vivo CT imaging in MDA-MA-231 tumor- bearing nude mice. After the i.t. injection of GdW10@BSA NCs in physiological saline (0.02 M, 50 μl), the signal at the tumor site was immediately and clearly seen. This indicates that the GdW10@BSA NCs have great potential as efﬁcient contrast agents for tumor CT imaging (Figure 3c). Thus, GdW10@BSA NCs have remarkable potential as dual-mode contrast agents for MR/CT imaging. NCs solution were used to obtain X-ray phantom images in vitro. Figure 3a reveals that the X-ray CT images become gradually brighter with increasing nanocluster concentration. The CT values, that is, HUs, increased linearly with GdW10@BSA NC concentration (Figure 3b). The measured X-ray CT absorption coefﬁcient of the GdW10@BSA NCs was 19.2 HU mM−1, which was three times higher than that of commercial iopromide (6.09 HU mM−1), indicating the good contrast efﬁcacy of the GdW10@BSA NCs for CT imaging. CT imaging of the synthesized GdW10@BSA NCs in vitro and in vivo CT imaging of the synthesized GdW10@BSA NCs in vitro and in vivo In addition to their MR imaging application due to its high-Z elements (that is, the W and Gd atoms), these GdW10@BSA NCs may have a strong X-ray attenuation ability. Thus, these nanoclusters could be used as contrast agents for CT imaging, which is one of the most commonly used imaging tools for medical diagnoses. CT imaging can be used to acquire high-resolution 3D structural details of the whole body. For the ﬁrst time, we evaluated the potential of the GdW10@BSA NCs as CT contrast agents. Varied concentrations of GdW10@BSA Figure 3 (a) In vitro CT phantom images of Iopromide and GdW10@BSA NCs with different concentrations (0, 6.25, 12.5, 25, 50 and 100 mM of Gd3+). (b) Hountsﬁed Unit values of GdW10@BSA NCs and Iopromide as the function of their concentrations. (c) In vivo CT images of a BALB/c mice bearing MDA- MB-231 tumor before and after intratumor injection of GdW10@BSA NCs. (dose: 0.02 M, 50 μl). Figure 3 (a) In vitro CT phantom images of Iopromide and GdW10@BSA NCs with different concentrations (0, 6.25, 12.5, 25, 50 and 100 mM of Gd3+). (b) Hountsﬁed Unit values of GdW10@BSA NCs and Iopromide as the function of their concentrations. (c) In vivo CT images of a BALB/c mice bearing MDA- MB-231 tumor before and after intratumor injection of GdW10@BSA NCs. (dose: 0.02 M, 50 μl). NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 4 (a) UV-vis absorption spectrum of GdW10 NCs solution before and after the reduction. Inset: the corresponding digital images he reduction. (b) The temperature evaluation of pure NaCl solution and reduced GdW10@BSA/NCs (1 mg ml−1) under the irradiati with a power density of 1 W cm−2 as a functional of irradiation time (0–10 min); (c) cell viability assays on Hela cells in d GdW10@BSA/NS solution (0, 62.5, 125, 250, 500 and 1000 μg ml−1) (c) before and (e) after 808 nm laser (1 W cm−2, 10 min); g studies using Calcein-AM and PI on Hela cells incubated with reduced GdW10@BSA/NS solution (1000 μg ml−1) (d) before and (f ion with 808 nm laser (1 W cm−2, 10 min). Error bars were based on s.d. of six parallel samples. P-values were obtained by the .01, *Po0.001. Y Yong et al 7 Figure 4 (a) UV-vis absorption spectrum of GdW10 NCs solution before and after the reduction. CT imaging of the synthesized GdW10@BSA NCs in vitro and in vivo Inset: the corresponding digital images of before and after the reduction. (b) The temperature evaluation of pure NaCl solution and reduced GdW10@BSA/NCs (1 mg ml−1) under the irradiation of 808 nm laser with a power density of 1 W cm−2 as a functional of irradiation time (0–10 min); (c) cell viability assays on Hela cells incubated with reduced GdW10@BSA/NS solution (0, 62.5, 125, 250, 500 and 1000 μg ml−1) (c) before and (e) after 808 nm laser (1 W cm−2, 10 min); live-dead cell staining studies using Calcein-AM and PI on Hela cells incubated with reduced GdW10@BSA/NS solution (1000 μg ml−1) (d) before and (f) after 10-min irradiation with 808 nm laser (1 W cm−2, 10 min). Error bars were based on s.d. of six parallel samples. P-values were obtained by the Student’s test: Po0.01, *Po0.001. the NIR photothermal properties that enable the nanoclusters to function as photothermal absorbing agents for NIR-triggered PTT. To assess these properties, the changes in temperature of aqueous dispersions of reduced GdW10@BSA NCs (1 mg ml1) were measured under 808 nm NIR laser irradiation (1 W cm−2, 10 min). A NaCl solution was used as the control (Figure 4b). The temperature of the GdW10@BSA NCs samples raised increased with the irradiation time and reached up to 55 °C after 10 min (1 W cm−2). The temperature of the NaCl solution only showed a mild temperature change (a change of ~ 5 °C). These results demonstrated that the GdW10@BSA NCs can rapidly and efﬁciently convert NIR light into heat. Due to the good photothermal heating effect, the photothermal therapeutic effects of reduced GdW10@BSA NCs on HeLa cells in vitro were studied. After incubation with reduced GdW10@BSA NCs, there was no noticeable cell apoptosis, even at concentrations up to 1.0 mg ml1. The HeLa cell viability remained greater than 90% (Figure 4c and Supplementary Figure S3). In contrast, after irradiation with 808 nm laser at a power density of 1 W cm−2 for 10 min, the cell viability substantially decreased to 5%, indicating the efﬁcient photothermal ablation of cancer cells (Figure 4e). These results were visually conﬁrmed by live-dead cell staining images, which showed almost no cells survived after the irradiation (Figures 4d and f). Therefore, our results suggested that the GdW10@BSA NCs could be used as an efﬁcient agent for the PTT of cancer cells. CT imaging of the synthesized GdW10@BSA NCs in vitro and in vivo NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al ated GdW10@BSA NCs-enhanced radiotherapy of nude mice eir PTT applications, the GdW10@BSA NCs could be sitizers for enhanced RT therapies of tumors due to the nanoclusters containing a large number of high GdW10@BSA particle (Gd and W atoms). RT commonly used treatments for cancer therapy energy X-ray or γ-ray radiation to kill cancer cell relative tumor volume curves of different group mice after different treatments: (i) PBS injection; (ii) GdW10@B GdW10@BSA NCs + X-ray. (b) The body weight of different group BEL-7402 tumor-bearing mice after various admin collected from each group tumor-bearing mice after radiotherapy. (d) The tumor inhibition rate of different group BE ntative images of different group BEL-7402 tumor-bearing mice after different administrations at the end of radio rom each group BEL-7402 tumor-bearing mice after various administrations. (g) H&E stained images of tumors co bars were based on s.d. of four parallel specimens. P-values were calculated by the Student’s test: Po0.01, *Po0 Y Yong et al 8 Figure 5 (a) The relative tumor volume curves of different group mice after different treatments: (i) PBS injection; (ii) GdW10@BSA NCs i.t. injection; (iii) X-ray only; (iv) GdW10@BSA NCs + X-ray. (b) The body weight of different group BEL-7402 tumor-bearing mice after various administrations. (c) Average weight of tumors collected from each group tumor-bearing mice after radiotherapy. (d) The tumor inhibition rate of different group BEL-7402 tumor-bearing mice. (e) Representative images of different group BEL-7402 tumor-bearing mice after different administrations at the end of radiotherapy. (f) Photos of tumors collected from each group BEL-7402 tumor-bearing mice after various administrations. (g) H&E stained images of tumors collected from different group mice. Error bars were based on s.d. of four parallel specimens. P-values were calculated by the Student’s test: Po0.01, ***Po0.001. In vivo as-fabricated GdW10@BSA NCs-enhanced radiotherapy of tumor-bearing nude mice In vivo as-fabricated GdW10@BSA NCs-enhanced radiotherapy of tumor-bearing nude mice the nanoclusters containing a large number of high Z elements in each GdW10@BSA particle (Gd and W atoms). RT is one of the most commonly used treatments for cancer therapy and employs high energy X-ray or γ-ray radiation to kill cancer cells without any depth In addition to their PTT applications, the GdW10@BSA NCs could be used as radiosensitizers for enhanced RT therapies of tumors due to NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 9 restrictions. For the ﬁrst time, we assessed the feasibility of GdW10@BSA NCs as radiosensitizers for the RT of tumors. BEL-7402 tumor-bearing nude mice were divided into four groups (a) PBS only; (b) GdW10@BSA NCs only; (c) RT only; and (d) GdW10@BSA NCs + RT. The tumor volume changes and body weights of all mice were recorded as functions of time (Figures 5a and b). As shown in Figure 5a, at 18 days after treatment, tumors in groups a and b grew rapidly with relative tumor volume changes (V/V0) of 8.29± 0.948 and 7.993 ± 1.042, respectively. This indicated that the single use of GdW10@BSA NCs without X-ray radiation had negligible effects on tumor growth (Supplementary Table S2). The mice treated with RT alone exhibited tumor growth inhibition, as indicated by the relative tumor volume change of 61.49 % (V/V0 = 4.103 ± 0.496) after 18 days of treatment (Figures 5c and d). The most efﬁcient tumor growth delay was achieved using the treatment of GdW10@BSA NCs +RT (group d). As shown in Figures 5e and f, compared with RT alone, the tumors in group d were thoroughly eliminated at 4 days after treatment without recurrence in the subsequent 14 days, thus indicating the efﬁcient sensitization effect of GdW10@BSA NCs on RT. These effects were conﬁrmed using H&E staining (Figure 5g). The HE staining images of tumor slices collected at 18 days after the treatments indicated that most of the tumor cells were destroyed under the GdW10@BSA NCs +RT treatment. Tumor cells treated with GdW10@BSA NCs or X-rays alone resulted in partial damages to the tumor tissues, which further showed the distinct GdW10@BSA NCs-enhanced RT effects. showed the potential for using GdW10@BSA NCs as a small but powerful platform for multimodal image-guided PTT/RT therapy, other POM-based multifunctional nanotheranostics could be developed for cancer diagnoses and therapy. CONFLICT OF INTEREST The authors declare no conﬂict of interest. ACKNOWLEDGEMENTS This work was supported by the National Basic Research Programs of China (973 programs, Nos. 2012CB932500 and 2015CB932104) and the NSFC (Nos. 21177128, 31751015, 21371173 and 21320102003). 1 Fernandez-Fernandez, A., Manchanda, R. & McGoron, A. Theranostic applications of nanomaterials in cancer: drug delivery, image-guided therapy, and multifunctional platforms. Appl. Biochem. Biotech. 165, 1628–1651 (2011). 2 Song, J., Huang, P., Duan, H. & Chen, X. Plasmonic vesicles of amphiphilic nanocrystals: optically active multifunctional platform for cancer diagnosis and therapy. Accounts Chem. Res. 48, 2506–2515 (2015). 3 Riehemann, K., Schneider, S., Luger, T., Godin, B., Ferrari, M. & Fuchs, H. Nanomedicine–challenge and perspectives. Angew. Chem. Int. Ed. 48, 872–897 (2009). 4 Rao, C., Matte, H. & Maitra, U. Graphene analogues of inorganic layered materials. Angew. Chem. Int. Ed. 52, 13162–13185 (2013). 5 Pu, K., Shuhendler, A., Jokerst, J., Mei, J., Gambhir, S., Bao, Z. & Rao, J. Semiconducting polymer nanoparticles as photoacoustic molecular imaging probes in living mice. Nat. Nanotechnol. 9, 233–239 (2014). Cytotoxicity Theranostic safety is an important factor for evaluating nanocluster biological applications. Firstly, we evaluated the stability of GdW10@BSA NCs at various pH value. No obvious changes have been noticed, indicating its well stability at current condition (Supplementary Figures S8 and S9). Next, to determine the biocompatibility of GdW10 NCs, the toxicity of the NCs was evaluated by various in vivo experiments, including studying changes in animal behavior, body weight, biodistribution, H&E staining, and blood biochemistry. After thirty days of infection, no mice had died and no obvious signs of abnormal changes in body weight, activity, eating, and drinking were observed. The biodistribution results showed that the GdW10 NCs mainly accumulated in the kidneys and blood (Supplementary Figure S4). Hematological and serum chemistry analyses revealed that all parameters, such as renal function markers (CRE, BUN), were within normal ranges (Supplementary Figure S5 and Supplementary Table S1). Moreover, the inﬂuence of GdW10 on the hemolytic behavior of red blood cells (RBCs) was also tested to further evaluate their biocompatibility. Negligible RBC hemolysis was detected, indicating that GdW10 possessed high blood compatibility (Supplementary Figure S6). Additionally, as shown in Supplementary Figure S7, no apparent tissue damage in the main organs was observed. These preliminary results showed that GdW10@BSA exhib- ited good biocompatibility at the tested dosage in the short term. g 6 Matte, H., Gomathi, A., Manna, A., Late, D., Datta, R., Pati, S. K. & Rao, C. N. MoS2 and WS2 analogues of graphene. Angew. Chem. Int. Ed. 49, 4059–4062 (2010). 7 Lucky, S., Soo, K. & Zhang, Y. Nanoparticles in photodynamic therapy. Chem. Rev. 115, 1990–2042 (2015). 8 Horcajada, P., Chalati, T., Serre, C., Gillet, B., Sebrie, C., Baati, T., Eubank, J. F., Heurtaux, D., Clayette, P., Kreuz, C., Chang, J. S., Hwang, Y. K., Marsaud, V., Bories, P. N., Cynober, L., Gil, S., Férey, G., Couvreur, P. & Gref, R. Porous metal-organic- framework nanoscale carriers as a potential platform for drug delivery and imaging. Nat. Mater. 9, 172–178 (2010). 9 Choi, H., Liu, W., Liu, F., Nasr, K., Misra, P., Bawendi, M. & Frangioni, J. V. Design considerations for tumour-targeted nanoparticles. Nat. Nanotechnol. 5, 42–47 (2010). 10 Chen, Y., Tan, C., Zhang, H. & Wang, L. Two-dimensional graphene analogues for biomedical applications. Chem. Soc. Rev. 44, 2681–2701 (2015). pp 11 Shao, K., Singha, S., Clemente-Casares, X., Tsai, S., Yang, Y. & Santamaria, P. Cytotoxicity 11 Shao, K., Singha, S., Clemente-Casares, X., Tsai, S., Yang, Y. & Santamaria, Nanoparticle-based immunotherapy for cancer. ACS Nano 9, 16–30 (2015). g g Nanoparticle-based immunotherapy for cancer. ACS Nano 9, 16–30 (2015). 12 Lammers, T., Kiessling, F., Hennink, W. & Storm, G. Nanotheranostics and image-guided drug delivery: current concepts and future directions. Mol. Pharm. 7, 1899–1912 (2010). 13 Lammers, T., Aime, S., Wim, E., Storm, G. & Kiessling, F. Theranostic nanomedicine. Accounts Chem. Res. 44, 1029–1038 (2011). 14 Xie, J., Lee, S. & Chen, X. Nanoparticle-based theranostic agents. Adv. Drug Deliver. Rev. 62, 1064–1079 (2010). , ( ) 15 Katz, E. & Willner, I. Integrated nanoparticle-biomolecule hybrid systems: synthesis, properties, and applications. Angew. Chem. Int. Ed. 43, 6042–6108 (2004). 16 Janib, S., Moses, A. & MacKay, J. Imaging and drug delivery using theranostic nanoparticles. Adv. Drug Deliver. Rev. 62, 1052–1063 (2010). 17 Xiao, Q., Zheng, X., Bu, W., Ge, W., Zhang, S., Chen, F., Xing, H., Ren, Q., Fan, W., Zhao, K., Hua, Y. & Shi, J. A core/satellite multifunctional nanotheranostic for in vivo imaging and tumor eradication by radiation/photothermal synergistic therapy. J. Am. Chem. Soc. 135, 13041–13048 (2013). 18 Wang, Y., Wang, K., Zhao, J., Liu, X., Bu, J., Yan, X. & Huang, R. Multifunctional mesoporous silica-coated graphene nanosheet used for chemo-photothermal synergistic targeted therapy of glioma. J. Am. Chem. Soc. 135, 4799–4804 (2013). In vivo as-fabricated GdW10@BSA NCs-enhanced radiotherapy of tumor-bearing nude mice Because POMs have varying sizes, structures and compositions, there are a number of potential avenues for developing other nanomedicines with better therapeutic outcomes and fewer side effects. restrictions. For the ﬁrst time, we assessed the feasibility of GdW10@BSA NCs as radiosensitizers for the RT of tumors. BEL-7402 tumor-bearing nude mice were divided into four groups (a) PBS only; (b) GdW10@BSA NCs only; (c) RT only; and (d) GdW10@BSA NCs + RT. The tumor volume changes and body weights of all mice were recorded as functions of time (Figures 5a and b). As shown in Figure 5a, at 18 days after treatment, tumors in groups a and b grew rapidly with relative tumor volume changes (V/V0) of 8.29± 0.948 and 7.993 ± 1.042, respectively. This indicated that the single use of GdW10@BSA NCs without X-ray radiation had negligible effects on tumor growth (Supplementary Table S2). The mice treated with RT alone exhibited tumor growth inhibition, as indicated by the relative tumor volume change of 61.49 % (V/V0 = 4.103 ± 0.496) after 18 days of treatment (Figures 5c and d). The most efﬁcient tumor growth delay was achieved using the treatment of GdW10@BSA NCs +RT (group d). As shown in Figures 5e and f, compared with RT alone, the tumors in group d were thoroughly eliminated at 4 days after treatment without recurrence in the subsequent 14 days, thus indicating the efﬁcient sensitization effect of GdW10@BSA NCs on RT. These effects were conﬁrmed using H&E staining (Figure 5g). The HE staining images of tumor slices collected at 18 days after the treatments indicated that most of the tumor cells were destroyed under the GdW10@BSA NCs +RT treatment. Tumor cells treated with GdW10@BSA NCs or X-rays alone resulted in partial damages to the tumor tissues, which further showed the distinct GdW10@BSA NCs-enhanced RT effects. CONCLUSION & But, S. Spectroscopic studies of Eu(III) and Nd(III) complexes with several polyoxometalates. J. Alloys Compd. 300–301, 370–376 (2000). 53 Lis, S. & But, S. Spectroscopic studies of Eu(III) and Nd(III) com 25 Melancon, M., Zhou, M. & Li, C. Cancer theranostics with near-infrared light-activatable multimodal nanoparticles. Accounts Chem. Res. 44, 947–956 (2011). polyoxometalates. J. Alloys Compd. 300–301, 370–376 (2000). 54 Lubal, P., Koprivova, H., Sedo, O., Havel, J., Lis, S. & But, S. Simultaneous determination of molybdenum(VI) and tungsten(VI) and its application in elemental analysis of polyoxometalates. Talanta 69, 800–806 (2006). 26 Kelkar, S. & Reineke, T. Theranostics: combining imaging and therapy. Bioconjugate Chem. 22, 1879–1903 (2011). 27 Huang, H., Barua, S., Sharma, G., Dey, S. & Rege, K. Inorganic nanoparticles for cancer imaging and therapy. J. Control. Release 155, 344–357 (2011). 55 AlDamen, M., Clemente-Juan, J., Coronado, E., Martı´-Gastaldo, C. & Gaita-Arin˜o, A. Mononuclear lanthanide single-molecule magnets based on polyoxometalates. J. Am. Chem. Soc. 130, 8874–8875 (2008). 28 Choi, K., Liu, G., Lee, S. & Chen, X. Theranostic nanoplatforms for simultaneous cancer imaging and therapy: current approaches and future perspectives. Nanoscale 4, 330–342 (2012). 56 Clemente-León, M., Coronado, E., Delhaes, P., Gómez-García, C. & Mingotaud, C. Hybrid Langmuir-Blodgett ﬁlms formed by alternating layers of magnetic polyoxometa- late clusters and organic donor molecules-towards the preparation of multifunctional molecular materials. Adv. Mater. 13, 574–577 (2001). 29 Ahmed, N., Fessi, H. & Elaissari, A. Theranostic applications of nanoparticles in cancer. Drug Discov. Today 17, 928–934 (2012). 30 Lux, F., Mignot, A., Mowat, P., Louis, C., Dufort, S., Bernhard, C., Denat, F., Boschetti, F., Brunet, C., Antoine, R., Dugourd, P., Laurent, S., Vander Elst, L., Muller, R., Sancey, L., Josserand, V., Coll, J. L., Stupar, V., Barbier, E., Rémy, C., Broisat, A., Ghezzi, C., Le Duc, G., Roux, S., Perriat, P. & Tillement, O. Ultrasmall rigid particles as multimodal probes for medical applications. Angew. Chem. Int. Ed. 50, 12299–12303 (2011). 57 Martinez-Perez, M., Montero, O., Evangelisti, M., Luis, F., Sese, J., Cardona-Serra, S. & Coronado, E. Fragmenting gadolinium: mononuclear polyoxometalate-based magnetic coolers for ultra-low temperatures. Adv. Mater. 24, 4301–4305 (2012). 58 Martínez-Pérez, M., Cardona-Serra, S., Schlegel, C., Moro, F., Alonso, P., Prima-García, H., Clemente-Juan, J. M., Evangelisti, M., Gaita-Ariño, A., Sesé, J., van Slageren, J., Coronado, E. & Luis, F. Gd-based single-ion magnets with tunable magnetic anisotropy: molecular design of spin qubits. Phys. Rev. Lett. 108, 247213–247218 (2012). CONCLUSION & Patzke, G. New perspectives on polyoxometalate catalysts: alcohol oxidation with Zn/Sb-polyoxotungstates. Chem-Eur. J. 18, 13293–13298 (2012). 39 Bassil, B. & Kortz, U. Divacant polyoxotungstates: reactivity of the gamma-decatungstates [γ-XW10O36]8 −(X= Si, Ge). Dalton Trans. 40, 9649–9661 (2011). 65 Liu, R., Li, S., Yu, X., Zhang, G., Zhang, S., Yao, J. & Zhi, L. A general green strategy for fabricating metal nanoparticles/polyoxometalate/graphene tri-component nanohybrids: enhanced electrocatalytic properties. J. Mater. Chem. 22, 3319–3322 (2012). 40 Zheng, L., Gu, Z., Ma, Y., Zhang, G., Yao, J., Keita, B. & Nadjo, L. Molecular interaction between europium decatungstate and histone H1 and its application as a novel biological labeling agent. J. Biol. Inorg. Chem. 15, 1079–1085 (2010). 66 Ju, F., VanderVelde, D. & Nikolla, E. Molybdenum-based polyoxometalates as highly active and selective catalysts for the epimerization of aldoses. ACS Catal. 4, 1358–1364 (2014). 41 Yang, H., Cheng, Y., Su, M., Xiao, Y., Hu, M., Wang, W. & Wang, Q. Polyoxometalate- biomolecule conjugates: a new approach to create hybrid drugs for cancer therapeutics. Bioorg. Med. Chem. Lett. 23, 1462–1466 (2013). 67 Carraro, M., Fiorani, G., Mognon, L., Caneva, F., Gardan, M., Maccato, C. & Bonchio, M. Hybrid polyoxotungstates as functional comonomers in new cross- linked catalytic polymers for sustainable oxidation with hydrogen peroxide. Chem-Eur. J. 18, 13195–13202 (2012). 42 Yamase, T. Anti-tumor-viral, and -bacterial activities of polyoxometalates for realizing an inorganic drug. J. Mater. Chem. 15, 4773–4782 (2005). 43 Wang, X., Liu, J., Li, J., Yang, Y., Liu, J., Li, B. & Pope, M. T. Synthesis and antitumor activity of cyclopentadienyltitanium substituted polyoxotungstate [CoW11O39(CpTi)]7 − (Cp = η5-C5H5). J. Inorg. Biochem. 94, 279–284 (2003). 68 Liu, Y., Shi, D., Zhao, J., Wang, X., Cheng, H., Chen, L., Ma, P. & Niu, J. A novel 2D organic-inorganic hybrid 3d–4f polyoxometalate built by [Gd(α-PW11O39)2]11 − units and [Cu(Dap)2]2+ bridges. Russ. J. Coord. Chem. 37, 712–717 (2011). 44 Wang, X., Li, F., Liu, S. & Pope, M. New liposome-encapsulated-polyoxometalates: synthesis and antitumoral activity. J. Inorg. Biochem. 99, 452–457 (2005). 69 Chai, W., Wang, S., Zhao, H., Liu, G., Fischer, K., Li, H., Wu, L. & Schmidt, M. Hybrid assemblies based on a gadolinium-containing polyoxometalate and a cationic polymer with spermine side chains for enhanced MRI contrast agents. Chem-Eur. J. 19, 13317–13321 (2013). 45 Ogata, A., Yanagie, H., Ishikawa, E., Morishita, Y., Mitsui, S., Yamashita, A., Hasumi, K., Takamoto, S., Yamase, T. & Eriguchi, M. CONCLUSION 31 Chen, H., Wang, G., Tang, W., Todd, T., Zhen, Z., Tsang, C., Hekmatyar, K., Cowger, T., Hubbard, R. B., Zhang, W., Stickney, J., Shen, B. & Xie, J. Gd-encapsulated carbonaceous dots with efﬁcient renal clearance for magnetic resonance imaging. Adv. Mater. 26, 6761–6766 (2014). 59 Cardona-Serra, S., Clemente-Juan, J., Coronado, E., Gaita-Arino, A., Camon, A., Evangelisti, M., Luis, F., Martínez-Pérez, M. J. & Sesé, J. Lanthanoid single-ion magnets based on polyoxometalates with a 5-fold symmetry: the series [LnP5W30O110] 12− (Ln3+ = Tb, Dy, Ho, Er, Tm, and Yb). J. Am. Chem. Soc. 134, 14982–14990 (2012). 32 Qi, W., Li, H. & Wu, L. A novel, luminescent, silica-sol–gel hybrid based on surfactant- encapsulated polyoxometalates. Adv. Mater. 19, 1983–1987 (2007). 33 Long, D., Tsunashima, R. & Cronin, L. Polyoxometalates: building blocks for functional nanoscale systems. Angew. Chem. Int. Ed. 49, 1736–1758 (2010). y g 34 Rhule, J., Hill, C. & Judd, D. Polyoxometalates in medicine. Chem. Rev. 98, 327–357 (1998). 60 Müller, A., Peters, F., Pope, M. & Gatteschi, D. Polyoxometalates very large clusters- nanoscale magnets. Chem. Rev. 98, 239–271 (1998). 35 Hussain, F., Conrad, F. & Patzke, G. A Gadolinium-bridged polytungstoarsenate(III) nanocluster: [Gd8As12W124O432(H2O)22]60−. Angew. Chem. Int. Ed. 48, 9088–9091 (2009). 61 Schaming, D., Farha, R., Xu, H., Goldmann, M. & Ruhlmann, L. Formation and photocatalytic properties of nanocomposite ﬁlms containing both tetracobalt dawson-derived sandwich polyanions and tetracationic porphyrins. Langmuir 27, 132–143 (2011). 36 Kong, X., Ren, Y., Zheng, P., Long, Y., Long, L., Huang, R. & Zheng, L. S. Construction of polyoxometalates-based coordination polymers through direct incorporation between polyoxometalates and the voids in a 2D network. Inorg. Chem. 45, 10702–10711 (2006). 62 Dolbecq, A., Mialane, P., Keita, B. & Nadjo, L. Polyoxometalate-based materials for efﬁcient solar and visible light harvesting: application to the photocatalytic degradation of azo dyes. J. Mater. Chem. 22, 24509–24521 (2012). 37 Canny, J., Tere, Z., Thouvenot, R. & Herve, G. Disubstituted tungstosilicates. 1. Synthesis, stability, and structure of the lacunary precursor polyanion γ-SiW10O368 −. Inorg. Chem. 25, 2114–2119 (1986). 63 Nisar, A., Zhuang, J. & Wang, X. Construction of amphiphilic polyoxometalate mesostructures as a highly efﬁcient desulfurization catalyst. Adv. Mater. 23, 1130–1135 (2011). 38 Fujita, Y., Ishihara, N., Fukaya, K., Srifa, A. & Naruke, H. Fabrication of lanthanide decatungstate [LnIII/IV(W5O18)2]8 −/9 −(Ln= CeIV/III, EuIII, and ErIII) thin ﬁlms using spin-coating from aqueous solutions. J. Alloys Compd. 519, 136–139 (2012). 64 Ni, L., Patscheider, J., Baldridge, K. CONCLUSION 19 Wang, Y., Wang, H., Liu, D., Song, S., Wang, X. & Zhang, H. Graphene oxide covalently grafted upconversion nanoparticles for combined NIR mediated imaging and photothermal/photodynamic cancer therapy. Biomaterials 34, 7715–7724 (2013). In summary, we developed a new theranostic agent based on GdW10@BSA NCs with an ultrasmall size and versatile properties. The as-prepared GdW10@BSA NCs signiﬁcantly enhanced CT/MR imaging signals and also elicited remarkable photothermal therapy (PTT)/radiotherapy therapy (RT) therapeutic effects. Moreover, H&E staining, blood hematology, biochemistry analysis and hemolytic analysis revealed no noticeable GdW10@BSA NCs in vivo toxicity, indicating biocompatibility. Most importantly, because our work 20 Sahu, A., Choi, W., Lee, J. & Tae, G. Graphene oxide mediated delivery of methylene blue for combined photodynamic and photothermal therapy. Biomaterials 34, 6239–6248 (2013). 21 Barua, S. & Mitragotri, S. Synergistic targeting of cell membrane, cytoplasm, and nucleus of cancer cells using rod-shaped nanoparticles. ACS Nano 7, 9558–9570 (2013). 22 Fan, Z., Dai, X., Lu, Y., Yu, E., Brahmbatt, N., Carter, N., Tchouwou, C., Singh, A. K., Jones, Y., Yu, H. & Ray, P. C. Enhancing targeted tumor treatment by near IR NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al Gadolinium polytungstate nanoclusters Y Y t l Y Yong et al 10 light-activatable photodynamic-photothermal synergistic therapy. Mol. Pharm. 11, 1109–1116 (2014). light-activatable photodynamic-photothermal synergistic therapy. Mol. Pharm. 11, 1109–1116 (2014). 51 Zhang, Y., Bo, X., Nsabimana, A., Munyentwali, A., Han, C., Li, M. & Guo, L. Green and facile synthesis of an Au nanoparticles@polyoxometalate/ordered mesoporous carbon tri-component nanocomposite and its electrochemical applications. Biosens. Bioelectron. 66, 191–197 (2015). 23 Tian, B., Wang, C., Zhang, S., Feng, L. & Liu, Z. Photothermally enhanced photodynamic therapy delivered by nano-graphene oxide. ACS Nano 5, 7000–7009 (2011). , 52 Stefan, L. & But, S. A new spectrophotometric method for the determination and simultaneous determination of tungsten and molybdenum in polyoxometalates and their 52 Stefan, L. & But, S. A new spectrophotometric method for the determination and simultaneous determination of tungsten and molybdenum in polyoxometalates and their Ln(III) complexes. J. Alloys Compd. 303–304, 132–136 (2000). , , p p simultaneous determination of tungsten and molybdenum in polyoxometala Ln(III) complexes. J. Alloys Compd. 303–304, 132–136 (2000). 24 Bardhan, R., Lal, S., Joshi, A. & Halas, A. Theranostic nanoshells from probe design to imaging and treatment of cancer. Accounts Chem. Res. 44, 936–946 (2011). Ln(III) complexes. J. Alloys Compd. 303–304, 132–136 (2000). 53 Lis, S. CONCLUSION Antitumour effect of polyoxomolybdates: induction of apoptotic cell death and autophagy in vitro and in vivo models. Brit. J. Cancer 98, 399–409 (2008). 70 Cao, J., Liu, S., Cao, R., Xie, L., Ren, Y., Gao, C. & Xu, L. Organic–inorganic hybrids assembled by bis(undecatungstophosphate) lanthanates and dinuclear copper (ii)–oxalate complexes. Dalton Trans. 115–120 (2008). 46 Ogata, A., Mitsui, S., Yanagie, H., Kasano, H., Hisa, T., Yamase, T. & Eriguchi, M. A novel anti-tumor agent, polyoxomolybdate induces apoptotic cell death in AsPC-1 human pancreatic cancer cells. Biomed. Pharmacother. 59, 240–244 (2005). 71 Sun, G., Feng, J., Wu, H., Pei, F., Fang, K. & Lei, H. Investigation of sandwiched gadolinium (III) complexes with tungstosilicates as potential MRI contrast agents. Magn. Reson. Imaging 22, 421–426 (2004). 47 Menon, D., Thomas, R., Narayanan, S., Maya, S., Jayakumar, R., Hussain, F., Lakshmanan, V. & Nair, S. A novel chitosan/polyoxometalate nano-complex for anti- cancer applications. Carbohyd. Polym. 84, 887–893 (2011). 72 Li, Z., Li, W., Li, X., Pei, F., Li, Y. & Lei, H. The gadolinium complexes with polyoxometalates as potential mri contrast agents. Magn. Reson. Imaging 25, 412–417 (2007). 48 Geng, J., Li, M., Ren, J., Wang, E. & Qu, X. Polyoxometalates as inhibitors of the aggregation of amyloid β peptides associated with Alzheimer’s disease. Angew. Chem. Int. Ed. 123, 4270–4274 (2011). 73 Da-liang, K., Bo, W., Sheng-yan, Z., Hai-shan, Y. & Yang, J. Volume-related efﬁciency of gadolinium polyoxometalates as MRI contrast agents. Chem. Res. Chin. Univ. 29, 1055–1058 (2013). 49 Gao, N., Sun, H., Dong, K., Ren, J., Duan, T., Xu, C. & Qu, X. Transition-metal- substituted polyoxometalate derivatives as functional anti-amyloid agents for Alzhei- mer’s disease. Nat. Commun. 5, 3422 (2014). 74 Feng, J., Li, X., Pei, F., Sun, G., Zhang, X. & Liu, M. An evaluation of gadolinium polyoxometalates as possible MRI contrast agent. Magn. Reson. Imaging 20, 407–412 (2002). 50 Zheng, L., Ma, Y., Zhang, G., Yao, J., Keita, B. & Nadjo, L. A multitechnique study of europium decatungstate and human serum albumin molecular interaction. Phys. Chem. Chem. Phys. 12, 1299–1304 (2010). NPG Asia Materials Gadolinium polytungstate nanoclusters Y Yong et al 11 75 Feng, J., Sun, G., Pei, F. & Liu, M. Comparison between GdDTPA and two gadolinium polyoxometalates as potential MRI contrast agents. J. Inorg. Biochem. 92, 193–199 (2002). 82 Zhou, Z., Kong, B., Yu, C., Shi, X., Wang, M., Liu, W., Sun, Y., Zhang, Y., Yang, H. NPG Asia Materials CONCLUSION & Yang, S. Tungsten oxide nanorods: an efﬁcient nanoplatform for tumor CT imaging and photothermal therapy. Sci. Rep. 4, 3653 (2014). 76 Lux, F., Sancey, L., Bianchi, A., Crémillieux, Y., Roux, S. & Tillement, O. Gadolinium- based nanoparticles for theranostic MRI-radiosensitization. Nanomedicine 10, 1801–1815 (2015). 83 Lee, N., Choi, S. & Hyeon, T. Nano-Sized CT Contrast agents. Adv. Mater. 25, 2641–2660 (2013). 84 Shigeta, S., Mori, S., Watanabe, J., Soeda, S., Takahashi, K. & Yamase, T. Synergistic anti-inﬂuenza virus A (H1N1) activities of PM-523 (polyoxometalate) and ribavirin in vitro and in vivo. Antimicrob. Agents. Chemother. 41, 1423–1427 (1997). 77 Kim, D., Seog, J., Kim, M., Yang, M., Gillette, E., Lee, S. & Han, S. W. Polyoxometalate-mediated one-pot synthesis of Pd nanocrystals with controlled morphologies for efﬁcient chemical and electrochemical catalysis. Chem-Eur. J. 21, 5387–5394 (2015). 85 Shigeta, S., Watanabe, J., Baba, M., Khenkin, A., HiII, C. & Schinazi, R. In vitro antimyxovirus and anti-human immunodeﬁciency virus activities of polyoxometalates. Antivir. Chem. Chemother. 6, 114–122 (1995). 78 AlDamen, M., Cardona-Serra, S., Clemente-Juan, J., Coronado, E., Gaita-Ariño, E., Martí-Gastaldo, C., Luis, F. & Montero, O. Mononuclear lanthanide single molecule magnets based on the polyoxometalates [Ln(W5O18)2]9 − and [Ln(2-SiW11O39)2]13−(LnIII) Tb, Dy, Ho, Er, Tm, and Yb. Inorg. Chem. 48, 3467–3479 (2009). This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license,userswillneedto obtainpermissionfromthelicense holderto reproduce the material. To view a copy of this license, visit http:// creativecommons.org/licenses/by/4.0/ 79 Goovaerts, V., Stroobants, K., Absillis, G. & Parac-Vogt, T. Molecular interactions between serum albumin proteins and Keggin type polyoxometalates studied using luminescence spectroscopy. Phys. Chem. Chem. Phys. 15, 18378–18387 (2013). 80 Hungerford, G., Suhling, K. & Green, M. Luminescence enhancement of a europium containing polyoxometalate on interaction with bovine serum albumin. Photochem. Photobiol. Sci. 7, 734–737 (2008). 81 Gao, P., Zhang, S., Li, H., Zhang, T., Wu, Y. & Wu, L. A two-step binding process of Eu-containing polyoxometalates to bovine serum albumin. Langmuir 31, 10888–10896 (2015). entary Information accompanies the paper on the NPG Asia Materials website (http://www.nature.com/am) NPG Asia Materials
https://openalex.org/W2606548622	https://minesparis-psl.hal.science/hal-00691350/document	English	null	On quality control procedures for solar radiation and meteorological measures, from subhourly to montly average time periods	HAL (Le Centre pour la Communication Scientifique Directe)	2,012	cc-by	2,065	To cite this version: Bella Espinar, Philippe Blanc, Lucien Wald, Carsten Hoyer-Klick, Marion Schroedter Homscheidt, et al.. On quality control procedures for solar radiation and meteorological measures, from subhourly to montly average time periods. EGU General Assembly 2012, Apr 2012, Vienne, Austria. hal-00691350 On quality control procedures for solar radiation and meteorological measures, from subhourly to montly average time periods average time periods Bella Espinar, Philippe Blanc, Lucien Wald, Carsten Hoyer-Klick, Marion Schroedter Homscheidt, Thomas Wanderer Bella Espinar, Philippe Blanc, Lucien Wald, Carsten Hoyer-Klick, Marion Schroedter Homscheidt, Thomas Wanderer Bella Espinar, Philippe Blanc, Lucien Wald, Carsten Hoyer-Klick, Marion Schroedter Homscheidt, Thomas Wanderer Distributed under a Creative Commons Attribution 4.0 International License TYPES OF QUALITY CONTROL PROCEDURES (QCPs) Monthly Daily Hourly Sub-hourly (1-minute average except for WS which is 2-minute) Global Horizontal Irradiance GHI (Wm−2) QCP based on extrema [3,7] 0.03 GHItoa < GHI < 1.2 I0 QCP based on rare observations [2,3,13] 0.03 GHItoa < GHI < GHItoa QCP based on extrema [3,7] 0.03 GHItoa < GHI < 1.2 I0 QCP based on rare observations [2,3,13] 0.03 GHItoa < GHI < GHItoa QCP based on extrema [3,4,7] 0.03 GHItoa < GHI < min (1.2 I0, 1.5 I0 cos(SZA)1.2 + 100) QCP based on rare observations [4,5,7] 0.03 GHItoa < GHI < 1.2 I0 cos(SZA)1.2 + 50 QCP based on extrema [3,4,7] 0.03 GHItoa < GHI < min (1.2 I0,1.5 I0 cos(SZA)1.2 + 100) QCP based on rare observations [3,4] 0.03 GHItoa < GHI < 1.2 I0 cos(SZA)1.2 + 50 Step QCP [14] Maximum step for two following measures: 1000 W m-2 Beam Normal Irradiance BNI (Wm−2) QCP based on extrema [4] 0 < BNI < I0 QCP based on extrema [4] 0 < BNI < I0 QCP based on extrema [4] 0 < BNI < I0 QCP based on rare observations [4] 0 < BNI < 0.95 I0 cos(SZA)0.2 + 10 QCP based on extrema [4] 0 < BNI < I0 QCP based on rare observations [4] 0 < BNI < 0.95 I0 cos(SZA)0.2 + 10 Diffuse Horizontal Irradiance DHI (Wm−2) QCP based on extrema [3,7] 0.03 GHItoa < DHI < 0.8 I0 QCP based on extrema [3,7] 0.03 GHItoa < DHI < 0.8 I0 QCP based on extrema [3,4,5,7] 0.03 GHItoa < DHI < min (0.8 I0, 0.95 I0 cos(SZA)1.2 + 50) QCP based on rare observations [3,4] 0.03 GHItoa < DHI < 0.75 I0 cos(SZA)1.2 + 30 QCP based on extrema [3,4,7] 0.03 GHItoa < DHI < min (0.8 I0, 0.95 I0 cos(SZA)1.2 + 50) QCP based on rare observations [3,4] 0.03 GHItoa < DHI < 0.75 I0 cos(SZA)1.2 + 30 Temperatutre Temp (°C) QCP based on extrema [10,14] -90 < Temp < + 60 QCP based on rare observations [10,14] -80 < Temp < + 50 QCP based on extrema [10,14] -90 < Temp < + 60 QCP based on rare observations [10,14] -80 < Temp < + 50 QCP based on extrema [10,14] -90 < Temp < + 60 QCP based on rare observations [10,14] -80 < Temp < + 50 Step QCP [8,9] Maximum step for two following measures: 8 °C QCP based on extrema [10,14] -90 < Temp < + 60 QCP based on rare observations [10,14] -80 < Temp < + 50 Step QCP [11,14] Maximum step for two following measures: 3 °C Minimum step over the past 60 minutes: 0.1 °C Relative Humidity Hum (%) QCP based on extrema [14] 0 < Hum < 100 QCP based on extrema [14] 0 < Hum < 100 QCP based on extrema [14] 0 < Hum < 100 Step QCP [9] Maximum step for two following values: 30 % QCP based on extrema [14] 0 < Hum < 100 Step QCP [14] Maximum step for two following values: 10 % Minimum step over the past 120 minutes: 0.1 % Wind Speed WS (m s-1) Step QCP [8,9] maximum step for two following values: 15 m s-1 QCP based on extrema (2-min average) [10] 0 < WS < 150 QCP based on rare observations (2-min average) [10] 0 < WS < 80 Step QCP [11,14] maximum step for two following values (2-min average): 20 m s-1 Minimum step over the past 60 minutes except for no wind periods (2-minute average): 0.5 m s-1 Consistency checks [7] DHI ≤ 1.1 GHI [7] DHI ≤ 1.1 GHI [1,5,6,13] For GHI > 20 W m-2 (if not, test not possible) \|1 - GHI / ( BHI + DHI)\|= 0.15 DHI ≤ 1.1 GHI [4,7,12] For GHI > 50 W m-2 (if not, test not possible): DHI/GHI < 1.05, for SZA < 75° DHI/GHI < 1.10, for 93° > SZA > 75° For DHI+BHI > 50 W m-2 (if not, test not possible) \|1- GHI / (BHI+DHI)\| ≤ 0.08, for SZA < 75° \|1 - GHI / (BHI+DHI)\| ≤ 0.15, for 75°< SZA <93° GHItoa = Horizontal Irradiance at top of atmosphere; I0 = normal irradiance at the top of the atmosphere corrected from Sun-Earth distance; SZA = Solar Zenithal Angle (rad); BHI = Beam Horizontal Irradiance QCPs in bold have been modified compared to their corresponding bibliographic references. HAL Id: hal-00691350 https://minesparis-psl.hal.science/hal-00691350v1 Submitted on 29 Nov 2014 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 92 Controlling the quality of measurements of meterological variables and solar radiation. From sub-hourly to monthly average time periods Authors: Espinar, B.(1), Blanc, P. (1), Wald, L. (1), Hoyer-Klick, C. (2), Schroedter-Homscheidt, M.(2), Wanderer, T. (2) (1) MINES ParisTech, CEP – Centre énergétique et procédés, BP 207, 1 rue Claude Daunesse, 06904 Sophia Antipolis Cedex France (2) DLR, German Aerospace Center DLR, Wessling, Germany E G i U i 2012 Vi (A t i ) 22 27 A il 2012 ENDORSE GA n° 262892 2011 - 2013 www.endorse-fp7.eu Controlling the quality of measurements of meterological variables and solar radiation. From sub-hourly to monthly average time periods ENDORSE GA n° 262892 2011 - 2013 www.endorse-fp7.eu Authors: Espinar, B.(1), Blanc, P. (1), Wald, L. (1), Hoyer-Klick, C. (2), Schroedter-Homscheidt, M.(2), Wanderer, T. (2) (1) MINES ParisTech, CEP – Centre énergétique et procédés, BP 207, 1 rue Claude Daunesse, 06904 Sophia Antipolis Cedex France (2) DLR, German Aerospace Center DLR, Wessling, Germany European Geosciences Union 2012, Vienna (Austria) 22 – 27 April 2012 Long-term radiation and meteorological measurements are available from a large number of ground measuring stations. However, close examination of the data often reveals a lack of quality, often for extended periods of time. Quality Control Procedures (QCPs) are a measure of how well data serve the purpose for which they were produced. We have searched in the bibliographic references the QCPs that are available for the solar radiation and meteorological data series which are of interest in the field of renewable energies, grouping them together by average time periods, checking them and modifying them if necessary. All of the QCPs that are presented in this document are applicable for all latitudes; they are not optimized regionally nor seasonably with the aim of being generic. The reseach leading to those results has partly received funding from the European Union’s Seventh Framework Programme (FP7/2007-2013). TYPES OF QUALITY CONTROL PROCEDURES (QCPs) •Range QCPs: verify that values are within a specific range (based on extrema for physically possible values and based on rare observations for possible but extremely rare values). thin a specific range (based on extrema for physically possible values and based on rare observations ) •Step QCPs: aimed at detecting unrealistic increments or stagnations in the time series, with respect to their sampling and integration period. •Consistency QCPs: verify the consistencies between two or more independent time series. TYPES OF QUALITY CONTROL PROCEDURES (QCPs) Sub-hourly (1-minute average except for WS which is 2-minute) QCP based on extrema [3,4,7] 0.03 GHItoa < GHI < min (1.2 I0,1.5 I0 cos(SZA)1.2 + 100) QCP based on rare observations [3,4] 0.03 GHItoa < GHI < 1.2 I0 cos(SZA)1.2 + 50 Step QCP [14] Maximum step for two following measures: 1000 W m-2 QCP based on extrema [3,4,7] 0.03 GHItoa < GHI < min (1.2 I0, 1.5 I0 cos(SZA)1.2 + 100) QCP based on rare observations [4,5,7] 0.03 GHItoa < GHI < 1.2 I0 cos(SZA)1.2 + 50 QCP based on extrema [3,4,7] 0.03 GHItoa < DHI < min (0.8 I0, 0.95 I0 cos(SZA)1.2 + 50) QCP based on rare observations [3,4] 0.03 GHItoa < DHI < 0.75 I0 cos(SZA)1.2 + 30 Wind Speed WS (m s-1) Consistency checks [1,5,6,13] For GHI > 20 W m-2 (if not, test not possible) \|1 - GHI / ( BHI + DHI)\|= 0.15 DHI ≤ 1.1 GHI GHItoa = Horizontal Irradiance at top of atmosphere; I0 = normal irradiance at the top of the atmosphere corrected from Sun-Earth distance; SZA = Solar Zenithal Angle (rad); BHI = Beam Horizontal Irradiance. QCPs in bold have been modified compared to their corresponding bibliographic references. mal irradiance at the top of the atmosphere corrected from Sun-Earth distance; SZA = Solar Zenithal Angle (rad); BHI = Beam Horizontal Irradiance. onding bibliographic references. GHItoa = Horizontal Irradiance at top of atmosphere; I0 = normal irradiance at the top of the atmosphere corrected from Sun-Earth distance; SZA = Solar Zenithal Angle (rad); BHI = Beam Horizontal Irradiance. QCPs in bold have been modified compared to their corresponding bibliographic references. Quality Control (QC) Output Report: After checking the plausibility of the data, some graphs, histograms and figures of flagged values may help in the interpretation of the result of QC and also to detect some other features of data. References [1] CIE 1994, report no. CIE-108, International Commission on Illumination. [2] ESRA 2000, European Solar Radiation Atlas, Les Presses de l’Ecole des Mines de Paris (FRA). [3] Geiger et al. 2002, , Sol. Energy, 73, 475-480. [4] Long and Dutton2002, BSRN Global Network recommended QC tests, V2.0. [5] De Miguel et al. 2001, , Sol. Energy, 70, 143-153. [6] Molineaux and Ineichen 2003, , International Daylight Measurement Programme, available at: http://idmp.entpe.fr/library/software/aqccie.pdf,. [7] Muneer and Fairooz 2002, , Build. Serv. Eng. Res. T., 23, 151-166. TYPES OF QUALITY CONTROL PROCEDURES (QCPs) [8] NDBC 2009, Handbook of Automated Data Quality Control Checks and Procedures, NOAA. [9] Vejen et al. 2002, Quality control of meteorological observations. Automatic methods used in the Nordic countries,Norwegian Meteorological Institute. [10] WMO No. 305 [11] WMO No. 488 [12] WMO No. 557 [13] Younes et al. 2005, Energy, 30, 1533-1549. [14] Zahumenský 2004, Guidelines on Quality Control Procedures for Data from Automatic Weather Stations, World Meteorological Organization. QCPs in bold have been modified compared to their corresponding bibliographic references. Quality Control (QC) Output Report: After checking the plausibility of the data, some graphs, histograms and figures of flagged values may help in the interpretation of the result of QC and also to detect some other features of data. References [1] CIE 1994, report no. CIE-108, International Commission on Illumination. [2] ESRA 2000, European Solar Radiation Atlas, Les Presses de l’Ecole des Mines de Paris (FRA). [3] Geiger et al. 2002, , Sol. Energy, 73, 475-480. [4] Long and Dutton2002, BSRN Global Network recommended QC tests, V2.0. [5] De Miguel et al. 2001, , Sol. Energy, 70, 143-153. [6] Molineaux and Ineichen 2003, , International Daylight Measurement Programme, available at: http://idmp.entpe.fr/library/software/aqccie.pdf,. [7] Muneer and Fairooz 2002, , Build. Serv. Eng. Res. T., 23, 151-166. [8] NDBC 2009, Handbook of Automated Data Quality Control Checks and Procedures, NOAA. [9] Vejen et al. 2002, Quality control of meteorological observations. Automatic methods used in the Nordic countries,Norwegian Meteorological Institute. [10] WMO No. 305 [11] WMO No. 488 [12] WMO No. 557 [13] Younes et al. 2005, Energy, 30, 1533-1549. [14] Zahumenský 2004, Guidelines on Quality Control Procedures for Data from Automatic Weather Stations, World Meteorological Organization. ). e d Quality Control (QC) Output Report: After checking the plausibility of the data, some graphs, histograms and figures of flagged values may help in the interpretation of the result of QC and also to detect some other features of data.
https://openalex.org/W2990805447	https://europepmc.org/articles/pmc6930547?pdf=render	English	null	A Subcellular Quantitative Proteomic Analysis of Herpes Simplex Virus Type 1-Infected HEK 293T Cells	Molecules/Molecules online/Molecules annual	2,019	cc-by	11,351	Received: 5 October 2019; Accepted: 16 November 2019; Published: 20 November 2019 Abstract: Herpes simplex virus type 1 (HSV-1) is widespread double-stranded DNA (dsDNA) virus that establishes life-long latency and causes diverse severe symptoms. The mechanisms of HSV-1 infection and HSV-1’s interactions with various host cells have been studied and reviewed extensively. Type I interferons were secreted by host cells upon HSV infection and play a vital role in controlling virus proliferation. A few studies, however, have focused on HSV-1 infection without the presence of interferon (IFN) signaling. In this study, HEK 293T cells with low toll-like receptor (TLR) and stimulator of interferon genes protein (STING) expression were infected with HSV-1 and subjected to a quantitative proteomic analysis. By using a subcellular fractionation strategy and high-performance mass spectrometry, a total of 6607 host proteins were quantiﬁed, of which 498 proteins were diﬀerentially regulated. A bioinformatics analysis indicated that multiple signaling pathways might be involved in HSV-1 infection. A further functional study indicated the role of Interferon-induced transmembrane protein 3 (IFITM3), Coiled-coil-helix-coiled-coil-helix domain-containing protein 2 (CHCHD2), and Tripartite motif-containing protein 27 (TRIM27) in inhibiting viral DNA replication and proliferation. Our data provide a global view of host responses to HSV-1 infection in HEK 293T cells and identify the proteins involved in the HSV-1 infection process. Keywords: quantitative proteomics; herpes simplex virus type 1; virus–host interaction; IFITM3; CHCHD2; TRIM27 molecules molecules A Subcellular Quantitative Proteomic Analysis of Herpes Simplex Virus Type 1-Infected HEK 293T Cells Weiwei Wan 1,2,†, Liangjie Wang 3,†, Xi Chen 4,5, Shenglin Zhu 1,2, Weijuan Shang 1, Gengfu Xiao 1,2,* and Lei-Ke Zhang 1,2,* Weiwei Wan 1,2,†, Liangjie Wang 3,†, Xi Chen 4,5, Shenglin Zhu 1,2, Weijuan Shang 1, Gengfu Xiao 1,2,* and Lei-Ke Zhang 1,2,* 1 State Key Laboratory of Virology, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan 430071, China; weiwei.wan.cas.whiov@gmail.com (W.W.); cqubiozsl@gmail.com (S.Z.); shangweijuan@wh.iov.cn (W.S.) 1 State Key Laboratory of Virology, Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan 430071, China; weiwei.wan.cas.whiov@gmail.com (W.W.); cqubiozsl@gmail.com (S.Z.); shangweijuan@wh.iov.cn (W.S.) g j 2 University of the Chinese Academy of Sciences, Beijing 100049, China 2 University of the Chinese Academy of Sciences, Beijing 100049, China University of the Chinese Academy of Sciences, Beijing 100049, China 3 Hubei Key Laboratory of Puriﬁcation and Application of Plant Anti-Cancer Active Ingredients, School of Chemistry and Life Sciences, Hubei University of Education, Wuhan 430205, China; wangliangjie@hue.edu.cn 3 Hubei Key Laboratory of Puriﬁcation and Application of Plant Anti-Cancer Active Ingredients, School of Chemistry and Life Sciences, Hubei University of Education, Wuhan 430205, China; wangliangjie@hue.edu.cn 4 Department of Biological Mass Spectrometry, Wuhan Institute of Biotechnology, Wuhan 430074, China; chenxi@spec-ally.com 5 Medical Research Institute, Wuhan University, Wuhan 430074, China * Correspondence: xiaogf@wh.iov.cn (G.X.); zhangleike@wh.iov.cn (L.-K.Z.) † These authors contributed equally to this work. 1. Introduction Herpes simplex virus type 1 (HSV-1), a member of the alpha-herpesvirus family, is a contagious human pathogen. According to a WHO report [1], HSV-1 is estimated to have infected 67% of the world’s population under the age of ﬁfty. HSV-1 infection establishes life-long latency with reactivation triggered by stimuli like UV light, tissue damage, or compromised immunity [2]. The severity of HSV-1 infection varies from mild symptoms, such as cold sores and watery blisters on the skin or mucous membranes, to life-threatening encephalitis in humans [3]. Molecules 2019, 24, 4215; doi:10.3390/molecules24234215 www.mdpi.com/journal/molecules www.mdpi.com/journal/molecules Molecules 2019, 24, 4215 2 of 13 As a double-stranded DNA virus, HSV-1 has a linear genome of 152 kbp surrounded by a protein layer called a tegument and a bilayer envelope with viral glycoproteins [4]. During infection, the virus ﬁrst binds the receptors present on host cell surfaces through glycoproteins of the viral envelope and then transitions from attachment to penetration, leading to the release of the viral capsid and tegument proteins into the cytoplasm. Later, viral capsids with tegument proteins use the microtubule network to transfer to the nuclear envelope, where they interact with nuclear pores and release their uncoated genomes into the nucleoplasm to perform viral transcription and replication. During these processes, the viruses will utilize various host cell systems to promote their own replication. As a powerful tool with high throughput and accuracy, MS-based proteomic analyses have been used increasingly in the analysis of virus–host interactions, aiming to discover new drug targets or unveil unknown mechanisms. Until now, proteomic analyses of HSV-1 have focused on virally induced changes in the cellular proteome [5–7] or protein post-translational modiﬁcations [8] and virus–protein interactions [9–11] during infection. In this study, a stable isotope-labeled amino acid culture (SILAC)-based quantitative proteomic analysis of HSV-1 infected HEK 293T cells was performed, aiming to explore the eﬀects of HSV-1 infection on the host proteome. An RNAi-based functional analysis was employed to explore the role of several regulated host proteins in viral replication. In addition, the host cell activates various antiviral responses to limit viral infection. The entry of HSV-1 into host cells triggers a series of both innate and adaptive immune responses, in which innate immunity plays a central role in determining the ﬁnal fate of infection [12]. 1. Introduction HSV-1 and Sendai virus (SeV) have been employed as models for DNA and RNA viruses, respectively, and have been extensively used in virus–host interaction studies, especially in studies investigating the mechanisms of the host’s immune response to DNA viruses and RNA viruses. A comparative analysis of the eﬀects of SeV and HSV-1 infection on the host proteome has never been reported. Our previous SILAC-based quantitative analysis of SeV-infected HEK 293T cells highlighted several biological processes that may be involved in the viral replication process and virus-induced innate immune response [13]. In this study, we also compared the eﬀects of SeV infection and HSV-1 infection on HEK 293T cells, and the biological processes regulated by both HSV-1 and SeV were identiﬁed. Our quantitative proteomic analysis of HSV-1- and SeV-infected cells can also help us to better understand HSV-1- or SeV-induced immune responses. markers of cytoplasmic and nuclear proteins, respectively. H: hours. 2 2 SILAC B d Q tit ti P t i A l i f HSV 1 i f t d HEK 293T C ll 2.2. SILAC-Based Quantitative Proteomic Analysis of HSV-1 Infected HEK 293T Cells 2.2. SILAC-Based Quantitative Proteomic Analysis of HSV-1 infected HEK 293T Cells To perform a quantitative proteomic analysis of HSV-1-infected host cells, HEK 293T cells were differentially labeled by being cultured in triple-SILAC media for seven passes, and then infected with HSV-1 at a multiplicity of infection (MOI) of 5 or mock treated (Figure 1B). Then, the infected cells were harvested at 4 h p.i. and 20 h p.i., and the mock cells were collected and mixed in an equal amount. The cells were then subcellularly fractionated into cytoplasmic and nuclear fractions, and the subcellular fractionation efficiency was confirmed by detecting lamin B1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using Western blot (Figure 1C). Then, the fractionated proteins were digested into peptides and further fractionated with strong cation To perform a quantitative proteomic analysis of HSV-1-infected host cells, HEK 293T cells were diﬀerentially labeled by being cultured in triple-SILAC media for seven passes, and then infected with HSV-1 at a multiplicity of infection (MOI) of 5 or mock treated (Figure 1B). Then, the infected cells were harvested at 4 h p.i. and 20 h p.i., and the mock cells were collected and mixed in an equal amount. The cells were then subcellularly fractionated into cytoplasmic and nuclear fractions, and the subcellular fractionation eﬃciency was conﬁrmed by detecting lamin B1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using Western blot (Figure 1C). Then, the fractionated proteins were digested into peptides and further fractionated with strong cation exchange chromatography (SCX) for MS analysis. exchange chromatography (SCX) for MS analysis. Three independent biological replicates were subjected to an MS analysis, and ProteinPilot v 5.0 was used for peptide identification and quantitation. Peptides with a confidence score above 95% and a ratio >0 and <99 were considered as being quantified. Proteins with at least one quantified peptide were included for further analyses. Using these criteria, a total of 6607 host proteins were quantified [false discovery rate (FDR) < 1%)], among which 5784 were from the cytoplasmic fraction d 3382 f th l f ti (Fi 2A) I th t l i f ti th b f Three independent biological replicates were subjected to an MS analysis, and ProteinPilot v 5.0 was used for peptide identiﬁcation and quantitation. Peptides with a conﬁdence score above 95% and a ratio >0 and <99 were considered as being quantiﬁed. Proteins with at least one quantiﬁed peptide were included for further analyses. Figure 1. Subcellular quantitative proteomic analysis of herpes simplex virus type 1 (HSV-1)-infected HEK 293T cells. (A) Intracellular levels of HSV-1 genome DNA and IFNB1 and ISG56 mRNAs in HSV-1 infected HEK 293T cells. Mock- or HSV-1 (MOI = 5)-infected HEK 293T cells were harvested at 4 h p.i. and 20 h p.i. Total RNA was extracted and reverse transcribed into the cDNA to quantify the intracellular mRNA level of IFNB1 and ISG56 using quantitative RT-PCR. The total DNA was extracted, and the intracellular DNA level of the HSV-1 genome was measured with quantitative RT-PCR. (B) The MS analysis workflow of the stable isotope-labeled amino acid culture (SILAC). (C) Confirming the subcellular fractionation efficiency by Western blot. Cytoplasmic and nuclear fractions from the three biological replicates (R1, R2, and R3) were subjected to Western blot analyses. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and lamin B1 were used as markers of cytoplasmic and nuclear proteins respectively H: hours Figure 1. Subcellular quantitative proteomic analysis of herpes simplex virus type 1 (HSV-1)-infected HEK 293T cells. (A) Intracellular levels of HSV-1 genome DNA and IFNB1 and ISG56 mRNAs in HSV-1 infected HEK 293T cells. Mock- or HSV-1 (MOI = 5)-infected HEK 293T cells were harvested at 4 h p.i. and 20 h p.i. Total RNA was extracted and reverse transcribed into the cDNA to quantify the intracellular mRNA level of IFNB1 and ISG56 using quantitative RT-PCR. The total DNA was extracted, and the intracellular DNA level of the HSV-1 genome was measured with quantitative RT-PCR. (B) The MS analysis workﬂow of the stable isotope-labeled amino acid culture (SILAC). (C) Conﬁrming the subcellular fractionation eﬃciency by Western blot. Cytoplasmic and nuclear fractions from the three biological replicates (R1, R2, and R3) were subjected to Western blot analyses. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and lamin B1 were used as markers of cytoplasmic and nuclear proteins, respectively. H: hours. 2.1. HSV-1 Replicates Eﬀectively in HEK 293T Cells without Inducing IFN-β Production To explore the eﬀects of HSV-1 infection on the proteome of HEK 293T cells, we performed a SILAC-based quantitative proteomic analysis of HSV-1-infected HEK 293T cells at 4 and 20 h post infection (p.i.), which represent the early and late phases of HSV-1 infection, respectively. HEK 293T cells are highly permissive to HSV-1 infection and have low TLR and STING expression, and thus their innate immune response during infection with a DNA virus is weak, resulting in weak activation of Interferon-stimulated genes (ISGs) and related biological processes. We performed a quantitative proteomic analysis on HEK 293T cells to minimize the eﬀect of interferons on the regulation of host proteins and biological processes. Twenty hours p.i. was chosen as the time point for late phase, because we found that the HSV-1 DNA level was much higher (Figure 1A), and no signiﬁcant apoptosis can be observed at this time point. This observation is consistent with the understanding that HSV-1 takes around 16–20 h to complete a lytic cycle [5]. However, the intracellular mRNA levels of IFNB1 and ISG56 were comparable between 4 and 20 h p.i., suggesting that type I IFN (IFN-I) production was not activated in HSV-1-infected HEK 293T cells, which is consistent with the literature [14,15]. 3 of 13 Molecules 2019, 24, 4215 markers of cytoplasmic and nuclear proteins, respectively. H: hours. 2 2 SILAC B d Q tit ti P t i A l i f HSV 1 i f t d HEK 293T C ll 2.2. SILAC-Based Quantitative Proteomic Analysis of HSV-1 Infected HEK 293T Cells In the nuclear fraction, 3225 proteins were quantified at 4 h p.i. and 3266 proteins were quantified at 20 h p.i. H: hours. (B–E) Figure 2. Overview of proteins quantiﬁed by SILAC-MS. (A) A Venn diagram of the quantiﬁed host proteins. A total of 6607 host proteins were quantiﬁed, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantiﬁed at 4 h p.i. and 20 h p.i., respectively. In the nuclear fraction, 3225 proteins were quantiﬁed at 4 h p.i. and 3266 proteins were quantiﬁed at 20 h p.i. H: hours. (B–E) Volcano plots of the diﬀerentially expressed proteins between the HSV-1- and mock-infected HEK 293T cells. Figure 2. Overview of proteins quantified by SILAC-MS. (A) A Venn diagram of the quantified host proteins. A total of 6607 host proteins were quantified, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantified at 4 h p.i. and 20 h p.i., respectively. In the nuclear fraction, 3225 proteins were quantified at 4 h p.i. and 3266 proteins were quantified at 20 h p.i. H: hours. (B–E) Figure 2. Overview of proteins quantiﬁed by SILAC-MS. (A) A Venn diagram of the quantiﬁed host proteins. A total of 6607 host proteins were quantiﬁed, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantiﬁed at 4 h p.i. and 20 h p.i., respectively. In the nuclear fraction, 3225 proteins were quantiﬁed at 4 h p.i. and 3266 proteins were quantiﬁed at 20 h p.i. H: hours. (B–E) Volcano plots of the diﬀerentially expressed proteins between the HSV-1- and mock-infected HEK 293T cells. Figure 2. Overview of proteins quantiﬁed by SILAC-MS. (A) A Venn diagram of the quantiﬁed host Volcano plots of the differentially expressed proteins between the HSV 1 and mock infected HEK 293T cells. 2.3. Validation of SILAC-MS Data by Western Blot We then performed Western blots on two differentially regulated proteins: Interferon regulatory factor 3 (IRF3) (category B) and IFITM3 (category C). markers of cytoplasmic and nuclear proteins, respectively. H: hours. 2 2 SILAC B d Q tit ti P t i A l i f HSV 1 i f t d HEK 293T C ll 2.2. SILAC-Based Quantitative Proteomic Analysis of HSV-1 Infected HEK 293T Cells As shown in Figure 3A, the Western blots indicated that IRF3 and IFITM3 were upregulated at 20 h p i but hardly at 4 h p i in The average values of the protein ratio from the three biological replicates and the p-value were calculated and plotted on a volcano plot (Figure 2B–E). Regulated proteins were divided into three categories: category A (high conﬁdence), category B (medium conﬁdence), and category C (low conﬁdence). Based on these criteria, 498 proteins were diﬀerentially regulated with high conﬁdence; 483 proteins were diﬀerentially regulated with medium conﬁdence; and 923 proteins were diﬀerentially regulated with low conﬁdence (Table S1). markers of cytoplasmic and nuclear proteins, respectively. H: hours. 2 2 SILAC B d Q tit ti P t i A l i f HSV 1 i f t d HEK 293T C ll 2.2. SILAC-Based Quantitative Proteomic Analysis of HSV-1 Infected HEK 293T Cells Using these criteria, a total of 6607 host proteins were quantiﬁed [false discovery rate (FDR) < 1%)], among which 5784 were from the cytoplasmic fraction and 3382 were from the nuclear fraction (Figure 2A). In the cytoplasmic fraction, the numbers of quantiﬁed 4 of 13 o three C (lo Molecules 2019, 24, 4215 calculated and plott proteins at 4 h p.i. and at 20 h p.i. were 5516 and 5661, respectively. In the nuclear fraction, 3225 quantiﬁed proteins were from 4 h p.i. and 3266 quantiﬁed proteins were from 20 h p.i. (Figure 2A). confidence). Based on these criteria, 498 proteins were differentially regulated with high confidence; 483 proteins were differentially regulated with medium confidence; and 923 proteins were differentially regulated with low confidence (Table S1). Figure 2. Overview of proteins quantified by SILAC-MS. (A) A Venn diagram of the quantified host proteins. A total of 6607 host proteins were quantified, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantified at 4 h p.i. and 20 h p.i., respectively. In the nuclear fraction, 3225 proteins were quantified at 4 h p.i. and 3266 proteins were quantified at 20 h p.i. H: hours. (B–E) V l l t f th diff ti ll d t i b t th HSV 1 d k i f t d HEK Figure 2. Overview of proteins quantiﬁed by SILAC-MS. (A) A Venn diagram of the quantiﬁed host proteins. A total of 6607 host proteins were quantiﬁed, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantiﬁed at 4 h p.i. and 20 h p.i., respectively. In the nuclear fraction, 3225 proteins were quantiﬁed at 4 h p.i. and 3266 proteins were quantiﬁed at 20 h p.i. H: hours. (B–E) Volcano plots of the diﬀerentially expressed proteins between the HSV-1- and mock-infected HEK 293T cells. Figure 2. Overview of proteins quantified by SILAC-MS. (A) A Venn diagram of the quantified host proteins. A total of 6607 host proteins were quantified, among which 5784 were from the cytoplasmic fraction (cyto) and 3382 were from the nuclear fraction (nucleo). In the cytoplasmic fraction, 5516 and 5661 proteins were quantified at 4 h p.i. and 20 h p.i., respectively. the cytoplasm. The SILAC-MS data indicated th th t l b t l t d t 20 h i B th 2.3. Validation of SILAC-MS Data by Western Blot Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not detected; H: hours. ; 2.4. Gene Ontology (GO) Analysis Reveals Biological Processes Regulated by HSV-1 Infection The differentially regulated proteins in both fractions at both time points were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) to perform a GO analysis based on “biological process” separately, with all the proteins quantified in all the fractions as the background (Table S2). As shown in Figure 4A–D, the categories “positive regulation of I-kappaB kinase/NF-kappaB signaling”, “positive regulation of GTPase activity”, “protein transport”, and “ion transmembrane transport” were over-represented by the regulated proteins at 4 h p.i. in the cytoplasmic fraction, while the categories “transcription from RNA polymerase II A quantitative RT-PCR analysis was performed on ﬁve regulated proteins. As demonstrated in Figure 3B, the intracellular mRNA level of CHCHD2 was downregulated at 20 h p.i. but hardly at 4 h p.i., while the intracellular mRNA levels of multidrug resistance-associated protein 1 (ABCC1), Caspase-10 (CASP10), and IFITM3 were upregulated at 20 h p.i., and the intracellular mRNA level CASP10 also increased at 4 h p.i. the cytoplasm. The SILAC-MS data indicated th th t l b t l t d t 20 h i B th 2.3. Validation of SILAC-MS Data by Western Blot the cytoplasm, but upregulated at 20 h p.i. Both the Western blots and SILAC-MS analysis did not detect IRF3 and IFITM3 in the nucleus. Although the confidence of regulation information of IRF3 and IFITM3 was low, our Western blot results confirmed their regulations. A quantitative RT-PCR analysis was performed on five regulated proteins. As demonstrated in Figure 3B, the intracellular mRNA level of CHCHD2 was downregulated at 20 h p.i. but hardly at 4 h p.i., while the intracellular mRNA levels of multidrug resistance-associated protein 1 (ABCC1), Caspase-10 (CASP10), and IFITM3 were upregulated at 20 h p.i., and the intracellular mRNA level CASP10 also increased at 4 h p.i. These regulations were consistent with the SILAC-MS data (Figure We then performed Western blots on two diﬀerentially regulated proteins: Interferon regulatory factor 3 (IRF3) (category B) and IFITM3 (category C). As shown in Figure 3A, the Western blots indicated that IRF3 and IFITM3 were upregulated at 20 h p.i., but hardly at 4 h p.i. in the cytoplasm. The SILAC-MS data indicated that IRF3 and IFITM3 were not regulated at 4 h p.i. in the cytoplasm, but upregulated at 20 h p.i. Both the Western blots and SILAC-MS analysis did not detect IRF3 and IFITM3 in the nucleus. Although the conﬁdence of regulation information of IRF3 and IFITM3 was low, our Western blot results conﬁrmed their regulations. 5 of 13 protein HSV-1 Molecules 2019, 24, 4215 level, which agreed w 0 [11]. Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not detected; H: hours. Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. the cytoplasm. The SILAC-MS data indicated th th t l b t l t d t 20 h i B th 2.3. Validation of SILAC-MS Data by Western Blot Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not detected; H: hours. [ ] Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not detected; H: hours. Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. Proteins from the cytoplasmic and nuclear fractions of the mock- or HSV-1-infected (MOI = 5) HEK 293T cells were extracted for Western blot analysis. Lamin B1 and GAPDH were used as internal controls for nuclear (nucleo) and cytoplasmic (cyto) proteins, respectively. (B) Quantitative RT-PCR analysis of the intracellular mRNA level of selected proteins. Mock- or HSV-1-infected (MOI = 5) HEK 293T cells were harvested, and the total mRNA was extracted and reverse transcribed into cDNA for quantitative RT-PCR analysis. The values are presented as the mean ± SD of three replicates. (C) SILAC-MS data for selected proteins. ND: not detected; H: hours. A quantitative RT-PCR analysis was performed on ﬁve regulated proteins. As demonstrated i [ ] [ ] Figure 3. Validation of the protein regulation data by quantitative RT-PCR and Western blots. (A) Western blot analysis of the host proteins. the cytoplasm. The SILAC-MS data indicated th th t l b t l t d t 20 h i B th 2.3. Validation of SILAC-MS Data by Western Blot These regulations were consistent with the SILAC-MS data (Figure 3C), suggesting that these proteins may be regulated at the transcriptional level. TRIM27 was identiﬁed by the MS data as being downregulated, while the intracellular mRNA level of TRIM27 did not change signiﬁcantly, suggesting that the downregulation of TRIM27 occurred at the protein level, which agreed with a recent study and indicated that TRIM27 is a degradation target of HSV-1 ICP0 [11]. promoter”, “viral budding via host endosomal sorting complexes required for transp complex” “toll like receptor signaling pathway” and “Wnt signaling path 2.4. Gene Ontology (GO) Analysis Reveals Biological Processes Regulated by HSV-1 Infection promoter”, “viral budding via host endosomal sorting complexes required for transp complex” “toll like receptor signaling pathway” and “Wnt signaling path 2.4. Gene Ontology (GO) Analysis Reveals Biological Processes Regulated by HSV-1 Infection complex , toll like receptor signaling pathway , and Wnt signaling pathway were over-represented by the regulated proteins at 4 h p.i. in the nuclear fraction. The categories “signal transduction”, “ion trans-membrane transport”, “response to oxidative stress”, and “response to interferon-gamma” were over-represented by the regulated proteins at 20 h p.i. in the cytoplasmic fraction, while “transcription from RNA polymerase II promoter”, “Wnt signaling pathway”, and The diﬀerentially regulated proteins in both fractions at both time points were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) to perform a GO analysis based on “biological process” separately, with all the proteins quantiﬁed in all the fractions as the background (Table S2). As shown in Figure 4A–D, the categories “positive regulation of I-kappaB kinase/NF-kappaB signaling”, “positive regulation of GTPase activity”, “protein transport”, and “ion transmembrane transport” were over-represented by the regulated proteins at 4 h p.i. in the cytoplasmic fraction, while the categories “transcription from RNA polymerase II promoter”, “viral budding via host endosomal sorting complexes required for transport (ESCRT) complex”, “toll-like receptor signaling pathway”, and “Wnt signaling pathway” were over-represented by the regulated 6 of 13 a 0 ocesses l i l Molecules 2019, 24, 4215 p g p.i. in the nuclear f proteins at 4 h p.i. in the nuclear fraction. The categories “signal transduction”, “ion trans-membrane transport”, “response to oxidative stress”, and “response to interferon-gamma” were over-represented by the regulated proteins at 20 h p.i. in the cytoplasmic fraction, while “transcription from RNA polymerase II promoter”, “Wnt signaling pathway”, and “positive regulation of T-cell proliferation” were over-represented by the regulated proteins at 20 h p.i. in the nuclear fraction. The over-represented categories suggest that these biological processes are more vulnerable under HSV-1 infection than other pathways. More regulated biological processes were identiﬁed in the nuclear fraction, possibly because the HSV-1 genome mainly replicates in the nucleus of infected cells. processes were identified in the nuclear fraction, possibly because the HSV-1 genome mainly replicates in the nucleus of infected cells. We also compared our data with previous quantitative proteomic analyses of SeV-infected HEK293T cells and found that 75 regulated proteins were common. promoter”, “viral budding via host endosomal sorting complexes required for transp complex” “toll like receptor signaling pathway” and “Wnt signaling path 2.4. Gene Ontology (GO) Analysis Reveals Biological Processes Regulated by HSV-1 Infection A GO analysis based on biological processes indicated that these proteins were enriched in the following categories: “cell migration”, “receptor-mediated endocytosis”, and “positive regulation of NF-kappaB transcription factor activity” (Figure 4E), suggesting that these biological processes can be regulated by both HSV-1 and SeV infections. Figure 4. Gene ontology analysis of differentially regulated proteins in HSV-1-infected HEK 293T cells. Regulated proteins identified at 4 h p.i. (A) or 20 h p.i. (C) in cytoplasmic (cyto) or regulated proteins identified at 4 h p.i. (B) or 20 h p.i. (D) in nuclear (nucleo) fractions were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) website for gene ontology (GO) analysis, respectively. All proteins quantified in this study were set as the background. H: hours. (E) Regulated proteins in both SeV- and HSV-1-infected cells were submitted Figure 4. Gene ontology analysis of diﬀerentially regulated proteins in HSV-1-infected HEK 293T cells. Regulated proteins identiﬁed at 4 h p.i. (A) or 20 h p.i. (C) in cytoplasmic (cyto) or regulated proteins identiﬁed at 4 h p.i. (B) or 20 h p.i. (D) in nuclear (nucleo) fractions were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) website for gene ontology (GO) analysis, respectively. All proteins quantiﬁed in this study were set as the background. H: hours. (E) Regulated proteins in both SeV- and HSV-1-infected cells were submitted to DAVID to perform a GO analysis based on biological processes. Figure 4. Gene ontology analysis of diﬀerentially regulated proteins in HSV-1-infected HEK 293T cells. Figure 4. Gene ontology analysis of differentially regulated proteins in HSV-1-infected HEK 293T cells. Regulated proteins identified at 4 h p.i. (A) or 20 h p.i. (C) in cytoplasmic (cyto) or regulated proteins identified at 4 h p.i. (B) or 20 h p.i. (D) in nuclear (nucleo) fractions were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) website for gene ontology (GO) analysis, respectively. All proteins quantified in this study were set as the background. H: hours. (E) Regulated proteins in both SeV- and HSV-1-infected cells were submitted Figure 4. Gene ontology analysis of diﬀerentially regulated proteins in HSV 1 infected HEK 293T cells. Regulated proteins identiﬁed at 4 h p.i. (A) or 20 h p.i. (C) in cytoplasmic (cyto) or regulated proteins identiﬁed at 4 h p.i. (B) or 20 h p.i. promoter”, “viral budding via host endosomal sorting complexes required for transp complex” “toll like receptor signaling pathway” and “Wnt signaling path 2.4. Gene Ontology (GO) Analysis Reveals Biological Processes Regulated by HSV-1 Infection (D) in nuclear (nucleo) fractions were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) website for gene ontology (GO) analysis, respectively. All proteins quantiﬁed in this study were set as the background. H: hours. (E) Regulated proteins in both SeV- and HSV-1-infected cells were submitted to DAVID to perform a GO analysis based on biological processes. p y g p 2.5. Roles of IFITM3, CHCHD2, and TRIM27 on Virus Proliferation To further explore the significance of our data, an interference RNA (RNAi)-based functional We also compared our data with previous quantitative proteomic analyses of SeV-infected HEK293T cells and found that 75 regulated proteins were common. A GO analysis based on biological processes indicated that these proteins were enriched in the following categories: “cell migration”, “receptor-mediated endocytosis”, and “positive regulation of NF-kappaB transcription factor activity” (Figure 4E), suggesting that these biological processes can be regulated by both HSV-1 and SeV infections. study on selected regulated proteins was performed. These p comparison with our pre ious quantitati e analysis of RNA iru 2.5. Roles of IFITM3, CHCHD2, and TRIM27 on Virus Proliferation The role of CHCHD2 in inhibiting HSV-1 replication has never been reported. Further studies on these proteins are necessary for further understanding of the underlying mechanisms and may help us discover new targets for antiviral treatments. Figure 5. Functional analysis of regulated proteins on virus proliferation. (A) The knockdown efficiency of siRNAs on targeted proteins in HEK 293T cells. HEK 293T cells were transfected with the indicated siRNAs, and at 24 h post transfection, the total mRNA was extracted and transcribed into cDNA for quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs were used as the negative control (NC). (B) The effects of the knockdown of selected proteins on virus DNA levels. HEK 293T cells were transfected with mock or siRNAs for selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Total DNA was harvested, and the intracellular level of the HSV-1 DNA was measured with quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs are indicated as NC. Y-axis: The intracellular level of HSV-1 DNA relative to NC. (C) The effects of the knockdown of selected proteins on virus titer in the supernatant. HEK 293T cells were transfected with mock or siRNAs for the selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Supernatants were collected and viral titers in the supernatants was measured. The values are presented as the mean ± SD of three replicates. The cells transfected with Figure 5. Functional analysis of regulated proteins on virus proliferation. (A) The knockdown eﬃciency of siRNAs on targeted proteins in HEK 293T cells. HEK 293T cells were transfected with the indicated siRNAs, and at 24 h post transfection, the total mRNA was extracted and transcribed into cDNA for quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs were used as the negative control (NC). (B) The eﬀects of the knockdown of selected proteins on virus DNA levels. HEK 293T cells were transfected with mock or siRNAs for selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Total DNA was harvested, and the intracellular level of the HSV-1 DNA was measured with quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs are indicated as NC. Y-axis: The intracellular level of HSV-1 DNA relative to NC. study on selected regulated proteins was performed. These p comparison with our pre ious quantitati e analysis of RNA iru 2.5. Roles of IFITM3, CHCHD2, and TRIM27 on Virus Proliferation comparison with our previous quantitative analysis of RNA virus infected host cells. CHCHD2 was To further explore the signiﬁcance of our data, an interference RNA (RNAi)-based functional study on selected regulated proteins was performed. These proteins were selected based on a comparison with our previous quantitative analysis of RNA virus-infected host cells. CHCHD2 was also regulated 7 of 13 8 of 13 Molecules 2019, 24, 4215 Molecules 2019, 24, x FOR by SeV [13], and IFITM3 was regulated by JEV [16], while TRIM27 was only regulated by HSV-1. HEK 293T cells were transfected with siRNAs against IFITM3, CHCHD2, and TRIM27, and 24 h later, the cells were collected, and intracellular mRNAs were harvested for quantitative RT-PCR to test the knockdown eﬃciency of these siRNAs (Figure 5A). All six pairs of siRNAs decreased gene expression signiﬁcantly and, therefore, were used for the following assays. Next, HEK 293T cells were transfected with the indicated siRNAs, and 24 h later, the cells were infected with HSV-1 at an MOI of 5. At 20 h p.i., both the cells and supernatants were collected, and the intracellular viral genomic DNA level was measured by quantitative RT-PCR. As shown in Figure 5B, the intracellular levels of HSV-1 DNA were signiﬁcantly higher in cells transfected with siRNAs against IFITM3, CHCHD2, or TRIM27, suggesting that the knockdown of IFITM3, CHCHD2, or TRIM27 increased the intracellular level of HSV-1 DNA. We also determined viral titers in the supernatants, and found that the knockdown of IFITM3, CHCHD2, or TRIM27 also increased the HSV-1 titer (Figure 5C). This implies that IFITM3, CHCHD2, and TRIM27 play positive roles in defending against HSV-1 infection. and may promote ZIKV replication and inhibit virus-induced IFN-I production [28]. We found that CHCHD2 was downregulated by HSV-1 infection in HEK 293T cells, and the knockdown of CHCHD2 can increase HSV-1 production in a cell culture supernatant (Figure 5). However, the detailed role of CHCHD2 on virus proliferation at the molecular level is still unknown, and more studies are needed. In sum, by using a quantitative proteomic analysis of HSV-1-infected HEK 293T cells at two time points, we provided a global view of host–viral protein interactions during infection in the absence of type I IFN. We identified multiple pathways that may participate in HSV-1 infection through a GO analysis. A functional assay suggested a positive role of three proteins: IFITM3, CHCHD2, and TRIM27 in the anti-HSV-1 process. study on selected regulated proteins was performed. These p comparison with our pre ious quantitati e analysis of RNA iru 2.5. Roles of IFITM3, CHCHD2, and TRIM27 on Virus Proliferation (C) The eﬀects of the knockdown of selected proteins on virus titer in the supernatant. HEK 293T cells were transfected with mock or siRNAs for the selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Supernatants were collected and viral titers in the supernatants was measured. The values are presented as the mean ± SD of three replicates. The cells transfected with scrambled siRNAs are indicated as NC. , p < 0.05; , p < 0.01. Figure 5. Functional analysis of regulated proteins on virus proliferation. (A) The knockdown efficiency of siRNAs on targeted proteins in HEK 293T cells. HEK 293T cells were transfected with the indicated siRNAs, and at 24 h post transfection, the total mRNA was extracted and transcribed into cDNA for quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs were used as the negative control (NC). (B) The effects of the knockdown of selected proteins on virus DNA levels. HEK 293T cells were transfected with mock or siRNAs for selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Total DNA was harvested, and the intracellular level of the HSV-1 DNA was measured with quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs are indicated as NC. Y-axis: The intracellular level of HSV-1 DNA relative to NC. (C) The effects of the knockdown of selected proteins on virus titer in the supernatant. HEK 293T cells were transfected with mock or siRNAs for the selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Supernatants were collected and viral titers in the supernatants was measured. The values are presented as the mean ± SD of three replicates. The cells transfected with Figure 5. Functional analysis of regulated proteins on virus proliferation. (A) The knockdown eﬃciency of siRNAs on targeted proteins in HEK 293T cells. HEK 293T cells were transfected with the indicated siRNAs, and at 24 h post transfection, the total mRNA was extracted and transcribed into cDNA for quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs were used as the negative control (NC). (B) The eﬀects of the knockdown of selected proteins on virus DNA levels. 3. Discussion HSV-1 is a widespread contagious pathogen and has been studied for decades. During the last few decades, MS-based proteomic methods have contributed signiﬁcantly to revealing more factors and mechanisms involved in HSV-1 infection and the corresponding host responses. A quantitative proteomic analysis using a 2D gel strategy was performed to explore the eﬀects of HSV-1 infection on the host proteome [5,6,17]. However, due to the limits of the gel-based strategy, the number of regulated proteins identiﬁed was low. Berard et al. also performed a SILAC-based quantitative proteomic analysis of HSV-1 infected cells [7], but they only used two replicates. In this study, a comprehensive analysis of HSV-1-infected HEK 239T cells was performed. To reduce the complexity of the proteins analyzed, cells were sub-fractionated into nuclear and cytoplasmic fractions, and three replicates were performed. A total of 6607 host proteins were quantiﬁed, of which 498 proteins were diﬀerentially regulated (Table S1). A GO analysis based on a biological process highlighted several proteins groups that were aﬀected by HSV-1 infection. The signiﬁcance of the NF-κB pathway [18] and the ubiquitin–proteasome system [19] on virus–host interactions during HSV infection have been reported repeatedly. In agreement with these reports, “NIK/NF-κB signaling” and “protein ubiquitination involved in ubiquitin-dependent protein catabolic process” were enriched in our GO data. The “positive regulation of GTPase activity”, “negative regulation of transcription from RNA polymerase II promoter”, and “negative regulation of cell proliferation” enriched in this study coincided with previous results that showed HSV can activate dynamin 2 GTPase [20] and alter the loading and positioning of RNA Polymerase II on host genes [21] and that the translation initiation factor eIF2α is inhibited by protein kinase RNA-activated (PKR) activation [22], respectively. IFN-inducible transmembrane (IFITM) protein 3 was a member of the ﬁrst identiﬁed ISGs. A wealth of data has suggested that IFITM3 strongly protects host cells against a broad range of viruses, including SARS coronavirus, inﬂuenza A virus, human immunodeﬁciency virus type 1, and Zika virus [23]. The antiviral eﬀects of IFITM3 might be attributed to the blockage of virus entry by inhibiting the fusion of the virus membrane, though the mechanisms are still unclear. One possible mechanism is that by binding with vesicle-associated membrane protein-associated protein A (VAPA), IFITM3 disrupts intracellular cholesterol homeostasis and blocks viral release to cytosol [24]. 3. Discussion This is consistent with our data and explains our ﬁnding that HSV-1 increases IFITM3 expression, and knockdown of this gene leads to higher virus proliferation. The role of the coiled-coil-helix-coiled-coil-helix domain containing 2 (CHCHD2) during herpesvirus infection has not been reported yet. CHCHD2 is a mitochondrial protein involved in apoptosis pathways by binding to BCL-xL and in stabilizing cytochrome c in the respiratory complex [25]. Its mutation has been linked to Parkinson’s disease [26]. Song et al. reported CHCHD2 expression in liver cancer and suggested that it might be a biomarker of hepatocellular carcinoma [27]. Our previous study found that CHCHD2 can be induced by Zika virus infection in HeLa cells and may promote ZIKV replication and inhibit virus-induced IFN-I production [28]. We found that CHCHD2 was downregulated by HSV-1 infection in HEK 293T cells, and the knockdown of CHCHD2 can increase HSV-1 production in a cell culture supernatant (Figure 5). However, the detailed role of CHCHD2 on virus proliferation at the molecular level is still unknown, and more studies are needed. In sum, by using a quantitative proteomic analysis of HSV-1-infected HEK 293T cells at two time points, we provided a global view of host–viral protein interactions during infection in the absence of type I IFN. We identiﬁed multiple pathways that may participate in HSV-1 infection through a GO analysis. A functional assay suggested a positive role of three proteins: IFITM3, CHCHD2, and TRIM27 in the anti-HSV-1 process. The role of CHCHD2 in inhibiting HSV-1 replication has never been reported. Further studies on these proteins are necessary for further understanding of the underlying mechanisms and may help us discover new targets for antiviral treatments. The role of the coiled-coil-helix-coiled-coil-helix domain containing 2 (CHCHD2) during herpesvirus infection has not been reported yet. CHCHD2 is a mitochondrial protein involved in apoptosis pathways by binding to BCL-xL and in stabilizing cytochrome c in the respiratory complex [25]. Its mutation has been linked to Parkinson’s disease [26]. Song et al. reported CHCHD2 expression in liver cancer and suggested that it might be a biomarker of hepatocellular carcinoma [27]. Our previous study found that CHCHD2 can be induced by Zika virus infection in HeLa cells and may promote ZIKV replication and inhibit virus-induced IFN-I production [28]. study on selected regulated proteins was performed. These p comparison with our pre ious quantitati e analysis of RNA iru 2.5. Roles of IFITM3, CHCHD2, and TRIM27 on Virus Proliferation HEK 293T cells were transfected with mock or siRNAs for selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Total DNA was harvested, and the intracellular level of the HSV-1 DNA was measured with quantitative RT-PCR analysis. Cells transfected with scrambled siRNAs are indicated as NC. Y-axis: The intracellular level of HSV-1 DNA relative to NC. (C) The eﬀects of the knockdown of selected proteins on virus titer in the supernatant. HEK 293T cells were transfected with mock or siRNAs for the selected genes, followed by a 24 h culture and then infected with HSV-1 at an MOI of 5. Supernatants were collected and viral titers in the supernatants was measured. The values are presented as the mean ± SD of three replicates. The cells transfected with scrambled siRNAs are indicated as NC. , p < 0.05; **, p < 0.01. Molecules 2019, 24, 4215 8 of 13 4.1. Cell and Virus HEK 293T cells obtained from CCTCC (China Center for Type Culture Collection, Wuhan, China) were grown in Dulbecco’s modiﬁed Eagle’s medium (DMEM, GIBCO, Grand Island, NE, USA) supplemented with 1% penicillin/streptomycin (GIBCO) and 10% fetal bovine serum (FBS, GIBCO) at 37 ◦C with 5% CO2. The HSV-1 strain 129 (H129) was used in this study, and virus was stocked at −80 ◦C. For SILAC labeling, HEK 293T cells were cultured in a SILAC medium (Thermo, Waltham, MA, USA) containing 10% dialyzed FBS (Thermo) and l-arginine and l-lysine (R0K0, “light”), l-arginine-13C6 and l-lysine-4,4,5,5-d4 (L) (R6K4, “medium”), or l-arginine-13C6, 15N4 and l-lysine-13C6, 15N2 (R10K8, “heavy”) (Sigma, St. Louis, MO, USA). Cells were cultured in the SILAC medium for at least seven doublings to ensure that the labeled amino acids were incorporated into cellular proteins. Cells cultured in diﬀerent mediums were infected with HSV-1 at an MOI of 5 or mock treated, and harvested at 4 or 20 h p.i. Then, mock-treated cells (light), HSV-1 infected cells harvested at 4 h p.i. (heavy), and HSV-1 infected cells harvested at 20 h p.i. (medium) were mixed in an equal amount. 3. Discussion We found that CHCHD2 was downregulated by HSV-1 infection in HEK 293T cells, and the knockdown of CHCHD2 can increase HSV-1 production in a cell culture supernatant (Figure 5). However, the detailed role of CHCHD2 on virus proliferation at the molecular level is still unknown, and more studies are needed. In sum, by using a quantitative proteomic analysis of HSV-1-infected HEK 293T cells at two time points, we provided a global view of host–viral protein interactions during infection in the absence of type I IFN. We identiﬁed multiple pathways that may participate in HSV-1 infection through a GO analysis. A functional assay suggested a positive role of three proteins: IFITM3, CHCHD2, and TRIM27 in the anti-HSV-1 process. The role of CHCHD2 in inhibiting HSV-1 replication has never been reported. Further studies on these proteins are necessary for further understanding of the underlying mechanisms and may help us discover new targets for antiviral treatments. 9 of 13 Molecules 2019, 24, 4215 4.3. Protein Extraction and Digestion Extracted proteins were collected and the protein concentration was measured with a BCA assay. Extracted proteins were reduced with 10 mM DTT at 56 ◦C for 30 min, alkylated with 40 mM IAA in the dark for 30 min, and left at room temperature for 1 h. The proteins were digested with trypsin (Promega, Madison, WI, USA) at a ratio of 1:50 (trypsin/protein w/w) at 37 ◦C overnight. Digested peptides were desalted with a SepPak C18 cartridge (Waters, Milford, CT, USA) and dried by SpeedVac (Thermo). Then, the digested peptides were fractionated with SCX, and the fractionated peptides were desalted with Ziptip (Millipore, Billerica, MA, USA) and stored at −80 ◦C before LC-MS/MS analysis. 4.2. Subcellular Fractionation Subcellular fractionation was performed with a nuclear and cytoplasmic protein extraction kit (Beyotime, Shanghai, China). The cell pellets were ﬁrst washed three times with pre-chilled PBS and then resuspended in pre-chilled nuclear and cytoplasmic protein extraction kit buﬀer A and incubated on ice for 15 min. Then, nuclear and cytoplasmic protein extraction kit buﬀer B was added, and the mixture was vortexed three times and centrifuged at 500 × g for 5 min. The supernatant was collected as the cytoplasmic fraction. The pellet was resuspended in nuclear and cytoplasmic protein extraction kit buﬀer C, incubated on ice for 15 min, and centrifuged at 13,200 × g for 15 min. The supernatant was then collected as the nuclear fraction. 4.5. Protein Identiﬁcation and Quantiﬁcation MS spectra were submitted to ProteinPilot version 5.0.1 (AB SCIEX) to perform peptide identiﬁcation and quantiﬁcation. The UniProt_Human database (download at 201701) containing HSV-1 proteins was used. The search parameters were as follows: sample type: SILAC (Lys+4D, Arg+6; Lys+8, Arg+10); cysteine alkylation: iodoacetamide; digestion: trypsin; instrument: tripleTOF 5600; miss cleavages tolerance: 2; ﬁxed modiﬁcation: carbamidomethyl Cys; variable modiﬁcation: none; MS1 initial mass error tolerance value: 0.05 Dalton; MS2 initial mass error tolerance value: 0.1 Dalton. The false discovery rate (FDR) analysis in ProteinPilot uses a “decoy database searching” strategy, and the FDRs of the ProteinPilot search results were all set as lower than 1% of the protein level. For protein quantiﬁcation, all quantiﬁed peptides were exported, and only peptides with a conﬁdence score >95% were kept for further analysis. In each replicate, the protein ratio was calculated by the weighted averaging ratios of its peptides, with peptide intensity as the weight. The protein ratio values used for the bioinformatic analysis were the weighted averages of the three biological replicates, while the P-value for the protein ratio was assessed using a student’s t-test (Table S1). All the quantiﬁed proteins that were presented contained at least one quantiﬁed peptide. One peptide quantitative information, as long as the peptide is unique to the protein, was also used as the quantitative information of the protein. The cut-oﬀfor diﬀerentially regulated proteins was set as described in the previous study [31]. Brieﬂy, the Gaussian distribution of the protein ratios was analyzed, and the values deviating from the mean of the normally distributed data by 1.96 standard deviations (SD) were considered as the cut-oﬀvalues. Regulated proteins were divided into three categories: category A: proteins with ratios greater than the upregulated or lesser than the downregulated cut-oﬀvalues and a p-value for a protein ratio of <0.05, and these proteins are high conﬁdence regulated proteins; category B: proteins with ratios greater than the upregulated or lesser than the downregulated cut-oﬀvalues in at least two independent replicates; category C: for proteins only quantiﬁed in one replicate, the ratios should be greater than the upregulated or lesser than the downregulated cut-oﬀvalues, and these are low conﬁdence regulated proteins. 4.4. Liquid Chromatography/Mass Spectroscopy (LC-MS/MS) Analysis To quantify the diﬀerentially expressed proteins in the samples, an LC-MS/MS analysis was performed as described previously [29]. A hybrid quadrupole-TOF LC/MS/MS mass spectrometer (TripleTOF 5600+, AB SCIEX) was coupled to a nanospray ion source. Peptides were ﬁrst loaded onto a 5 µm-C18 trap column (5 × 0.3 mm; Agilent Technologies, Santa Clara, CA, USA) and then separated on a C18 analytical column (75 µm × 150 mm, 3 µm particle size, 100 Å pore size; Eksigent, Dublin, OH, USA). A 100 min gradient was established using mobile phase A (3% DMSO, 96.9% H2O, and 0.1% formic acid) and mobile phase B (3% DMSO, 96.9% ACN, and 0.1% formic acid), with a constant ﬂow rate of 300 nL/min. For the MS/MS analysis, each scan cycle consisted of one full-scan mass spectrum (with the m/z ranging from 350 to 1500 and charge states from 2 to 5) followed by 20 MS/MS events. The threshold count was set to 120 to activate MS/MS accumulation, and the former target ion exclusion was set at 18 s. The tandem mass spectra were extracted by Peakview version 2.0 (AB SCIEX, Framingham, MA, USA). The mass spectrometry proteomics data 10 of 13 Molecules 2019, 24, 4215 were deposited to the ProteomeXchange Consortium via the PRIDE [30] partner repository with the dataset identiﬁer PXD015887. 4.6. Gene Ontology Analysis To perform the GO analysis, diﬀerentially regulated proteins were submitted to DAVID [32], with all the quantiﬁed proteins in this study set as the background. Proteins were classiﬁed into diﬀerent categories based on their roles in biological processes, and a statistical overrepresentation test was performed. P-values were assessed with a binomial test and corrected for multiple tests using a Bonferroni procedure. Only those categories with a p-value of 0.05 were considered over- or under-represented. 4.7. Cell Transfection For siRNA transfection, an siRNA was added in Opti-MEM and then mixed with Opti-MEM containing RNAiMAX (Invitrogen). The mixture was incubated at room temperature for 5 min, and 50 µL mixed solution was added to pre-seeded HEK 293T cells in 24-well plates per well. At 24 h post transfection, the cells were infected with HSV-1. At 20 h p.i., both the cell and supernatant were collected, and the viral genome and titer were measured. The ﬁnal concentration of the siRNA used was 40 nM. The siRNA sequences (5’–3’) were as follows: TRIM27#1-GAAUUAAGAGAGGCUCAGUUA; TRIM27#2-GCCCUACUUCAGUCUGAGUUA; IFITM3#1-UCGUCAUCCCAGUGCUGAU; IFITM3#2- CCCACGUACUCCAACUUCC; CHCHD2#1-ACAGAGCUUGAUGUCACCCUG; CHCHD2#2- CAGUGGAGGAAGUAAUGCUGA. 11 of 13 Molecules 2019, 24, 4215 11 of 13 Author Contributions: Conceptualization, L.-K.Z.; data curation, W.W.; funding acquisition, G.X. and L.-K.Z.; investigation, W.W., L.W., X.C., S.Z. and W.S.; methodology, W.W., L.W. and S.Z.; project administration, G.X. and L.-K.Z.; supervision, G.X. and L.-K.Z.; validation, W.W. and L.W.; writing—original draft, L.W.; writing—review and editing, W.W. and L.W. 4.9. Viral DNA Puriﬁcation and Quantitationn Viral DNA was puriﬁed with the TIANamp Genomic DNA Kit (Tiangen, Beijing, China). HSV-1-infected HEK 293T cells were collected and lysed with Buﬀer A from the TIANamp Genomic DNA Kit, and then the DNA was puriﬁed following the manufacturer’s protocol. The relative viral DNA level was measured with quantitative RT-PCR. SYBR Premix Ex Taq™(Applied Biosystems, Foster City, CA, USA) was used as the ﬂuorescent dye on Applied Biosystems 7500 Real-Time PCR Systems. The primers used were synthesized by Sangon (Shanghai, China) and the sequences were as follows (5’–3’): HSV-1, Forward: CGGCCGTGTGACACTATCG, Reverse: CTCGTAAAATGGCCCCTCC; GAPDH, Forward: GAAGGTGAAGGTCGGAGTC, Reverse: GAAGATGGTGATGGGATTTC. The concentration of the primers used for RT-PCR was 400 nM. Funding: This research was funded by the State Key Laboratory of Virology (grant no. Y606031GF1) and Hubei Provincial Department of Education (grant no. Q20193004). Acknowledgments: We thank The Center for Instrumental Analysis and Metrology and the Core Facility and Technical Support, Wuhan Institute of Virology for providing technical assistance. 4.8. Western Blot Analysis Cell samples were lysed with a radio immunoprecipitation assay (RIPA) lysis buﬀer (Beyotime, China) and subjected to SDS-PAGE. Proteins in the gel were transferred onto a polyvinylidene ﬂuoride (PVDF) membrane (Millipore) and blocked with 5% non-fat milk in TBST buﬀer. The blots on the membrane were detected by primary antibodies and corresponding horseradish peroxidase-conjugated secondary antibodies (ProteinTech, Wuhan, China). These antibodies were diluted in a TBST buﬀer containing 5% non-fat milk, and the membrane was washed with a TBST buﬀer containing 0.1% Tween-20. Finally, the membrane was visualized by enhanced chemiluminescence (ECL) (Millipore), and protein bands were visualized by image using the ChemDoc MP Imaging System (Bio-Rad, Hercules, CA, USA). A mouse monoclonal antibody against GAPDH (ABclonal, Wuhan, China), a rabbit monoclonal antibody against IFITM3 (CST, Danvers, MA, USA), and a rabbit polyclonal antibody against IRF3 (Proteintech) and Lamin B1 (Proteintech) were purchased from indicated companies. 4.9. Viral DNA Puriﬁcation and Quantitationn 4.10. Intracellular RNA Extraction and Quantitation After removal of the supernatant and washing with PBS, the cell samples were lysed with a Trizol reagent (Promega) and total RNA was extracted. The residual genomic DNA was removed by DNAase, and total RNA was reverse transcribed to cDNA using M-MLV reverse transcriptase (Promega). The relative mRNA expression was quantiﬁed by quantitative RT-PCR. SYBR Premix Ex Taq™(Applied Biosystems) was used as the ﬂuorescent dye on Applied Biosystems 7500 Real-Time PCR Systems. The primers used were synthesized by Sangon (Shanghai, China), and the sequences were as follows (5’–3’): After removal of the supernatant and washing with PBS, the cell samples were lysed with a Trizol reagent (Promega) and total RNA was extracted. The residual genomic DNA was removed by DNAase, and total RNA was reverse transcribed to cDNA using M-MLV reverse transcriptase (Promega). The relative mRNA expression was quantiﬁed by quantitative RT-PCR. SYBR Premix Ex Taq™(Applied Biosystems) was used as the ﬂuorescent dye on Applied Biosystems 7500 Real-Time PCR Systems. The primers used were synthesized by Sangon (Shanghai, China), and the sequences were as follows (5’–3’): CHCHD2, Forward: TACCAGGAGCCTCAGGGAAC, Reverse: CAAGTCGGCACTGTTTCAGC; TRIM27, Forward: GTACTTCGCAGAGCCCATGA, Reverse: TACCAGTTGGGTCACGTTGG; ABCC1, Forward: ATCACCTTCTCCATCCCCGA, Reverse: TTCTGAATCCAGGCCTGCTG; CASP10, Forward: ATCCTTTCGGCATGTGGAGG, Reverse: GGCTGGGGCATCTGTTTCTT; IFITM3, Forward: CACTGTCCAAACCTTCTTCTCTC, Reverse: TCACGTCGCCAACCATCTTC. The concentration of the primers used for RT-PCR was 400 nM. p y y g g q CHCHD2, Forward: TACCAGGAGCCTCAGGGAAC, Reverse: CAAGTCGGCACTGTTTCAGC; TRIM27, Forward: GTACTTCGCAGAGCCCATGA, Reverse: TACCAGTTGGGTCACGTTGG; ABCC1, Forward: ATCACCTTCTCCATCCCCGA, Reverse: TTCTGAATCCAGGCCTGCTG; CASP10, Forward: ATCCTTTCGGCATGTGGAGG, Reverse: GGCTGGGGCATCTGTTTCTT; IFITM3, Forward: CACTGTCCAAACCTTCTTCTCTC, Reverse: TCACGTCGCCAACCATCTTC. The concentration of the primers used for RT-PCR was 400 nM. CACTGTCCAAACCTTCTTCTCTC, Reverse: TCACGTCGCCAACCATCTTC. The concentration of the primers used for RT-PCR was 400 nM. Supplementary Materials: The following are available online. Table S1: HSV-1/mock ratios of quantiﬁed prot n HSV-1-infected HEK 293T cells. Author Contributions: Conceptualization, L.-K.Z.; data curation, W.W.; funding acquisition, G.X. and L.-K.Z.; investigation, W.W., L.W., X.C., S.Z. and W.S.; methodology, W.W., L.W. and S.Z.; project administration, G.X. and L.-K.Z.; supervision, G.X. and L.-K.Z.; validation, W.W. and L.W.; writing—original draft, L.W.; writing—review and editing, W.W. and L.W. Funding: This research was funded by the State Key Laboratory of Virology (grant no. Y606031GF1) and Hubei Provincial Department of Education (grant no. Q20193004). Funding: This research was funded by the State Key Laboratory of Virology (grant no. Y606031GF1) and Hubei Provincial Department of Education (grant no. Q20193004). Conﬂicts of Interest: The authors declare no conﬂicts of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; and in the decision to publish the results. References Conwell, S.E.; White, A.E.; Harper, J.W.; Knipe, D.M. Identiﬁcation of TRIM27 as a novel degradation target of herpes simplex virus 1 ICP0. J. Virol. 2015, 89, 220–229. [CrossRef] 12. Paludan, S.R.; Bowie, A.G.; Horan, K.A.; Fitzgerald, K.A. Recognition of herpesviruses by the innate immune system. Nat. Reviews Immunol. 2011, 11, 143–154. [CrossRef] [PubMed] 13. Zhu, S.L.; Chen, X.; Wang, L.J.; Wan, W.W.; Xin, Q.L.; Wang, W.; Xiao, G.; Zhang, L.K. Global quantitative proteomic analysis proﬁles host protein expression in response to Sendai virus infection. Proteomics 2017. [CrossRef] [PubMed] 14. Christensen, M.H.; Jensen, S.B.; Miettinen, J.J.; Luecke, S.; Prabakaran, T.; Reinert, L.S.; Mettenleiter, T.; Chen, Z.J.; Knipe, D.M.; Sandri-Goldin, R.M.; et al. HSV-1 ICP27 targets the TBK1-activated STING signalsome to inhibit virus-induced type I IFN expression. EMBO J. 2016, 35, 1385–1399. [CrossRef] [PubMed] 15. Hornung, V.; Rothenfusser, S.; Britsch, S.; Krug, A.; Jahrsdorfer, B.; Giese, T.; Endres, S.; Hartmann, G. Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 2002, 168, 4531–4537. [CrossRef] 16. Zhang, L.K.; Chai, F.; Li, H.Y.; Xiao, G.; Guo, L. Identiﬁcation of host proteins involved in Japanese encephalitis 15. Hornung, V.; Rothenfusser, S.; Britsch, S.; Krug, A.; Jahrsdorfer, B.; Giese, T.; Endres, S.; Hartmann, G. Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 2002, 168, 4531–4537. [CrossRef] Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 2002, 168, 4531–4537. [CrossRef] 16. Zhang, L.K.; Chai, F.; Li, H.Y.; Xiao, G.; Guo, L. Identiﬁcation of host proteins involved in Japanese encephalitis virus infection by quantitative proteomics analysis. J. Proteome Res. 2013, 12, 2666–2678. [CrossRef] y g y 16. Zhang, L.K.; Chai, F.; Li, H.Y.; Xiao, G.; Guo, L. Identiﬁcation of host proteins involved in Japanese encephalitis virus infection by quantitative proteomics analysis. J. Proteome Res. 2013, 12, 2666–2678. [CrossRef] 17. Santamaria, E.; Mora, M.I.; Potel, C.; Fernandez-Irigoyen, J.; Carro-Roldan, E.; Hernandez-Alcoceba, R.; Prieto, J.; Epstein, A.L.; Corrales, F.J. Identiﬁcation of replication-competent HSV-1 Cgal+ strain signaling targets in human hepatoma cells by functional organelle proteomics. Mol. Cell. Proteom. 2009, 8, 805–815. [CrossRef] 18. Marino-Merlo, F.; Papaianni, E.; Medici, M.A.; Macchi, B.; Grelli, S.; Mosca, C.; Borner, C.; Mastino, A. References 1. Herpes Simplex Virus. Available online: https://www.who.int/en/news-room/fact-sheets/detail/herpes- simplex-virus#hsv1 (accessed on 17 November 2019). 1. Herpes Simplex Virus. Available online: https://www.who.int/en/news-room/fact-sheets/detail/herpes- simplex-virus#hsv1 (accessed on 17 November 2019). 2. Luecke, S.; Paludan, S.R. Innate recognition of alphaherpesvirus DNA. Adv. Virus Res. 2015, 92, 63–100. [CrossRef] [PubMed] 2. Luecke, S.; Paludan, S.R. Innate recognition of alphaherpesvirus DNA. Adv. Virus Res. 2015, 92, 63–100. [CrossRef] [PubMed] . Klapper, P.E.; Cleator, G.M. Herpes simplex virus. Intervirology 1997, 40, 62–71. [CrossRef] [PubMed] 4. Grunewald, K.; Desai, P.; Winkler, D.C.; Heymann, J.B.; Belnap, D.M.; Baumeister, W.; Steven, A.C. Three-dimensional structure of herpes simplex virus from cryo-electron tomography. Science 2003, 302, 1396–1398. [CrossRef] [PubMed] 4. Grunewald, K.; Desai, P.; Winkler, D.C.; Heymann, J.B.; Belnap, D.M.; Baumeister, W.; Steven, A.C. Three-dimensional structure of herpes simplex virus from cryo-electron tomography. Science 2003, 302, 1396–1398. [CrossRef] [PubMed] 5. Antrobus, R.; Grant, K.; Gangadharan, B.; Chittenden, D.; Everett, R.D.; Zitzmann, N.; Boutell, C. Proteomic analysis of cells in the early stages of herpes simplex virus type-1 infection reveals widespread changes in the host cell proteome. Proteomics 2009, 9, 3913–3927. [CrossRef] [PubMed] 5. Antrobus, R.; Grant, K.; Gangadharan, B.; Chittenden, D.; Everett, R.D.; Zitzmann, N.; Boutell, C. Proteomic analysis of cells in the early stages of herpes simplex virus type-1 infection reveals widespread changes in the host cell proteome. Proteomics 2009, 9, 3913–3927. [CrossRef] [PubMed] 6. Sanchez-Quiles, V.; Mora, M.I.; Segura, V.; Greco, A.; Epstein, A.L.; Foschini, M.G.; Dayon, L.; Sanchez, J.C.; Prieto, J.; Corrales, F.J.; et al. HSV-1 Cgal+ infection promotes quaking RNA binding protein production and induces nuclear-cytoplasmic shuttling of quaking I-5 isoform in human hepatoma cells. Mol Cell Proteom. 2011, 10, M111.009126. [CrossRef] 7. Berard, A.R.; Coombs, K.M.; Severini, A. Quantiﬁcation of the host response proteome after herpes simplex virus type 1 infection. J. Proteome Res. 2015, 14, 2121–2142. [CrossRef] [PubMed] 8. Sloan, E.; Tatham, M.H.; Groslambert, M.; Glass, M.; Orr, A.; Hay, R.T.; Everett, R.D. Analysis of the SUMO2 Proteome during HSV-1 Infection. PloS Pathog. 2015, 11, e1005059. [CrossRef] 9. Lin, A.E.; Greco, T.M.; Dohner, K.; Sodeik, B.; Cristea, I.M. A proteomic perspective of inbuilt viral protein regulation: pUL46 tegument protein is targeted for degradation by ICP0 during herpes simplex virus type 1 infection. Mol. Cell. Proteom. 2013, 12, 3237–3252. [CrossRef] 10. Rowles, D.L.; Terhune, S.S.; Cristea, I.M. Discovery of host-viral protein complexes during infection. Methods Mol. Biol. (Cliftonn. J.) 2013, 1064, 43–70. [CrossRef] 11. 4.10. Intracellular RNA Extraction and Quantitation Acknowledgments: We thank The Center for Instrumental Analysis and Metrology and the Core Facility and Technical Support, Wuhan Institute of Virology for providing technical assistance. Conﬂicts of Interest: The authors declare no conﬂicts of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; and in the decision to publish the results. 12 of 13 12 of 13 Molecules 2019, 24, 4215 References HSV-1-induced activation of NF-kappaB protects U937 monocytic cells against both virus replication and apoptosis. Cell Death Dis. 2016, 7, e2354. [CrossRef] 19. Sun, H.; Zhang, Q.; Jing, Y.Y.; Zhang, M.; Wang, H.Y.; Cai, Z.; Liuyu, T.; Zhang, Z.D.; Xiong, T.C.; Wu, Y.; et al. USP13 negatively regulates antiviral responses by deubiquitinating STING. Nat. Commun. 2017, 8, 15534. [CrossRef] 13 of 13 13 of 13 Molecules 2019, 24, 4215 20. Martin, C.; Leyton, L.; Hott, M.; Arancibia, Y.; Spichiger, C.; McNiven, M.A.; Court, F.A.; Concha, M.I.; Burgos, P.V.; Otth, C. Herpes Simplex Virus Type 1 Neuronal Infection Perturbs Golgi Apparatus Integrity through Activation of Src Tyrosine Kinase and Dyn-2 GTPase. Front. Cell. Infect. Microbiol. 2017, 7, 371. [CrossRef] 21. Birkenheuer, C.H.; Danko, C.G.; Baines, J.D. Herpes Simplex Virus 1 Dramatically Alters Loading and Positioning of RNA Polymerase II on Host Genes Early in Infection. J. Virol. 2018, 92. [CrossRef] 22. Vandevenne, P.; Sadzot-Delvaux, C.; Piette, J. Innate immune response and viral interference strategies developed by human herpesviruses. Biochem. Pharm. 2010, 80, 1955–1972. [CrossRef] [PubMed] 23. Li, C.; Du, S.; Tian, M.; Wang, Y.; Bai, J.; Tan, P.; Liu, W.; Yin, R.; Wang, M.; Jiang, Y.; et al. The Host Restriction Factor Interferon-Inducible Transmembrane Protein 3 Inhibits Vaccinia Virus Infection. Front. Immunol. 2018, 9, 228. [CrossRef] [PubMed] 24. Amini-Bavil-Olyaee, S.; Choi, Y.J.; Lee, J.H.; Shi, M.; Huang, I.C.; Farzan, M.; Jung, J.U. The antiviral eﬀector IFITM3 disrupts intracellular cholesterol homeostasis to block viral entry. Cell Host Microbe 2013, 13, 452–464. [CrossRef] [PubMed] 25. Imai, Y.; Meng, H.; Shiba-Fukushima, K.; Hattori, N. Twin CHCH Proteins, CHCHD2, and CHCHD10: Key Molecules of Parkinson’s Disease, Amyotrophic Lateral Sclerosis, and Frontotemporal Dementia. Int. J. Mol. Sci. 2019, 20, 908. [CrossRef] [PubMed] 26. Funayama, M.; Ohe, K.; Amo, T.; Furuya, N.; Yamaguchi, J.; Saiki, S.; Li, Y.; Ogaki, K.; Ando, M.; Yoshino, H.; et al. CHCHD2 mutations in autosomal dominant late-onset Parkinson’s disease: A genome-wide linkage and sequencing study. Lancet Neurol. 2015, 14, 274–282. [CrossRef] 27. Song, R.; Yang, B.; Gao, X.; Zhang, J.; Sun, L.; Wang, P.; Meng, Y.; Wang, Q.; Liu, S.; Cheng, J. Cyclic adenosine monophosphate response element-binding protein transcriptionally regulates CHCHD2 associated with the molecular pathogenesis of hepatocellular carcinoma. Mol. Med. Rep. 2015, 11, 4053–4062. [CrossRef] 28. Xin, Q.L.; Deng, C.L.; Chen, X.; Wang, J.; Wang, S.B.; Wang, W.; Deng, F.; Zhang, B.; Xiao, G.; Zhang, L.K. References Quantitative Proteomic Analysis of Mosquito C6/36 Cells Reveals Host Proteins Involved in Zika Virus Infection. J. Virol. 2017, 91. [CrossRef] 29. Zhu, S.; Wan, W.; Zhang, Y.; Shang, W.; Pan, X.; Zhang, L.K.; Xiao, G. Comprehensive Interactome Analysis Reveals that STT3B is Required for the N-Glycosylation of Lassa Virus Glycoprotein. J. Virol. 2019. [CrossRef] 29. Zhu, S.; Wan, W.; Zhang, Y.; Shang, W.; Pan, X.; Zhang, L.K.; Xiao, G. Comprehensive Interactome Analysis Reveals that STT3B is Required for the N-Glycosylation of Lassa Virus Glycoprotein. J. Virol. 2019. [CrossRef] 30. Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D.J.; Inuganti, A.; 29. Zhu, S.; Wan, W.; Zhang, Y.; Shang, W.; Pan, X.; Zhang, L.K.; Xiao, G. Comprehensive Interactome Analysis Reveals that STT3B is Required for the N-Glycosylation of Lassa Virus Glycoprotein. J. Virol. 2019. [CrossRef] 30. Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D.J.; Inuganti, A.; Griss, J.; Mayer, G.; Eisenacher, M.; et al. The PRIDE database and related tools and resources in 2019: Improving support for quantiﬁcation data. Nucleic Acids Res. 2019, 47, D442–D450. [CrossRef] 30. Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D.J.; Inuganti, A.; Griss, J.; Mayer, G.; Eisenacher, M.; et al. The PRIDE database and related tools and resources in 2019: Improving support for quantiﬁcation data. Nucleic Acids Res. 2019, 47, D442–D450. [CrossRef] 31. Emmott, E.; Goodfellow, I. Identiﬁcation of protein interaction partners in mammalian cells using SILAC-immunoprecipitation quantitative proteomics. J. Vis. Exp. 2014, e51656. [CrossRef] 31. Emmott, E.; Goodfellow, I. Identiﬁcation of protein interaction partners in mammalian cells using SILAC-immunoprecipitation quantitative proteomics. J. Vis. Exp. 2014, e51656. [CrossRef] 32. DAVID Bioinformatics Resources. Available online: https://david.ncifcrf.gov/ (accessed on 17 November 2019). 32. DAVID Bioinformatics Resources. Available online: https://david.ncifcrf.gov/ (accessed on 17 November 2019). Sample Availability: Samples of the compounds (HSV-1) are available from the authors. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4392643072	https://www.whp-journals.co.uk/CCH/article/download/952/694	English	null	Large Fires and Climatic Variability in Urban Europe, 1500–1800	Deleted Journal	2,024	cc-by	11,421	David Garrioch David Garrioch ABSTRACT: Several hundred large urban fires occurred in Europe during the early modern period, but they did not take place randomly. This article charts their incidence and reveals a peak in the seventeenth century, coinciding with some of the coolest periods of the Little Ice Age. This apparent paradox can best be explained by climatic variability, since overall cooling was accompanied by numerous warm, dry anomalies. While the cause of fires was usually human activity, and small fires happened frequently, this paper shows that many of the largest conflagrations of the early modern period took place in years of such hot and/or dry climatic anomalies, and closer analysis of individual fires con firms that these meteorological conditions facilitated their spread. This strongly suggests that climatic variability associated with the Little Ice Age was a major determinant of the timing of large fires in Europe. Over the same broad period, climatic disasters linked to cooler and damper conditions contributed to social and political instability, and there is evidence that this in turn undermined fire prevention and control and thus further increased the likelihood of small fires becoming large ones. KEYWORDS: History of fire; Little Ice Age; early modern Europe; urban history; climatic variability T he history of large fires in early modern European towns offers a strik ing paradox. Charting their incidence across a long period reveals that the greatest number, and many of the most disastrous ones, occurred dur ing the coldest periods of the Little Ice Age. Five of the eight largest fires to strike European towns in peacetime, during the entire early modern period (1500–1800), took place in the thirty years between 1656 and 1686, and the seventeenth century as a whole witnessed a disproportionate number of major conflagrations. These included the Great Fire of London of 1666, the most dev astating urban fire in Europe since Ancient times, and the second-largest early modern urban fire, the 1656 conflagration in Aachen that destroyed around 4,600 houses. This article argues that climatic factors provide much of the explanation for this apparent paradox, and that they do so in two ways. Firstly, a great number of large fires, right across the early modern period but particu larly in the seventeenth century, coincided with anomalous hot, dry seasons. Climates and Cultures in History 1 (2024): 00–00. doi: 10.3197/whpcch.63842135436332 ISSN: 2635-1331 CC BY 4.0 © 2024 The White Horse Press. 1. Exceptions are Mauelshagen 2010 and Zwierlein 2021: 90–99. 2. The key study is Frost and Jones 1989. There has been debate about the chronology pro posed by Frost and Jones, and Shane Ewen has argued that the concept of the ‘fire gap’ is methodologically flawed. See particularly Pearson 2004: 33–38; Zwierlein 2021: 69; Ewen 2006. David Garrioch Secondly, the climatic disasters of the Little Ice Age contributed to political, economic and social instability that undermined the capacity of urban govern ments and populations to prevent and manage large fires. T Climates and Cultures in History 1 (2024): 00–00. doi: 10.3197/whpcch.63 ISSN: 2635-1331 CC BY 4.0 © 2024 The White Horse Press. David Garrioch Until recently, neither urban historians nor environmental historians have paid much attention to climatic factors in accounting for disastrous fires.1 Although large parts of individual European towns and cities burned at in tervals, destructive fires have largely been taken as a given, accidents waiting to happen. Since flame was used for many day-to-day purposes, there was an ever-present risk of wildfire in towns built primarily of wood, with narrow streets and densely-packed housing. Only across the nineteenth century did urban fires that burned hundreds or even thousands of houses disappear from Europe, a change generally attributed to the introduction of less flammable building materials and to suburbanisation.2 Environmental history has not shown great interest in early modern urban fire, despite Stephen Pyne’s ob servation that, as long as towns were predominantly built of materials taken from the lands that surrounded them, primarily wood and thatch, urban fires behaved in the same ways as wildfires elsewhere. He too sees the nineteenth century as the major turning point in urban fire regimes, as industrialisation brought new forms of building and new ways of controlling and using fire (Pyne 2001: 101–18). Only in the last few years have historians begun to recognise that acciden tal urban fires did not occur randomly in the preindustrial period. In 2012, the editors of an important collective volume, covering many parts of the world, argued that fire regimes have shifted significantly, particularly since the expan sion of global trade in the seventeenth century, which drove the creation of new forms of wealth and social control. They suggested that in cities dominated by imperial and mercantile interests, by autocratic rulers, or by colonial elites, new political regimes combined with local environmental conditions to change the way fires took place and were managed (Bankoff et al. 2012: Introduction; Garrioch 2019a).l The influence of climatic factors, however, has only begun to be explored. The most important work to date is Cornel Zwierlein’s monumental study of fires in German-speaking Central European towns during the past millennium (2021). Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 to the long period of cooling were major factors in the conflagrations both of 1540 and 1666 (Zwierlein 2021: 90–99).3 Drawing on this work, Geoffrey Parker (2013: 62–63) also connects drought with exceptionally large fires in the mid-seventeenth century, pointing to several examples from around the world, and suggests that climate was the ‘true culprit’ in the fires of 1666. The present article builds on these insights, examining the record of large urban fires across most of Europe from 1500 to 1800, and drawing on the work of cli mate historians to examine the role of meteorological conditions and broader climatic factors in the outbreak and spread of such fires.i The sources for early modern fires are incomplete and often imprecise, and this must particularly be borne in mind in any comparative study. Early chronicles, compiled in religious houses or by local scholars, were often more concerned to demonstrate God’s purpose than to construct an accurate chronol ogy. They frequently tell us only that a large fire took place, rarely give the precise date, and are sometimes vague even about the year. We know, for ex ample, that Venice experienced two bad fires in 1505/6, and that one burned ‘the entire district of Casselaria’, while the second destroyed ‘the whole island of Rialto’ (Gallicciolli 1795: Vol. 2, 237). Many descriptions of fires, even at later dates, were based on hearsay, rather than eyewitness information, and where first-hand accounts exist, they are often impressionistic, referring to a ‘great’, ‘dreadful’ or ‘impetuous’ fire. Many accounts, in pamphlets or local histories, simply assert that half, a third, or even the whole of a town burned. Even where more detail is given, the information needs to be treated with caution, as different sources often offer conflicting accounts. Estimates of the destruction in the French town of Issoudun in 1651, for example, range from 600 to 1200 houses destroyed (Pérémé 1847: 205). Many accounts of damage, as in this example, offer suspiciously rounded figures. Wherever possible, we must compare different sources. Municipal records sometimes mention pay ments to firefighters and court rulings awarding compensation to victims, or plans drawn up for reconstruction, often indicate that earlier reports of the scale of a fire were greatly exaggerated. Archaeological or architectural re cords sometimes help to evaluate the accuracy of the original accounts. 3. For corroborating evidence on the 1666 anomaly, see Mauelshagen 2010: 128. Climates and Cultures in History 1 (2024) David Garrioch It identifies 1666, unusually warm and dry, as one of the worst peace time fire years in the continent’s history, second only to 1540, which climate historians have shown to be one of the hottest and driest of the last 500 years. Zwierlein also points out that the two decades of the 1650s and 1660s not only witnessed a wave of fires in parts of Europe, but also huge blazes in Edo (Tokyo) and Constantinople (Istanbul). He concludes that anomalies linked Climates and Cultures in History 1 (2024) 2 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Where the sources offer different estimates, the minimum estimate of damage has been adopted here.i We rarely possess much information on fire behaviour, unfortunately, since that might also indicate the role of meteorological conditions. Only occasion ally, as in the London fire of 1666, did witnesses record spotting, changes of fire direction and sudden increases in intensity, or details such as shifts in wind direction, strength and gusting. This is very useful information. But there was no agreement on terminology, and neither the scientific knowledge nor the 3 David Garrioch instruments of the time allowed the collection of information on the intensity of flames, or even accurate observations of wind speed and direction.i The most common statistical indication of fire damage is the number of houses destroyed, which is commonly used by historians of urban fire for com parative purposes. Figures are very frequently provided in diaries, municipal requests for assistance and newspaper reports, and these estimates offer some sense of scale. Admittedly, this method of evaluating the size of a fire has its drawbacks. Public rumour and claims for compensation often inflated the losses. Contemporary accounts do not always distinguish between dwellings, warehouses and other buildings. In some places, local measures were used, as in Scandinavia, where fire damage was often expressed as the number of gårdar that burned: literally ‘yards’, the buildings around a central courtyard. And even where estimates seem fairly accurate, the dimensions of a ‘house’ varied considerably from one place to another. Small towns typically contained single- or two-storey houses, whereas the larger cities had many multi-storey buildings. The most extreme example is eighteenth-century Edinburgh, which was built on a confined site and where some residential buildings reached ten or even fifteen storeys. In London or Vienna, most of the buildings were no more than four storeys but, as in most European towns of any size, there was a significant difference between the centre and the suburbs. Even within indi vidual neighbourhoods the size of houses might vary greatly, with large houses inhabited by noble families sitting alongside far humbler dwellings. There were also major changes over time. Even the largest sixteenth-century towns had a very low skyline, with few residential buildings more than two storeys tall, but by 1800 they contained many apartment buildings that were not only higher but had a much greater surface area. Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 abundant, has been shown to have omitted a great many fires (Pearson 2004: 33). Blazes that destroyed over 100 houses, on the other hand, are far more likely to be traceable in the historical record, to be reported in the international press as well as in local sources, and to be recalled in local histories. It is im portant to note, however, that this threshold is arbitrary and has no absolute value. It serves simply to identify the greatest conflagrations, which exceeded all hope of control with the resources available to early modern urban popu lations, destroyed significant parts of a city or town, and caused serious social and economic dislocation. I have followed Jan de Vries’s study of European Urbanization in exclud ing the eastern parts of Europe (Russia, the Baltic states, the eastern half of Poland, Hungary, Slovakia, Ukraine and the Balkans), areas which in any case had fewer towns than many other parts of Europe. This enables a comparison with his estimates of population growth during the early modern period (de Vries 1984). I have, however, adopted a minimal definition of ‘urban’ that includes places with a population of at least 2,000, although in the absence of reliable censuses it is often difficult to be sure. The database also excludes fires that we know were deliberately lit as weapons of war, since their chronology, as Zwierlein has shown, primarily reflects shifts in military strategies. In the seventeenth century, armies routinely burned towns, either as a siege tactic or to punish the inhabitants, but in the eighteenth century this practice was largely abandoned (Zwierlein 2021: 87–88). Between 1500 and 1800, in the regions covered by this survey, I have used a wide range of primary and secondary sources, including many archival ones, to identify 574 peacetime fires that burned 100 or more houses.4 They took place in 417 towns in different parts of Europe. Their approximate size is given in Table 1. The largest, by far, was the Great Fire of London of 1666, which destroyed some 13,200 houses and 87 churches. Next in size was the 1656 Aachen fire, which devastated seven-eighths of the town’s buildings: some 4,425 houses and twenty churches (Porter 1998; Kraus 2007: 35, 48). A blaze of 1684 burned over 1,700 houses in Hamburg, then a city of 60,000 people. 4. This includes, for German-speaking Central Europe, considerable relevant material gener ously provided by Cornel Zwierlein. Comprehensive data on Poland are now available in Karpińsky 2021. Large Fires and Climatic Variability in Urban Europe, 1500–1800 Nevertheless, the number of houses destroyed is the best measure we have, and this study uses the threshold of 100 houses, for what I will simply call ‘large’ fires. This is much higher than the benchmark adopted by Eric Jones and his collaborators, who took the loss of ten houses as the minimum thresh old for a ‘major’ fire, one that Cornel Zwierlein also broadly adopted in his survey of fires in German-speaking Central European towns (Jones et al. 1984; Zwierlein 2021: 74). My reason for using a higher figure is pragmatic. The nature and quality of the sources vary considerably from place to place. The primary ones used by Zwierlein, the extraordinary Städtebücher, were orig inally a product of the Nazi regime’s desire to document the history of the German Volk, and do not exist for other parts of Europe (Zwierlein 2021: 65– 69). Elsewhere, gathering equivalent data would require visits to the surviving archives in hundreds of towns, and would even then be seriously incomplete. Even the admirable Gazetteer compiled by Jones, Porter and Turner, which attempted to cover only England, where newspapers and local histories are Climates and Cultures in History 1 (2024) 4 Large Fires and Climatic Variability in Urban Europe, 1500–1800 continued to occur, with the 1720s and the 1750s worse than most decades of the preceding century. The nineteenth century is not included in this study, but blazes of the kind discussed here continued to occur until the 1840s, although the general trend was for major fires to be confined to large single buildings and industrial complexes rather than destroying wide residential areas (Ewen 2010). In terms of the simple numbers of dwellings burned, what we might call the early modern era of great urban fires lasted from the early seventeenth century to the early nineteenth. Yet, within this period, the seventeenth century stands out. In the years 1600–1699 we know of 241 large peacetime fires, compared with 105 for the sixteenth century, an increase of just under 230 per cent. Even allowing for a possible under-recording of large fires in the earlier period, this was a dramatic change. There were also slightly more large fires recorded for the seventeenth century than for the eighteenth, when 228 are identifiable. That was despite the growth in the numbers and size of towns and cities in the 1700s, which meant there were many more buildings that could have burned. The exceptional character of the 1600s is also apparent if we consider the incidence of very large fires (those that burned >500 houses – see Figure 1). Only two (if we accept a suspiciously rounded figure of 500 houses lost in Łowicz in 1529) occurred between 1500 and 1599, but 21 were recorded be tween 1600 and 1699. In 1700 to 1799, despite there being more large cities, fourteen such fires took place. As already noted, the second half of the sev enteenth century also witnessed the most devastating fires: the Great Fire of London of 1666 and that of Aachen in 1656. Only the 1842 fire in Hamburg, which burned 4,200 houses, came close to these in size. Within the seventeenth century, the decade of the 1680s was the worst. It was unusual in witnessing one or more large fires recorded every year, and these included some very extensive ones: around 700 houses burned in Turku in 1681, more than 1,000 buildings in the London suburb of Wapping in 1682, over 1,700 houses in Hamburg in 1684, some 1,300 in Stockholm in 1686, and 568 in Mühlhausen in 1689. Climates and Cultures in History 1 (2024) Large Fires and Climatic Variability in Urban Europe, 1500–1800 Stockholm too ex perienced many large fires, including one that burned around 1,800 buildings, mostly small, single-storey dwellings. A fire of similar size burned nearly a third of Copenhagen in 1728. Devastating fires also completed the almost total destruction of Lisbon, a city of around 180,000 people, in the aftermath of the earthquake of 1755 (Zwierlein 2021: 200; Stockholm City Archives [hence forth SSA], John Swensk Collection, BI–4; Gamrath 1999; Molesky 2012). Climates and Cultures in History 1 (2024) 5 David Garrioch David Garrioch Table 1. Number of fires, 1500–1800, by scale of destruction. Number of houses destroyed Number of fires >1000 12 500–999 24 300–499 82 100–299 456 Total 574 Figure 1. Peacetime fires that burned 100–499 houses (light) and >500 houses (dark), 1500–1800. Table 1. Number of fires, 1500–1800, by scale of destruction. Number of houses destroyed Number of fires >1000 12 500–999 24 300–499 82 100–299 456 Total 574 Figure 1. Peacetime fires that burned 100–499 houses (light) and >500 houses (dark), 1500–1800. Table 1. Number of fires, 1500–1800, by scale of destruction. Number of houses destroyed Figure 1. Peacetime fires that burned 100–499 houses (light) and >500 houses (dark), 1500–1800. The chronology of reported large fires, by decade, is given in Figure 1. Their incidence fluctuated across the sixteenth century, with peaks in the 1540s and the 1580s. The beginning of the seventeenth century then saw a sudden jump in their frequency, rising to peaks in the 1650s and especially the 1680s. Not only were the overall numbers of these fires higher in those years, but there were many more very large conflagrations (over 500 houses burned). After a relative lull in the 1690s, the number of large fires recorded then rose steadily to another peak in the 1720s. For the rest of the eighteenth century, it remained at levels broadly similar to, but often below, those of the 1600s. Very large fires Climates and Cultures in History 1 (2024) 6 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Further fires in Passau, Dresden, Bergen, Gera, Elberfeld and Sagan destroyed over 300 houses each. i The graphs reveal, therefore, that both the numbers of large fires, and their scale, were greatest during some of the coldest periods of the Little Ice Age, which is generally dated from around 1300 to roughly 1850 (White et al. 2018: 338–44). It was one of the coolest eras of the last two millennia, although the decline in average temperatures was not constant: in Europe, Alpine glaciers advanced rapidly in the fourteenth century, again from the 1580s to around 1660, and in the nineteenth century. The first half of the sixteenth century was mild, but it was followed by markedly colder, wetter conditions that per sisted until the early seventeenth century. There was then a brief respite before 7 7 David Garrioch cooling resumed. The middle decades of the century, from the 1640s to the 1670s, punctuated by some very large volcanic eruptions, brought many ex ceptionally cold, wet years across much of the northern hemisphere. European winter temperatures were again low during what is termed the late Maunder Minimum (1684–1715), associated with reduced solar activity, and the 1690s were very cold and wet, leading to famines in some parts of Europe. Warmer conditions gradually returned in the first half of the eighteenth century but, after 1740, average winter temperatures were again cooler in much of the con tinent, other seasons more favourable (Pfister et al 2018; Parker 2013: 3–8, 15).i This broad pattern is now well known, but finding out what was happening in a particular year, in a specific region, is more difficult. No meteorological measurements were taken before the mid-seventeenth century, and thereafter only in a few places. Even after they began, the instruments used were often not accurate and the recording was not always systematic. Before the eight eenth century, too, the records kept by different observers rarely cover the same periods, making us reliant on often idiosyncratic methods of measure ment. Fortunately, many early modern chroniclers and diarists commented on extreme weather events, albeit in very subjective terms, and this has enabled the reconstruction of certain weather and climate patterns. 5. For a survey of the sources, the methods used to analyse them, and the reconstructions they make possible, see White, Pfister and Mauelshagen 2018: 27–148. Large Fires and Climatic Variability in Urban Europe, 1500–1800 and at that Time very strongly’ (Strype 1720: vol. 1, 227). Eyewitness reports tell of the wind carrying embers several hundred metres ahead of the fire front, allowing the flames to jump firebreaks. This may indicate the formation of convection currents, also suggested by the observation of ash falling many kilometres away from the city and the smell of smoke 100 miles distant (Bell 1920: 68, 98–9). i Similar weather conditions prevailed at the time of other great fires. The one in Aachen in May 1656 followed a long period of warm, dry weather and, on the day of the fire, a strong westerly wind blew embers across the town from the suburb where the fire started, producing secondary fires (Kraus 2007: 42- 7). The year 1728 also had an unseasonably warm summer, and in October that year some 1,670 houses burned in Copenhagen. Such weather dried out the buildings and sometimes reduced the availability of water with which to fight fires: both in Copenhagen in 1728 and in London in 1666, there were some dif ficulties getting enough water, especially in the crucial early stages of the fire (Gamrath 1999: 297–98).i ( ) The years when the greatest numbers of large fires were recorded were also, in the majority of cases, exceptionally hot and dry. During the three centuries covered by this study, the worst individual years were 1684, 1723 and 1794, each of which saw seven large fires. The summer of 1684 was hot and dry, and there were fires in towns across northern Germany and Estonia, including one in Hamburg that burned 1,714 houses (Zwierlein 2021: 200; Glaser 2001: 168). In 1723, the spring was unusually warm across much of Europe, and in May, a big fire on the south shore of Stockholm burned some 400 houses. Driven by strong winds, it even leapt the harbour, a distance of nearly a kilo metre (Pfister et al. 2018: 280. SSA, John Swensk Collection, BI–4). In June, one of the largest urban fires in France during the early modern period took place in Châteaudun on a hot and dry Sunday, burning around 1,000 houses. It too was hastened by strong and changeable winds, and only stopped when it reached the barrier of the town walls (Robreau 2009: vol. 2, 180). In 1794, the spring was unusually warm and the summer was hot. Large Fires and Climatic Variability in Urban Europe, 1500–1800 In addition, paleocli matologists have come up with a wide range of proxy measures, such as tree ring data, the dates of harvests, evidence of advances and retreats of glaciers, river and lake depths and archaeological material.5 Most attention, understand ably, has been devoted to the overall cooling trend and to its consequences for European society, and only recently have droughts begun to attract attention (Garnier 2019; Pribyl 2020; Przybylak et al. 2020; Leijonhufvud and Retsö 2021). Unfortunately, the nature and quality of the sources varies considerably from place to place, and more research has been done in some regions than in others. There was much variation in local conditions across the continent, and these are often very difficult to establish when looking at fires in individual towns. Such gaps make it impossible to establish statistically valid correlations between meteorological conditions and large fires.l It is nevertheless clear that the largest conflagrations occurred overwhelm ingly in hot, dry and windy weather, particularly during summer and autumn. The summer of 1666 was unusually warm, and there was no rain in London in the six weeks before the Great Fire began on 2 September (Clark 1891–1900: vol. 2, 82; Evelyn 1955: vol. 3, 451; Macadam 2012: 237; Manley 1974: 393). This ‘great Drought’, wrote John Strype in 1720, ‘had so dryed the Timber, that it was never more apt to take Fire’. He added that ‘an Easterly Wind, (which is the dryest of all others) had blown for several Days together before; Climates and Cultures in History 1 (2024) Climates and Cultures in History 1 (2024) 8 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Four of the five fires whose dates are known took place in those months (Pfister et al. 2018: 281; Glaser 2001: 176). Climates and Cultures in History 1 (2024) David Garrioch Table 2. worst years for large fires recorded 1500–1800, and known weather conditions in the areas affected. Year No. fires Season where known (no. of fires) Conditions 1540 6 Summer (5) ‘Megadrought’, very hot in most of Europe all year 1590 5 Spring (1), summer (2) Drought year 1616 6 Summer (1), autumn (1) Very hot summer across Europe 1624 6 Autumn/winter (2) Unexceptional in most regions 1628 5 Spring (1) Cold and wet in regions where fires occurred 1634 5 Autumn (1) Drought in some regions 1652 5 Spring (1), summer (2), autumn (1) Very hot summer, drought across Europe 1664 5 Spring (2), summer (1) Unexceptional year 1682 5 Summer (1), autumn (1) Unexceptional year 1684 7 Summer (4) Hot, dry summer 1686 5 Spring (2), summer (1), autumn (1) Very dry spring/summer in much of Central Europe 1719 6 Spring (1), summer (4), autumn (1) Warm summer, extreme drought and heat in France 1723 7 Late winter (1), spring (4), summer (1) Unusually warm, dry spring. Strong winds 1725 5 Spring (3), summer (1) Mostly wet year 1731 6 Spring (1), summer (4) Dry in some regions affected 1794 7 Winter (1), spring (1), summer (3) Warm spring, very hot dry summer 1795 5 Late spring (2), summer (3) Unexceptional year Table 2. worst years for large fires recorded 1500–1800, and known weather conditions in the areas affected. As Table 2 shows, there were a further five years when six large fires oc curred: 1540, 1616, 1624, 1719 and 1731. While we do not know the precise dates of all of these blazes, in four of the five years there appears to be a clear link to climatic conditions: only in 1624 is there no evidence of particularly hot or dry conditions. The summer of 1540, the hottest and driest of the entire early modern period, has already been mentioned. 6. Gai, Mahytka and Otte 2019: 213; F.L. Siegel, ‘Historische und statistische Notizen über die Wichtigsten der abgebrannten Gebäuden ... und einer Schilderung des grossen Brandes derselben im Jahr 1616’: http://www.oschatz-damals.de/siegel_07.html (accessed 27 Jan. 2023). Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 some 365 houses were destroyed in the north-eastern French town of Sainte Menehould, where the fire jumped wide streets, as well as the river that ran through the town. The dry conditions extended to Switzerland and to the west ern parts of Germany, where 400 houses burned in Frankfurt-am-Main in June, and there was insufficient water to fight the flames. The same year, two large fires took place in Stockholm, in July and September, burning some 200 and 300 houses respectively (Grasset 1988; Pfister 1988: 129; Herden 2005: 69; SSA, John Swensk Collection, BI–4; Forsbom 1950: 9). In 1731, once again, the summer and autumn were unusually dry and quite warm, both in England and in the Paris region. Early in June, a big fire in Blandford, in the west of England, burned 337 houses, and a day later one in Tiverton, not far away, destroyed 298. Both fires were driven by strong winds.7 Figure 2. the summer of 1719 (JJA), compared to the average over the refer ence period of 1900–1960. Map 2a shows the anomalies of mean temperature in ºC. Map 2b shows the anomalies of precipitation, in millimetres. Sources: created using the KNMI Climate Explorer (https://climexp.knmi.nl), historical reconstructions. Temperature data from Luterbacher et al. 2004 and Xoplaki et al. 2005; precipitation data from Pauling et al. 2006. Figure 2. the summer of 1719 (JJA), compared to the average over the refer ence period of 1900–1960. Map 2a shows the anomalies of mean temperature in ºC. Map 2b shows the anomalies of precipitation, in millimetres. Sources: created using the KNMI Climate Explorer (https://climexp.knmi.nl), historical reconstructions. Temperature data from Luterbacher et al. 2004 and Xoplaki et al. 2005; precipitation data from Pauling et al. 2006. Some of the years when there were five large fires were also unusually warm and/or dry. In 1590, large fires occurred in Wolverhampton and in Poznan, as severe drought affected England and Western Poland (Brázdil et al. 2016: 108; Pribyl 2020: 1031; Przybylak et al. 2020: 648) Most of Europe again had an exceptionally hot, dry summer in 1652, which in Sweden was called ‘the Great Drought Year’. 7. Snell 1892: 72; https://www.dorsetlife.co.uk/2009/03/when-blandford-burnt/ (accessed 5 Aug. 2023); Garnier 2019: 59. Large Fires and Climatic Variability in Urban Europe, 1500–1800 That of 1616 was very hot, and two of the fires were very significant: 440 houses burned in Oschatz in July, 375 in Paderborn in September.6 1719 had one of the few warm summers in the entire eighteenth century, and in France and central Germany was a year of extreme drought and heat, 1.8 to 2ºC higher than the average for 1900–1960 (see Figure 2a) (Garnier 2019: 58–59; Pfister et al. 2018: 280–81). In August, 6. Gai, Mahytka and Otte 2019: 213; F.L. Siegel, ‘Historische und statistische Notizen über die Wichtigsten der abgebrannten Gebäuden ... und einer Schilderung des grossen Brandes derselben im Jahr 1616’: http://www.oschatz-damals.de/siegel_07.html (accessed 27 Jan. 2023). Climates and Cultures in History 1 (2024) 10 Large Fires and Climatic Variability in Urban Europe, 1500–1800 It saw five large fires that we know of, including some excep tional ones: 500 houses burned in Glasgow in July, at least 600 in two separate fires in Stockholm in March and September, and 433 in the German town of 11 Climates and Cultures in History 1 (2024) David Garrioch David Garrioch Luckau (Pribyl 2020: 1034; Pfister 1988: 140; Rácz 2013: 131; Leijonhufvud and Retsö 2021: 2018; SSA, John Swensk Collection, BI–4; Cleland 1820: 4). Luckau (Pribyl 2020: 1034; Pfister 1988: 140; Rácz 2013: 131; Leijonhufvud and Retsö 2021: 2018; SSA, John Swensk Collection, BI–4; Cleland 1820: 4). Figure 3. The summer of 1686 (JJA), compared to the average over the ref erence period 1900–1960. Map 3a shows the anomalies of mean temperature in ºC. Map 3b shows the anomalies of precipitation, in millimetres. In grey scale image, the darker the colour, the higher the temperature. Sources: as for Figure 2. Figure 3. The summer of 1686 (JJA), compared to the average over the ref erence period 1900–1960. Map 3a shows the anomalies of mean temperature in ºC. Map 3b shows the anomalies of precipitation, in millimetres. In grey scale image, the darker the colour, the higher the temperature. Sources: as for Figure 2. The spring of 1686 was the driest in 500 years and, in late March, a large fire destroyed 358 houses in the substantial German town of Gera. The summer was also hot and dry across much of Europe (see Figure 3). In June, a huge blaze burned 1,300 houses in Stockholm while, in September, another large fire took place in Bergen, in Norway, reportedly burning 231 blocks (Pauling et al. 2006: 393. Glaser 2001: 169; Hahn 1855: vol. 1, 659; SSA, John Swensk Collection, BI–3; Nielsen 1877: 372). 1634 was also very dry in England, Finland, and some of the Czech lands, although it is not singled out in climate histories of the German-speaking areas where large fires took place (Pribyl 2020: 1031; Garnier 2019: 51; Vesajoki and Tornberg 1994: 52; Brázdil et al. 2016: 108). Although climate data are missing for some years and for some regions, therefore, it is clear that a great many large fires, including a majority of the most disastrous ones, occurred in hot and dry seasons that conspicuously de parted from the general climatic trend. Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 of drought that sometimes lasted for several years (Pfister 2018; Garnier 2019; Pribyl 2020). Studies of the late Maunder Minimum and of the eighteenth and nineteenth centuries suggest that these anomalous seasons were related to at mospheric circulation in the Atlantic, which determined wind direction over much of Europe, and influenced both temperature and precipitation. These anomalies also appear to be linked to abnormal wind strength, as well as direc tion, which are both crucial factors in the incidence of large fires (Frich and Frydendahl 1994; Mellado-Cano et al. 2018; Mellado-Cano et al. 2020; Pfister 1994: 311). They continued into the eighteenth century but became less fre quent after about 1750, although with variations across the continent (Pfister et al. 2018: 273–83; Frenzel, Pfister and Gläser (eds) 1994). It seems clear that they played a significant role in the incidence of large urban fires. Yet we must also note the exceptions to this general pattern. There were a few unusually warm and dry summers, such as that of 1747, when very few large fires seem to have occurred (Pfister et al. 2018: 280. Le Roy Ladurie 2004–06: vol. 2, p. 24). Nor were heat and drought a necessary precondition for large fires. In 1628, for example, five such conflagrations took place in different Central European towns, making it one of the dozen worst years be tween 1500 and 1800. Yet across most of that region the year was cold and wet (Pfister et al. 2018: 280). In 1651, too, Issoudun lost over 600 houses to fire, but conditions do not seem to have been either hot or dry. Similarly, a huge fire in the London docks – admittedly driven by a very strong wind – destroyed over a thousand buildings in 1682, even though the year was very wet (Pérémé 1847: 205; Blackstone 1957: 56). A great many towns, constructed of wood and subject to the same weather conditions, never experienced a large fire, even at times when others in the same region burned. But of course we do not expect all forests to burn simultaneously, even when the fire risk is high. Hot, dry weather and high winds greatly increased the likelihood that a small fire would spread and become unstoppable, but did not make huge conflagrations inevitable. Climates and Cultures in History 1 (2024) Large Fires and Climatic Variability in Urban Europe, 1500–1800 Climate historians have established that the Little Ice Age, like the Anthropocene, was characterised by consider able seasonal variability and by extreme weather events. While most years were cool and damp, by comparison with earlier periods and with the later nineteenth and twentieth centuries, some winters were very cold and a few summers exceptionally hot. The summer of 1540 rivalled the heat wave of 2003 in Europe, and at intervals across the early modern era there were periods Climates and Cultures in History 1 (2024) 12 Large Fires and Climatic Variability in Urban Europe, 1500–1800 In any case, aside from lightning strikes, meteorological conditions were not in themselves a cause of urban fires, which were overwhelmingly produced by human activity. It is worth asking, therefore, in what circum stances climatic variability was important in the incidence of large urban fires. One obvious factor was the nature of the buildings in any given location, since they provided the primary fuel for urban wildfires. In Mediterranean Western Europe, where early deforestation made timber scarce and expensive and where stone was often plentiful and suitable for construction, there were very few large fires. By the sixteenth century, most Italian, Spanish and south ern French towns were built of stone and had tiled roofs. In the Low Countries, too, brick became the usual building material quite early and, there too, few large fires occurred. Elsewhere in Europe, the replacement of wood and straw 13 David Garrioch with brick, stone, tiles or slates was often required by building regulations, but these were very unevenly observed. The construction of firewalls, in stone or brick, was obligatory in some places, and appears to have been effective, particularly where they projected above the roofline. In Paris and some other places, the widespread use of plaster – an effective fire retardant – was an im portant factor in preventing small fires from spreading (Keene 1999: 196–98; Garrioch 2019b). Historians have documented a trend towards the use of less flammable materials, especially in the eighteenth century (Frost and Jones 1989; Borsay 1989: 18). Many historians have suggested that rapid population increases – at least as a proxy for urban growth – were a major factor in the occurrence of major fires. However, the data do not bear this out. Across the areas included in this study, Jan de Vries has estimated that the number of town dwellers grew by roughly 51 per cent between 1500 and 1599, by just over 24 per cent from 1600 to 1699, and by a little more than 44 per cent from 1700 to 1799. It was to accelerate dramatically after 1800. Yet, as we have seen (Figure 1), the number of large fires actually fell in the second half of the sixteenth century, rose strongly during the seventeenth-century period of slower urban growth, then levelled off in the eighteenth century, even though the urban population was then increasing more rapidly. 8. de Vries 1984: 50. Linear regression analysis, using de Vries’s figures, indicates only a moderate association (R2 = 0.64). I am grateful to Kathleen Neal for this calculation. 9. de Vries 1984: 272; Aubert 2010: 96; Ericson 2001: 100. Zwierlein 2021 (102) makes similar points for Hamburg. Large Fires and Climatic Variability in Urban Europe, 1500–1800 of the numerous large fires in those areas (Power 1972). The same was al most certainly true in seventeenth-century Stockholm, where very large fires in 1625, 1640, 1652 and 1686 all coincided with a rapid increase in population and the extension of suburbs that contained mainly timber buildings, some times with thatched roofs (Ericson 2001: 98–100). Population pressure could also deplete groundwater supplies, making the task of firefighters even harder. There is evidence that the London water table fell in the early modern period, due to overuse (van Lieshout 2016: 790). ( ) It is often forgotten, however, that buildings were not the only source of fuel for urban fires. London is, again, a conspicuous example. The frequency of very large fires there resulted partly from its functions as a port and a manufacturing centre. The Great Fire of 1666, according to witness reports, only became dis astrous after it reached old wooden warehouses along the river, because they contained ‘the most combustible matter of Tarr, Pitch, Hemp, Rosen, and Flax which was all layd up thereabouts so that in six houres it became a huge stream of fire at least a mile long and could not possibly be approach’d or quencht’ (Nicolson and Hutton (eds) 1992: 276). The 1682 Wapping fire engulfed not only houses but also many warehouses and shipyards along the river, which contained similarly flammable substances (Sad and Lamentable News from Wapping). Between 1666 and 1800, of sixteen fires in London that burned more than 100 houses, all but three took place in port areas. This reflects the wider growth of world trade, and particularly that of the Atlantic. Hamburg had a similar port economy and experienced a very large fire in 1684. There too, many smaller but still dangerous fires destroyed factories and warehouses containing sugar, rum, and other flammable products linked to European and Atlantic trade (Garrioch 2016; Zwierlein 2021: 99–108). Yet, once again, pre ventative measures helped determine whether the presence of such potential fuels contributed to large fires. From the late seventeenth century, the authori ties in Paris and Vienna regulated dangerous industries reasonably effectively, removing them from densely-packed residential areas to the urban outskirts (Le Roux 2011: 46–68; Weigl 2001–2006: vol. 2, 157). Wide avenues were created in many cities, especially in the late seventeenth and eighteenth cen turies. Climates and Cultures in History 1 (2024) Large Fires and Climatic Variability in Urban Europe, 1500–1800 Overall, the statistical correlation between population increase and the number of large fires was relatively low.8 In individual places, too, population growth did not necessarily accompany frequent large fires. Paris grew far more rapidly than most European cities between 1600 and 1700, but experienced only one fire that burned around 100 houses. Amsterdam also grew very fast in the first half of that century but ap pears to have had no fires of this size. By contrast, some of the most disastrous early modern fires took place in towns that had very low, even negative, rates of growth. Aachen, where the second-largest European fire of the period occurred in 1656, had a lower population in 1650 than in 1500. In Rennes, the popula tion had stagnated for thirty years before the fire of 1720. Eighteenth-century Stockholm, whose economy and population grew slowly, if at all, nevertheless continued to experience large fires, notably in 1719, 1723 and 1759.9 This does not mean that population growth was irrelevant. It resulted in high demand for housing, which in turn encouraged poor-quality timber construction, if there were no well-enforced building regulations. That was certainly true for early modern London, especially in the port districts that housed a large labouring population, and probably contributed to the gravity Climates and Cultures in History 1 (2024) 14 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Even when the motive was to demonstrate the power of the ruler or to improve traffic flow, the open spaces prevented fires from spreading.iiii Once a fire did break out, the efficacy of firefighting was crucial in deter mining whether it was extinguished or developed into a large conflagration. Controlling a dangerous fire depended particularly on early detection and rapid action. In many places, the introduction of night watch services, or of fire-spot ters installed on church towers, enabled fires to be caught before they spread. Such arrangements assisted Vienna to avoid large fires after a terrible confla gration in 1627 (Pils 1999: 183). They were not failsafe, though, since similar Climates and Cultures in History 1 (2024) 15 David Garrioch David Garrioch measures did not prevent huge fires in Stockholm, for example – although it is possible that without them there might have been even more blazes.i Most accidental fires were put out quickly, even though very few towns engaged paid firefighters before 1800. The inhabitants were well aware of the danger, and some groups – notably building workers and sometimes other oc cupational categories – developed a lot of expertise in fighting fires. The tools were quite simple: buckets, fire-axes, hooks, ladders and, by the late seven teenth century, hand-operated fire pumps. Thus equipped, people doused the flames and dampened nearby sources of fuel. They watched for flying embers and quickly extinguished them. Knowing that fires usually spread from roof to roof, they removed or dampened wooden slats and shingles. If a fire con tinued to grow, they demolished buildings to create firebreaks. Such measures often enabled even multi-house blazes to be controlled and put out. One in the French town of Rennes in 1661, for example, destroyed around fifteen houses in a densely built block, but was prevented from spreading further. Another in Paris, in 1718, destroyed about twenty houses before being brought under con trol (Fillaut 1999: 32; Bibliothèque nationale, Paris, France. Collection Joly de Fleury, MS 1324, fol. 127). i Many disastrous fires, by contrast, began late at night, when there were few people in the streets. This happened in Aachen in 1656, in London in 1666, in the London borough of Southwark in 1676, in Rennes in 1720, and in Copenhagen in 1728. Fires that broke out on holidays, when many volun teer firefighters were absent, were also more likely to spread. Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 on the part of town authorities, both in providing equipment and training and in ensuring adequate supplies of water. In most towns, where there were no permanent firefighters, the experience and capacity of local officials was also crucial. Amid the confusion and smoke of large fires, they directed operations and decided when and where to create firebreaks. Drought, heat and unfavourable meteorological conditions were there fore far from being the only factors that helped determine when and where small fires grew into large ones. Yet some of these other variables, notably the preparedness of city governments in the face of the fire risk, were indi rectly influenced by the unfavourable climatic conditions of the Little Ice Age. These had a hugely disruptive impact on European society. Harvest failures produced by untimely cold, excessive rain, or even drought, led to population decline and economic crisis. High prices for basic necessities were disastrous at a time when much of the European population lived permanently on the edge of poverty. One result was very high mortality: it has been estimated that between 1691 and 1701 a million people died in France of climate-related fam ine and disease, in an overall population of around 20 million. Some 600,000 more were to die because of the harsh winter of 1709 (Parker 2013: 589). Undernourishment made them more vulnerable to diseases that flourished in very cold and wet conditions. But drought was also a killer. In early modern Britain, many very dry years were accompanied by outbreaks of plague, fevers and smallpox, and the great drought of 1719 in France saw high mortality from dysentery, spread by polluted water (Pribyl 2020:1035; Garnier 2019: 59). Food shortages, death and disease in turn produced serious social disrup tion and political instability. The seventeenth century was marked by huge numbers of local revolts. These have been closely studied in France, where the very harsh winter of 1662 was followed by near famine in many towns, pro ducing a record number of riots. This pattern recurred at the end of the century, from 1692 to 1710, when harvests failed repeatedly. The same years witnessed a huge spike in revolts against tax collectors (Nicolas 2008: 87, 346–67). Climates and Cultures in History 1 (2024) Large Fires and Climatic Variability in Urban Europe, 1500–1800 The great fire in Rennes in 1720 started on a Sunday night just before Christmas and, when the alarm was given, few people arrived to fight the flames before they took hold (Aubert and Provost (eds) 2020: 28–29). Even so, such accidents were less likely where preventative measures were enforced, and where firefighting services were well organised and funded. In Amsterdam, the fire service seems to have been particularly effective. Many potentially serious fires broke out there, and a few consumed multiple build ings, but well-trained firefighters generally arrived promptly, armed with the best equipment available, before the flames could spread further. This was also true of Venice, where squads from the city Arsenal were employed as firefight ers, becoming experienced in reaching fires and extinguishing them quickly (van der Heyden 1996; Svalduz 2006: 62). In these two cities, of course, the widespread use of brick for building made their task easier, and canals pro vided ready access to abundant water. Elsewhere, firefighters depended on public fountains and conduits, which were reliable only if the city govern ment built and maintained them. That was done in Vienna and Paris, but in London firefighters sometimes had difficulty obtaining water (Pearson 2004: 83–84). Effective prevention and firefighting required considerable foresight Climates and Cultures in History 1 (2024) 16 Large Fires and Climatic Variability in Urban Europe, 1500–1800 It is arguable, too, that the number and extreme brutality of the civil conflicts and wars of the seventeenth century were also in part a product of harsh economic conditions linked to climatic variability.i Certainly, the economic consequences of these crises made it more diffi cult for urban authorities to invest in fire prevention and control. Even though firefighting was mostly done by volunteers, they required buckets, fire pumps and other tools. These were expensive, but tax revenues fell as populations and production declined, and municipalities had difficulty finding funds. In Rennes, when the great fire of 1720 began, there were no buckets or other firefighting equipment in the municipal store (Aubert and Provost (eds) 2020: 82). This was not a direct result of climatic factors, but these contributed to the 17 David Garrioch pressures on local government. In many other towns, parishes and trade guilds were responsible for providing firefighting equipment, and these same institu tions were generally the primary providers of poor relief. Demands on their resources grew at precisely the moment when they had less to spare. Social and political crises may also have undermined the local leadership and social cohesion that were crucial in the event of a large fire. In London in 1666, the failure of the mayor to take decisive action has often been identified as a major factor in preventing the fire from being controlled in its early stages, and some of the local aldermen also appear to have opposed the destruction of houses to create firebreaks (Bell 1920: 29–30, 346). This might have been partly a product of inexperience. The political disruption of the years before the Great Fire – Civil War, followed by republican government until 1660, then the return of the monarchy – had left the city in the hands of less experienced administrators. The situation was exacerbated by a major plague epidemic, in 1665, that killed many people and drove others from the city. The mayor himself was relatively young, had come into office largely thanks to royal patronage, and had far less experience than most previous occupants of the po sition (Harding 2019; Porter 1998: 58). The case of London was exceptional, but economic and social dislocation and political instability were particularly widespread in the second half of the seventeenth century. Large Fires and Climatic Variability in Urban Europe, 1500–1800 the scale of urban fires. Towns where extensive use was made of non-flam mable building materials, and where dangerous activities were removed from residential areas, were far less affected by large fires. Those where firefighting was well funded and well organised were also often able to avoid large fires, even when meteorological conditions were unfavourable. These were crucial factors in reducing the overall number of large fires during the eighteenth cen tury, despite quite rapid urban growth. For precisely this reason, the social and economic disruptions brought by climatic conditions during the Little Ice Age, and the political dislocation that accompanied them, had an impact on the ability of some urban communities to deal with wildfire. Even towns where, in normal times, fires were usually con trolled quickly now had fewer resources available to invest in less flammable building materials, in prevention, and in training and equipping firefighters. There is circumstantial evidence that in some places these disruptions to local government contributed to reduced enforcement of preventive measures and to poor handling of dangerous fires. The early modern European experience offers a sobering warning of the potential impact of climatic variability on urban wildfire. Today’s global warm ing is producing forest fires of unprecedented intensity, which in Australia, the United States and parts of southern Europe have converged on urban periph eries. But the risk is greater in areas of overcrowded housing where building materials are more vulnerable to fire. For example, the favelas around São Paolo, in Brazil, experienced massive fires in 2009, 2012, 2014 and 2016. Similarly huge blazes destroyed slums and shantytowns in New Delhi and in Freetown, Sierra Leone, in 2018 and, in 2021, a similar one hit Dhaka, in Bangladesh. Equally vulnerable urban zones exist in many large cities, notably in tropical and subtropical areas. The reporting of these blazes has focused on the ramshackle nature of the buildings, the absence of fire safety measures, and often on the likelihood that arson was involved, but climatic factors have received less attention. In addition to the direct effects of heat and drought, the early modern example of social and political disruption that was directly or indirectly produced by climatic factors is worth considering in the twenty- first century context. Large Fires and Climatic Variability in Urban Europe, 1500–1800 This did not auto matically lead to poor fire preparedness or management, but the failure of local officials to provide equipment, and to act effectively when fire disasters took place, is well documented. This article has offered evidence of a causal link between large peacetime fires in early modern European towns and climatic variability linked to the Little Ice Age. The seventeenth century emerges as the worst period for such fires, both in the number that occurred and in their extent. It includes many of the most devastating ones, notably those of 1656 in Aachen and 1666 in London. This is at first surprising, given that climatic conditions were in general colder and wetter than in the Middle Ages or in the nineteenth and twentieth centu ries. Yet the individual years and seasons when large fires were most numerous were, with a few exceptions, hot and dry, and analysis of individual large and very large fires reveals that drought, heat, and high winds contributed directly to their gravity. These weather conditions were anomalies, which appear to be a feature of periods of climatic variability. While not directly causing the fires, such anomalies greatly contributed to their spread and intensity. They were not the only relevant variable. The impact of climatic conditions on wildfire, as we know from recent experience, is not independent of human activity. Just as increased housing on the fringes of cities has contributed to losses in recent years, so in early modern times the forms of building, zoning, and other preventative measures – or their absence – significantly influenced Climates and Cultures in History 1 (2024) 18 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Urban planning, building regulation, fire prevention and firefighting are, in most parts of the world, publicly funded and run, and risk dislocation if budgets are cut and enforcement disregarded. Some of this was already clear in the Grenfell Tower fire of 2017 in London. It seems, alas, that the age of great urban fires is far from over. Climates and Cultures in History 1 (2024) 19 David Garrioch Acknowledgments This article has benefited greatly from generous comments by Nigel Tapper, Jan Pinder and the anonymous reviewers for Climates and Cultures in History. My thanks to Kathleen Neal for statistical advice. The Australian Research Council provided funding that enabled the research, and I am immensely grateful to Cornel Zwierlein for shar ing some of his data on German-speaking Central Europe. Jana Verhoeven and Libby Burrell provided excellent research assistance, and I acknowledge assistance from ar chivists in many cities. Any errors are my own. Bibliography Aubert, G. 2010. ‘De la ville médiévale à la ville des Lumières’. In G. Aubert, A. Croix and M. Denis (eds). Histoire de Rennes. 2nd ed. Rennes: Apogée/Presses universi taires de Rennes. pp. 89–107. Aubert, G. and G. Provost (eds). 2020. Le grand incendie de Rennes, 1720. Rennes: Presses universitaires de Rennes. Bankoff, G., U. Lübken and J. Sand (eds). 2012. Flammable Cities: Urban Conflagration and the Making of the Modern World. Madison: University of Wisconsin Press. Bell, W. 1920. The Great Fire of London in 1666. London: John Lane. Blackstone, G.V. 1957. A History of the British Fire Service. London: Routledge and Kegan Paul. Borsay, P. 1989. The English Urban Renaissance: Culture and Society in the Provincial Town, 1660–1770. Oxford: Clarendon Press. Brázdil, R. P. Dobrovolný, M. Trnka, U. Büntgen, L. Rezníčková, O. Kotyza, H. Valášek and P. Štěpánek. 2016. ‘Documentary and instrumental-based drought in dices for the Czech lands back to AD 1501’, Climate Research 70: 103–1. https:// doi.org/10.3354/cr01380 Clark, A. (ed.) 1891–1900. The Life and Times of Anthony Wood, Antiquary, of Oxford, 1632–1695, Described by Himself. 5 vols. Oxford: Printed for the Oxford Historical Society, at the Clarendon Press. Clark, P. and P. Slack (eds). 1972. Crisis and Order in English Towns, 1500–1700. London: Routledge and Kegan Paul. Cleland, J. 1820. Enumeration of the Inhabitants of the City of Glasgow. Glasgow: Printed by James Hedderwick. Ericson, L. 2001. Stockholms historia under 750 år. Lund: Historiska media. Evelyn, J. 1955. The Diary of John Evelyn. 6 vols. Oxford: Clarendon Press. https://doi. org/10.1093/oseo/instance.00006908 Ewen, S. 2010. Fighting Fires. Creating the British Fire Service, 1800–1978. Houndmills, Basingstoke: Palgrave Macmillan. Climates and Cultures in History 1 (2024) 20 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Ewen, S. 2006. ‘The problem of fire in nineteenth-century British cities: the case of Glasgow’. In M. Dunkeld (ed.) Proceedings of the Second International Congress on Construction History. 4 vols. Cambridge: Construction History Society. vol. 1, pp. 1061–74. Fillaut, R. 1999. Rennes, Des combattants du feu aux techniciens du risque. Rennes: Amicale des sapeurs-pompiers de Rennes/Maury. Forsbom, A. 1950. ‘Eldsvådorna i Stockholm: kronologisk översikt’, Stockholms Borgargilles Årsbok: 3–18. Frenzel, B., C. Pfister and B. Gläser (eds). 1994. Climatic Trends and Anomalies in Europe (1675–1715). Stuttgart: Gustav Fischer Verlag. Frich, P. and K. Frydendahl. 1994. ‘The summer climate in the Øresund region of Denmark, A.D. 1675 to 1715’. In Frenzel, Pfister and Gläser (eds), Climatic Trends, pp. 33–41. Frost, L.E. and E.L. Jones. 1989. ‘The fire gap and the greater durability of nineteenth century cities’. Planning Perspectives 4: 333–47. https://doi. org/10.1080/02665438908725687 Gai, S. R. Mahytka and R. Otte. 2019. ‘Im Schatten der Domburg – 2200 Jahre Siedlungsgeschichte im Paderborner Stadtkern’, Archäologie in Westfalen-Lippe: 212–16. Gallicciolli, G. 1795. Delle memorie venete antiche profane ed ecclesiastiche raccolte da Giambattista Gallicciolli. 8 vols. Venice: Appresso Domenico Fracasso. Gamrath, H. 1999. ‘The Great Fire of Copenhagen in 1728’. In Körner, Bartlome and Flückiger (eds). Stadtzerstörung und Wiederaufbau, vol. 1, pp. 293–302. Garnier, E. 2019. ‘Historic drought from archives: beyond the instrumental record: science and policy’. In Iglesias, Assimacopoulos and Van Lanen (eds), Drought: Science and Policy. pp. 45–67. https://doi.org/10.1002/9781119017073.ch3 Garrioch, D. 2019a. ‘Towards a fire history of European cities, 1500–1850’. Urban History 46: 202–24. https://doi.org/10.1017/S0963926818000275 Garrioch, D. 2019b. ‘Why didn’t Paris burn in the seventeenth and eighteenth centuries?’ French Historical Studies 42 (1): 35–65. https://doi.org/10.1215/00161071-7205197 Garrioch, D. 2016. ‘1666 and London’s fire history: a re-evaluation’, Historical Journal 59 (2): 319–38. https://doi.org/10.1017/S0018246X15000382 Glaser, R. 2001. Klimageschichte Mitteleuropas. 1000 Jahre Wetter, Klima, Katastrophen. Darmstadt: Primus Verlag. Grasset, L. 1988. ‘L’incendie de Sainte Menehould, 7 au 8 août 1719’. Horizons d’Ar gonne 57: 9–24. Hahn, F. 1855. Geschichte von Gera und dessen nächster Umgebung Gera. 2 vols. Gera: Verlag des Verfassers. Harding, V. 2019. ‘Disaster management: the City of London and the Great Fire of 1666’. Paper presented at Firefighting and the Narratives of Great Fires in Early Modern Europe, workshop organized by DISCOMPOSE (Disasters, Communication and Politics in Southwestern Europe) and DECIvE (Dealing with the Collective Interest in early modern Europe), Naples, 22 January. Large Fires and Climatic Variability in Urban Europe, 1500–1800 21 Climates and Cultures in History 1 (2024) David Garrioch David Garrioch Herden, R.B. 2005. Roter Hahn und rotes Kreuz: Chronik der Geschichte des Feuerlösch- und Rettungswesens. Bad Rippoldsau-Schapbach and Kehl am Rhein: Books on Demand. van der Heyden, J. 1996 [1690]. A Description of Fire Engines with Water Hoses and the Method of Fighting Fires Now Used in Amsterdam. Translated and with intro duction by L.S. Multhauf. Canton, Mass.: Science History Publications. Iglesias, A., D. Assimacopoulos and H.A.J. Van Lanen (eds). 2019. Drought: Science and Policy. Hoboken, NJ: John Wiley & Sons. https://doi.org/10.1002/9781119017073 Jones, E.L., S. Porter and M. Turner. 1984. A Gazetteer of English Urban Fire Disasters, 1500–1900, Historical Geography Research Group of the Institute of British Geographers, Research Paper Series, No. 13. Norwich: Geo Books. Karpińsky, A. 2021. Pożary w miastach Rzeczypospolitej w XVI–XVIII wieku i ich następstwa ekonomiczne, społeczne i kulturowe. Katalog. Warsaw: Wydawnictwa Uniwersytetu Warszawskiego. https://doi.org/10.31338/uw.9788323543787 Keene, D. 1999. ‘Fire in London: destruction and reconstruction, A.D. 982–1676’. In Körner, Bartlome and Flückiger (eds), Stadtzerstörung und Wiederaufbau, pp. 187–211. Kraus, T.R. 2007. ‘Der Aachener Stadtbrand vom 2. Mai 1656 und seine Folgen’. Zeitschrift des Aachener Geschichtsvereins 109: 35–99. Körner, M., N. Bartlome and E. Flückiger (eds). 1999. Stadtzerstörung und Wiederaufbau: Zerstörungen durch Erdbeben, Feuer und Wasser. Destruction and Reconstruction of Towns: Destruction by Earthquakes, Fire and Water. 3 vols. Bern and Vienna: P. Haupt. Leggewie, C. and F. Mauelshagen (eds). 2018. Climate and Culture, Climate Change and Cultural Transition in Europe. Leiden: Brill. https://doi. org/10.1163/9789004356825 Leijonhufvud, L. and D. Retsö. 2021. ‘Documentary evidence of droughts in Sweden between the Middle Ages and ca. 1800 CE’. Climate of the Past 17: 2015–29. https://doi.org/10.5194/cp-17-2015-2021 Le Roux, T. 2011. Le Laboratoire des pollutions industrielles. Paris, 1770–1830. Paris: Albin Michel. pp. 46–68. Le Roy Ladurie, E. 2004–2006. Histoire humaine et comparée du climat. 2 vols. Paris: Fayard. https://doi.org/10.3406/edyte.2004.886 van Lieshout, C. 2016. ‘Droughts and dragons: Geography, rainfall, and eighteenth-cen tury London’s water systems’. Technology and Culture 57 (4): 780–805. https://doi. org/10.1353/tech.2016.0107 Luterbacher, J., D. Dietrich, E. Xoplaki, M. Grosjean and H. Wanner. 2004. ‘European seasonal and annual temperature variability, trends, and extremes Since 1500’. Science 303 (5 March): 1499–1503. https://doi.org/10.1126/science.1093877 Macadam, J. 2012. ‘English weather: the seventeenth-century diary of Ralph Josselin’. Journal of Interdisciplinary History 43: 221–46. https://doi.org/10.1162/ JINH_a_00380 Climates and Cultures in History 1 (2024) 22 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Manley, G. 1974. ‘Central England temperatures: monthly means 1659 to 1973’. Quarterly Journal of the Royal Meteorological Society 100: 389–405. https://doi. org/10.1002/qj.49710042511 Mauelshagen, F. 2010. Klimageschichte der Neuzeit, 1500–1900. Darmstadt: WBG. Mellado-Cano, J. D. Barriopedro, R. García-Herrera and R.M. Trigo. 2020. ‘New obser vational insights into the atmospheric circulation over the Euro-Atlantic sector since 1685’. Climate Dynamics 54: 823–41. https://doi.org/10.1007/s00382-019-05029-z Mellado-Cano, J., D. Barriopedro, R. García-Herrera, R.M. Trigo and M.C. Alvarez-Castro. 2018. ‘Euro-Atlantic atmospheric circulation during the late Maunder Minimum’. Journal of Climate 31: 3849–63. https://doi.org/10.1175/ JCLI-D-17-0261.1 Molesky, M. 2012. ‘The Great Fire of Lisbon, 1755’. In Bankoff, Lübken and Sand (eds), Flammable Cities, pp. 147–69. Nicolas, J. 2008 [2002]. La Rébellion française, 1661–1789. Paris: Gallimard. https:// doi.org/10.14375/NP.9782070359714 Nicolson, M.H. and S. Hutton (eds). 1992. The Conway Letters: The Correspondence of Anne, Viscountesse Conway, Henry More, and Their Friends, 1642–1684. Oxford: Oxford University Press, revised edition. https://doi.org/10.1093/ac trade/9780198248767.book.1 Nielsen, Y. 1877. Bergen fra de ældste tider indtil nutiden. En historisk-topografisk skildring. Christiania, C. Tønsberg. Parker, G. 2013. Global Crisis. War, Climate Change & Catastrophe in the Seventeenth Century. New Haven & London: Yale University Press. pp. 62–63. https://doi. org/10.2307/j.ctt32bksk Pauling, A., J. Luterbacher, C. Casty and H. Wanner. 2006. ‘Five hundred years of grid ded high-resolution precipitation reconstructions over Europe and the connection to large-scale circulation’. Climate Dynamics 26: 387–405. https://doi.org/10.1007/ s00382-005-0090-8 Pearson, R. 2004. Insuring the Industrial Revolution. Fire Insurance in Great Britain, 1700–1850. Aldershot: Ashgate. pp. 33–38. Pérémé, A. 1847. Recherches historiques et archéologiques sur la ville d’Issoudun. Paris: Duprat. Pfister, C. 2018. ‘The “Black Swan” of 1540: aspects of a European megadrought’. In C. Leggewie and F. Mauelshagen (eds). Climate and Culture, Climate Change and Cultural Transition in Europe. Leiden: Brill. pp. 156–93. https://doi. org/10.1163/9789004356825_007 Pfister, C. et al. 2018. ‘Early Modern Europe’. In White, Pfister and Mauelshagen (eds), Palgrave Handbook, pp. 265–95. https://doi.org/10.1057/978-1-137-43020-5_23 Pfister, C. 1994. ‘Spatial patterns of climatic change in Europe A.D. 1675 to 1715’. In Frenzel, Pfister and Gläser (eds), Climatic Trends, pp. 287–317. Pfister, C. 1988. Klimageschichte der Schweiz: 1525–1860. Das Klima der Schweiz von 1525 – 1860 und seine Bedeutung in der Geschichte von Bevölkerung und Landwirtschaft. 3rd ed. 2 vols in 1. Bern and Stuttgart: Paul Haupt. Climates and Cultures in History 1 (2024) 23 David Garrioch David Garrioch Pils, S.C. 1999. ‘“... Damit nur an Wasser khain Menngl erscheine ... Large Fires and Climatic Variability in Urban Europe, 1500–1800 ”: Vom Umgang der Stadt Wien mit dem Feuer in der frühen Neuzeit’. In Körner, Bartlome and Flückiger (eds), Stadtzerstörung und Wiederaufbau, vol. 1, pp. 173–86. Porter, S. 1998 [1996]. The Great Fire of London. Godalming, Surrey: Bramley Books. Power, M.J. 1972. ‘East London housing in the seventeenth century’. In Clark and Slack (eds), Crisis and Order in English Towns, 1500–1700. pp. 237–62. Pribyl, K. 2020. ‘A survey of the impact of summer droughts in southern and eastern England, 1200–1700’. Climate of the Past 16: 1027–41. https://doi.org/10.5194/ cp-16-1027-2020 Przybylak, K. et al. 2020. ‘Droughts in the area of Poland in recent centuries in the light of multi-proxy data’. Climate of the Past 16: 627–61. https://doi.org/10.5194/ cp-16-627-2020 Pyne, S.J. 2001. Fire. A Brief History. Seattle: University of Washington Press. Rácz, L. 2013. The Steppe to Europe: An Environmental History of Hungary in the Traditional Age. Cambridge: The White Horse Press. Robreau, B. 2009. Histoire du pays dunois. 2 vols. Châteaudun: Société dunoise d’archéologie. Sad and Lamentable News from Wapping. 1682. London: Printed for J. Clarke, Seignior. Snell, F.J. 1892. The Chronicles of Twyford, Being a New and Popular History of the Town of Tiverton. Tiverton and London: Gregory, Son, and Tozer/Elliot Stock. Strype, J. 1720. A Survey of the Cities of London and Westminster. 2 vols. London: A. Churchill et al. Svalduz, E. 2006. ‘Dal fuoco si rinasce. Gli incendi a Venezia dal xv al xviii secolo’. In D. Calabi (ed.), Venezia in Fumo. I grandi incendi della città-fenice. Bergamo: Leading Edizioni. pp. 41–82. Vesajoki, H. and M. Tornberg. 1994. ‘Outlining the climate in Finland during the pre- instrumental period on the basis of documentary sources’. In Frenzel, Pfister and Gläser (eds), Climatic Trends and Anomalies in Europe. pp. 51–60. de Vries, J. 1984. European Urbanization, 1500–1800. Cambridge, Mass.: Harvard University Press. Weigl, A. 2001–2006. ‘Gewerbliche Konjunkturen’. In P. Csendes and F. Opll (eds), Wien: Geschichte einer Stadt. 3 vols. Vienna: Böhlau. vol. 2, pp. 146–61. White, S., C. Pfister and F. Mauelshagen (eds). 2018. The Palgrave Handbook of Climate History. London: Palgrave. https://doi.org/10.1057/978-1-137-43020-5 White, S., J. Brooke, and C. Pfister. 2018. ‘Climate, weather, agriculture, and food’, in White, Pfister and Mauelshagan (eds), The Palgrave Handbook of Climate History, pp. 331–53. https://doi.org/10.1057/978-1-137-43020-5_27 Xoplaki, E., J. Luterbacher, H. Paeth, D. Dietrich, N. Steiner, M. Grosjean and H. Wanner. 2005. Large Fires and Climatic Variability in Urban Europe, 1500–1800 ‘European spring and autumn temperature variability and change of extremes over the last half millennium’. Geophysical Research Letters 32. https:// doi.org/10.1029/2005GL023424 Climates and Cultures in History 1 (2024) 24 Large Fires and Climatic Variability in Urban Europe, 1500–1800 Zwierlein, C. 2021 [2011]. Prometheus Tamed: Fire, Security, and Modernities, 1400 to 1900. Leiden and Boston: Brill. https://doi.org/10.1163/9789004431225 Zwierlein, C. 2021 [2011]. Prometheus Tamed: Fire, Security, and Modernities, 1400 to 1900. Leiden and Boston: Brill. https://doi.org/10.1163/9789004431225 Archives Bibliothèque nationale, Paris, France. Collection Joly de Fleury. Stockholm City Archives (SSA), John Swensk Collection. Climates and Cultures in History 1 (2024) THE AUTHOR David Garrioch is Emeritus Professor of History at Monash University. He works on the social history of eighteenth-century Paris and of other European cities in the early modern period. His most recent book is The Huguenots of Paris and the coming of Religious Freedom (Cambridge UP, 2014). Email: David.Garrioch@monash.edu ORCID: https://orcid.org/0000-0003-0827-309X David Garrioch is Emeritus Professor of History at Monash University. He works on the social history of eighteenth-century Paris and of other European cities in the early modern period. His most recent book is The Huguenots of Paris and the coming of Religious Freedom (Cambridge UP, 2014). Climates and Cultures in History 1 (2024) Climates and Cultures in History 1 (2024) 25
W2051823049.txt	https://www.ams.org/tran/2008-360-10/S0002-9947-08-04484-X/S0002-9947-08-04484-X.pdf	en		Topological structure of (partially) hyperbolic sets with positive volume	Transactions of the American Mathematical Society	2,008	public-domain	10,075	TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 360, Number 10, October 2008, Pages 5551–5569 S 0002-9947(08)04484-X Article electronically published on April 28, 2008 TOPOLOGICAL STRUCTURE OF (PARTIALLY) HYPERBOLIC SETS WITH POSITIVE VOLUME JOSÉ F. ALVES AND VILTON PINHEIRO Abstract. We consider both hyperbolic sets and partially hyperbolic sets attracting a set of points with positive volume in a Riemannian manifold. We obtain several results on the topological structure of such sets for diﬀeomorphisms whose diﬀerentiability is larger than one. We show in particular that there are no partially hyperbolic horseshoes with positive volume for such diffeomorphisms. We also give a description of the limit set of almost every point belonging to a hyperbolic set or a partially hyperbolic set with positive volume. 1. Introduction Since the 60s hyperbolic sets have played an important role in the development of the Theory of Dynamical Systems. These are invariant (by a smooth map) compact sets over which the tangent bundle splits into two invariant subbundles, one of them contracting and the other one expanding under the action of the derivative of the map. In this work we are concerned with discrete dynamical systems (smooth transformations of a manifold), but our techniques also proved useful in the continuous setting (vector ﬁelds in a manifold), especially for the study of singularhyperbolic sets as done in [2]. In the last decades an increasing emphasis has been put on the dynamics of partially hyperbolic sets. These are compact invariant sets for which the tangent bundle splits into two invariant subbundles having contracting/expanding behavior in one direction and the other one being dominated by it. Precise deﬁnitions of all these objects will be given in the next section. In this context a special role has been played by the horseshoes, which have been introduced by Smale and, as shown in [17], always exist near a transverse homoclinic point associated to some hyperbolic periodic point of saddle type, i.e. a point whose orbit asymptotically approaches that saddle point, both in the past and in the future. Horseshoes can be used to show that transverse homoclinic points are always accumulated by periodic points, but the dynamical richness of these objects goes far beyond the initial application by Smale, and since then many Received by the editors June 5, 2006 and, in revised form, January 8, 2007. 2000 Mathematics Subject Classiﬁcation. Primary 37Dxx. Key words and phrases. Hyperbolic set, partially hyperbolic set, horseshoe. This work was carried out at the Federal University of Bahia, University of Porto and IMPA. The ﬁrst author was partially supported by CMUP, by a grant of FCT and by POCI/MAT/61237/2004. The second author was partially supported by PADCT/CNPq and by POCI/MAT/61237/2004. c 2008 American Mathematical Society Reverts to public domain 28 years from publication 5551 5552 JOSÉ F. ALVES AND VILTON PINHEIRO other results have been proved using horseshoes. These are Cantor sets which are, in dynamical terms, topologically conjugated to full shifts. Special interest lies in the horseshoes that appear when one unfolds a homoclinic tangency. Knowing how fat these horseshoes are can have several implications in the dynamical behavior after the homoclinic bifurcation. In this setting we mention the thickness, which has been used by Newhouse [10] to prove the existence of inﬁnitely many sinks, and the Hausdorﬀ dimension, which has been used by Palis, Takens and Yoccoz to study the prevalence of hyperbolicity after the unfolding of a homoclinic tangency; see [11, 12, 13]. One interesting issue we will address is the volume of horseshoes. As shown by Bowen in [6], there are C 1 diﬀeomorphisms with hyperbolic horseshoes of positive volume. On the other hand, Bowen has proved in [7, Theorem 4.11] that a basic set (locally maximal hyperbolic set with a dense orbit) of a C 2 diﬀeomorphism which attracts a set with positive volume, necessarily attracts a neighborhood of itself. In particular, the unstable manifolds through points of this set must be contained in it, and consequently C 2 diﬀeomorphisms have no horseshoes with positive volume. For diﬀeomorphisms whose diﬀerentiability is greater than one, we prove the nonexistence of horseshoes with positive volume in a more general context of sets with some partially hyperbolic structure. Using our framework in the context of hyperbolic sets, we are able to show that Bowen’s result still holds without the local maximality assumption, i.e. a transitive hyperbolic set which attracts a set with positive volume necessarily attracts a neighborhood of itself. Furthermore, we are able to prove that there are no proper transitive hyperbolic sets with positive volume for diﬀeomorphisms whose diﬀerentiability is greater than one. Similar results for sets with nonempty interior have already been obtained in [1, Theorem 1] and in [8, Theorem 1.1]. On the other hand, as described in [1, Remark 2.1] or in [8, Example 2], there exist (nontransitive) hyperbolic sets with positive volume which do not attract neighborhoods of themselves. Let us mention two more important results in this direction. It follows from [18, Theorem 2] that proper uniformly partially hyperbolic sets supporting a unique equilibrium state and attracting open neighborhoods of themselves necessarily have zero volume. In the conservative setting, [5, Theorem 15] gives that a hyperbolic set for a volume preserving C 2 diﬀeomorphism either has zero volume or coincides with the whole manifold. In this work we also give a good description of the limit set of almost every point in a hyperbolic set with positive volume: there is a ﬁnite number of basic sets for which the ω-limit set of Lebesgue almost every point is contained in one of these basic sets. We are also able to prove in a partially hyperbolic setting that these ω-limit sets are contained in the closure of ﬁnitely many hyperbolic periodic points. 2. Statement of results Let f : M → M be a diﬀeomorphism of a compact connected Riemannian manifold M . We say that f is C 1+ if f is C 1 and Df is Hölder continuous. We use Leb to denote a normalized volume form on the Borel sets of M that we call Lebesgue measure. Given a submanifold γ ⊂ M we use Lebγ to denote the measure on γ induced by the restriction of the Riemannian structure to γ. A set Λ ⊂ M is said to be invariant if f (Λ) = Λ, and positively invariant if f (Λ) ⊂ Λ. PARTIALLY HYPERBOLIC SETS 5553 2.1. Partially hyperbolic sets. Let K be a positively invariant compact set, and deﬁne f n (K). Λ= n≥0 Suppose that there is a continuous splitting TK M = E cs ⊕ E cu of the tangent bundle restricted to K, and assume that this splitting is Df -invariant over Λ. We say that this is a dominated splitting (over Λ) if there is a constant 0 < λ < 1 such that for some choice of a Riemannian metric on M Df \| Excs · Df −1 \| Efcu(x) ≤ λ, for every x ∈ Λ. We call E cs the centre-stable bundle and E cu the centre-unstable bundle. Λ is said to be partially hyperbolic if additionally E cs is uniformly contracting or E cu is uniformly expanding, meaning that there exists 0 < λ < 1 such that Df \| Excs ≤ λ, for every x ∈ Λ, or Df −1 \| Efcu(x) ≤ λ, for every x ∈ Λ. We say that f is nonuniformly expanding along the centre-unstable direction in K if there is c > 0 such that for Lebesgue almost every x ∈ K 1 log Df −1 \| Efcuj (x) < −c. n j=1 n (NUE) lim inf n→+∞ Condition NUE means that the derivative has expanding behavior in the centreunstable direction in average over the orbit of x for an inﬁnite number of times. If condition NUE holds for every point in a compact invariant set Λ, then E cu is uniformly expanding in the centre-unstable direction in Λ. This is not necessarily the case if NUE occurs only Lebesgue almost everywhere. A class of diﬀeomorphisms with a dominated splitting T M = E cs ⊕ E cu for which NUE holds Lebesgue almost everywhere in M and E cu is not uniformly expanding can be found in [4, Appendix A]. We say that an embedded disk γ ⊂ M is an unstable manifold, or an unstable disk, if dist(f −n (x), f −n (y)) → 0 exponentially fast as n → ∞, for every x, y ∈ γ. Similarly, γ is called a stable manifold, or a stable disk, if dist(f n (x), f n (y)) → 0 exponentially fast as n → ∞, for every x, y ∈ γ. It is well known that every s (x) and a local point in a hyperbolic set possesses a local stable manifold Wloc u s u unstable manifold Wloc (x) which are disks tangent to Ex and Ex at x respectively. A compact invariant set Λ is said to be horseshoe-like if there are local stable and local unstable manifolds through all its points which intersect Λ in a Cantor set. Theorem A. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a forward invariant compact set with a continuous splitting TK M = E cs ⊕ E cu n dominated over Λ = n≥0 f (K). If NUE holds for a positive Lebesgue set of points x ∈ K, then Λ contains some local unstable disk. The next result is a direct consequence of Theorem A whenever E cu is uniformly expanding. If, on the other hand, E cs is uniformly contracting, then we just have to apply Theorem A to f −1 . 5554 JOSÉ F. ALVES AND VILTON PINHEIRO Corollary B. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a compact invariant set with Leb(K) > 0 having a continuous splitting TK M = E cs ⊕ E cu for which Λ = n≥0 f n (K) is partially hyperbolic. (1) If E cs is uniformly contracting, then Λ contains a local stable disk. (2) If E cu is uniformly expanding, then Λ contains a local unstable disk. In particular, C 1+ diﬀeomorphisms have no partially hyperbolic horseshoe-like sets with positive volume. The same conclusion holds for partially hyperbolic sets intersecting a local stable disk or a local unstable disk in a positive Lebesgue measure subset, as Corollary D below shows. Theorem C. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a forward invariant compact set with a continuous splitting TK M = E cs ⊕ E cu dominated over Λ = n≥0 f n (K). Assume that there is a local unstable disk γ such that NUE holds for every x in a positive Lebγ subset of γ ∩ K. Then Λ contains some local unstable disk. The next result is an immediate consequence of Theorem C, in the case that E cu is uniformly expanding, and a consequence of the same theorem applied to f −1 when E cs is uniformly contracting. Actually, we shall prove a stronger version of this result in Theorem 5.1. Corollary D. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a forward invariant compact set having a continuous splitting TK M = E cs ⊕ E cu dominated over Λ = n≥0 f n (K). (1) If E cs is uniformly contracting and there is a local stable disk γ such that Lebγ (γ ∩ K) > 0, then Λ contains a local stable disk. (2) If E cu is uniformly expanding and there is a local unstable disk γ such that Lebγ (γ ∩ K) > 0, then Λ contains a local unstable disk. Using the previous results we are able to give a description of the ω-limit of Lebesgue almost every point in a partially hyperbolic whose centre-unstable direction displays nonuniform expansion in a subset with positive volume. Recall that the ω-limit of x ∈ M is the set of accumulation points of its orbit. Theorem E. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a forward invariant compact set with Leb(K) > 0 having a continuous splitting TK M = E cs ⊕ E cu for which Λ = n≥0 f n (K) is partially hyperbolic. Assume that E cs is uniformly contracting and NUE holds for Lebesgue almost every x ∈ K. Then there are hyperbolic periodic points p1 , . . . , pk ∈ Λ such that: (1) W u (pi ) ⊂ Λ for each 1 ≤ i ≤ k; (2) for Leb almost every x ∈ K there is 1 ≤ i ≤ k with ω(x) ⊂ W u (pi ). Moreover, if E cu has dimension one, then for each 1 ≤ i ≤ k (3) W u (pi ) attracts an open neighborhood of itself. This last conclusion also holds whenever E cs is uniformly contracting. Actually, more can be said in the case of uniformly hyperbolic sets with positive volume as we shall see in the next subsection. PARTIALLY HYPERBOLIC SETS 5555 2.2. Hyperbolic sets. We say that a compact invariant set Λ is hyperbolic if there is a Df -invariant splitting TΛ M = E s ⊕ E u of the tangent bundle restricted to Λ and a constant λ < 1 such that (for some choice of a Riemannian metric on M ) for every x ∈ Λ Df \| Exs < λ and Df −1 \| Exu < λ. We are able to prove that transitive hyperbolic sets with positive volume necessarily coincide with the whole manifold, i.e. the diﬀeomorphism is Anosov. Theorem F. Let f : M → M be a C 1+ diﬀeomorphism and let Λ ⊂ M be a transitive hyperbolic set. (1) If Λ has positive volume, then Λ = M . (2) If Λ attracts a set with positive volume, then Λ attracts a neighborhood of itself. The main reason why we cannot generalize the results in this subsection to the context of partially hyperbolic sets is that the length of local stable/unstable manifolds may shrink to zero when iterated back/forth, respectively. The next result gives a description of the ω-limit of Lebesgue almost every point in a hyperbolic set with positive volume. Taking f −1 , a similar decomposition holds for α-limits. Theorem G (Spectral decomposition). Let f : M → M be a C 1+ diﬀeomorphism and let Λ ⊂ M be a hyperbolic set with positive volume. There are hyperbolic sets Ω1 , . . . , Ωq ⊂ Λ such that: (1) (2) (3) (4) for Leb almost every x ∈ Λ there is 1 ≤ i ≤ q such that ω(x) ⊂ Ωi ; Ωj attracts a neighborhood of itself in M , for each 1 ≤ j ≤ q; f \|Ωk is transitive; Per(f ) is dense in Ωj , for each 1 ≤ j ≤ q. Moreover, for each 1 ≤ k ≤ q there is a decomposition of Ωk into disjoint hyperbolic sets Ωk = Ωk,1 ∪ · · · ∪ Ωk,nk such that: (5) f (Ωk,i ) = Ωk,i+1 , for 1 ≤ i < nk , and f (Ωk,nk ) = Ωk,1 ; (6) f nk : Ωk,i → Ωk,i is topologically mixing for every 1 ≤ i ≤ nk . 2.3. Overview. This paper is organized in the following way. In Section 3 we present some results from [4] on the Hölder control of the tangent direction of certain submanifolds, and in Section 4 we derive some bounded distortion results. Theorem A and Theorem C are actually corollaries of a slightly more general result that we present at the beginning of Section 5. Let us mention that the results in Section 5 (especially Lemma 5.4) are not a consequence of the results in [4], since we are using a weaker form of nonuniform expansion in NUE. Theorem E is proved in Section 6. Finally, in Section 7 we prove Theorem F and Theorem G. 3. Hölder control of tangent direction In this section we present some results from [4, Section 2] concerning the Hölder control of the tangent direction of submanifolds. Though the results in [4] are stated for C 2 diﬀeomorphims, they are valid for diﬀeomorphisms of class C 1+ , as observed in [4, Remark 2.3]. 5556 JOSÉ F. ALVES AND VILTON PINHEIRO Let K be a positively invariant compact set for which there is a continuous splitting TK M = E cs ⊕ E cu of the tangent bundle restricted to K which is Df invariant over Λ= f n (K). n≥0 We ﬁx continuous extensions of the two bundles E cs and E cu to some compact neighborhood U of Λ, that we still denote by E cs and E cu . Replacing K by a forward iterate of it, if necessary, we may assume that K ⊂ U . Given 0 < a < 1, we deﬁne the centre-unstable cone ﬁeld (Cacu (x))x∈U of width a by (1) Cacu (x) = v1 + v2 ∈ Excs ⊕ Excu such that v1 ≤ av2 . We deﬁne the centre-stable cone ﬁeld (Cacs (x))x∈U of width a in a similar way, just reversing the roles of the subbundles in (1). We ﬁx a > 0 and U small enough so that, up to slightly increasing λ < 1, the domination condition remains valid for any pair of vectors in the two cone ﬁelds, i.e. Df (x)v cs · Df −1 (f (x))v cu ≤ λv cs v cu , for every v cs ∈ Cacs (x), v cu ∈ Cacu (f (x)), and any x ∈ U ∩ f −1 (U ). Note that the centre-unstable cone ﬁeld is positively invariant: Df (x)Cacu (x) ⊂ Cacu (f (x)), whenever x, f (x) ∈ U . Indeed, the domination property together with the invariance of E cu over Λ implies that (2) cu Df (x)Cacu (x) ⊂ Cλa (f (x)) for every x ∈ Λ. This extends to any x ∈ U ∩ f −1 (U ) just by continuity, slightly increasing λ < 1, if necessary. Remark 3.1. The invariance of the splitting TK M = E cs ⊕ E cu is used in [4] to derive conclusions for the points in the small neighborhood U of K. Though we are taking here the invariance of the splitting restricted to Λ, since we are assuming K ⊂ U , where U is a small neighborhood of Λ, the results of [4, Section 2.1] are still valid in our situation. See also [4, Remark 2.1]. We say that an embedded C 1 submanifold N ⊂ U is tangent to the centreunstable cone ﬁeld if the tangent subspace to N at each point x ∈ N is contained in the corresponding cone Cacu (x). Then f (N ) is also tangent to the centre-unstable cone ﬁeld, if it is contained in U , by the domination property. We choose δ0 > 0 small enough so that the inverse of the exponential map expx is deﬁned on the δ0 neighborhood of every point x in U . From now on we identify this neighborhood of x with the corresponding neighborhood Ux of the origin in Tx N , through the local chart deﬁned by exp−1 x . Reducing δ0 , if necessary, we may suppose that Excs is contained in the centrestable cone Cacs (y) of every y ∈ Ux . In particular, the intersection of Cacu (y) with Excs reduces to the zero vector. Then, the tangent space to N at y is parallel to the graph of a unique linear map Ax (y) : Tx N → Excs . Given constants C > 0 and 0 < ζ ≤ 1, we say that the tangent bundle to N is (C, ζ)-Hölder if for every y ∈ N ∩ Ux and x ∈ V0 (3) Ax (y) ≤ Cdx (y)ζ , PARTIALLY HYPERBOLIC SETS 5557 where dx (y) denotes the distance from x to y along N ∩ Ux , deﬁned as the length of the shortest curve connecting x to y inside N ∩ Ux . Recall that we have chosen the neighborhood U and the cone width a suﬃciently small so that the domination property remains valid for vectors in the cones Cacs (z), Cacu (z), and for any point z in U . Then, there exist λ1 ∈ (λ, 1) and ζ ∈ (0, 1] such that (4) Df (z)v cs · Df −1 (f (z))v cu 1+ζ ≤ λ1 < 1 for every norm 1 vectors v cs ∈ Cacs (z) and v cu ∈ Cacu (z), at any z ∈ U . Then, up to reducing δ0 > 0 and slightly increasing λ1 < 1, condition (4) remains true if we replace z by any y ∈ Ux , with x ∈ U (taking · to mean the Riemannian metric in the corresponding local chart). We ﬁx ζ and λ1 as above. Given a C 1 submanifold N ⊂ U , we deﬁne (5) κ(N ) = inf{C > 0 : the tangent bundle of N is (C, ζ)-Hölder}. The next result appears in [4, Corollary 2.4]. Proposition 3.2. There exists C1 > 0 such that, given any C 1 submanifold N ⊂ U tangent to the centre-unstable cone ﬁeld, (1) there exists n0 ≥ 1 such that κ(f n (N )) ≤ C1 for every n ≥ n0 such that f k (N ) ⊂ U for all 0 ≤ k ≤ n; (2) if κ(N ) ≤ C1 , then the same is true for every iterate f n (N ) such that f k (N ) ⊂ U for all 0 ≤ k ≤ n; (3) in particular, if N and n are as in (2), then the functions Jk : f k (N ) x −→ log \| det Df \| Tx f k (N ) \|, 0 ≤ k ≤ n, are (L, ζ)-Hölder continuous with L > 0 depending only on C1 and f . 4. Hyperbolic times and bounded distortion Let K ⊂ M be a forward invariant compact set and let Λ ⊂ K ⊂ U be as in Section 3. The following notion will allow us to derive uniform behaviour (expansion, distortion) from the nonuniform expansion. Deﬁnition 4.1. Given σ < 1, we say that n is a σ-hyperbolic time for x ∈ K if n Df −1 \| Efcuj (x) ≤ σ k , for all 1 ≤ k ≤ n. j=n−k+1 If a > 0 is taken suﬃciently small in the deﬁnition of our cone ﬁelds, and we choose δ1 > 0 also small so that the δ1 -neighborhood of K should be contained in U , then by continuity (6) 1 Df −1 (f (y))v ≤ √ Df −1 \|Efcu(x) v, σ whenever x ∈ K, dist(x, y) ≤ δ1 and v ∈ Cacu (f (y)). Given any disk ∆ ⊂ M , we use dist∆ (x, y) to denote the distance between x, y ∈ ∆ measured along ∆. The distance from a point x ∈ ∆ to the boundary of ∆ is dist∆ (x, ∂∆) = inf y∈∂∆ dist∆ (x, y). 5558 JOSÉ F. ALVES AND VILTON PINHEIRO Lemma 4.2. Take any C 1 disk ∆ ⊂ U of radius δ, with 0 < δ < δ1 , tangent to the centre-unstable cone ﬁeld. There is n0 ≥ 1 such that for x ∈ ∆ ∩ K with dist∆ (x, ∂∆) ≥ δ/2 and n ≥ n0 a σ-hyperbolic time for x, then there is a neighborhood Vn of x in ∆ such that: (1) f n maps Vn diﬀeomorphically onto a disk of radius δ1 around f n (x) tangent to the centre-unstable cone ﬁeld; (2) for every 1 ≤ k ≤ n and y, z ∈ Vn , distf n−k (Vn ) (f n−k (y), f n−k (z)) ≤ σ k/2 distf n (Vn ) (f n (y), f n (z)); (3) for every 1 ≤ k ≤ n and y ∈ Vn , n Df −1 \| Efcuj (y) ≤ σ k/2 . j=n−k+1 Proof. First we show that f n (∆) contains some disk of radius δ1 around f n (x), as long as (7) n>2 log(δ/(2δ1 )) . log(σ) Deﬁne ∆1 as the connected component of f (∆) ∩ U containing f (x). For k ≥ 1, we inductively deﬁne ∆k+1 ⊂ f k+1 (∆) as the connected component of f (∆k ) ∩ U containing f k+1 (x). We shall prove that ∆n contains some disk of radius δ1 around f n (x), for n as in (7). Observe that since ∆j ⊂ U , the invariance (2) gives that for every j ≥ 1 Tw ∆j ⊂ Cλcuj a (w), (8) for every w ∈ ∆j . Let η0 be a curve of minimal length in ∆n connecting f n (x) to f n (y) ∈ ∆n for which dist∆n (f n (x), f n (y)) < δ1 . For 0 ≤ k ≤ n, writing ηk = f −k (η0 ) we have ηk ⊂ ∆n−k . We prove by induction that length(ηk ) < σ k/2 δ1 , for 0 ≤ k ≤ n. Let 1 ≤ k ≤ n and assume that length(ηj ) < σ j/2 δ1 , for 0 ≤ j ≤ k − 1. Denote by η̇0 (w) the tangent vector to the curve η0 at the point w. Using the fact that ηk ⊂ ∆n−k and (8) we have Df −j (w)η̇0 (w) ∈ Cλcun−j a (f −j (w)) ⊂ Cacu (f −j (w)). Then, by the choice of δ1 in (6) and the deﬁnition of σ-hyperbolic time, Df −k (w)η̇0 (w) ≤ σ −k/2 η̇0 (w) n Df −1 \|Efcuj (x) ≤ σ k/2 η̇0 (w). j=n−k+1 Hence, length(ηk ) ≤ σ k/2 length(η0 ) < σ k/2 δ1 . This completes our induction. In particular we have length(ηn ) < σ n/2 δ1 . Moreover, the k preimage of the ball of radius δ1 in ∆n centered at f n (x) is contained in U for each 1 ≤ k ≤ n. If ηn is a curve in ∆ connecting x to y ∈ ∂∆, then we must have n<2 log(δ/(2δ1 )) . log(σ) Hence f n (∆) contains some disk of radius δ1 around f n (x) for n as in (7). PARTIALLY HYPERBOLIC SETS 5559 Now let D1 be the disk of radius δ1 around f n (x) in f n (∆) and let Vn = f −n (D1 ), for n as in (7). Take any y, z ∈ Vn and let η0 be a curve of minimal length in D1 connecting f n (y) to f n (z). Deﬁning ηk = f −n+k (η0 ), for 1 ≤ k ≤ n, and arguing as before, we inductively prove that for 1 ≤ k ≤ n length(ηk ) ≤ σ k/2 length(η0 ) = σ k/2 distf n (Vn ) (f n (y), f n (z)), which implies that for 1 ≤ k ≤ n distf n−k (Vn ) (f n−k (y), f n−k (z)) ≤ σ k/2 distf n (Vn ) (f n (y), f n (z)). This completes the proof of the ﬁrst two items of the lemma. Given y ∈ Vn we have dist(f j (x), f j (y)) ≤ δ1 for every 1 ≤ j ≤ n, which together with (6) gives n n Df −1 \| Efcuj (y) ≤ σ −k/2 Df −1 \| Efcuj (x) ≤ σ k/2 . j=n−k+1 j=n−k+1 Recall that f (x) ∈ K for every j, and n is a σ-hyperbolic time for x. j We shall sometimes refer to the sets Vn as hyperbolic pre-balls and to their images f n (Vn ) as hyperbolic balls. Notice that the latter are indeed balls of radius δ1 . Corollary 4.3 (Bounded distortion). There exists C2 > 1 such that given ∆ as in Lemma 4.2 with κ(∆) ≤ C1 , and given any hyperbolic pre-ball Vn ⊂ ∆ with n ≥ n0 , then for all y, z ∈ Vn 1 \| det Df n \| Ty ∆\| ≤ C2 . ≤ C2 \| det Df n \| Tz ∆\| Proof. For 0 ≤ i < n and y ∈ ∆, we denote Ji (y) = \| det Df \| Tf i (y) f i (∆)\|. Then, log n−1 \| det Df n \| Ty ∆\| = log Ji (y) − log Ji (z) . n \| det Df \| Tz ∆\| i=0 By Proposition 3.2, log Ji is (L, ζ)-Hölder continuous, for some uniform constant L > 0. Moreover, by Lemma 4.2, the sum of all distf j (∆) (f j (y), f j (z))ζ over 0 ≤ j ≤ n is bounded by 2δ1 /(1 − σ ζ/2 ). Then it suﬃces to take C2 = exp(2δ1 L/(1 − σ ζ/2 )). 5. A local unstable disk inside Λ Now we are able to prove Theorems A and C. These will be obtained as corollaries of the next slightly more general result, as we shall see. Take K ⊂ M as a forward invariant compact set and let Λ ⊂ K ⊂ U be as before. Theorem 5.1. Let f : M → M be a C 1+ diﬀeomorphism and let K ⊂ M be a forward invariant compact set with a continuous splitting TK M = E cs ⊕ E cu dominated over Λ = n≥0 f n (K). Assume that there is a disk ∆ tangent to the centre-unstable cone ﬁeld intersecting K in a positive Leb∆ set of points where NUE holds. Then Λ contains some local unstable disk. Let us show that Theorem 5.1 implies Theorem A. Assume that NUE holds for Lebesgue almost every x ∈ K with Leb(K) > 0. Choosing a Leb density point of K, we laminate a neighborhood of that point into disks tangent to the centre-unstable cone ﬁeld contained in U . Since the relative Lebesgue measure of the intersections 5560 JOSÉ F. ALVES AND VILTON PINHEIRO of these disks with K cannot be all equal to zero, we obtain some disk ∆ as in the assumption of Theorem 5.1. For showing that Theorem 5.1 implies Theorem C, we just have to observe that local unstable manifolds are tangent to the centre-unstable subspaces and these vary continuously with the points in K, thus being tangent to the centre-unstable cone ﬁeld. In the remainder of this section we shall prove Theorem 5.1. Let H ⊂ K be the set of points where NUE holds. Let ∆ be a disk tangent to the centre-unstable cone ﬁeld intersecting H in a positive Leb∆ . Since NUE remains valid under positive iteration, by Proposition 3.2 we may assume that κ(∆) < C1 . It is no restriction to assume that H intersects the sub-disk of ∆ of radius δ/2, for some 0 < δ < δ1 , in a positive Leb∆ subset, and we do so. The following lemma is due to Pliss [15], and a proof of it in this precise form can be found in [4, Lemma 3.1]. Lemma 5.2. Given A ≥ c2 > c1 > 0 there exists θ > 0 such that for any real numbers a1 , . . . , aN with aj ≤ A and N aj ≥ c2 N, for every 1 ≤ j ≤ N, j=1 there are l > θN and 1 < n1 < · · · < nl ≤ N so that ni aj ≥ c1 (ni − n), for every 0 ≤ n < ni and 1 ≤ i ≤ l. j=n+1 Corollary 5.3. There is σ > 0 such that every x ∈ H has inﬁnitely many σhyperbolic times. Proof. Given x ∈ H, by NUE we have inﬁnitely many positive integers N for which N log Df −1 \|Efcuj (x) ≤ −cN. j=1 Then it suﬃces to take c1 = c/2, c2 = c, A = sup log Df −1 \|E cu , and aj = − log Df −1 \|Efcuj (x) in the previous lemma. Note that under assumption NUE we are unable to prove the existence of a positive frequency of hyperbolic times at inﬁnity, as in [4, Corollary 3.2]. This would be possible if we had taken lim sup instead of lim inf in the deﬁnition of NUE. The existence of inﬁnitely many hyperbolic times is enough for what follows. Lemma 5.4. Let O be an open set in ∆ such that Leb∆ (O ∩ H) > 0. Given any small ρ > 0 there is a hyperbolic time n, a hyperbolic pre-ball V ⊂ O and W ⊂ V such that ∆n = f n (W ) is a disk of radius δ1 /4 tangent to the centre-unstable cone ﬁeld, and Leb∆n (f n (H)) ≥ (1 − ρ)Leb∆n (∆n ). Proof. Take a small number > 0. Let C be a compact subset of O ∩ H and let A be an open neighborhood of O ∩ H in ∆ such that Leb∆ (A \ C) < Leb∆ (C). It follows from Corollary 5.3 and Lemma 4.2 that we can choose for each x ∈ C a σ-hyperbolic time n(x) and a hyperbolic pre-ball Vx such that Vx ⊂ A. Recall PARTIALLY HYPERBOLIC SETS 5561 that Vx is the neighborhood of x which is mapped diﬀeomorphically by f n(x) onto a ball Bδ1 (f n(x) (x)) of radius δ1 around f n(x) (x), tangent to the centre-unstable cone ﬁeld. Let Wx ⊂ Vx be the pre-image of the ball Bδ1 /4 (f n(x) (x)) of radius δ1 /4 under this diﬀeomorphism. By compactness there are x1 , ..., xm ∈ C such that C ⊂ Wx1 ∪ ... ∪ Wxs . Writing (9) {n1 , ..., ns } = {n(x1 ), ..., n(xm )}, with n1 < n2 < ... < ns , let I1 ⊂ N be a maximal set of {1, ..., m} such that if i ∈ I1 then n(xi ) = n1 and Wxi ∩ Wxj = ∅ for all j ∈ I1 with j = i. Inductively we deﬁne Ik for 2 ≤ k ≤ s as follows: Supposing that Ik−1 has already been deﬁned, let Ik ⊂ N be a maximal set of {1, . . . , m} such that if i ∈ Ik , then n(xi ) = nk and Wxi ∩ Wxj = ∅ for all j ∈ Ik with j = i, and also Wxi ∩ Wxj = ∅ for all j ∈ I1 ∪ ... ∪ Ik−1 . Let I = I1 ∪ . . . ∪ Is . By maximality, each Wxj , for 1 ≤ j ≤ m, intersects some Wxi with i ∈ I and n(xj ) ≥ n(xi ). Thus, given any 1 ≤ j ≤ m and taking i ∈ I such that Wxj ∩ Wxi = ∅ and n(xj ) ≥ n(xi ), we get f n(xi ) (Wxj ) ∩ Bδ1 /4 (f n(xi ) (xi )) = ∅. Lemma 4.2 assures us that diam(f n(xi ) (Wxj )) ≤ δ1 (n(xj )−n(xi ))/2 δ1 σ ≤ , 2 2 and so f n(xi ) (Wxj ) ⊂ Bδ1 (f n(xi ) (xi )). This implies that Wxj ⊂ Vxi . Hence {Vxi }i∈I is a covering of C. It follows from Corollary 4.3 that there is a uniform constant γ > 0 such that Leb∆ (Wxi ) ≥ γ, Leb∆ (Vxi ) for every i ∈ I. Hence Leb∆ Wxi = i∈I Leb∆ (Wxi ) i∈I ≥ γ Leb∆ (Vxi ) i∈I ≥ γ Leb∆ Vxi i∈I ≥ γ Leb∆ (C). Setting ρ = min Leb∆ (Wxi \ C) :i∈I , Leb∆ (Wxi ) 5562 JOSÉ F. ALVES AND VILTON PINHEIRO we have ε Leb∆ (C) ≥ Leb∆ (A \ C) Wxi \ C ≥ Leb∆ ≥ i∈I Leb∆ (Wxi \ C) i∈I ≥ ρ Leb∆ Wxi i∈I ≥ ργ Leb∆ (C). This implies that ρ < ε/γ. Since ε > 0 can be taken arbitrarily small, we may choose Wxi with the relative Lebesgue measure of C in Wxi arbitrarily close to 1. Then, by bounded distortion, the relative Lebesgue measure of f n(xi ) (H) ⊃ f n(xi ) (C) in f n(xi ) (Wxi ), which is a disk of radius δ1 /4 around f n(xi ) (xi ) tangent to the centreunstable cone ﬁeld, is also arbitrarily close to 1. Observe that since points in H have inﬁnitely many σ-hyperbolic times, we may take the integer n(xi ) arbitrarily large, as long as n1 in (9) is also taken large enough. Proposition 5.5. There are a sequence of sets W1 ⊃ W2 ⊃ · · · and a sequence of positive integers n1 ≤ n2 ≤ · · · such that: (1) Wk is contained in some hyperbolic pre-ball with hyperbolic time nk ; (2) ∆k = f nk (Wk ) is a disk of radius δ1 /4, centered at some point xk , tangent to the centre-unstable cone ﬁeld; (3) f nk (Wk+1 ) is contained in the disk of radius δ1 /8 centered at xk ; Leb∆k (f nk (H)) (4) lim = 1. k→∞ Leb∆k (∆k ) Proof. Take a constant 0 < ρ < 1 such that for any disk D of radius δ1 /4 centered at some point x tangent to the centre-unstable cone ﬁeld the following holds: if LebD (A) ≥ (1 − ρ)LebD (D) for some A ⊂ D, then we must have LebD∗ (A) > 0, where D∗ ⊂ D is the disk of radius δ1 /8 centered at the same point x. Note that it is possible to make a choice of ρ in these conditions only depending on the radius of the disk and the dimension of the disk. Surely, once we have chosen some ρ satisfying the required property, then any smaller number still has that property. We shall use Lemma 5.4 successively in order to deﬁne the sequence of sets (Wk )k and hyperbolic times (nk )k inductively. Let us start with O = ∆ and 0 < ρ < 1 with the property above. By Lemma 5.4 there are n1 ≥ 1 and W1 ⊂ V1 ⊂ O, where V1 is a hyperbolic pre-ball with hyperbolic time n1 , such that ∆1 = f n1 (W1 ) is a disk of radius δ1 /4 centered at some point x1 , tangent to the centre-unstable cone ﬁeld, with Leb∆1 (f n1 (H)) ≥ 1 − ρ. Leb∆1 (∆1 ) Considering ∆∗1 ⊂ ∆1 the disk of radius δ1 /8 centered at x1 , then by the choice of ρ we have Leb∆∗1 (f n1 (H)) > 0. Let O1 ⊂ W1 be the part of W1 which is sent by f n1 diﬀeomorphically onto ∆∗1 . We have Leb∆ (O1 ∩ H) > 0. PARTIALLY HYPERBOLIC SETS 5563 Next we apply Lemma 5.4 to O = O1 and ρ/2 in the place of ρ. Then we ﬁnd a hyperbolic time n2 and W2 ⊂ O1 such that ∆2 = f n2 (W2 ) satisﬁes ρ Leb∆2 (f n2 (H)) ≥1− . Leb∆2 (∆2 ) 2 Observe that W2 ⊂ O1 ⊂ W1 . Then we take O2 ⊂ W2 as that part of W2 which is sent by f n2 diﬀeomorphically onto the disk ∆∗2 of radius δ1 /8 and proceed inductively. The next proposition gives the conclusion of Theorem 5.1. Proposition 5.6. The sequence (∆k )k has a subsequence converging to a local unstable disk ∆∞ of radius δ1 /4 inside Λ. Proof. Let (∆k )k be the sequence of disks given by Proposition 5.5 and let (xk )k be the sequence of points at which these disks are centered. Up to taking subsequences, we may assume that the centers of the disks converge to some point x. Using AscoliArzela, a subsequence of the disks converges to some disk ∆∞ centered at x. We necessarily have ∆∞ ⊂ Λ. Note that each ∆k is contained in the nk -iterate of ∆, which is a disk tangent to the centre-unstable cone ﬁeld. The domination property implies that the angle between ∆k and E cu goes uniformly to 0 as n → ∞. In particular, ∆∞ is tangent to E cu at every point in ∆∞ ⊂ Λ. By Lemma 4.2, given any n ≥ 1, then f −n is a σ n/2 -contraction on ∆k for every large k. Passing to the limit, we get that f −n is a σ n/2 -contraction in the E cu direction over ∆∞ for every n ≥ 1. The fact that the Df -invariant splitting TΛ M = E cs ⊕ E cu is dominated implies that any expansion Df may exhibit along the complementary direction E cs is weaker than the expansion in the E cu direction. Then there exists a unique unstable manifold u (x) tangent to E cu and which is contracted by the negative iterates of f ; see Wloc [14]. Since ∆∞ is contracted by every f −n , and all its negative iterates are tangent u (x). to the centre-unstable cone ﬁeld, then ∆∞ is contained in Wloc 6. Existence of hyperbolic periodic points Here we prove Theorem E. By Proposition 5.5 there exist a sequence of sets W1 ⊃ W2 ⊃ · · · contained in ∆ and a sequence of positive integers n1 ≤ n2 ≤ · · · such that: (1) Wk is contained in some hyperbolic pre-ball with hyperbolic time nk ; (2) ∆k = f nk (Wk ) is a disk of radius δ1 /4, centered at some point xk , tangent to the centre-unstable cone ﬁeld; (3) f nk (Wk+1 ) is contained in the disk ∆∗k of radius δ1 /8 centered at xk . Taking a subsequence, if necessary, we have by Proposition 5.6 that the sequence of disks (∆k )k accumulates on a local unstable disk ∆∞ of radius δ1 /4 which is contained in Λ. Our aim now is to prove that Λ contains the unstable manifold of some periodic point. Similarly to (6), we choose δ > 0 small so that Wδs (z) is deﬁned for every z ∈ Λ, the 2δ-neighborhood of Λ is contained in U , and (10) Df −1 (f (y))v ≤ σ −1/4 Df −1 \|Efcu(x) v, whenever x ∈ U , dist(x, y) ≤ 2δ, and v ∈ Cacu (y). 5564 JOSÉ F. ALVES AND VILTON PINHEIRO Proposition 6.1. Given Λ1 ⊂ Λ with Leb(Λ1 ) > 0, there exist a hyperbolic periodic point p ∈ Λ and δ2 > 0 (not depending on p) such that: (1) (2) (3) (4) W u (p) ⊂ Λ; u the size of Wloc (p) is at least δ2 ; u u (p) almost every point in W LebWloc loc (p) belongs to H; there is x ∈ Λ1 with ω(x) ⊂ W u (p). Proof. Let x denote the center of the accumulation disk ∆∞ . Let us consider the cylinder Cδ = Wδs (y), y∈∆∞ and the projection along local stable manifolds π : Cδ −→ ∆∞ . Slightly diminishing the radius of the disk ∆∞ , if necessary, we may assume that there is a positive integer k0 such that for every k ≥ k0 π(∆k ∩ Cδ ) = ∆∞ (11) and ∆∗k ⊂ Cδ . For each k ≥ k0 let πk : ∆∞ −→ ∆k be the projection along the local stable manifolds. Notice that these projections are continuous and π ◦ πk = id∆∞ . Take a positive integer k1 > k0 suﬃciently large so that 1 (12) π(∆k1 ∩ Cδ/2 ) = ∆∞ and λnk1 −nk0 ≤ . 4 We have ∆k1 = f nk1 (Wk1 ) ⊂ f nk1 −nk0 (f nk0 (Wk0 +1 )) ⊂ f nk1 −nk0 (∆∗k0 ), which together with (11) and (12) implies that there is some disk ∆0 ⊂ ∆∞ such that π ◦ f nk1 −nk0 ◦ πk0 (∆0 ) = ∆∞ . Thus there must be some z ∈ ∆0 ⊂ ∆∞ which is a ﬁxed point for the continuous map π ◦ f nk1 −nk0 ◦ πk0 . This means that there are zk0 , zk1 ∈ Wδs (z) with zk0 ∈ ∆k0 and zk1 ∈ ∆k1 such that f nk1 −nk0 (zk0 ) = zk1 . Letting γ = Wδs (z), we have distγ (w, zk1 ) ≤ 2δ for every w ∈ γ. This implies that distγ (f nk1 −nk0 (w), zk1 ) = distγ (f nk1 −nk0 (w), f nk1 −nk0 (zk0 )) ≤ 2δλnk1 −nk0 , which together with (12) gives distγ (f nk1 −nk0 (w), z) ≤ distγ (f nk1 −nk0 (w), zk1 ) + distγ (zk1 , z) ≤ δ. We conclude that f nk1 −nk0 (Wδs (z)) ⊂ Wδs (z). Since Wδs (z) is a topological disk, this implies that Wδs (z) must necessarily contain some periodic point p of period m = nk1 − nk0 . As z ∈ ∆∞ and p ∈ Wδs (z), it follows that p ∈ Λ, by closeness of Λ. Let us now prove that p is a hyperbolic point. As p ∈ Wδs (z), it is enough to show that Df −m \| Efcum (p) < 1. Let q = Wδs (z) ∩ f nk0 (Wk1 ). Observe that since p ∈ Λ ∩ Wδs (z), then q belongs to the 2δ-neighborhood of Λ, which is contained in U . Since Wk1 is contained in some hyperbolic pre-ball with hyperbolic time nk1 , it follows from Lemma 4.2 that for every 1 ≤ j ≤ nk1 and y ∈ Wk1 , Df −j \| Efcunk1 (y) ≤ σ j/2 . PARTIALLY HYPERBOLIC SETS 5565 In particular, taking j = m = nk1 − nk0 and y = f −nk0 (q), we have Df −m \| Efcum (q) ≤ σ m/2 . The choice of δ in (10) together with the fact that p, q ∈ Wδs (z) implies that (13) Df −m \| Efcum (p) ≤ m Df −1 \| Efcuj (p) j=1 ≤ σ −m/4 m Df −1 \| Efcuj (q) j=1 ≤ σ (14) m/4 . Thus we have proved the hyperbolicity of p. u Now, since p is a hyperbolic periodic point, there is Wloc (p) a local unstable manifold through p tangent to the centre-unstable bundle. As ∆∞ transversely cuts the local stable manifold through p, then using the λ-lemma we deduce that the positive iterates of ∆∞ accumulate on the unstable manifold through p. Since these iterates are all contained in Λ and Λ is a closed set, we must have W u (p) ⊂ Λ, which then implies that W u (p) ⊂ Λ. Thus we have proved the ﬁrst part of the result. By (13) and (14) we deduce that every multiple of m is a σ 1/4 -hyperbolic time for p. Then we choose δ2 > 0 such that an inequality as in (6) holds with δ2 in u the place of δ1 and σ 1/8 in the place of σ 1/2 . Using Lemma 4.2 with Wloc (p) in the place of ∆ and taking a suﬃciently large σ 1/4 -hyperbolic time for p we deduce u (p). This imples that its image by that there is a hyperbolic pre-ball inside Wloc the hyperbolic time, which is a disk of radius δ2 around p, is contained in the local unstable manifold of p. This gives the second part of the result. Observe that as long as we take the local unstable manifold through p small u (p) belongs to the local stable manifold of some enough, then every point in Wloc point in ∆∞ . By construction, ∆∞ is accumulated by the disks ∆k = f nk (Wk ) which, by Proposition 5.5, satisfy (15) Leb∆k (f nk (H)) = 1. k→∞ Leb∆k (∆k ) lim Since H is positively invariant, we have Leb∆k (H) = 1. k→∞ Leb∆k (∆k ) lim Now let ϕ : Λ → R be the continuous function given by ϕ(x) = log Df −1 \| Excu . Since Birkhoﬀ’s time averages are constant for points in a same local stable manifold and the local stable foliation is absolutely continuous, we deduce that Leb∆∞ (H) = 1. Leb∆∞ (∆∞ ) The same conclusion holds for the local unstable manifold of p in the place of ∆∞ by the same reason. Now let us prove the last item. Since H has full Lebesgue measure in Λ and Λ1 ⊂ Λ has positive Lebesgue measure, we may start our construction with the set H1 = H ∩ Λ1 in the place of H intersecting the disk ∆ in a positive Leb∆ measure 5566 JOSÉ F. ALVES AND VILTON PINHEIRO set of points. Although we do not have invariance of H1 , by (15) we still have the property that the iterates of H1 ⊂ Λ1 accumulate on the whole ∆∞ . Since the u stable manifolds through points in Wloc (p) intersect ∆∞ , there must be points in u Λ1 accumulating on Wloc (p). Let p1 be a hyperbolic periodic point as in Proposition 6.1. Let B1 be the basin of W u (p1 ), i.e. the set of points x whose ω-limit is contained in W u (p1 ). If Leb(Λ \ B1 ) = 0, then we have proved the theorem. Otherwise, let Λ1 = Λ \ B1 . Using again Proposition 6.1 we obtain a point p2 ∈ Λ such that the basin B2 of W u (p2 ) attracts some point of Λ1 . By deﬁnition of Λ1 we must have W u (p1 ) = W u (p2 ). We proceed inductively, thus obtaining periodic points p1 , . . . , pn ∈ Λ with W u (pi ) = W u (pj ) for every i = j. This process must stop after a ﬁnite number of steps. Actually, if there were an inﬁnite sequence of points as above, by compactness, choosing pi1 , pi2 suﬃciently close, using the inclination lemma we would get W u (pi1 ) = W u (pi2 ). So far we have proved the ﬁrst two items of Theorem E. Assume now that E cu has dimension one. We want to show that each W u (pi ) attracts an open set containing W u (pi ). Given 1 ≤ i ≤ k, by Proposition 6.1 we can ﬁnd at least one point on each connected component of W u (pi ) \ {pi } belonging to H. Since these points have inﬁnitely many hyperbolic times, then each connected component of W u (pi ) \ {pi } must necessarily have inﬁnite arc length; recall Lemma 4.2. This implies that each point x ∈ W u (pi ) has an unstable arc γ u (x) ⊂ W u (pi ) of a ﬁxed length passing through it. Let Wδs (y). B(x) = y∈γ u (x) By domination, the angles of γ u (x) and the local stable manifolds Wδs (y) with y ∈ γ u (x) are uniformly bounded away from zero. Thus, B(x) must contain some ball of uniform radius (not depending on x), and so the set x∈W u (pi ) B(x) is a neighborhood of W u (pi ). Since, for each x ∈ W u (pi ), the points in B(x) have their ω-limit set contained in W u (pi ), we are done. 7. Hyperbolic sets with positive volume In this section we prove Theorem F and Theorem G. Since in the present situation Df \|Exu is uniformly expanding, then we have NUE for every x ∈ Λ. 7.1. Transitive case. Assume ﬁrst that Λ has positive volume. It follows from Corollary B that Λ must contain some local unstable disk. The ﬁrst item of Theorem F is a consequence of the following folklore lemma whose proof we give here for the sake of completeness. Lemma 7.1. If Λ is a transitive hyperbolic set containing the local unstable manifold of some point, then Λ contains the local unstable manifolds of all its points. Proof. Take δ > 0 small such that Wδs (x) and Wδu (y) intersect at most in one point, for every x, y ∈ Λ, and assume that Wδu (x0 ) ⊂ Λ for some x0 ∈ Λ. Let z ∈ Λ be a point with dense orbit in Λ. It is no restriction to assume that Wδs (z) intersects Wδu (x0 ), and we do so. Let x1 = Wδs (z) ∩ Wδu (x0 ). We also have Wδu (x1 ) ⊂ Λ. Given any point y ∈ Λ, we take a sequence of integers 0 = n1 < n2 < · · · such PARTIALLY HYPERBOLIC SETS 5567 that f nk (z) → y, when k → ∞. Since x1 ∈ W s (z), we also have xk := f nk (x1 ) → y, when k → ∞. The local unstable manifolds through the points x1 , x2 , ... are necessarily contained in Λ and accumulate on a disk D(y) contained in Λ and containing y. Since the local unstable disks are tangent to the unstable spaces, u for every w ∈ D(y). By the continuity of these spaces implies that Tw D(y) = Ew uniqueness of the unstable foliation, we must have D(y) coinciding with the local unstable manifold through y. Using the previous lemma applied to f −1 , we have that Λ must also contain the stable manifolds through its points. Then we easily deduce that every point in Λ belongs in the interior of Λ, thus showing that Λ is an open set. Since Λ is assumed to be closed, we conclude that Λ = M , thus having proved the ﬁrst part of Theorem F. Lemma 7.2. Let Λ be a hyperbolic set attracting a set with positive volume. Then there is a point in Λ whose local unstable manifold is contained in Λ. Proof. We ﬁx continuous extensions (not necessarily continuous) of the two bundles E cs and E cu to some neighborhood U of Λ. Let A be the set of points which are attracted to Λ under positive iteration. Since A has positive volume, there must be some compact set C ⊂ A with positive volume, and some N ∈ N such that f n (C) ⊂ U for every n ≥ N . Letting f n (C) ∪ Λ, K= n≥N we have that K is a compact forward invariant set with positive volume for which Λ= f n (K). n≥1 The conclusion of the lemma then follows from Theorem E. The second part of Theorem F can now be easily deduced from Lemma 7.1 and Lemma 7.2. Actually, it follows from the lemmas that Wδs (x) x∈Λ is a neighborhood of Λ whose points are attracted to Λ under positive iteration. 7.2. Nontransitive case. Here we consider the case of hyperbolic sets with positive volume not necessarily transitive and prove Theorem G. Let Σ = W u (p) ⊂ Λ, where p is a hyperbolic periodic point given by Proposition 6.1. We claim that Σ contains the local unstable manifolds of all its points. Indeed, if x ∈ Σ, then there is a sequence (xn )n of points in W u (p) converging to x. The continuous variation of the local unstable manifolds gives that the local unstable manifolds of the points xn , which are contained in Σ, accumulate on the local unstable manifold of x. By closeness, the local unstable manifold of x must be contained in Σ. Thus, deﬁning Wδs (x), A= x∈Σ we have that A is a neighborhood of Σ whose points have their ω-limit set contained in Σ. Since Σ is a hyperbolic set with a local product structure attracting an open 5568 JOSÉ F. ALVES AND VILTON PINHEIRO neighborhood of itself, then by [9, Theorem 18.3.1] there are hyperbolic invariant sets Ω1 , . . . , Ωs ⊂ Σ ⊂ Λ verifying (3)-(6) of Theorem G. Moreover, their union is the set of nonwandering points of f in Σ, N W (f \|Σ) = Ω1 ∪ · · · ∪ Ωs . Since L(f \|Σ) ⊂ N W (f \|Σ), this implies that ω(x) ⊂ Ω1 ∪ · · · ∪ Ωs for every x ∈ A. Recall that every point in A belongs to the stable manifold of some point in Σ. Now since Ω1 , . . . , Ωs are disjoint compact invariant sets, given x ∈ A, we must even have ω(x) ⊂ Ωi for some 1 ≤ i ≤ s. Reordering these sets if necessary, let Ω1 , . . . , Ωq , for some q ≤ s, be those which attract a set with positive Lebesgue measure. By Theorem F and transitivity, each Ω1 , . . . , Ωq attracts a neighborhood of itself. Acknowledgement We are grateful to M. Viana for valuable discussions and references on these topics. References [1] F. Abdenur, C. Bonatti, L. J. Dı́az, Non-wandering sets with non-empty interior, Nonlinearity 17 (2004), 175-191. MR2023438 (2004m:37031) [2] J. F. Alves, V. Araújo, M. J. Paciﬁco, V. Pinheiro, On the volume of singular-hyperbolic sets, Dyn. Syst., to appear. [3] J. F. Alves, V. Araújo, B. Saussol, On the uniform hyperbolicity of some nonuniformly hyperbolic systems, Proc. Amer. Math. Soc. 131, no. 4 (2003), 1303-1309. MR1948124 (2003k:37046) [4] J. F. Alves, C. Bonatti, M. Viana, SRB measures for partially hyperbolic systems whose central direction is mostly expanding, Invent. Math. 140 (2000), 351-398. MR1757000 (2001j:37063b) [5] J. Bochi, M. Viana, Lyapunov exponents: How frequently are dynamical systems hyperbolic?, Advances in Dynamical Systems. Cambridge University Press, 2004. MR2090775 (2005g:37060) [6] R. Bowen, A horseshoe with positive measure, Invent. Math. 29 (1975), 203-204. MR0380890 (52:1787) [7] R. Bowen, Equilibrium states and the ergodic theory of Axiom A diﬀeomorphisms, Lecture Notes in Mathematics 470 (1975), Springer. MR0442989 (56:1364) [8] T. Fisher, Hyperbolic sets with nonempty interior, Disc. Cont. Dynam. Syst., 15 no. 2 (2006), 433-446. MR2199438 (2007a:37035) [9] A. Katok, B. Hasselblat, Introduction to the modern theory of dynamical systems, Cambridge Univ. Press, 1995. MR1326374 (96c:58055) [10] S. Newhouse, The abundance of wild hyperbolic sets and non-smooth stable sets for diﬀeomorphisms, Publ. Math. IHES 50 (1979), 101-151. MR556584 (82e:58067) [11] J. Palis, F. Takens, Hyperbolicity and the creation of homoclinic orbits, Ann. Math. 125 (1987), 337-374. MR881272 (89b:58118) [12] J. Palis, J-C. Yoccoz, Homoclinic tangencies for hyperbolic sets of large Hausdorﬀ dimension, Acta Mathematica 172 (1994), 91-136. MR1263999 (95a:58088) [13] J. Palis, J-C. Yoccoz, Fers à cheval non uniformément hyperboliques engendrés par une bifurcation homocline et densité nulle des attracteurs, C. R. Acad. Sci. Paris, 333 Série I (2001), 867-871. MR1873226 (2003h:37078) [14] Ya. Pesin, Families of invariant manifolds corresponding to non-zero characteristic exponents, Math. USSR Izv. 10 (1976), 1261-1302. [15] V. Pliss. On a conjecture due to Smale, Diﬀ. Uravnenija 8 (1972), 262-268. MR0299909 (45:8957) [16] M. Shub, Global stability of dynamical systems, Springer Verlag, 1987. MR869255 (87m:58086) PARTIALLY HYPERBOLIC SETS 5569 [17] S. Smale, Diﬀeomorphisms with many periodic points, Diﬀ. and Comb. Topology, Princeton Univ. Press (1965), 63-80. MR0182020 (31:6244) [18] L.-S. Young, Large deviations in dynamical systems, Trans. Amer. Math. Soc. 318, no. 2, (1990), 525-543. MR975689 (90g:58069) Departamento de Matemática Pura, Faculdade de Ciências do Porto, Rua do Campo Alegre 687, 4169-007 Porto, Portugal E-mail address: jfalves@fc.up.pt URL: http://www.fc.up.pt/cmup/jfalves Departamento de Matemática, Universidade Federal da Bahia, Av. Ademar de Barros s/n, 40170-110 Salvador, Brazil E-mail address: viltonj@ufba.br
https://openalex.org/W2767137300	https://www.scielo.br/j/jpe/a/rwwRmWFzLvzJX8XQ9gCvbvS/?lang=en&format=pdf	English	null	A SYSTEMATIC REVIEW OF CROSS-CULTURAL ADAPTATION AND VALIDATION OF BORG'S RATING OF PERCEIVED EXERTION SCALE	Journal of Physical Education	2,017	cc-by	7,488	ABSTRACT The aim of this systematic review was to analyze the cross-cultural adaptation and validation processes of Borg’s Rating of Perceived Exertion Scale (6-20) and its instructions during exercise in adults and the elderly. The study was conducted using databases that include the areas of health, sports and psychology (Pubmed, ScienceDirect, Bireme (BVS), PsycINFO, SportDiscus and Scopus), and descriptors in the English and Portuguese languages were standardized for all databases ("exercise" / "exercicio", "validity" / "validade", "borg scale" / "escala de borg"; "perceived exertion" / "percepção de esforço" and "validation" / "validação"). Among the 1,488 items found, 14 articles were selected for analysis. The main findings were that cross-cultural adaptation (Cantonese and Japanese) was reported in only four articles. The lowest-scored criteria for methodological quality were related to the concept of perceived exertion, and the instructions for use and familiarization with Borg’s scale. The results indicate that the majority of the studies did not apply the cross-cultural adaptation process and additionally, that concurrent validity was the only validation method reported. Therefore, it is necessary to carry out the cross-cultural adaptation of Borg’s Scale (6-20), mainly in the Portuguese language, in addition to other types of validation in order to ensure more accurate results. Keywords: Exercise. Perception. Validity of tests. Translation. RESUMO RESUMO O objetivo dessa revisão sistemática foi analisar os processos de adaptação transcultural e validação da Escala de Percepção Subjetiva de Esforço de Borg (6-20) e de suas instruções durante o exercício em adultos e idosos. A busca foi realizada nas bases de dados da área da saúde, esporte e psicologia (Pubmed, ScienceDirect, Bireme (BVS), PsycINFO, SportDiscus e Scopus) e os descritores em inglês e português foram padronizados para todas as bases de dados (“exercise”/”exercício”; “validity”/”validade”; “borg scale”/”escala de borg”; “perceived exertion”/”percepção de esforço”; e “validation”/”validação”). Entre os 1488 artigos encontrados, foram selecionados 14 artigos para análise. Os principais resultados encontrados foram que a adaptação transcultural (Cantonês e Japonês) foi reportada em apenas quatro artigos. Os critérios mais baixos de qualidade metodológica foram aqueles relacionados a definição de percepção de esforço, a instrução e familiarização com a Escala. Os resultados indicam que a maioria dos estudos não realizou o processo de adaptação transcultural e além disso, só a validade concorrente foi utilizada como método de validação. Portanto, é necessário realizar adaptação transcultural da Escala de Borg (6-20) principalmente na língua portuguesa, além dos diferentes tipos de validação para se obter resultados mais precisos. p p Palavras-chave: Exercício. Percepção. Validade dos testes. Tradução. 1Universidade Federal do Paraná, Curitiba-PR, Brasil. 1Universidade Federal do Paraná, Curitiba-PR, Brasil. DOI: 10.4025/jphyseduc.v28i1.2853 DOI: 10.4025/jphyseduc.v28i1.2853 DOI: 10.4025/jphyseduc.v28i1.2853 Review Article uana Loss Cabral1, Paula Born Lopes1, Renata Wolf1, Joice Mara Facco Stefanello1 e Gleber Pereira1 uana Loss Cabral1, Paula Born Lopes1, Renata Wolf1, Joice Mara Facco Stefanello1 e Gleber Pereira1 Introduction g p In spite of the recurrent and continuous use of Borg’s scale (6-20), it has been suggested that few studies have performed the cross-cultural adaptation of the instrument according to the guidelines found in the literature, with the use of methods such as the application of back translation14, the analysis of equivalences with the original version15, and the analysis of the target version by a committee of specialists16. This process becomes important, since, if the meaning of the terms on the instrument were altered during the translation to a target language, the instructions given could modify the construct that was originally proposed. In the same manner, if the meaning of the quantitative semantics is altered, the numerical categories related to these semantics will be interpreted incorrectly by the individual and, with this, the perceived exertion reported will be less accurate. In addition to this, when the instrument is not validated empirically (validation of content, construct, and criteria), its cross-cultural adaptation alone is not sufficient to ensure that the interpretations and conclusions of the phenomena related to a determined construct will be adequate17. For example, when the intensity of exercise increases, consequently, it is expected that the physiological responses will be higher and that perceived exertion will also increase, due to being an response to external load. However, it is necessary to develop processes for validation after cross-cultural adaptation in order to ensure that the translated Borg’s scale (6- 20) is truly measuring what it proposes, which is perceived exertion. The growing number of publications about adults and the elderly utilizing Borg’s scale (6-20)18 is due to the expressive increase and incentive for the practice of physical exercise for this population, with the intuition of monitoring the intensity of the exercise through the perceived exertion19. As such, it has become more important to verify if Borg’s scale (6-20) and its instructions have undergone the cross-cultural adaptation process for different languages, along with its respective validation for adults and the elderly. Therefore, the objective of this study was to analyze the cross-cultural adaptation and validation processes of Borg’s Rating of Perceived Exertion Scale (6-20) and of its instructions during exercise, for adults and the elderly. Introduction Perceived exertion can be defined as the perception of how difficult and extenuating a physical activity is1. It is a psycho-physiological measure whose objective is quantifying the subjective perceived exertion originating throughout all the bodily systems1,2. Various studies have demonstrated interest in understanding how individuals perceive exertion during physical exercise (internal load) from a determined work load (external load)3. This concept was introduced to academia for the first time by Borg’s scale4,5, which demonstrated a correlation with cardiac frequency4 and intensity of exercise6. Borg’s scale (6-20) is a categorical matrix of numbered intervals with equal distances between different perceptions of exertion5. The relation between the numerical categories and the verbal descriptors was J. Phys. Educ. v. 28, e2853, 2017. Cabral et al. Page 2 of 13 Page 2 of 13 Cabral et al. established with the use of quantitative semantics, which allow for the adequate judgment of perceived exertion by the individual with the numerical categories on the scale7. The scale also includes instructions related to its use, in addition to the concept related to the construct to be evaluated, with the purpose of facilitating the understanding of the individual and increasing the precision of reporting of perceived exertion. Due to the fact that this instrument is easy to understand and apply, and is of low cost1,8, various researchers have utilized this scale for diverse populations9. However, few studies took the time to carry out the cross- cultural adaptation and validation of this instrument (of the scale and its instructions for use) for different cultures, age groups, and socioeconomic levels. Cross-cultural adaptation is a method utilized for the translation and adaptation of an instrument for a sample different from the original population where it was applied10,11. This process has been frequently utilized in studies due to being faster and more economical in comparison with the processes involved in creating a new instrument. In addition to translating the instrument, it is necessary to take into consideration the differences between the language of origin and the target language13, due to the cultural differences between the populations14. Therefore, cross-cultural adaptation allows samples of individuals with different ethnicities to present provide similar responses for the same test, due to a similar understanding of the instrument between samples. Introduction The results of this systematic review will instruct future studies with regards to the necessity for applying the processes of cross-cultural adaptation of Borg’s scale (6-20) for different languages or for applying validation processes in studies with different protocols, exercise modalities, or populations different from the original study. J. Phys. Educ. v. 28, e2853, 2017. A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Page 3 of 13 A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Research Strategies This systematic review followed methodological guidelines based on the peer-reviewed literature20. The study was completed using six electronic databases, which include the areas of health, sports, and psychology (Pubmed, ScienceDirect, Bireme (BVS), PsycINFO, SportDiscus and Scopus). The searches were carried out between May and July of 2016, with no data restriction. Borg’s Rating of Perceived Exertion Scale (6-20) is a one-dimensional instrument that evaluates the perceived exertion during the performance of a physical exercise1. This scale presents the definition of the construct of perceived exertion and after that provides instruction for how the exertion should be perceived and how it needs to be reported using the numerical values on the scale. The numerical descriptors, with respective quantitative semantics, facilitate the understanding of the participants. The values on the scale go from 6 (“no exertion at all”) to 20 (“maximum exertion”). There is also an explanation of the numerical descriptors which are important for increasing the accuracy of the result, such that 9 corresponds to “very light” exercise, 13 corresponds to “somewhat hard” exercise, 17 corresponds to “very hard” exercise, and 19 corresponds to “extremely hard” exercise. And finally, there are instructions explaining the importance of utilizing the scale for greater accuracy, independent of external opinions. Criteria for inclusion and exclusion Articles in English or Portuguese were included, with searches for the keywords in syntactic form carried out in both languages. Two descriptors were standardized by the Medical Subject Headings (MeSH) (“exercise” and “validity”) in English, and other terms were those frequently used in this area of study (“borg scale”, “perceived exertion” and “validation”). The descriptors in Portuguese were “exercício” (exercise), “validade” (validity), “escala de borg” (Borg’s scale), “percepção de esforço” (perceived exertion), and “validação” (validation), searched for in the same way as for the descriptors in English. The logical operators “AND” and “OR” were used for the combination of descriptors and the terms used in the search for articles. Studies conducted with adults (18 to 60 years) and the elderly (over 60 years of age) that applied Borg’s scale (6-20) to exercise and that also presented some process of validation were included. Validation processes included the validation of content (an analysis of the scale’s content by specialists in the area), the validation of concurrent criteria (the correlation between the perceived exertion and physiological variables), and the validation of predictive criteria (the correlation between the perceived effort and physiological variables at determined moments). Cross-cultural adaptation was considered when the study applied at least one protocol or used as a reference another study which applied the cross-cultural adaptation procedures accepted in the literature, which include: the translation, synthesis, and back translation14, the analysis of equivalences between the original version and the target version, the evaluation of the target version by a committee of specialists, and a pilot study of the target population13,16. After the completion of these phases, Borg’s scale (6-20) should be adapted cross-culturally in order to ensure that the quantitative semantics of the scale have not been altered and that the instructions regarding the use of the scale have the same meaning as the original version. Review articles, theses, dissertations, book chapters, articles about groups with special health conditions (ex.: diabetes, multiple sclerosis, or coronary heart disease), articles unrelated to the proposed theme, articles that used scales other than Borg’s scale (6-20) for J. Phys. Educ. v. 28, e2853, 2017. Cabral et al. Page 4 of 13 Page 4 of 13 Cabral et al. measuring perceived exertion, articles that did not apply validation processes, and articles that used the perceived exertion of a session as a measurement were excluded from the review. Criteria for inclusion and exclusion The use of perceived exertion of a session is distinct method for analyzing perceived exertion, in which the calculation is carried out after a complete exercise session; thus, the result cannot be compared with the results of perceived exertion during exercise. The systematic review was carried out by two independent reviewers who completed the following stages: (1) the search for articles using the descriptors in the databases and the extraction of the titles; (2) the exclusion of repeated titles; (3) the reading of the titles and selection of abstracts; (4) the reading of the abstracts and selection of the articles for full reading; (5) the full reading of the articles; and (6) the inclusion of the articles for the systematic review analysis. The stages of the inclusion or exclusion of articles followed the pre-established criteria described above. At the end of each stage, a consensus meeting was held in order to determine the articles to be included in the next stage of the analysis. In the case of divergence, a third reviewer was contacted in order to break the tie of the stage. g g The association of the descriptors generated 1,488 articles (228 articles in PubMed, 847 articles in ScienceDirect, 81 articles in BVS, 33 articles in PsycINFO, 217 articles in SportDiscus, and 82 articles in Scopus), totaling 1,202 after the exclusion of repeated articles. Of these, 1,051 were excluded for being unrelated to the topic. Finally, 42 articles were selected for a full reading. Figure 1 presents the criteria for the search, selection, and exclusion of articles for this review. After the inclusion of articles found by the systematic review, three studies found in the references of the selected studies were also included. Articles found n=1488 n=286 repeated articles Reading of titles n=1202 n=1051 articles excluded for being unrelated to the topic Reading of abstracts n=151 n=109 articles excluded for not meeting the criteria for inclusion and exclusion Full reading of the articles n=42 n= 31 articles excluded for not meeting the criteria for inclusion and exclusion Articles selected for the review n= 11 n= 3 articles included found in the references of the selected articles Relevant articles included in the Systematic Review n = 14 Figura 1. Methodological Quality The studies included in this review were analyzed with respect to their methodological quality (Table 1). The criteria utilized included the following items: 1) the proposal of the study corresponds to the results found; 2) the inclusion of the statistics utilized; 3) the inclusion of the experimental design; 4) the inclusion of the general characteristics of the sample; 5) the inclusion of the nationality of the sample; 6) the inclusion of the definition of perceived exertion; 7) the inclusion of instructions for the use of Borg’s scale (6-20); 8) a description of how familiarization with the scale was carried out; and 9) a conclusion responding to the objective of the study. In the case of doubts regarding the quality of the studies, a third reviewer was contacted. The maximum rating was 9 points, such that a higher rating represented a higher level of quality of the articles. Criteria for inclusion and exclusion Flowchart of the search, selection, and exclusion of the articles for review Source: The authors n=286 repeated articles n=1051 articles excluded for being unrelated to the topic Reading of abstracts n=151 n=109 articles excluded for not meeting the criteria for inclusion and exclusion selected for the review n= 11 Articles selected for the review n= 11 Relevant articles included in the Systematic Review n = 14 Figura 1. Flowchart of the search, selection, and exclusion of the articles for review Source: The authors J. Phys. Educ. v. 28, e2853, 2017. Page 5 of 13 A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Results 1,488 articles in total were found, and after the process the inclusion and exclusion of studies, 14 articles were included (Table 2). The total sample of participants (the summation of all the studies) was a group of 426 individuals, consisting of the following: 333 adults (average (DP): 31.3 (5.81 years), 128 men and 205 women; and 93 elderly individuals without special health conditions (70.1 (4.05) years), 16 men and 77 women. Of the 14 articles analyzed, only three studied an elderly population21-23. The 14 articles included in the revision spanned a time period from 1976 to 2015. Table 1. Table of the methodological quality of the articles included in the review. Author/Year 1 2 3 4 5 6 7 8 9 Total (max.=9) Baildon et al., 1994 x x x x x 5 Chen et al., 2013 x x x x x x 6 Chung et al., 2015 x x x x 4 Edwards et al., 1976 x x x x 4 Karavatas & Travakol, 2005 x x x x x 5 Leung et al., 2004 x x x x x x x 7 Lollgen et al., 1977 x x 2 López-Miñarro & Muyor, 2010 x x x x x 5 Muyor, 2013 x x x x x 5 Myuor & López-miñarro, 2012 x x x x x 5 Pollock et al., 2013 x x x x 3 Robertson et al., 1996 x x x x x 5 Shigematsu et al., 2004 x x x x x x x x 8 Wenos et al., 1996 x x x x x 5 Description of the criteria for methodological quality: 1) proposal of the study corresponds to the results found; 2) inclusion of the statistics utilized; 3) inclusion of experimental design; 4) inclusion of the general characteristics of the sample; 5) inclusion of the nationality of the sample; 6) inclusion of the definition of perceived exertion; 7) inclusion of instructions for the use of Borg’s scale (6-20); 8) description of how familiarization with the scale was carried out; and 9) inclusion of a conclusion responding to the objective of the study. Source: The authors Table 1. Table of the methodological quality of the articles included in the review. Author/Year 1 2 3 4 5 6 7 8 9 Total (max able of the methodological quality of the articles included in the review. Page 6 of 13 Page 6 of 13 Cabral et al. 6 and only one of these articles did not use the scale adapted to the culture of the sample21,23-26 and only one of these articles did not use the scale adapted to the culture of the study26. y The exercise modalities utilized in the studies were cycle ergometer (38%)21,23,25,27-29, treadmill (19%)22,24,26, resistance exercises (13%)24, indoor cycling (19%)30,31,32, Nintendo Wii games (6%)33, and cycle ergometer adapted for the pool (6%)24. Among the protocols utilized in the articles, four studies applied test protocols with an incremental load (25%), one study applied a decreasing load (6%), four studies applied a continuous load (25%), and seven studies applied an interval load (44%). pp After the analysis of the methodological quality of the studies, a divergence in the manner of application of Borg’s scale was found, as the majority of the studies lost points for the criteria referring to the definition of perceived exertion and the instructions for the use and familiarization of the scale, and only one article achieved points for all three criteria23. The average rating of the articles was five points for the criteria for methodological quality (57.1%). The article that achieved the highest rating (8 points) was Shigematsu et. al. (2004)23 and the article which achieved the lowest rating (2 points) was Lollgen et al. (1977)29. Cross-cultural adaptation was developed in only one study25 and three other studies (21%)21,23,24 utilized Borg’s scale with the cross-cultural adaptation developed by previous studies25,35. The article by Edwards et. al. (1976)28 did not follow the guidelines for cross- cultural adaptation proposed by the literature15,16 , when translating Borg’s scale from English to the Swedish language. Among the different methods used to determine the validity of the scale, the articles in the review only applied concurrent validity as a part of the study. The subjective perceived exertion was correlated with cardiac frequency (CF), potency, rate of ventilation, oxygen consumption, gross metabolic efficiency, and blood lactate concentration. Perceived exertion and cardiac frequency (CF) were correlated in ten articles, such that one study presented a very strong correlation (r = 0.90  1.00)23, five studies presented strong correlation (r = 0.70  0.90)21,22,25,27,28, two studies presented moderate correlation (r = 0.50  0.70)26,29, one study presented weak correlation (r = 0.30  0.50)33, and one study presented very weak correlation (r = 0.00  0.30)32. J. Phys. Educ. v. 28, e2853, 2017. Results Description of the criteria for methodological quality: 1) proposal of the study corresponds to the results found; 2) inclusion of the statistics utilized; 3) inclusion of experimental design; 4) inclusion of the general characteristics of the sample; 5) inclusion of the nationality of the sample; 6) inclusion of the definition of perceived exertion; 7) inclusion of instructions for the use of Borg’s scale (6-20); 8) description of how familiarization with the scale was carried out; and 9) inclusion of a conclusion responding to the objective of the study. Source: The authors For the analysis of the affiliated country of the authors of the articles included in this review, it was verified that 28.6% of the studies originated in the United States, 21.5% in Spain, 14.3% in Germany, 7.1% in Switzerland, and 28.6% in Asian countries (China, Taiwan and Japan). Among these articles, only 35.7% (n=5) of the studies reported the nationality of J. Phys. Educ. v. 28, e2853, 2017. Page 6 of 13 Page 6 of 13 Furthermore, one study presented a high regression coefficient for the treadmill test (70%, R2=0.70) and a relatively weak coefficient for resistance exercises (27%, R2=0.27)24. There was also elevated correlation between perceived exertion and maximum consumption of oxygen (r = 0.774 to 0.998)23. Potency was correlated in four studies, such that these studies presented correlations from moderate to very strong (r = 0.50  ≥ 0.90). In addition to this, weak values of correlation were found between perceived exertion and percentage of reserve cardiac frequency (r = 0.41; r = 0.18; r = 0.37) in different studies30-32. The analysis of the 35 correlations for different exercise protocols presented the following results. In the four studies that applied incremental tests, 17% of the correlations were between strong and very strong, and 9% of the correlations were moderate. For the seven articles that applied interval tests, 20% of the correlations were weak to very weak, 11% of the correlations were moderate, and 14% of the correlations were strong to very strong. For the analyses in the four studies that applied continuous tests, 9% of the correlations were moderate and 20% of the correlations were strong to very strong. J. Phys. Educ. v. 28, e2853, 2017. Page 7 of 13 A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Table 3. Results of the studies included in the review able 3. Translated without including the protocol for execution Page 6 of 13 Results of the studies included in the review es Affiliated country of the authors Sample Protocol/Exercise modality Type of Validation Results Cross-cultural adaptation al., Germany 10 men 22.6 (1.9) years Incremental/decreasing tests on cycle ergometer (maximum and minimum exertion) Concurrent Validity Correlation between RPE and potency after 1.5 min (r=0.83) RPE and CF after 1.5 min (r=0.78) Not cited 2013 Taiwan 12 men 22.9 (1.3) years Continuous tests at different intensities on treadmill (dynamic exercise) and with partially dynamic and isometric resistance exercises Concurrent Validity Coefficient of determination between RPE and CF for dynamic (R²= 0.70) partially dynamic (R²= 0.27) and static (R²= 0.10) exercise Utilized the version from Cantonese de Leung et al (2004)25 for Chinese participants al., China 40 elderly individuals (16 men and 24 women) 69.8 (4.5) years Interval tests on cycle ergometer Concurrent Validity Correlation between RPE and CF (r =0.70) between PSE and VO2 (r = 0.51) Utilized the version from Cantonese de Leung et al (2004)25 for Chinese participants al., Switzerland 3 men 28.3 (3.8) years Continuous and interval tests on cycle ergometer Concurrent Validity Correlation between RPE and potency (C: r=0.96; I: r=0.94) RPE and O2 consumption (C: r=0.97; I: r=0.91) RPE and CF (C: r=0.87; I: r=0.85) RPE and respiratory rate (C: r=0.66; I: r=-0.39) RPE and rate of ventilation (C: r=0.93; I: r=0.89) RPE and lactate (C: r=0.77; I: r=0.63) Translated without including the protocol for execution Results Karavatas & Travakol, 2005 United States 12 adults (6 men and 6 women) 26.8 (7.4) years) Incremental tests on treadmill Concurrent Validity Correlation between (r= 0.58) Leung et al., 2004 China 54 adults (25 men and 29 women) 22.2 (4.7) years Incremental tests on cycle ergometer Concurrent Validity Correlation between (r ≥0.73) Correlation and potency (r ≥ 0.69 between RPE and VO Lollgen et al., 1977 Germany 4 men 29 (8.5) years Interval tests on cycle ergometer Concurrent Validity Correlation between (r=0.71) Correlation between (r=0.20) Correlation between (r=0.63) López-Miñarro & Muyor, 2010 Spain 59 adults (13 men and 46 women) 32.1 (10.2) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation betwee reserve %CF (r Muyor & López- Miñarro, 2012 Spain 80 adults (19 men and 61 women) 32.3 (9.5) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation betwee reserve %CF (r Muyor, 2013 Spain 53 adults (25 men and 28 women) 28.8 (6) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation between (r=0.29) Correlation b and reserve %CF & 005 United States 12 adults (6 men and 6 women) 26.8 (7.4) years) Incremental tests on treadmill Concurrent Validity Correlation between RPE and CF (r= 0.58) Not cited l., China 54 adults (25 men and 29 women) 22.2 (4.7) years Incremental tests on cycle ergometer Concurrent Validity Correlation between RPE and CF (r ≥0.73) Correlation between RPE and potency (r ≥ 0.69) Correlation between RPE and VO2max (r ≥0.68) Adaptation based on Brisilin (1970)14 al., Germany 4 men 29 (8.5) years Interval tests on cycle ergometer Concurrent Validity Correlation between RPE and WL (r=0.71) Correlation between RPE and PR (r=0.20) Correlation between RPE and CF (r=0.63) Not cited ro & 10 Spain 59 adults (13 men and 46 women) 32.1 (10.2) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation between RPE and reserve %CF (r=0.41) Not cited pez- 12 Spain 80 adults (19 men and 61 women) 32.3 (9.5) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation between RPE and reserve %CF (r=0.18) Not cited 13 Spain 53 adults (25 men and 28 women) 28.8 (6) years Interval tests during indoor cycling (SpinningR) Concurrent Validity Correlation between RPE and CF (r=0.29) Correlation between RPE and reserve %CF (r=0.37) Not cited Correlation between RPE and CF (r= 0.58) Not cited Results Correlation between RPE and CF (r =0.70) between PSE and VO2 (r = 0.51) Correlation between RPE and CF (r =0.70) between PSE and VO2 (r = 0.51) Correlation between RPE and potency (C: r=0.96; I: r=0.94) RPE and O2 consumption (C: r=0.97; I: r=0.91) RPE and CF (C: r=0.87; I: r=0.85) RPE and respiratory rate (C: r=0.66; I: r=-0.39) RPE and rate of ventilation (C: r=0.93; I: r=0.89) RPE and lactate (C: r=0.77; I: r=0.63) J. Phys. Educ. v. 28, e2853, 2017. Cabral et al. Page 8 of 13 Page 8 of 13 Page 8 of 13 Cabral et al. J. Phys. Educ. v. 28, e2853, 2017. Discussion The objective of the present study was to analyze the validation and cross-cultural adaptation processes of Borg’s Rating of Perceived Exertion Scale (6-20) and of its instructions during exercise in adults and the elderly. The main results of this systematic review were that cross-cultural adaptation was reported in only four studies (for only the Cantonese and Japanese languages), and concurrent validity was the only validation method used, such that the majority of the studies used adults in the sample. In addition to this, the criteria for methodological quality that were less frequently achieved were those related to the application of Borg’s Rating of Perceived Exertion Scale (6-20). Cross-cultural adaptation is a necessary method for the translation and adaptation of an instrument for different cultures15. However, few adaptations for Borg’s scale (6-20) were found in this systematic review. Failing to complete the adaptation process can also influence the use of the scale. Furthermore, few articles reported the nationality of the sample. Evidence in the literature has indicated that there are cultural differences between different countries, as well as between different regions of the same country12,15,16. Situations were also found in which the instrument was originally developed for young adults and applied to the elderly21. In addition to a demand for caution during the entire process of adaptation of the linguistic terms, the validation process should determine if the translated scale accurately measures the construct to which it refers, for different age groups. Thus, the need for adaptation and validation of the instrument in accordance with the particularities and specificities of each sample (ex.: age group, language, and culture) is fundamental so that the results are understandable in relation to the new culture, without losing the original meaning16. No studies carrying out the cross-cultural adaptation of Borg’s scale (6-20) and its instructions were found for the Portuguese language, even though there is widespread use of the scale in this language. The translation of the scale and its instructions into Portuguese were only found in Borg’s book (2000)36, without a description of the cross-cultural adaptation process for this language. It should be noted that this scale has been frequently referenced in studies carried out in Brazil37. Changes to the qualitative semantics in the original scale can occur when cross-cultural adaptation is not applied, and this can directly influence the values reported when the translated scale is used. Page 10 of 13 Cabral et al. Correlation between RPE and CF (r= 0.58) Not cited Correlation between RPE and CF (r ≥0.73) Correlation between RPE and potency (r ≥ 0.69) Correlation between RPE and VO2max (r ≥0.68) Correlation between RPE and WL (r=0.71) Correlation between RPE and PR (r=0.20) Correlation between RPE and CF (r=0.63) J. Phys. Educ. v. 28, e2853, 2017. A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Page 9 of 13 Pollock et al., 2013 United States 13 adults (2 men and 11 women) 53.5 (5.4) years Interval tests with a variation of virtual games for the Nintendo Wii Concurrent Validity Correlation between RPE and CF (r=0.32) Not cited Robertson et al., 1996 United States 9 men 28.2 (4) years Continuous tests at different intensities on cycle ergometer adapted for arms and legs in the pool Concurrent Validity Correlation between RPE and potency (r= 0.56 to 0.79) Correlation between RPE and GME (r=0.55 to 0.68) Not cited Shigematsu et al., 2004 Japan 24 women 46.9 (7) years and 29 elderly women 75.5 (3.8) years Incremental tests on cycle ergometer Concurrent Validity Correlation between RPE and VO2 (r= 0.77-0.99) RPE and CF (r= 0.96 youth) (r= 0.95 elderly women) Utilized the Japanese version from Onodera and Miyashita (1976)35 Wenos et al., 1996 United States 24 elderly women 65 (3.8) years Continuous tests at different intensities on treadmill Concurrent Validity Correlation between RPE and CF (r=0.85) Not cited RPE = rating of perceived exertion; CF= cardiac frequency; VO2 = maximum volume of oxygen; %MR = percentage of maxium repetitions; WL: workload; PR: pedaling rate; O2= oxygen; GME = gross metabolic efficiency; C = continuous; I = intermittent. Source: The authors Page 9 of 13 J. Phys. Educ. v. 28, e2853, 2017. Page 10 of 13 Discussion Thus, there is a gap in the development of the cross-cultural adaptation process for Borg’s scale (6-20) for various languages, including Portuguese. One aspect to be taken into consideration is that the searches for the articles were only conducted in English and Portuguese, which could limit the results found, such that there is a chance that other studies which carried out the cross- cultural adaptation and validation processes for Borg’s scale (6-20) in different languages are not represented in this systematic review. The stage following cross-cultural adaption is that of validation. The articles found in this systematic review only utilized concurrent validity as a method of validation for Borg’s scale (6-20). This type of validity is the one most frequently found in correlations with ratings of perceived exertion, probably due to the ease of relating the scale with physiological measures (ex.: cardiac frequency, maximum volume of oxygen, rate of ventilation, among others) that have already been validated for evaluating the intensity of exercise6. Another possible motive for the use of this type of validity in the fact that Borg’s scale (6-20) was developed to correspond to the linear increase of psycho-physical functions with physiological response and intensity of exercise1. However, it can be observed that the relationship between the physiological variables and the exercise protocol used must be taken J. Phys. Educ. v. 28, e2853, 2017. A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale Page 11 of 13 Page 11 of 13 into consideration. This is observed, for example, in the CF analyses of the incremental and interval tests. For interval exercises, to the extent to which the intensity changes (decreasing or increasing), CF accompanies this variation, but perceived exertion continues to increase over time, even though the relationship of exertion/pause is maintained through the session35. For incremental tests, CF and perceived exertion increase in a linear manner with the elevation of the intensity of exercise. The motive for this discrepancy between exercise protocols and perceived exertion responses and physiological variations might be the origin of perceived exertion, sensory afferent pathways, or in the brain3,39. However, this discussion is outside the scope of this article. Thus, it is important to verify if the exercise protocol for which the scale is validated is similar to that which the researcher will develop in their study, in addition to the exercise modality. Discussion The methodological quality of the application of Borg’s scale was lacking in the studies reviewed due to the absence of the definition of the construct of perceived exertion and instructions for the use and familiarization of the instrument. The concept of perceived exertion is the first aspect to be considered when Borg’s scale (6-20) is applied to a population, and only five studies clarified this concept. Studies have related that there are variations in the definition of perceived exertion42, as the perception of general body fatigue or the perception of discomfort observed during the exercise1,41. Shigematsu et al (2004)23 reports the definition as “how difficult the individual feels the exercise is.” It is also possible to distinguish perceived exertion by way of other sensations not mentioned in the definitions above, such as, for example, thermal pain and discomfort41. With this, the discrepancy that studies have presented regarding the definition of ratings of perceived exertion can be observed. Thus, when studies use Borg’s scale (6-20) and its instructions, there should be a standard for the definition proposed by Borg (1998)1 (“perception of how hard and extenuating a physical task is”), so that the construct is not altered, and as a result, incorrect interpretations are not made which make the comparison of results between studies impossible40. The instructions for Borg’s scale (6-20) were provided in only five of the articles in this systematic review. Borg (1998)1 explains that the instructions for application of the scale should be clear and easy to understand, making necessary a standardized format among individuals to be tested, in order to obtain a more accurate result. The way in which instructions are given and questions are asked during the use of the scale of perceived exertion can influence the results and hinder the individual’s understanding of the scale42. Another problem is the difficulty of comparison between studies, when the instructions and analyses of perceived exertion are given in different ways for different studies. Therefore, it is important to consider the definition and the instructions given to the individuals in the sample, so that perceived exertion can be obtained accurately41. A session for familiarization with the scale ensures that individuals will be able to differentiate perceived exertion from other sensations and avoid underestimating or overestimating the indicated score41. References 1. Borg G. Borg's perceived exertion and pain scales. Champaign: Human Kinetcs, 1998. 2. Smirmaul BDPC. Sense of effort and other unpleasant sensations during exercise: clarifying concepts and mechanisms. Br J Sports Med. 2012;46(5):308–11. 2. Smirmaul BDPC. Sense of effort and other unpleasant sensations during exercise: clarifying concepts and mechanisms. Br J Sports Med. 2012;46(5):308–11. . Azevedo A, De Matos LF, Nakamura FY, Pereira G. Perception of effort monitors internal load during compounded circuit training. Motriz Rev Educ Fis. 2016;22(1):90–3. p g 4. Borg G. Physical Performance and Perceived Exertion. Studia Psychologica Et Pedagogica. 1962. p 5. Borg G. Perceived exertion as an indicator of somatic stress. Scand J Rehab Med. 1970;2-3:92-98. 6. Chen MJ, Fan X, Moe ST. Criterion-related validity of the Borg ratings of perceived exertion scale in healthy individuals: a meta-analysis. J Sports Sci. 2002;20(11):873–99. 7. Noble BJ, Robertson RJ. Perceived exertion. Champaign: Human Kinetics, 1996. obertson RJ. Perceived exertion. Champaign: Human 8. Faulkner J, Eston RG. Perceived Exertion Research in the 21 St C Entury : Developments, Reflections and Questions for the Future. J Exerc Sci Fit. 2008;6(1):1–14. 8. Faulkner J, Eston RG. Perceived Exertion Research in the 2 Questions for the Future. J Exerc Sci Fit. 2008;6(1):1–14. 9. Psycharakis SG. A longitudinal analysis on the validity and reliability of ratings of perceived exertio elite swimmers. J Strength Cond Res. 2011 25(2): 420–426. 10. Weeks A, Swerissen H, Belfrage J. Study Instruments. Eval Rev. 2007;153–65. 11. Borsa JC, Damásio BF, Bandeira DR. Cross-Cultural Adaptation and Validation of Psychological Instruments: Some Considerations. Pandéia, 2012, 22(53), 423-432. 12. Pasquali L. Instrumentação psicológica: fundamentos e práticas. Porto Alegre: Artmed, 2010. 13. Hambleton RK, Zenisky AL. Translating and adapting tests for cross cultural assessments. In D. Matsumoto & F. J. R. van de Vijver (Eds.), Cross-cultural research methods in psychology. New York: Cambridge University Press; 2011;46-74. 14. Brislin RW. Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology 1970 14. Brislin RW. Back translation for cross cultural research. Journal of Cross Cultural Psychology 1970;1: 185. 15. Beaton DE, Bombardier C, Guillemin F, Ferraz MB. Guidelines for the process of cross-cultural adaptation of self-report measures. Spine (Philadelphia). 2000;25(24):3186–91. 16. Guillemin F, Bombardier C, Beaton D. Cross-cultural adaptation of health-related quality of life measures: Literature review and proposed guidelines. J Clin Epidemiol. 1993;46(12):1417–32. 17. Shohamy E. The role of language tests in the construction and validation of secon-language aquisition theories. In: tarone, e. Discussion In addition to this, familiarization with the scale is indispensable for the valid collection of data, since the utilization of pre-tests for perceived exertion involves memory recall and verbal associations43. However, only five studies reported familiarization with the scale. Therefore, familiarization with Borg’s scale (6-20) should be carried out so that the individual understands perceived exertion, reports accurate scores, and also allows for the comparison between evaluations from different instances. J. Phys. Educ. v. 28, e2853, 2017. Page 12 of 13 Cabral et al. Conclusions Future studies that apply Borg’s Rating of Perceived Exertion Scale (6-20) for individuals that speak Portuguese as their native language need to develop cross-cultural adaptation and validation procedures for the scale (concurrent criteria, predictive criteria and content) for this language. Thus, after the development of a version of Borg’s scale (6-20) adapted to this language, and in cases in which the population evaluated or the exercise protocols or modalities are different from the original study, processes of validation should be applied in order to ensure that the scale evaluates perceived exertion adequately. J. Phys. Educ. v. 28, e2853, 2017. Page 13 of 13 A systematic review of cross-cultural adaptation and validation of Borg's rating of perceived exertion scale 22. Wenos DL, Wallace JP, Surburg PR, Morris HH. Reliability and comparison of RPE during variable and constant exercise protocols performed by older women./Fiabilite et comparaison du niveau de perception de l’ effort pendant des protocoles a dominante variable ou constante d’ exercices physiques e. Int J Sports Med. 1996;17(3):193–8. 23. Shigematsu R, Ueno LM, Nakagaichi M. Rate of Perceived Exertion as a Tool to Monitor Cycling Exercise Intensity in Older Adults. 2004. p. 3–9. 24. Chen Y-L, Chen C-C, Hsia P-Y, Lin S-K. Relationships of borg’s RPE 6-20 scale and heart rate in dynamic and static exercises among a sample of young taiwanese men. Percept Mot Skills. 2013;117(3):971–82. 24. Chen Y-L, Chen C-C, Hsia P-Y, Lin S-K. Relationships of borg’s RPE 6-20 scale and heart rate in dynamic and static exercises among a sample of young taiwanese men. Percept Mot Skills. 2013;117(3):971–82. 25. Leung RW, Leung ML, Chung PK. Validity and reliability of a Cantonese-translated rating of perceived and static exercises among a sample of young taiwanese men. Percept Mot Skills. 2013;117(3):971–82. 25. Leung RW, Leung ML, Chung PK. Validity and reliability of a Cantonese-translated rating of perceived exertion scale among Hong Kong adults. Percept Mot Ski. 2004;98:725–35. 25. Leung RW, Leung ML, Chung PK. Validity and reliability of a Cantonese-translated rating of perceived exertion scale among Hong Kong adults. Percept Mot Ski. 2004;98:725–35. 26. Karavatas SG, Tavakol K. Concurrent validity of Borg’s Rating of Perceived Exertion in African-American young adults, employing heart rate as the standard. J Allied Heal Sci Pract. 2005;3(1):7p. 7. Baildon RWA, Ulmer H. Time course of effort perception (RPE) and influence of strain. 1994;13:25–9 28. Edwards RHT, Hesser CM, Wigertz O, Melcher A, Ekelund LG. Physiological correlates of perceived exertion in continuous and intermittent exercise with same average power output. Eur J Clin Invest. 1972;2(2):108–14. ( ) 29. Lollgen H, Ulmer HV, Nieding GV. Heart rate and perceptual response to exercise with different pedalling spees in normal subjects and patients. 1977;304:297–304. 30. López-Miñarro PA, Muyor RJM. Heart rate and overall ratings of perceived exertion during Spinning® cycle indoor session in novice adults. Science & Sports. 2010;25, 238—244 31. Muyor JM, López-Miñarro PA. Overall ratings of perceived exertion and heart rate during indoor cycling session in non expert subjects. Med Sport, 2012: 65: 145-54. 32. Muyor M. Author address: Luana Loss Cabral. Address: Rua Coração de Maria, 92, Jardim Botânico, CEP 80210-132, Curitiba - PR. E-mail: luanalcabral@yahoo.com.br References E.; gass, s. M.; cohen, a. D. (Org.). Research methodology in second language aquisition. New Jersey: Lawrence Erlbaum Associates publishers, 1994. g ( g ) aquisition. New Jersey: Lawrence Erlbaum Associates publishers, 1994. 18. Soriano-Maldonado A, Romero L, Femia P, Roero C, Ruiz JR, Gutierrez A. A Learning Protocol Improves the Validity of the Borg 6–20 RPE Scale During Indoor Cycling. Int J Sports Med. 2014; 35:379–384. the Validity of the Borg 6–20 RPE Scale During Indoor Cycling. Int J Sports Med. 2014; 35:379–384. 19. Russell WD. On the current status of rated perceived exertion. Percept Mot Skills. 1997;84(3 Pt 1):799–808. 20. Sampaio RF, Mancini M C. Estudos de revisão sistemática: um guia para síntese criteriosa da evidência científica. Rev Bras Fisioter. 2007;11(1):83–9. 19. Russell WD. On the current status of rated perceived exertion. Percept Mot Skills. 1997;84(3 Pt 1):7 p p 20. Sampaio RF, Mancini M C. Estudos de revisão sistemática: um guia para síntese criteriosa da evidência científica. Rev Bras Fisioter. 2007;11(1):83–9. 21. Chung P-K, Zhao Y, Liu J-D, Quach B. A brief note on the validity and reliability of the rating of perceived exertion scale in monitoring exercise intensity among chinese older adults in hong kong. Percept Mot Skills. 2015;121(3):805–9. J. Phys. Educ. v. 28, e2853, 2017. Page 13 of 13 Page 13 of 13 Page 13 of 13 Exercise intensity and validity of the ratings of perceived exertion (Borg and OMNI Scales) in an indoor cycling session by. 2013;39(December):93–101. 33. Pollock BS, Barkley JE, Potenzini N, DeSalvo RM, Buser SL, Otterstetter R, et al. Validity of borg ratings of perceived exertion during active video game play. Int J Exerc Sci. 2013;6(2):164–70. 34. Robertson R, Goss F, Michael T, Moyna N, Gordon P, Visich P, Kang J, Angelopoulos T, Dasilva S, Metz K. Percep Mot Skills, 1996,83,3-13. 35. Onodera K, Miyashita M. A study of Japanese scale for rating of perceived exertion in endurance exercise. Japanese Journal of Physical Education, 1976; 21, 191-203. 36. Borg G. Escalas de borg para dor e esforço percebido. São Paulo: Editora Manole, 2000. 37. Afonso LDS, Dos Santos JFB, Lopes JR, Tambelli R, Santos EHR, Back FA, et al. Frequência cardíaca máxima em esteira ergométrica em diferentes horários. Rev Bras Med do Esporte. 2006;12(6):318–22. 38. Nakamura FY, Silva KRT, Okuno NM, Pereira G, Kokubun E. Determinação da sobrecarga anaeróbia em exercício intermitente a partir do modelo de potência crítica. Motriz. 2007;13(4):298–304. 39. Marcora SM. Perception of effort during exercise is independent of afferent feedback from skeletal muscles, heart, and lungs. J Appl Phyiology. 2009;106(May 2008):2060–2. 40. Pereira G, de Souza DM, Reichert FF, Smirmaul BPC. RBCDH and mechanisms : a literature review. Rev Bras Cineant Desemp Hum. 2014;16(5):579–87. 41. Pageaux B. Perception of effort in Exercise Science: Definition, measurement and perspectives. Eur J Sport Sci. 2016;16(8):1–10. 2. Abbiss CR, Peiffer JJ, Meeusen R, Skorski S. Role of ratings of perceived exertion during self-paced exercise: What are we actually measuring? Sport Med. 2015;45(9):1235–43. 3. Eston R, Coquart J, Lamb K, & Parfitt G. Misperception: No evidence to dismiss RPE as regulator of moderate-intensity exercise. Medicine and Science in Sports and Exercise, 2015;47(12), 2676. J. Phys. Educ. v. 28, e2853, 2017.
https://openalex.org/W2990331509	https://europepmc.org/articles/pmc6873726?pdf=render	English	null	Association of IKZF1 SNPs in cold medicine-related Stevens–Johnson syndrome in Thailand	Clinical and translational allergy	2,019	cc-by	3,105	Chantaren et al. Clin Transl Allergy (2019) 9:61 https://doi.org/10.1186/s13601-019-0300-9 Chantaren et al. Clin Transl Allergy (2019) 9:61 https://doi.org/10.1186/s13601-019-0300-9 Clinical and Translational Allergy Open Access © The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abstract Purpose: Our meta-analysis of several ethnic groups (Japanese, Korean, Indian, Brazilian) revealed a significant genome-wide association between cold medicine-related SJS/TEN (CM-SJS/TEN) with severe ocular complications (SOC) and IKZF1 SNPs, suggesting that IKZF1 might be a potential marker for susceptibility to CM-SJS/TEN with SOC. In this study, we examined the association between CM-SJS/TEN with SOC and the IKZF1 SNPs in the Thai population. Methods: 57 CM-SJS/TEN with SOC and 171 control samples were collected at Chulalongkorn University and Mahidol University. Genomic DNA samples were genotyped for the IKZF1 SNPs at Kyoto Prefectural University of Medi- cine in Japan using the TaqMan SNP genotyping assay. Results: The four SNPs previously reported to be associated with CM-SJS/TEN with SOC in the Japanese were exam- ined in the Thai samples. Although the number of Thai cases (n = 57) was small, a significant association between CM-SJS/TEN with SOC and IKZF1 SNPs which included rs4917014 (T vs G, OR = 2.9, p = 0.0012, Pc = 0.0049), rs4917129 (T vs C, OR = 2.8, p = 0.0026, Pc = 0.010) and rs10276619 (G vs A, OR = 1.8, p = 0.012, Pc = 0.048) was identified. Conclusion: In addition to the Japanese, Korean and Indian populations, Thai cases with CM-SJS/TEN and SOC were significantly associated with IKZF1 SNPs. With our previous report of the critical role of IKZF1 in mucocutaneous inflammation, these results suggest that IKZF1 is important in the pathogenesis of CM-SJS/TEN with SOC. Keywords: Stevens–Johnson syndrome, Cold medicine, Severe ocular complications, SNPs, IKZF1 Association of IKZF1 SNPs in cold medicine‑related Stevens–Johnson syndrome in Thailand Patchima Chantaren1†, Passara Jongkhajornpong2†, Mayumi Ueta3† , Vilavun Puangsricharern1, Kaevalin Lekhanont2, Phattrawan Pisuchpen1, Pinnita Prabhasawat4, Kanya Suphapeetiporn5 and Shigeru Kinoshita3 To the editor All experimental pro- cesses were complied with the principles set forth in the Helsinki declaration. The diagnostic criteria of SJS/ TEN were based on a confirmed history of acute onset of high fever, skin eruption with at least two sites of serious mucocutaneous involvement including the oral mucosa and the ocular surface [5]. Thai healthy volun- teers were used as controls. CM was defined as the drug that patients took for relieving cold symptoms includ- ing nonsteroidal anti-inflammatory drugs (NSAIDS), acetaminophen, and other multi-ingredient cold medi- cations [6]. We previously reported that in Japanese, for SJS/TEN with SOC, acetaminophen was a main drug of CM; 48% of the patients have taken acetaminophen before developing SJS/TEN with SOC [6]. In Thailand, paracetamol (which is equal to acetaminophen) might be also important causative drug of CM, because 20 of 57 (35%) patients have taken paracetamol before developing it. The patients were classified as having SOC if the fol- lowing manifestations were detected; severe conjunctivi- tis, pseudomembrane, and epithelial defect on the ocular surface in the acute stage and/or ocular sequelae such as dry eye, trichiasis, symblepharon, and conjunctival inva- sion into the cornea in the chronic stage [6]. Our functional analysis of SNPs of the IKZF1 gene revealed that the ratio of the splicing isoforms Ik2/Ik1 could be affected by IKZF1 SNPs significantly associated with susceptibility to CM-SJS/TEN with SOC [5]. The quantity of the Ik2 isoform is increased in disease-protec- tive genotypes of IKZF1 (rs4917014 G/G and rs10276619 A/A) [5]. As Ikaros 2, an Ik2 isoform lacks the DNA- binding ability and seems to be dominant-negative. It is possible that the function of Ikaros, the protein of IKZF1, is enhanced in CM-SJS/TEN with SOC [5]. Ikaros is a transcription factor that regulates numer- ous biological events. It was reported that Ikaros-null mice lack B-lineage cells, NK cells, peripheral lymph node- and fetal T-cells, thus Ikaros family members regulate important cell-fate decisions in the develop- ment of the adaptive immune system [8]. On the other hand, we have reported that epithelium might be con- tribute to the pathobiology of CM-SJS/TEN with SOC [9]. Thus, we produced K5-Ikzf1-EGFP transgenic mice (Ikzf1 Tg) by introducing the Ik1 isoform into cells expressing keratin 5, which is expressed in epithe- lial tissues such as the epidermis and conjunctiva and found that mucocutaneous inflammation was exacer- bated in Ikzf1-Tg mice. To the editor of several ethnic groups (Japanese, Korean, Indian, Bra- zilian) revealed a significant genome-wide association between CM-SJS/TEN with SOC and IKZF1, suggesting that IKZF1 might be a potential marker for susceptibility to CM-SJS/TEN with SOC [5]. In this study, we examined the association between Thai CM-SJS/TEN with SOC and the IKZF1 SNPs, known to be associated with the Japanese CM-SJS/TEN with SOC. of several ethnic groups (Japanese, Korean, Indian, Bra- zilian) revealed a significant genome-wide association between CM-SJS/TEN with SOC and IKZF1, suggesting that IKZF1 might be a potential marker for susceptibility to CM-SJS/TEN with SOC [5]. In this study, we examined the association between Thai CM-SJS/TEN with SOC and the IKZF1 SNPs, known to be associated with the Japanese CM-SJS/TEN with SOC. The 4 SNPs previously reported to be associated with CM-SJS/TEN with SOC in the Japanese were examined in the Thai samples. Although the number of Thai cases (n = 57) was small, we again found a significant associa- tion between CM-SJS/TEN with SOC and 3 IKZF1 SNPs which included rs4917014 (T vs G, OR = 2.9, p = 0.0012, Pc = 0.0049), rs4917129 (T vs C, OR = 2.8, p = 0.0026, Pc = 0.010) and rs10276619 (G vs A, OR = 1.8, p = 0.012, Pc = 0.048) (Table 1). Our previous results of meta-analy- sis in the Japanese, Korean, Indian and Brazilian showed the significant associations in 3 IKZF1 SNPs [rs4917014 (T vs G, OR = 2, p = 8.5 × 10−11) (which is equal to (G (minor allele) vs T (major allele), OR = 0.5) in the previ- ous paper), rs4917129 (T vs C, OR = 2, p = 8.0 × 10−9) (which is equal to (C (minor allele) vs T (major allele), OR = 0.5 in the previous paper)) and rs10276619 (G vs A, OR = 1.8, p = 4.3 × 10−9)] [5]. Present results in the Thai population are in concordance with our previous report. The CM-SJS/TEN with SOC and control samples were collected at Chulalongkorn University (King Chulalong- korn Memorial Hospital) and Mahidol University (Ram- athibodi Hospital and Siriraj Hospital). Genomic DNA samples were genotyped for the IKZF1 SNPs at Kyoto Prefectural University of Medicine in Japan. The study was approved by the institutional review board of all institutes. Protocol explanation and obtaining written informed consent were done in all participants before starting experimental procedures. To the editor medications and non-steroid anti-inflammatory drugs (NSAIDs) were the main causative drugs of SJS/TEN with SOC [2]. In the acute stage, in addition to skin eruption and erosion, SJS/TEN with SOC patients manifest severe conjunctivitis with corneal and conjunctival erosion and pseudo-membranes. Despite healing of the skin lesions, in the chronic stage of SJS/TEN with SOC, ocular surface sequelae such as severe dry eye, symblepharon, trichiasis, scaring of palpebra conjunctiva, and conjunctival inva- sion into the cornea may persist [3]. While the reported annual incidence of SJS/TEN is very rare (only 1–6/106 individuals), its mortality rate is high (3% for SJS and 27% for TEN) [4]. We previously reported that the IKZF1 gene was strongly associated with CM-SJS/TEN with SOC in Japanese patients [5]. In addition, a meta-analysis Stevens–Johnson syndrome (SJS) and its severe type, toxic epidermal necrolysis (TEN), are acute inflamma- tory vesiculobullous reactions of the skin and mucosa including the ocular surface, oral cavity, and genitals. Severe ocular complications (SOC) appear in about half of SJS/TEN patients diagnosed by dermatologists [1]. Cold medicines (CM), including multi-ingredient cold Correspondence: mueta@koto.kpu‑m.ac.jp †Patchima Chantaren, Passara Jongkhajornpong and Mayumi Ueta contributed equally as first authors 3 Department of Frontier Medical Science and Technology for Ophthalmology, Kyoto Prefectural University of Medicine, 465 Kajiicho, Hirokoji, Kawaramachi, Kamigyoku, Kyoto 602‑0841, Japan Full list of author information is available at the end of the article Correspondence: mueta@koto.kpu‑m.ac.jp †Patchima Chantaren, Passara Jongkhajornpong and Mayumi Ueta contributed equally as first authors 3 Department of Frontier Medical Science and Technology for Ophthalmology, Kyoto Prefectural University of Medicine, 465 Kajiicho, Hirokoji, Kawaramachi, Kamigyoku, Kyoto 602‑0841, Japan Full list of author information is available at the end of the article Correspondence: mueta@koto.kpu‑m.ac.jp †Patchima Chantaren, Passara Jongkhajornpong and Mayumi Ueta contributed equally as first authors 3 Department of Frontier Medical Science and Technology for Ophthalmology, Kyoto Prefectural University of Medicine, 465 Kajiicho, Hirokoji, Kawaramachi, Kamigyoku, Kyoto 602‑0841, Japan Full list of author information is available at the end of the article Page 2 of 4 Chantaren et al. Clin Transl Allergy (2019) 9:61 Chantaren et al. Clin Transl Allergy (2019) 9:61 Chantaren et al. Clin Transl Allergy The TaqMan SNP genotyping assay (Applied Biosystems, Foster City, CA) was used for the genotypes of the IKZF1 gene as previously reported [5]. Chi squared test was applied to a two-by-two contingency table for the allele frequency and the dominant and recessive models. Acknowledgements We thank all the patients and volunteers who enrolled in our study. We also thank Ms. Hiromi Nishigaki for technical assistance. Consent for publication Protocol explanation and obtaining written informed consent were done in all participants before starting experimental procedures. 6. Ueta M, Kaniwa N, Sotozono C, Tokunaga K, Saito Y, Sawai H, Miyadera H, Sugiyama E, Maekawa K, Nakamura R, et al. Independent strong association of HLA-A02:06 and HLA-B44:03 with cold medicine-related 6. Ueta M, Kaniwa N, Sotozono C, Tokunaga K, Saito Y, Sawai H, Miyadera H, Sugiyama E, Maekawa K, Nakamura R, et al. Independent strong association of HLA-A02:06 and HLA-B44:03 with cold medicine-related Ethics approval and consent to participate The study was approved by the institutional review board of Chulalongkorn University (King Chulalongkorn Memorial Hospital), Mahidol University 5. Ueta M, Sawai H, Sotozono C, Hitomi Y, Kaniwa N, Kim MK, Seo KY, Yoon KC, Joo CK, Kannabiran C, et al. IKZF1, a new susceptibility gene for cold medicine-related Stevens–Johnson syndrome/toxic epidermal necrolysis with severe mucosal involvement. J Allergy Clin Immunol. 2015;135(6):1538–45. (Ramathibodi Hospital and Siriraj Hospital) and Kyoto Prefectural University of Medicine. To the editor They developed dermatitis with some having blepharoconjunctivitis [10]. Histologi- cal analysis showed not only dermatitis but also tissue inflammation in the blepharoconjunctiva, tongue, and paronychia [10], similar to the findings in patients in the acute state of SJS/TEN with SOC [9]. Our studies demonstrated that IKZF1 could play a critical role in Of 57 CM-SJS/TEN with SOC, 23 were male and 34 were female; their age ranged from 6 to 73 years [median 42.3 ± 15.6 (SD) years]. The age at SJS/TEN onset ranged from 2 to 54 years (median 24.8 ± 13.8 years). The con- trols were 85 males and 86 females; their median age was 39.5 ± 14.3 years. Some of the CM-SJS/TEN patients and some of the controls had been included in our earlier studies [7]. Subjects were obtained DNA extraction from whole peripheral blood using the PAX gene blood DNA kits (Qiagen, Hilden, Germany) or from saliva using Ora- gene DNA (Kyodou International, Kanagawa, Japan). Chantaren et al. Clin Transl Allergy (2019) 9:61 Page 3 of 4 Chantaren et al. Clin Transl Allergy Table 1 Woolf’s correction rs number of IKZF1 SNP Genotypes Case (N = 57) Control (N = 171) Allele 1 vs 2 Genotype 11 vs 12 + 22 Genotype 11 + 12 vs 22 p-value Corrected p-value OR (95%CI) p-value Corrected p-value OR (95%CI) p-value Corrected p-value OR (95%CI) rs897693 11 T/T 36/57 (63.2%) 114/171 (66.7%) 0.69 0.63 1.00 12 T/C 17/57 (29.8%) 45/171 (26.3%) – – – 22 C/C 4/57 (7%) 12/171 (7%) – – – rs4917014 11 T/T 46/57 (80.7%) 100/171 (58.5%) 1.22.E−03 2.46.E−03 0.0619 12 G/T 11/57 (19.3%) 61/171 (35.7%) 4.89.E−03 9.86.E−03 – 22 G/G 0/57 (0%) 10/171 (5.8%) 2.91 (1.49–5.68) 2.97 (1.44–6.13) – rs4917129 11 T/T 47/57 (82.5%) 109/171 (63.7%) 2.59.E−03 8.48.E−03 4.97.E−02 12 T/C 10/57 (17.5%) 51/171 (29.8%) 1.03.E−02 3.39.E−02 0.199 22 C/C 0/57 (0%) 11/171 (6.4%) 2.82 (1.40–5.68) 2.67 (1.26–5.66) 8.24* (0.48–142.08) rs10276619 11 G/G 25/57 (43.9%) 52/171 (30.4%) 1.21.E−02 0.063 2.97.E−02 12 G/A 26/57 (45.6%) 78/171 (45.6%) 4.84.E−02 – 0.119 22 A/A 6/57 (10.5%) 41/171 (24%) 1.76 (1.13–2.74) – 2.68 (1.07–6.70) Table 1 Woolf’s correction maintaining mucocutaneous homeostasis [10] and sug- gested that it might be implicated in the aggravation of mucocutaneous inflammation seen in the presence of CM-SJS/TEN with SOC [10]. Funding h k g This work was supported by grants-in-aid from the Ministry of Education, Cul- ture, Sports, Science and Technology of the Japanese government, and by the JSPS Core-to-Core Program, A. Advanced Research Networks and also partly supported by grants-in-aids for scientific research from the Japanese Ministry of Health, Labor and Welfare. 3. Sotozono C, Ang LP, Koizumi N, Higashihara H, Ueta M, Inatomi T, Yokoi N, Kaido M, Dogru M, Shimazaki J, et al. New grading system for the evalu- ation of chronic ocular manifestations in patients with Stevens–Johnson syndrome. Ophthalmology. 2007;114(7):1294–302. 4. Power WJ, Ghoraishi M, Merayo-Lloves J, Neves RA, Foster CS. Analysis of the acute ophthalmic manifestations of the erythema multiforme/Ste- vens–Johnson syndrome/toxic epidermal necrolysis disease spectrum. Ophthalmology. 1995;102(11):1669–76. 4. Power WJ, Ghoraishi M, Merayo-Lloves J, Neves RA, Foster CS. Analysis of the acute ophthalmic manifestations of the erythema multiforme/Ste- vens–Johnson syndrome/toxic epidermal necrolysis disease spectrum. Ophthalmology. 1995;102(11):1669–76. Author details 1 Department of Ophthalmology, Faculty of Medicine, Chulalongkorn Univer- sity and Excellence Center for Cornea and Limbal Stem Cell Transplantation, King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bangkok, Thailand. 2 Department of Ophthalmology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand. 3 Department of Frontier Med- ical Science and Technology for Ophthalmology, Kyoto Prefectural University of Medicine, 465 Kajiicho, Hirokoji, Kawaramachi, Kamigyoku, Kyoto 602‑0841, Japan. 4 Department of Ophthalmology, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand. 5 Department of Pediatrics, Faculty of Medicine, Chulalongkorn University and Excellence Center for Medical Genetics, King Chulalongkorn Memorial Hospital, the Thai Red Cross Society, Bangkok, Thailand. In addition to the Japanese, Korean and Indian in our previous report [5], CM-SJS/TEN with SOC was sig- nificantly associated with IKZF1 SNPs in the Thai cases. With our previous report of the critical role of IKZF1 in mucocutaneous inflammation [8], these results suggest that IKZF1 is important in the pathogenesis of CM-SJS/ TEN with SOC. Received: 20 June 2019 Accepted: 13 November 2019 Received: 20 June 2019 Accepted: 13 November 2019 Authors’ contributions 1. Sotozono C, Ueta M, Nakatani E, Kitami A, Watanabe H, Sueki H, Iijima M, Aihara M, Ikezawa Z, Aihara Y, et al. Predictive factors associated with acute ocular involvement in Stevens–Johnson syndrome and toxic epidermal necrolysis. Am J Ophthalmol. 2015;160(2):228–37. 1. Sotozono C, Ueta M, Nakatani E, Kitami A, Watanabe H, Sueki H, Iijima M, Aihara M, Ikezawa Z, Aihara Y, et al. Predictive factors associated with acute ocular involvement in Stevens–Johnson syndrome and toxic epidermal necrolysis. Am J Ophthalmol. 2015;160(2):228–37. MU made the experimental design, analyze the SNPs, wrote the text, and prepared the table. PC, PJ, PP and PP collected samples. PC, PJ, VP, KL, KS and SK contributed to the research and reviewed the manuscript. All authors read and approved the final manuscript. 2. Ueta M. Results of detailed investigations into Stevens–Johnson syndrome with severe ocular complications. Invest Ophthalmol Vis Sci. 2018;59(14):DES183–91. 2. Ueta M. Results of detailed investigations into Stevens–Johnson syndrome with severe ocular complications. Invest Ophthalmol Vis Sci. 2018;59(14):DES183–91. Competing interests Competing interests The authors declare that they have no competing interests. The authors declare that they have no competing interests. Chantaren et al. Clin Transl Allergy (2019) 9:61 Chantaren et al. Clin Transl Allergy (2019) 9:61 Page 4 of 4 10. Ueta M, Hamuro J, Nishigaki H, Nakamura N, Shinomiya K, Mizushima K, Hitomi Y, Tamagawa-Mineoka R, Yokoi N, Naito Y, et al. Mucocutaneous inflammation in the Ikaros family zinc finger 1-keratin 5-specific trans- genic mice. Allergy. 2018;73(2):395–404. 7. Jongkhajornpong P, Lekhanont K, Pisuchpen P, Chantaren P, Puangsrich- arern V, Prabhasawat P, Suphapeetiporn K, Kinoshita S, Ueta M. Associa- tion between HLA-B44:03-HLA-C07:01 haplotype and cold medicine- related Stevens–Johnson syndrome with severe ocular complications in Thailand. Br J Ophthalmol. 2018;102(9):1303–7. Stevens–Johnson syndrome with severe mucosal involvement. Sci Rep. 2014;4:4862. 7. Jongkhajornpong P, Lekhanont K, Pisuchpen P, Chantaren P, Puangsrich- arern V, Prabhasawat P, Suphapeetiporn K, Kinoshita S, Ueta M. Associa- tion between HLA-B44:03-HLA-C07:01 haplotype and cold medicine- related Stevens–Johnson syndrome with severe ocular complications in Thailand. Br J Ophthalmol. 2018;102(9):1303–7. 8. John LB, Ward AC. The Ikaros gene family: transcriptional regulators of hematopoiesis and immunity. Mol Immunol. 2011;48(9–10):1272–8. 9. Ueta M, Kinoshita S. Ocular surface inflammation is regulated by innate immunity. Prog Retin Eye Res. 2012;31(6):551–75. 9. Ueta M, Kinoshita S. Ocular surface inflammation is regulated by innate immunity. Prog Retin Eye Res. 2012;31(6):551–75. Chantaren et al. Clin Transl Allergy (2019) 9:61 8. John LB, Ward AC. The Ikaros gene family: transcriptional regulators of hematopoiesis and immunity. Mol Immunol. 2011;48(9–10):1272–8. 10. Ueta M, Hamuro J, Nishigaki H, Nakamura N, Shinomiya K, Mizushima K, Hitomi Y, Tamagawa-Mineoka R, Yokoi N, Naito Y, et al. Mucocutaneous inflammation in the Ikaros family zinc finger 1-keratin 5-specific trans- genic mice. Allergy. 2018;73(2):395–404. Stevens–Johnson syndrome with severe mucosal involvement. Sci Rep. 2014;4:4862. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research ? Choose BMC and benefit from:
https://openalex.org/W3045704485	https://vbn.aau.dk/ws/files/362868828/Ramilowski_et_al_2020_Functional_annotation_of_human_long_noncoding_RNAs_via_molecular_phenotyping.pdf	English	null	Functional annotation of human long noncoding RNAs via molecular phenotyping	Genome research	2,020	cc-by	13,144	Aalborg Universitet © 2020 Ramilowski et al. This article, published in Genome Research, is avail- able under a Creative Commons License (Attribution 4.0 International), as de- scribed at http://creativecommons.org/licenses/by/4.0/. 47These authors contributed equally to this work. Corresponding authors: michiel.dehoon@riken.jp, jay.shin@riken.jp, carninci@riken.jp Article published online before print. Article, supplemental material, and publi- cation date are at http://www.genome.org/cgi/doi/10.1101/gr.254219.119. Freely available online through the Genome Research Open Access option. Functional annotation of human long noncoding RNAs via molecular phenotyping Ramilowski, Jordan A; Yip, Chi Wai; Agrawal, Saumya; Chang, Jen-Chien; Ciani, Yari; Kulakovskiy, Ivan V; Mendez, Mickaël; Ooi, Jasmine Li Ching; Ouyang, John F; Parkinson, Nick; Petri, Andreas; Roos, Leonie; Severin, Jessica; Yasuzawa, Kayoko; Abugessaisa, Imad; Akalin, Altuna; Antonov, Ivan V; Arner, Erik; Bonetti, Alessandro; Bono, Hidemasa; Borsari, Beatrice; Brombacher, Frank; Cameron, Christopher JF; Cannistraci, Carlo Vittorio; Cardenas, Ryan; Cardon, Melissa; Chang, Howard; Dostie, Josée; Ducoli, Luca; Favorov, Alexander; Fort, Alexandre; Garrido, Diego; Gil, Noa; Gimenez, Juliette; Guler, Reto; Handoko, Lusy; Harshbarger, Jayson; Hasegawa, Akira; Hasegawa, Yuki; Hashimoto, Kosuke; Hayatsu, Norihito; Heutink, Peter; Hirose, Tetsuro; Imada, Eddie L; Itoh, Masayoshi; Kaczkowski, Bogumil; Kanhere, Aditi; Kawabata, Emily; Kawaji, Hideya; Kawashima, Tsugumi; Kelly, S Thomas; Kojima, Miki; Kondo, Naoto; Koseki, Haruhiko; Kouno, Tsukasa; Kratz, Anton; Kurowska-Stolarska, Mariola; Kwon, Andrew Tae Jun; Leek, Jeffrey; Lennartsson, Andreas; Lizio, Marina; López-Redondo, Fernando; Luginbühl, Joachim; Maeda, Shiori; Makeev, Vsevolod J; Marchionni, Luigi; Medvedeva, Yulia A; Minoda, Aki; Müller, Ferenc; Muñoz-Aguirre, Manuel; Murata, Mitsuyoshi; Nishiyori, Hiromi; Nitta, Kazuhiro R; Noguchi, Shuhei; Noro, Yukihiko; Nurtdinov, Ramil; Okazaki, Yasushi; Orlando, Valerio; Paquette, Denis; Parr, Callum J C; Rackham, Owen J L; Rizzu, Patrizia; Sánchez Martinez, Diego Fernando; Sandelin, Albin; Sanjana, Pillay; Semple, Colin A M; Shibayama, Youtaro; Sivaraman, Divya M; Suzuki, Takahiro; Szumowski, Suzannah C; Tagami, Michihira; Taylor, Martin S; Terao, Chikashi; Thodberg, Malte; Thongjuea, Supat; Tripathi, Vidisha; Ulitsky, Igor; Verardo, Roberto; Vorontsov, Ilya E; Yamamoto, Chinatsu; Young, Robert S; Baillie, J Kenneth; Forrest, Alistair R R; Guigó, Roderic; Hoffman, Michael M; Hon, Chung Chau; Kasukawa, Takeya; Kauppinen, Sakari; Kere, Juha; Lenhard, Boris; Schneider, Claudio; Suzuki, Harukazu; Yagi, Ken; de Hoon, Michiel J L; Shin, Jay W; Carninci, Piero Published in: DOI (link to publication from Publisher): 10.1101/gr.254219.119 Creative Commons License CC BY 4.0 Publication date: 2020 Document Version Publisher's PDF, also known as Version of record Link to publication from Aalborg University DOI (link to publication from Publisher): 10.1101/gr.254219.119 Document Version Publisher's PDF, also known as Version of record Link to publication from Aalborg University Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org oaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org oaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org oaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org oaded from Functional annotation of human long noncoding RNAs via molecular phenotyping RNAs via molecular phenotyping Jordan A. Ramilowski,1,2,47 Chi Wai Yip,1,2,47 Saumya Agrawal,1,2 Jen-Chien Chang,1,2 Yari Ciani,3 Ivan V. Kulakovskiy,4,5 Mickaël Mendez,6 Jasmine Li Ching Ooi,2 John F. Ouyang,7 Nick Parkinson,8 Andreas Petri,9 Leonie Roos,10,11 Jessica Severin,1,2 Kayoko Yasuzawa,1,2 Imad Abugessaisa,1,2 Altuna Akalin,12 Ivan V. Antonov,13 Erik Arner,1,2 Alessandro Bonetti,2 Hidemasa Bono,14 Beatrice Borsari,15 Frank Brombacher,16,17 Christopher J.F. Cameron,18,23,46 Carlo Vittorio Cannistraci,19,20 Ryan Cardenas,21 Melissa Cardon,1 Howard Chang,22 Josée Dostie,23 Luca Ducoli,24 Alexander Favorov,25,26 Alexandre Fort,2 Diego Garrido,15 Noa Gil,27 Juliette Gimenez,28 Reto Guler,16,17 Lusy Handoko,2 Jayson Harshbarger,2 Akira Hasegawa,1,2 Yuki Hasegawa,2 Kosuke Hashimoto,1,2 Norihito Hayatsu,1 Peter Heutink,29 Tetsuro Hirose,30 Eddie L. Imada,26 Masayoshi Itoh,2,31 Bogumil Kaczkowski,1,2 Aditi Kanhere,21 Emily Kawabata,2 Hideya Kawaji,31 Tsugumi Kawashima,1,2 S. Thomas Kelly,1 Miki Kojima,1,2 Naoto Kondo,2 Haruhiko Koseki,1 Tsukasa Kouno,1,2 Anton Kratz,2 Mariola Kurowska-Stolarska,32 Andrew Tae Jun Kwon,1,2 Jeffrey Leek,26 Andreas Lennartsson,33 Marina Lizio,1,2 Fernando López-Redondo,1,2 Joachim Luginbühl,1,2 Shiori Maeda,1 Vsevolod J. Makeev,25,34 Luigi Marchionni,26 Yulia A. Medvedeva,13,34 Aki Minoda,1,2 Ferenc Müller,21 Manuel Muñoz-Aguirre,15 Mitsuyoshi Murata,1,2 Hiromi Nishiyori,1,2 Kazuhiro R. Nitta,1,2 Shuhei Noguchi,1,2 Yukihiko Noro,2 Ramil Nurtdinov,15 Yasushi Okazaki,1,2 Valerio Orlando,35 Denis Paquette,23 Callum J.C. Parr,1 Owen J.L. Rackham,7 Patrizia Rizzu,29 Diego Fernando Sánchez Martinez,26 Albin Sandelin,36 Pillay Sanjana,21 Colin A.M. Semple,37 Youtaro Shibayama,1,2 Divya M. Sivaraman,1,2 Takahiro Suzuki,1,2 Suzannah C. Szumowski,2 Michihira Tagami,1,2 Martin S. Taylor,37 Chikashi Terao,1 Malte Thodberg,36 Supat Thongjuea,2 Vidisha Tripathi,38 Igor Ulitsky,27 Roberto Verardo,3 Ilya E. Vorontsov,25 Chinatsu Yamamoto,2 Robert S. Young,39 J. Kenneth Baillie,8 Alistair R.R. Forrest,1,2,40 Roderic Guigó,15,41 Michael M. Hoffman,42 Chung Chau Hon,1,2 Takeya Kasukawa,1,2 Sakari Kauppinen,9 Juha Kere,33,43 Boris Lenhard,10,11,44 Claudio Schneider,3,45 Harukazu Suzuki,1,2 Ken Yagi,1,2 Michiel J.L. de Hoon,1,2 Jay W. Shin,1,2 and Piero Carninci1,2 RNAs via molecular phenotyping Jordan A. Ramilowski,1,2,47 Chi Wai Yip,1,2,47 Saumya Agrawal,1,2 Jen-Chien Chang,1,2 Yari Ciani,3 Ivan V. Kulakovskiy,4,5 Mickaël Mendez,6 Jasmine Li Ching Ooi,2 John F. Ouyang,7 Nick Parkinson,8 Andreas Petri,9 Leonie Roos,10,11 Jessica Severin,1,2 Kayoko Yasuzawa,1,2 Imad Abugessaisa,1,2 Altuna Akalin,12 Ivan V. Antonov,13 Erik Arner,1,2 Alessandro Bonetti,2 Hidemasa Bono,14 Beatrice Borsari,15 Frank Brombacher,16,17 Christopher J.F. Cameron,18,23,46 Carlo Vittorio Cannistraci,19,20 Ryan Cardenas,21 Melissa Cardon,1 Howard Chang,22 Josée Dostie,23 Luca Ducoli,24 Alexander Favorov,25,26 Alexandre Fort,2 Diego Garrido,15 Noa Gil,27 Juliette Gimenez,28 Reto Guler,16,17 Lusy Handoko,2 Jayson Harshbarger,2 Akira Hasegawa,1,2 Yuki Hasegawa,2 Kosuke Hashimoto,1,2 Norihito Hayatsu,1 Peter Heutink,29 Tetsuro Hirose,30 Eddie L. Imada,26 Masayoshi Itoh,2,31 Bogumil Kaczkowski,1,2 Aditi Kanhere,21 Emily Kawabata,2 Hideya Kawaji,31 Tsugumi Kawashima,1,2 S. 30:1060–1072 Published by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/20; www.genome.org FANTOM6 pilot study 1RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa 230-0045, Japan; 2RIKEN Center for Life Science Technologies, Yokohama, Kanagawa 230-0045, Japan; 3Laboratorio Nazionale Consorzio Interuniversitario Biotecnologie (CIB), Trieste 34127, Italy; 4Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, Moscow 119991, Russia; 5Institute of Protein Research, Russian Academy of Sciences, Pushchino 142290, Russia; 6Department of Computer Science, University of Toronto, Toronto, Ontario M5S 1A1, Canada; 7Program in Cardiovascular and Metabolic Disorders, Duke-National University of Singapore Medical School, Singapore 169857, Singapore; 8Roslin Institute, University of Edinburgh, Edinburgh EH25 9RG, United Kingdom; 9Center for RNA Medicine, Department of Clinical Medicine, Aalborg University, Copenhagen 9220, Denmark; 10Institute of Clinical Sciences, Faculty of Medicine, Imperial College London, London W12 0NN, United Kingdom; 11Computational Regulatory Genomics, MRC London Institute of Medical Sciences, London W12 0NN, United Kingdom; 12Berlin Institute for Medical Systems Biology, Max Delbrük Center for Molecular Medicine in the Helmholtz Association, Berlin 13125, Germany; 13Institute of Bioengineering, Research Center of Biotechnology, Russian Academy of Sciences, Moscow 117312, Russia; 14Graduate School of Integrated Sciences for Life, Hiroshima University, Higashi-Hiroshima City 739-0046, Japan; 15Centre for Genomic Regulation (CRG), The Barcelona Institute of Science and Technology, Barcelona, Catalonia 08003, Spain; 16International Centre for Genetic Engineering and Biotechnology (ICGEB), University of Cape Town, Cape Town 7925, South Africa; 17Institute of Infectious Diseases and Molecular Medicine (IDM), Department of Pathology, Division of Immunology and South African Medical Research Council (SAMRC) Immunology of Infectious Diseases, Faculty of Health Sciences, University of Cape Town, Cape Town 7925, South Africa; 18School of Computer Science, McGill University, Montréal, Québec H3G 1Y6, Canada; 19Biomedical Cybernetics Group, Biotechnology Center (BIOTEC), Center for Molecular and Cellular Bioengineering (CMCB), Center for Systems Biology Dresden (CSBD), Cluster of Excellence Physics of Life (PoL), Department of Physics, Technische Universität Dresden, Dresden 01062, Germany; 20Center for Complex Network Intelligence (CCNI) at the Tsinghua Laboratory of Brain and Intelligence (THBI), Department of Bioengineering, Tsinghua University, Beijing 100084, China; 21Institute of Cancer and Genomic Sciences, College of Medical and Dental Sciences, University of Birmingham, Birmingham B15 2TT, United Kingdom; 22Center for Personal Dynamic Regulome, Stanford University, Stanford, California 94305, USA; 23Department of Biochemistry, Rosalind and Morris Goodman Cancer Research Center, McGill University, Montréal, Québec H3G 1Y6, Canada; 24Institute of Pharmaceutical Sciences, Swiss Federal Institute of Technology, Zurich 8093, Switzerland; 25Department of Computational Systems Biology, Vavilov Institute of General Genetics, Russian Academy of Sciences, Moscow 119991, Russia; 26Department of Oncology, Johns Hopkins University, Baltimore, Maryland 21287, USA; 27Department of Biological Regulation, Weizmann Institute of Science, Rehovot 76100, Israel; 28Epigenetics and Genome Reprogramming Laboratory, IRCCS Fondazione Santa Lucia, Rome 00179, Italy; 29Genome Biology of Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Tübingen 72076, Germany; 30Graduate School of Frontier Biosciences, Osaka University, Suita 565-0871, Japan; 31RIKEN Preventive Medicine and Diagnosis Innovation Program (PMI), Saitama 351-0198, Japan; 32Institute of Infection, Immunity, and Inflammation, University of Glasgow, Glasgow, Scotland G12 8QQ, United Kingdom; 33Department of Biosciences and Nutrition, Karolinska Institutet, Huddinge 14157, Sweden; 34Moscow Institute of Physics and Technology, Dolgoprudny 141701, Russia; 35Biological and Environmental Sciences and Engineering Division, King Abdullah University of Science and Technology, Thuwal 23955-6900, Kingdom of Saudi Arabia; 36Department of Biology and BRIC, University of Copenhagen, Denmark, Copenhagen N DK2200, Denmark; 37MRC Human Genetics Unit, University of Edinburgh, Edinburgh EH4 2XU, United Kingdom; 38National Centre for Cell Science, Pune, Maharashtra 411007, India; 39Centre for Global Health Research, Usher Institute, University of Edinburgh, Edinburgh EH8 9AG, United Kingdom; 40Harry Perkins Institute of Medical Research, QEII Medical Centre and Centre for Medical Research, The University of Western Australia, Nedlands, Perth, Western Australia 6009, Australia; 41Universitat Pompeu Fabra (UPF), Barcelona, Catalonia 08002, Spain; 42Princess Margaret Cancer Centre, Toronto, Ontario M5G 1L7, Canada; 43Stem Cells and Metabolism Research Program, University of Helsinki and Folkhälsan Research Center, 00290 Helsinki, Finland; 44Sars International Centre for Marine Molecular Biology, University of Bergen, Bergen N-5008, Norway; 45Department of Medicine and Consorzio Interuniversitario Biotecnologie p.zle Kolbe 1 University of Udine, Udine 33100, Italy; 46Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06510, USA Long noncoding RNAs (lncRNAs) constitute the majority of transcripts in the mammalian genomes, and yet, their func- tions remain largely unknown. Functional annotation of human long noncoding RNAs via molecular phenotyping Thomas Kelly,1 Miki Kojima,1,2 Naoto Kondo,2 Haruhiko Koseki,1 Tsukasa Kouno,1,2 Anton Kratz,2 Mariola Kurowska-Stolarska,32 Andrew Tae Jun Kwon,1,2 Jeffrey Leek,26 Andreas Lennartsson,33 Marina Lizio,1,2 Fernando López-Redondo,1,2 Joachim Luginbühl,1,2 Shiori Maeda,1 Vsevolod J. Makeev,25,34 Luigi Marchionni,26 Yulia A. Medvedeva,13,34 Aki Minoda,1,2 Ferenc Müller,21 Manuel Muñoz-Aguirre,15 Mitsuyoshi Murata,1,2 Hiromi Nishiyori,1,2 Kazuhiro R. Nitta,1,2 Shuhei Noguchi,1,2 Yukihiko Noro,2 Ramil Nurtdinov,15 Yasushi Okazaki,1,2 Valerio Orlando,35 Denis Paquette,23 Callum J.C. Parr,1 Owen J.L. Rackham,7 Patrizia Rizzu,29 Diego Fernando Sánchez Martinez,26 Albin Sandelin,36 Pillay Sanjana,21 Colin A.M. Semple,37 Youtaro Shibayama,1,2 Divya M. Sivaraman,1,2 Takahiro Suzuki,1,2 Suzannah C. Szumowski,2 Michihira Tagami,1,2 Martin S. Taylor,37 Chikashi Terao,1 Malte Thodberg,36 Supat Thongjuea,2 Vidisha Tripathi,38 Igor Ulitsky,27 Roberto Verardo,3 Ilya E. Vorontsov,25 Chinatsu Yamamoto,2 Robert S. Young,39 J. Kenneth Baillie,8 Alistair R.R. Forrest,1,2,40 Roderic Guigó,15,41 Michael M. Hoffman,42 Chung Chau Hon,1,2 Takeya Kasukawa,1,2 Sakari Kauppinen,9 Juha Kere,33,43 Boris Lenhard,10,11,44 Claudio Schneider,3,45 Harukazu Suzuki,1,2 Ken Yagi,1,2 Michiel J.L. de Hoon,1,2 Jay W. Shin,1,2 and Piero Carninci1,2 60 Genome Research www.genome.org 30:1060–1072 Published by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/20; www.genome.org 30:1060–1072 Published by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/20; www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from FANTOM6 pilot study [Supplemental material is available for this article.] Ramilowski et al. Ramilowski et al. Over 50,000 loci in the human genome transcribe long noncoding RNAs (lncRNAs) (Iyer et al. 2015; Hon et al. 2017), which are de- fined as transcripts at least 200 nucleotides (nt) long with low or no protein-coding potential. Although lncRNA genes outnumber protein-coding genes in mammalian genomes, they are compara- tively less conserved (Ulitsky 2016), lowly expressed, and more cell-type-specific (Hon et al. 2017). However, the evolutionary conservation of lncRNA promoters (Carninci et al. 2005) and the structural motifs of lncRNAs (Chu et al. 2015; Xue et al. 2016) suggest that lncRNAs are fundamental biological regulators. To date, only a few hundred human lncRNAs have been extensively characterized (de Hoon et al. 2015; Quek et al. 2015; Volders et al. 2015; Ma et al. 2019), revealing their roles in regulating tran- scription (Engreitz et al. 2016b), translation (Carrieri et al. 2012), and chromatin state (Gupta et al. 2010; Guttman et al. 2011; Guttman and Rinn 2012; Quinn and Chang 2016; Ransohoff et al. 2018). Table S1; Fig. 1A–C). Using RNA-seq profiling of fractionated RNA, we annotated the lncRNA subcellular localization as the chromatin-bound (35%), nucleus-soluble (27%), or cytoplasmic (38%) (Fig. 1D). We then designed a minimum of five non-over- lapping antisense oligonucleotides against each lncRNA (Supple- mental Methods; Supplemental Table S2; Fig. 1E,F) and transfected them individually using an automated cell culture platform to minimize experimental variability (Fig. 1G). The overall knockdown efficiencies across 2021 ASOs resulted in me- dian value of 45.4%, and we could successfully knockdown 879 out of 2021 (43.5%) ASOs (>40% knockdown efficiency in at least two primer pairs or >60% in one primer pair) (Supplemental Ta- ble S2). ASOs targeting exons or introns were equally effective, and knockdown efficiencies were independent of the genomic class, expression level, and subcellular localization of the lncRNA (Supplemental Fig. S1A–D). Our recent FANTOM5 computational analysis showed that 19,175 (out of 27,919) human lncRNA loci are functionally impli- cated (Hon et al. 2017). Yet, genomic screens are necessary to com- prehensively characterize each lncRNA. One common approach of gene knockdown followed by a cellular phenotype assay typically characterizes a small percentage of lncRNAs for a single observable phenotype. For example, a recent large-scale screening using CRISPR interference (CRISPRi) found that ∼3.7% of targeted lncRNA loci are essential for cell growth or viability in a cell-type- specific manner (Liu et al. 2017). A subset of lncRNAs are associated with cell growth and morphology changes To evaluate the effect of each lncRNA knockdown on cell growth and morphology, we imaged ASO-transfected HDFs in duplicate every 3 h for a total of 48 h (Supplemental Table S3) and estimat- ed their growth rate based on cell confluence measurements (Fig. 2A,B). First, we observed across all ASOs that changes in cell growth and morphological parameters were significantly correlat- ed with knockdown efficiency (Supplemental Fig. S1E). Consider- ing both successful knockdown and significant growth inhibition (Student’s two-sided t-test FDR ≤0.05), 246 out of 879 ASOs (∼28%) showed cellular phenotype (Fig. 2C; Supplemental Table S3). To assess globally whether the observed growth inhibition is lncRNA-specific, we used all 194 lncRNAs successfully targeted by at least two ASOs (Supplemental Fig. S2A) and found that ASOs tar- geting the same lncRNA were significantly more likely to have a concordant growth response than ASOs targeting different lncRNA (empirical P = 0.00037) (Supplemental Methods; Supple- mental Fig. S2B). However, different ASOs targeting the same lncRNA typically showed different effects on growth, possibly due to variable knockdown efficiencies or differences in targeted lncRNA isoforms, as well as off-target effects. To reliably identify target-specific cellular phenotype, we applied conditional cutoffs based on the number of successful ASOs per each lncRNA (Supple- mental Methods; Supplemental Fig. S2C) and identified 15/194 lncRNAs (7.7%) with growth phenotype (adjusted background <5%) (Supplemental Fig. S2D). We validated A1BG-AS1, which was previously implicated in cell growth (Bai et al. 2019), CATG00000089639, RP11-195F19.9, and ZNF213-AS1 by measur- ing the MKI67 proliferation protein marker upon knockdown with siRNAs and with selected ASOs (Fig. 2D; Supplemental Fig. S2E). y g p y As a part of the FANTOM6 pilot project, we established an au- tomated high-throughput cell culture platform to suppress 285 lncRNAs expressed in human primary dermal fibroblasts (HDFs) using antisense LNA-modified GapmeR antisense oligonucleotide (ASO) technology (Roux et al. 2017). We then quantified the effect of each knockdown on cell growth and morphology using real- time imaging, followed by Cap Analysis Gene Expression (CAGE) (Murata et al. 2014) deep sequencing to reveal molecular pathways associated with each lncRNA. In contrast to cellular phenotyping, molecular phenotyping provides a detailed assessment of the re- sponse to a lncRNA knockdown at the molecular level, allowing bi- ological pathways to be associated to lncRNAs even in the absence of an observable cellular phenotype. Ramilowski et al. In addition, CRISPR-Cas9 experi- ments targeting splice sites identified ∼2.1% of lncRNAs that affect growth of K562 (Liu et al. 2018), and a CRISPR activation study re- vealed ∼0.11% lncRNAs to be important for drug resistance in mel- anoma (Joung et al. 2017). However, many of these studies target the genomic DNA, potentially perturbing the chromatin architec- ture, or focus on a single cellular assay, possibly missing other rele- vant functions and underlying molecular pathways. A subset of lncRNAs are associated with cell growth and morphology changes All data and analysis results are publicly available (see Data access), and results can be interac- tively explored using our in-house portal (https://fantom.gsc.riken .jp/zenbu/reports/#FANTOM6). In addition to cell growth, we also explored changes in cell morphology (Fig. 2E). Using a machine learning-assisted workflow (Methods), each cell was segmented and its morphological features representing various aspects of cell shapes and sizes were quanti- fied (Fig. 2F; Supplemental Table S3; Carpenter et al. 2006). As an example, knockdown of 14/194 lncRNAs (7.2%) affected the spindle-like morphology of fibroblasts, as indicated by a consistent decrease in their observed eccentricity without reducing the cell number, suggesting possible cellular transformation toward epi- thelial-like states. Collectively, we observed 59/194 lncRNAs (∼30%) affecting cell growth and/or morphological parameters (Fig. 2G; Supplemental Table S3). Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from 1062 Genome Research www.genome.org FANTOM6 pilot study As part of the FANTOM6 project, we systematically knocked down the expression of 285 lncRNAs in human dermal fibroblasts and quantified cellular growth, morphological changes, and transcriptomic re- sponses using Capped Analysis of Gene Expression (CAGE). Antisense oligonucleotides targeting the same lncRNAs exhib- ited global concordance, and the molecular phenotype, measured by CAGE, recapitulated the observed cellular phenotypes while providing additional insights on the affected genes and pathways. Here, we disseminate the largest-to- date lncRNA knockdown data set with molecular phenotyping (over 1000 CAGE deep-sequencing libraries) for further exploration and highlight functional roles for ZNF213-AS1 and lnc-KHDC3L-2. [Supplemental material is available for this article.] Genome Research www.genome.org 1061 Selection and ASO-mediated knockdown of lncRNA targets Selection and ASO-mediated knockdown of lncRNA targets Human dermal fibroblasts are nontransformed primary cells that are commonly used for investigating cellular reprogramming (Takahashi et al. 2007; Ambasudhan et al. 2011), wound healing (Li and Wang 2011), fibrosis (Kendall and Feghali-Bostwick 2014), and cancer (Kalluri 2016). Here, an unbiased selection of lncRNAs expressed in HDFs was performed to choose 285 lncRNAs for functional interrogation (Methods; Supplemental 1062 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from FANTOM6 pilot study E F B A C D G Figure 1. Selection of lncRNA targets, theirproperties, and the studyoverview.(A) CAGEexpression levelsat log2TPM (tagsper million) and human dermal fibroblasts (HDFs) specificity of lncRNAs in the FANTOM CAT catalog (Hon et al. 2017) (N = 62,873; gray), lncRNAs expressed in HDFs (N = 6125; blue), and targeted lncRNAs (N = 285; red). The dashed vertical line indicates most lowly expressed lncRNA target (∼0.2 TPM). (B) Gene conservation levels of lncRNAs in the FANTOM CAT catalog (gray), lncRNAs expressed in HDFs (blue), and targeted lncRNAs (red). Crossbars indicate the median. No significant difference is observed when comparing targeted and expressed in HDF lncRNAs (Wilcoxon P = 0.11). (C) Similar to that in B but for genomic classes of lncRNAs. Most of the targeted lncRNAs and those expressed in HDFs are expressed from divergent promoters. (D) Subcellular localization (based on relative abundances from RNA-seq fractionation data) for targeted lncRNAs. Chromatin-bound (N = 98; blue); nuclear soluble (N = 76; green); cytoplasmic (N = 108; red). Black con- tours represent the distribution of all lncRNAs expressed in HDFs. (E) Example of ZNF213-AS1 loci showing transcript model, CAGE, and RNA-seq signal along with targeting ASOs. (F) Number of ASOs for target lncRNAs and controls used in the experiment. (G) Schematics of the study. B A C A B D E G D E F G F Figure 1. Selection of lncRNA targets, theirproperties, and the studyoverview.(A) CAGEexpression levelsat log2TPM (tagsper million) and human dermal fibroblasts (HDFs) specificity of lncRNAs in the FANTOM CAT catalog (Hon et al. 2017) (N = 62,873; gray), lncRNAs expressed in HDFs (N = 6125; blue), and targeted lncRNAs (N = 285; red). The dashed vertical line indicates most lowly expressed lncRNA target (∼0.2 TPM). Selection and ASO-mediated knockdown of lncRNA targets (B) Gene conservation levels of lncRNAs in the FANTOM CAT catalog (gray), lncRNAs expressed in HDFs (blue), and targeted lncRNAs (red). Crossbars indicate the median. No significant difference is observed when comparing targeted and expressed in HDF lncRNAs (Wilcoxon P = 0.11). (C) Similar to that in B but for genomic classes of lncRNAs. Most of the targeted lncRNAs and those expressed in HDFs are expressed from divergent promoters. (D) Subcellular localization (based on relative abundances from RNA-seq fractionation data) for targeted lncRNAs. Chromatin-bound (N = 98; blue); nuclear soluble (N = 76; green); cytoplasmic (N = 108; red). Black con- tours represent the distribution of all lncRNAs expressed in HDFs. (E) Example of ZNF213-AS1 loci showing transcript model, CAGE, and RNA-seq signal along with targeting ASOs. (F) Number of ASOs for target lncRNAs and controls used in the experiment. (G) Schematics of the study. Molecular phenotyping by CAGE recapitulates cellular phenotypes and highlights functions of lncRNAs (Supplemental Fig. S3A), with ∼57% of ASOs showing at least 10 dif- ferentially expressed genes (FDR≤0.05; abs[log2FC] > 0.5). For 84 divergent-antisense lncRNAs (targeted by 186 independent ASOs) (Supplemental Methods), we found their partner gene to be general- ly unchanged (median abs[log2FC] = ∼0.13), with an exception of two significantly down-regulated and three significantly up-regulat- ed genes (FDR≤0.05) (Supplemental Fig. S3B). We have, however, noticed a common response in a large number of ASOs (∼30%– 35% of all responding ASOs), such as down-regulation of cell- cycle-related pathways, up-regulated stress genes and pathways, or altered cell metabolism and energetics (Supplemental Fig. S3C,D). Next, we selected 340 ASOs with high knockdown efficiencies (mostly >50%; median 71.4%) and sequenced 970 CAGE libraries to analyze 154 lncRNAs (Fig. 3A; Supplemental Table S4). To assess functional implications by individual ASOs, we performed differ- ential gene expression, Motif Activity Response Analysis (MARA) (The FANTOM Consortium et al. 2009), and Gene Set Enrichment Analysis (GSEA) (Fig. 3B–F; Subramanian et al. 2005), and com- pared them with cellular phenotype. When comparing knockdown-mediated molecular and cellu- lar response, we found that transcription factor motifs that We globally observed significant knockdown-mediated tran- scriptomic changes (which generally correlated with KD efficiency) 1063 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Ramilowski et al. e cell growth, including TFDP1, E2F1,2,3, and EP300, were S6). Moreover, correlations of growth with GSEA pathways ( E F B A C D G 2. Cell growth and morphology assessment. (A) Selected example (PTPRG1-AS1) showing the normalized growth rate estimation using a ma A (negative control). (B) Correlation of the normalized growth rate for technical duplicates across 2456 Incucyte samples. (C) Density distribu alized growth rates (technical replicates averaged) 252 ASOs targeting lncRNAs with successful knockdown (KD) and growth phenotype (b nt in two replicates (FDR < 0.05 as compared to matching NC_A; 246 ASOs inhibited growth), 627 ASOs targeting lncRNAs with successfu , 270 negative control (NC_A) samples (gray), and 90 mock-transfected cells (Lipofectamine only) samples (yellow). (D) MKI67 staining (gro n validation) for four selected lncRNA targets after siRNA and ASOs suppression. (E) Incucyte cell images of selected distinct cell morpholo upon an lncRNA KD. (F) An overview of the cell morphology imaging processing pipeline using a novel lncRNA target, CATG000089639.1 mple. S6). Moreover, correlations of growth with GSEA pathways (Fig. 3F; Supplemental Fig. S4B; Supplemental Table S6) or with FANTOM5 coexpression clusters (Supplemental Fig. S4C) showed that cell growth and replication-related pathways were positively correlated with the measured growth rate, whereas those related S6). Moreover, correlations of growth with GSEA pathways (Fig. 3F; Supplemental Fig. S4B; Supplemental Table S6) or with FANTOM5 coexpression clusters (Supplemental Fig. S4C) showed that cell growth and replication-related pathways were positively correlated with the measured growth rate, whereas those related FANTOM6 pilot study FANTOM6 pilot study 1 1,000 2,000 3,000 4,000 Pathways, ranked FDR-adjusted p-value for Spearman’s correlation Pos. correlation Neg. correlation HALLMARK GO GROWTH GO POS. REGULAT. HALLMARK INTERFERON GAMMA RESPONSE HALLMARK APOPTOSIS GO CHROMATIN GO DNA CONFORMATION CHANGE GO CHROMATIN REMODELING GO DNA PACKAGING GO MITOTIC CYTOKINESIS REACTOME CELL CYCLE Pathway keywords: Cell cycle, growth, replication, mitosis, G0/1/2/S/M phase, DNA packaging, chromatin, nucleosome, centrosome Aptoptosis, TP53 Immunity, defense, leukocyte, cytokine, interferon, tumor necrosis Other OF STAT CASCADE FACTOR ACTIVITY TP53 ACTIVITY A C E B D F 0.00 0.25 0.50 0.75 1.00 0 25 50 75 100 Knockdown efficiency, % Cumulative density Transfected (other) (1,646 ASOs) Selected for CAGE (375 ASOs) 1 20 40 60 80 100 Motifs, ranked FDR-adjusted p-value for Spearman’s correlation Pos. correlation Neg. correlation PPARG SREBF1,2 STAT2,4,6 FOS; FOS B,L1; JUN B,D RXRA; VDR dimer EP300 MYB ZBTB14 E2F1..5 TFDP1 -4 −2 0 2 4 -log10(p) × sign(NES). 104 103 102 101 0 101 102 103 104 1 50 100 150 200 250 300 340 ASOs, ranked Differentially expressed genes, count Motifs Z-score on relative motif actvity 340 ASOs STAT2,4,6 TP53 PPARG SREBF1,2 JUN E2F1..5 MYB RXRA_VDR{dimer} ZBTB14 EGR1..3 TFDP1 SMAD1..7,9 IRF1,2,7 TGIF1 EP300 −2 0 2 4 −4 10-16 10-12 10-8 10-4 1 10-4 10-8 10-12 10-16 10-10 10-6 10-2 1 10-2 10-6 10-10 10-8 10-4 10-8 10-4 Up-regulated Down-regulated Figure 3. CAGE predicts cellular phenotypes. (A) RT-qPCR knockdown efficiency for 2021 ASO-tran fected samples (targeted lncRNAs only). Gray dashed line indicates 50% KD efficiency generally require for CAGE selection. Purple dashed lines indicate median KD efficiency (71.5%) for 375 ASOs selected fo CAGE sequencing. After quality control, 340 ASOs targeting lncRNAs were included for further analys (B) Distribution of significantly differentially expressed genes (up-regulated: FDR < 0.05, Z-score > 1.64 log2FC > 0.5; and down-regulated: FDR < 0.05, Z-score < −1.645, log2FC < −0.5) across all 340 ASOs. (C Motif Response Activity Analysis (MARA) across 340 ASOs. Scale indicates Z-score of the relative motif a tivity (the range was set to abs[Z-score] = <5 for visualization purposes). (D) Correlation between norma ized growth rate and motif activities across 340 ASOs targeting lncRNAs with highlighted example Motif sizes shown are scaled based on the HDF expression of their associated TFs (range 1 to ∼60 TPM). (E) Enriched biological pathways across 340 ASOs. Scale indicates GSEA enrichment value calc lated as −log10(p) × sign(NES). FANTOM6 pilot study (F) Same as in D but for selected GSEA pathways. Pathways sizes are scale based on the number of associated genes. A 0.00 0.25 0.50 0.75 1.00 0 25 50 75 100 Knockdown efficiency, % Cumulative density Transfected (other) (1,646 ASOs) Selected for CAGE (375 ASOs) B 104 103 102 101 0 101 102 103 104 1 50 100 150 200 250 300 340 ASOs, ranked Differentially expressed genes, count Up-regulated Down-regulated A B Cell radius and axis length were associat- ed with GSEA categories related to actin arrangement and cilia, whereas cell com- pactness was negatively correlated with apoptosis. The extensive molecular phe- notyping analysis also revealed pathways not explicitly associated with cell growth and cell morphology, such as transcrip- tion, translation, metabolism, develop- ment, and signaling (Fig. 3E). Next, to globally assess whether in- dividual ASO knockdowns lead to lncRNA-specific effects, we scaled the ex- pression change of each gene across the whole experiment and compared differ- entially expressed genes (Fig. 3B) of all possible ASO pairs targeting the same lncRNA target versus different lncRNAs (Supplemental Methods; Supplemental Table S5). We found that the concor- dance of the same target group was sig- nificantly greater than that of the different target group (comparing the Jaccard indices across 10,000 permuta- tions) (Supplemental Fig. S5A), suggest- ing that ASO knockdowns are nonrandom and lead to more lncRNA specific effects than the nontargeting ASO pairs. Further, by requiring at least five common DEGs (FDR ≤0.05, abs [log2FC] > 0.5, abs[Z-score] > 1.645) and ASO-pairs significantly above the non- targeting ASO pairs background (P ≤ 0.05), we identified 16 ASO pairs, target- ing 13 lncRNAs, exhibiting reproducible knockdown-mediated molecular re- sponses in human dermal fibroblasts (Supplemental Fig. S5B). Corresponding GSEA pathways and MARA motifs of these 16 ASO pairs are shown in Supplemental Figure S5C. D 1 20 40 60 80 100 Motifs, ranked FDR-adjusted p-value for Spearman’s correlation Pos. correlation Neg. correlation PPARG SREBF1,2 STAT2,4,6 FOS; FOS B,L1; JUN B,D RXRA; VDR dimer EP300 MYB ZBTB14 E2F1..5 TFDP1 10-10 10-6 10-2 1 10-2 10-6 10-10 10-8 10-4 10-8 10-4 C Motifs Z-score on relative motif actvity 340 ASOs STAT2,4,6 TP53 PPARG SREBF1,2 JUN E2F1..5 MYB RXRA_VDR{dimer} ZBTB14 EGR1..3 TFDP1 SMAD1..7,9 IRF1,2,7 TGIF1 EP300 −2 0 2 4 −4 D C 1 1,000 2,000 3,000 4,000 Pathways, ranked FDR-adjusted p-value for Spearman’s correlation Pos. correlation Neg. correlation HALLMARK GO GROWTH GO POS. FANTOM6 pilot study REGULAT. HALLMARK INTERFERON GAMMA RESPONSE HALLMARK APOPTOSIS GO CHROMATIN GO DNA CONFORMATION CHANGE GO CHROMATIN REMODELING GO DNA PACKAGING GO MITOTIC CYTOKINESIS REACTOME CELL CYCLE Pathway keywords: Cell cycle, growth, replication, mitosis, G0/1/2/S/M phase, DNA packaging, chromatin, nucleosome, centrosome Aptoptosis, TP53 Immunity, defense, leukocyte, cytokine, interferon, tumor necrosis Other OF STAT CASCADE FACTOR ACTIVITY TP53 ACTIVITY F 10-16 10-12 10-8 10-4 1 10-4 10-8 10-12 10-16 E -4 −2 0 2 4 -log10(p) × sign(NES). F E −2 0 2 -log10(p) × sign(NES). Figure 3. CAGE predicts cellular phenotypes. (A) RT-qPCR knockdown efficiency for 2021 ASO-trans- fected samples (targeted lncRNAs only). Gray dashed line indicates 50% KD efficiency generally required for CAGE selection. Purple dashed lines indicate median KD efficiency (71.5%) for 375 ASOs selected for CAGE sequencing. After quality control, 340 ASOs targeting lncRNAs were included for further analysis. (B) Distribution of significantly differentially expressed genes (up-regulated: FDR < 0.05, Z-score > 1.645, log2FC > 0.5; and down-regulated: FDR < 0.05, Z-score < −1.645, log2FC < −0.5) across all 340 ASOs. (C) Motif Response Activity Analysis (MARA) across 340 ASOs. Scale indicates Z-score of the relative motif ac- tivity (the range was set to abs[Z-score] = <5 for visualization purposes). (D) Correlation between normal- ized growth rate and motif activities across 340 ASOs targeting lncRNAs with highlighted examples. Motif sizes shown are scaled based on the HDF expression of their associated TFs (range 1 to ∼600 TPM). (E) Enriched biological pathways across 340 ASOs. Scale indicates GSEA enrichment value calcu- lated as −log10(p) × sign(NES). (F) Same as in D but for selected GSEA pathways. Pathways sizes are scaled based on the number of associated genes. Molecular phenotyping by CAGE recapitulates cellular phenotypes and highlights functions of lncRNAs (G) lncRNAs (N = 59) significantly (FDR < 0.05) and consistently (after adjusting for the number of successfully targeting ASOs) affec wth (N = 15) and cell morphologies (N = 44). owski et al. B A C B A B B A C C B D D D D G G E F E F E F Figure 2. Cell growth and morphology assessment. (A) Selected example (PTPRG1-AS1) showing the normalized growth rate estimation using a match- ing NC_A (negative control). (B) Correlation of the normalized growth rate for technical duplicates across 2456 Incucyte samples. (C) Density distribution of normalized growth rates (technical replicates averaged) 252 ASOs targeting lncRNAs with successful knockdown (KD) and growth phenotype (blue) consistent in two replicates (FDR < 0.05 as compared to matching NC_A; 246 ASOs inhibited growth), 627 ASOs targeting lncRNAs with successful KD (purple), 270 negative control (NC_A) samples (gray), and 90 mock-transfected cells (Lipofectamine only) samples (yellow). (D) MKI67 staining (growth inhibition validation) for four selected lncRNA targets after siRNA and ASOs suppression. (E) Incucyte cell images of selected distinct cell morphologies changes upon an lncRNA KD. (F) An overview of the cell morphology imaging processing pipeline using a novel lncRNA target, CATG000089639.1, as an example. (G) lncRNAs (N = 59) significantly (FDR < 0.05) and consistently (after adjusting for the number of successfully targeting ASOs) affecting cell growth (N = 15) and cell morphologies (N = 44). promote cell growth, including TFDP1, E2F1,2,3, and EP300, were positively correlated with the measured cell growth rate, whereas transcription factor motifs known to inhibit growth or induce apoptosis (e.g., PPARG, SREBPF, and STAT2,4,6) were negatively correlated (Fig. 3D; Supplemental Fig. S4A; Supplemental Table 1064 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from siRNA validation experiments To evaluate whether the lncRNA-specific effects can be measured by other knock- down technologies, nine lncRNAs, with relatively mild growth phenotype, were subjected to siRNA knockdown. Measur- ing transcriptional response, we noted that higher concordance was observed for ASO modality alone (Supplemental Fig. S5D). The observed discrepancies in the transcriptional response between ASO- and siRNA-mediated knockdowns could be contributed by their mode of action and var- iable activities in different subcellular compartments. Next, a con- cordant response was found for (5/36) ASO-siRNA pairs targeting three lncRNAs (Supplemental Fig. S5E; Supplemental Table S5), enriched in the cytoplasm (MAPKAPK5-AS1), soluble nuclear fraction (LINC02454), and in the chromatin-bound fraction (A1BG-AS1). Although we cannot completely exclude the techni- cal artifacts of each technology, concordant cellular response R < 0.05, Z score > 1.645, 5) across all 340 ASOs. (C) re of the relative motif ac- relation between normal- th highlighted examples. ed TFs (range 1 to ∼600 A enrichment value calcu- . Pathways sizes are scaled To evaluate whether the lncRNA-specific effects can be measured by other knock- down technologies, nine lncRNAs, with relatively mild growth phenotype, were subjected to siRNA knockdown. Measur- ing transcriptional response, we noted that higher concordance was observed for ASO modality alone (Supplemental Fig. S5D). The observed discrepancies in to immunity, and cell stress and cell death were negatively correlat- ed. We found that among 53 ASOs implicated in a growth-inhibi- tion pathway based on the CAGE profiles, only 43% of them showed growth inhibition in the real-time imaging. This might suggest better sensitivity of transcriptomic profiling when detect- ing phenotypes as compared to live cell imaging methods, which are more prone to a delayed cellular response to the knockdown. Additionally, morphological changes were reflected in the Additionally, morphological changes were reflected in the molecular phenotype assessed by CAGE (Supplemental Fig. S4D). y, p g g molecular phenotype assessed by CAGE (Supplemental Fig. S4D). 1065 Genome Research www.genome.org Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Ramilowski et al. E B A C D gure 4. ZNF213-AS1 regulates cell growth, migration, and proliferation. (A) Normalized growth rate across four distinct ASOs (in duplicate) targeting NF213-AS1 as compared to six negative control samples (shown in gray). siRNA validation experiments (B) Enrichment of biological pathways associated with growth, proliferation, ound healing, migration, and adhesion for ASO_02 and ASO_05. (C) Most consistently down- and up-regulated transcription factor binding motifs in- uding those for transcription factors known to modulate growth, migration, and proliferation such as for example EGR family, EP300, GTF2I. D) Knockdown efficiency measured by RT-qPCR after wound closure assay (72 h posttransfection) showing sustained suppression (65%–90%) of NF213-AS1. (E) Transfected, replated, and mitomycin C (5 µg/mL)-treated HDF cells were scratched and monitored in the Incucyte imaging system. elative wound closure rate calculated during the 24 h postscratching shows 40%–45% reduction for the two targeting ASOs (ASO_02 [N = 10] and SO_05 [N = 13]) as compared to NC_A transfection controls (N = 33, shown in gray) and the representative images of wound closure assay 16 h ostscratching B A B B C A D E D E Figure 4. ZNF213-AS1 regulates cell growth, migration, and proliferation. (A) Normalized growth rate across four distinct ASOs (in duplicate) targeting ZNF213-AS1 as compared to six negative control samples (shown in gray). (B) Enrichment of biological pathways associated with growth, proliferation, wound healing, migration, and adhesion for ASO_02 and ASO_05. (C) Most consistently down- and up-regulated transcription factor binding motifs in- cluding those for transcription factors known to modulate growth, migration, and proliferation such as for example EGR family, EP300, GTF2I. (D) Knockdown efficiency measured by RT-qPCR after wound closure assay (72 h posttransfection) showing sustained suppression (65%–90%) of ZNF213-AS1. (E) Transfected, replated, and mitomycin C (5 µg/mL)-treated HDF cells were scratched and monitored in the Incucyte imaging system. Relative wound closure rate calculated during the 24 h postscratching shows 40%–45% reduction for the two targeting ASOs (ASO_02 [N = 10] and ASO_05 [N = 13]) as compared to NC_A transfection controls (N = 33, shown in gray) and the representative images of wound closure assay 16 h postscratching. exhibited by using ASOs alone suggests that lncRNAs, in part, are essential regulatory elements in cells. Yet, our study generally war- rants a careful assessment of specific findings from different knockdown technologies, including CRISPR-inhibition, and dem- onstrates a requirement of using multiple replicates in a given tar- get per each modality. S7), the molecular phenotype revealed suppression of GSEA path- ways related to cell growth, as well as to cell proliferation, motility, and extracellular structure organization (Fig. 4B). siRNA validation experiments We also observed consistent down-regulation of motifs related to the observed cellu- lar phenotype, for example, EGR1, EP300, SMAD1…7,9 (Fig. 4C). As cell motility pathways were affected by the knockdown, we tested whether ZNF213-AS1 could influence cell migration. Based on the wound-closure assay after transient cell growth inhibition (mitomycin C and serum starvation) (Supplemental Fig. S2F,G), we observed a substantial reduction of wound closure rate (∼40% over a 24-h period) in the ZNF213-AS1-depleted HDFs (Fig. 4D, E). The reduced wound healing rate should thus mainly reflect re- duced cell motility, further confirming affected motility pathways predicted by the molecular phenotype. 1066 Genome Research www.genome.org RP11-398K22.12 (KHDC3L-2) regulates KCNQ5 in cis Next, we investigated in detail RP11-398K22.12 (ENSG00000229 852), where the knockdowns by two independent ASOs (ASO_03, ASO_05) successfully reduced the expression of the tar- get lncRNA (67%–82% knockdown efficiency, respectively) and further down-regulated its neighboring genes, KCNQ5 and its divergent partner novel lncRNA CATG00000088862.1 (Fig. 5A). Although the two genomic loci occupy Chromosome 6 and are 650 kb away, Hi-C analysis (Supplemental Methods; Supple- mental Fig. S7; Supplemental Table S8) showed that they are locat- ed within the same topologically associated domain (TAD) and spatially colocalized (Fig. 5B). Moreover, chromatin-enrichment and single molecule RNA-FISH of RP11-398K22.12 (Fig. 5C; Supplemental Table S9) suggested its highly localized cis-regulato- ry role. We observed a reduction in cell growth for ∼7.7% of our tar- get lncRNA genes, which is in line with previous experiments us- ing CRISPRi-pooled screening, which reported 5.9% (in iPS cells) of lncRNAs exhibiting a cell growth phenotype (Liu et al. 2017). Although these rates are much lower than for protein-coding genes (Sokolova et al. 2017), recurrent observations of cell growth phe- notypes (including cell death) strongly suggest that a substantial fraction of lncRNAs play an essential role in cellular physiology and viability. Further, when applying image-based analysis, we found that lncRNAs affect cell morphologies (Fig. 2G), which has not been so far thoroughly explored. y In FANTOM5 (Hon et al. 2017), expression levels of RP11- 398K22.12, KCNQ5, and CATG00000088862.1 were enriched in brain and nervous system samples, whereas GTEx (The GTEx Consortium 2015) showed their highly specific expression in the brain, particularly in the cerebellum and the cerebellar hemisphere (Fig. 5D). GTEx data also showed that expression of RP11- 398K22.12 was highly correlated with the expression of KCNQ5 and CATG00000088862.1 across neuronal tissues (Fig. 5E,F), with the exception of cerebellum and the cerebellar hemisphere, potentially due to relatively lower levels of KCNQ5 and CATG00000088862.1, whereas levels of RP11-398K22.12 re- mained relatively higher. Additionally, we found an eQTL SNP (rs14526472) overlapping with RP11-398K22.12 and regulating ex- pression of KCNQ5 in brain caudate (P = 4.2 × 10−6; normalized ef- fect size −0.58). All these findings indicate that RP11-398K22.12 is implicated in the nervous system by maintaining the expression of KCNQ5 and CATG000 00088862.1 in a cis-acting manner. Several lncRNAs such as MALAT1, NEAT1, and FIRRE have been reported to orchestrate transcription, RNA processing, and gene expression (Kopp and Mendell 2018) but are not essential for mouse development or viability. RP11-398K22.12 (KHDC3L-2) regulates KCNQ5 in cis These observations advocate for assays that can comprehensively profile the molecular changes inside perturbed cells. Therefore, in contrast to cell-based assays, functional elucidation via molecular phenotyping provides com- prehensive information that cannot be captured by a single phe- notypic assay. Herein, the number of overlapping differentially expressed genes between two ASOs of the same lncRNA targets in- dicated that 10.9% of lncRNAs exert a reproducible regulatory function in HDF. Although the features of selected lncRNAs are generally simi- lar to those of other lncRNAs expressed in HDFs (Fig. 1B–D), the cell-type-specific nature of lncRNAs and the relatively small sam- pling size (119 lncRNAs with knockdown transcriptome profiles) used in our study may not fully represent the whole extent of lncRNA in other cell types. However, lncRNA targets that did not exhibit a molecular phenotype may be biologically relevant in oth- er cell types or cell states (Li and Chang 2014; Liu et al. 2017). At the same time, our results showed that particular lncRNAs ex- pressed broadly in other tissues (e.g., in the human brain) were functional in HDFs (such as RP11-398K22.12). Although the exact molecular mechanisms of RP11-398K22.12 are not yet fully under- stood, its potential role in HDFs suggests that lncRNAs may be functionally relevant across multiple tissues in spite of the cell- type-specific expression of lncRNAs. FANTOM6 pilot study FANTOM6 pilot study FANTOM6 pilot study in the expression response of divergent promoters (Supplemental Fig. S3B). Recent studies suggest that transcription of lncRNA loci that do not overlap with other transcription units may influence RNA polymerase II occupancy on neighboring promoters and gene bodies (Engreitz et al. 2016a; Cho et al. 2018). Thus, it is plau- sible that transcription of targeted lncRNA was maintained, de- spite suppression of mature or nascent transcripts using ASOs. This further suggests that the functional responses described in this study are due to interference of processed transcripts present either in the nucleus, the cytoplasm, or both. Although it is argu- able that ASOs may interfere with general transcription by target- ing the 5′-end of nascent transcripts and thus releasing RNA polymerase II, followed by exonuclease-mediated decay and tran- scription termination (aka “torpedo model”) (Proudfoot 2016), most of the ASOs were designed across the entire length of the transcript. Since we did not broadly observe dysregulation in near- by genes, interference of transcription or splicing activity is less likely to occur. differentiation and proliferation of tumors (Supplemental Fig. S6B–E). Further, univariate Cox proportional hazard analysis as well as Kaplan-Meier curves for LGG were significant and consis- tent with our findings (HR = 0.61, BH FDR = 0.0079). The same survival analysis on LAML showed a weak association with poor prognostic outcome, but the results were not significant (Supplemental Fig. S6F,G). ZNF213-AS1 is associated with cell growth and migration Extensive molecular and cellular phenotype data for each ASO knockdown can be explored using our portal https://fantom.gsc .riken.jp/zenbu/reports/#FANTOM6. As an example of an lncRNA associated withcell growthand morphology(Fig. 2G), weshowcase ZNF213-AS1 (RP11-473M20.14). This lncRNA is highly conserved in placental mammals, moderately expressed (∼eight CAGE tags per million) in HDFs, and enriched in the chromatin-bound frac- tion. Four distinct ASOs (ASO_01, ASO_02, ASO_05, and ASO_06) stronglysuppressedexpressionof ZNF213-AS1, whereasexpression of the ZNF213 sense gene was not significantly affected in any of the knockdowns. The four ASOs caused varying degrees of cell growth inhibition (Fig. 4A). ASO_01 and ASO_06 showed a reduc- tion in cell number, as well as an up-regulation of apoptosis and im- mune and defense pathways in GSEA, suggesting cell death. While cell growth inhibition observed for ASO_02 and ASO_05 was con- firmed by MKI67 marker staining (Fig. 2D; Supplemental Table As these results indicated a potential role of ZNF213-AS1 in cell growth and migration, we used FANTOM CAT Recount 2 at- las (Imada et al. 2020), which incorporates The Cancer Genome Atlas (TCGA) data set (Collado-Torres et al. 2017), and found rel- atively higher expression of ZNF213-AS1 in acute myeloid leuke- mia (LAML) and in low-grade gliomas (LGG) as compared to other cancers (Supplemental Fig. S6A). In LAML, the highest ex- pression levels were associated with mostly undifferentiated states, whereas in LGG, elevated expression levels were found in oligodendrogliomas, astrocytomas, and in IDH1 mutated tu- mors, suggesting that ZNF213-AS1 is involved in modulating 1066 Genome Research www.genome.org 1066 Genome Research www.genome.org 1066 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Discussion This study systematically annotates lncRNAs through molecular and cellular phenotyping by selecting 285 lncRNAs from human dermal fibroblasts across a wide spectrum of expression, conserva- tion levels and subcellular localization enrichments. Using ASO technology allowed observed phenotypes to be associated to the lncRNA transcripts, whereas, in contrast, CRISPR-based approach- es may synchronically influence the transcription machinery at the site of the divergent promoter or affect regulatory elements of the targeted DNA site. Knockdown efficiencies obtained with ASOs were observed to be independent of lncRNA expression lev- els, subcellular localization, and of their genomic annotation, al- lowing us to apply the same knockdown technology to various classes of lncRNAs. Further, we used siRNA technology to knockdown lncRNA targets as a method for independent validation. When comparing the transcriptomes perturbed by ASOs and siRNAs, concordance was observed only for three out of nine lncRNAs. This discrepancy is likely due to different modes of actions of the two technologies. Whereas ASOs invoke RNase H-mediated cleavage, primarily active We investigated the cis-regulation of nearby divergent pro- moters, which has been reported as one of the functional roles of lncRNA (Luo et al. 2016). However, in agreement with previous studies (Guttman et al. 2011), we did not observe general patterns Genome Research www.genome.org 1067 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Ramilowski et al. E F B A C D e 5. RP11-398K22.12 down-regulates KCNQ5 and CATG00000088862.1 in cis. (A) Changes in expression levels of detectable genes in topologically associated domain (TAD) as RP11-398K22.12 based on Hi-C analysis. Both KCNQ5 and CATG00000088862.1 are down- (P < 0.05) upon the knockdown of RP11-398K22.12 by two independent ASOs in CAGE analysis (left) as further confirmed with RT-qP . (B) (Top) Representation of the chromatin conformation in the 4-Mb region proximal to the TAD containing RP11-398K22.12, follow e locus gene annotation, CAGE, RNA-seq, and ATAC-seq data for native HDFs. (Bottom) Schematic diagram showing Hi-C predicted c of RP11-398K22.12 (blue) and KCNQ5 (gray) (25-kb resolution, frequency ≥5) in HDF cells. Red line indicates RP11-398K22.12 and KCN ct. (C) FISH image for RP11-398K22.12, suggesting proximal regulation. TUG1 FISH image (suggesting trans regulation) is included a arison; (bar = 10 µm). ASO transfection for real-time imaging The HDF cells were transfected manually in 96-well plates to facil- itate high-throughput real-time imaging. The cells were seeded 24 h before transfection at a density of 5200 cells per well. A final concentration of 20 nM ASO and 2 µL Lipofectamine RNAiMAX (Thermo Fisher Scientific) were mixed in 200 µL Opti-MEM (Thermo Fisher Scientific). After incubating at room temperature for 5 min, 18 µL of the transfection mix was added to 90 µL com- plete medium in each well. The ASOs were divided into 14 runs and transfected in duplicate. Each plate accommodated six wells of NC_A control, two wells of MALAT1 ASO control, and two wells of mock-transfection (Lipofectamine alone) control. Automated cell culturing, ASO transfection, and cell harvesting Robotic automation (Hamilton) was established to provide a stable environment and accurate procedural timing control for cell cul- turing and transfection. In brief, trypsin-EDTA detachment, cell number and viability quantification, cell seeding, transfection, and cell harvesting were performed with automation. All transfec- tions were divided into 28 runs on a weekly basis. ASO transfection was performed with duplication. In each run, there were 16 inde- pendent transfections with ASO negative control A (NC_A, Exiqon) and 16 wells transfected with an ASO targeting MALAT- 1 (Exiqon). The HDF cells were seeded in 12-well plates with 80,000 cells in each well 24 h prior to the transfection. A final concentration of 20 nM ASO and 2 µL Lipofectamine RNAiMAX (Thermo Fisher Scientific) were mixed in 200 µL Opti-MEM (Thermo Fisher Scientific). The mixture was incubated at room temperature for 5 min and added to the cells, which were maintained in 1 mL com- plete medium. The cells were harvested 48 h posttransfection by adding 200 µL RLT buffer from the RNeasy 96 kit (Qiagen) after PBS washing. The harvested lysates were kept at −80°C. RNA was extracted from the lysate for real-time quantitative RT-PCR (Supplemental Methods). In summary, our study highlights the functional importance of lncRNAs regardless of their expression, localization, and conser- vation levels. Molecular phenotyping is a powerful and generally more sensitive to knockdown-mediated changes platform to reveal the functional relevance of lncRNAs that cannot be observed based on the cellular phenotypes alone. With additional molecular pro- filing techniques, suchas RNAduplex mapsinlivingcellsto decode common structural motifs (Lu et al. 2016), and Oxford Nanopore Technology (ONT) to annotate the full-length variant isoforms of lncRNAs (Hardwick et al. 2019), the structure-to-functional rela- tionship of lncRNAs may be elucidated further in the future. FANTOM6 pilot study FANTOM6 pilot study are generally more suitable for functional screenings of our lncRNA (62% found in the nuclear compartment). Besides, the dy- namics of secondary effects mediated by different levels of knock- down from different technologies are likely to be observed as discordance when considering the whole transcriptome, where this kind of discordance has been reported previously (Stojic et al. 2018). In contrast, in the MKI67 assay, where only a single feature such as growth phenotype is assayed, siRNA knockdown re- vealed higher reproducibility with ASO knockdown. This suggest- ed that the growth phenotype might be triggered by different specific pathways in ASO- and siRNA-knockdowns. are generally more suitable for functional screenings of our lncRNA (62% found in the nuclear compartment). Besides, the dy- namics of secondary effects mediated by different levels of knock- down from different technologies are likely to be observed as discordance when considering the whole transcriptome, where this kind of discordance has been reported previously (Stojic et al. 2018). In contrast, in the MKI67 assay, where only a single feature such as growth phenotype is assayed, siRNA knockdown re- vealed higher reproducibility with ASO knockdown. This suggest- ed that the growth phenotype might be triggered by different specific pathways in ASO- and siRNA-knockdowns. as an experimental control. For details, please refer to the Supplemental Methods. as an experimental control. For details, please refer to the Supplemental Methods. ASO design ASOs were designed as RNase H-recruiting locked nucleic acid (LNA) phosphorothioate gapmers with a central DNA gap flanked by 2–4 LNA nucleotides at the 5′ and 3′ ends of the ASOs. For de- tails, please refer to the Supplemental Methods. p p y Previous studies suggest that lncRNAs regulate gene expres- sion in trans epigenetically, via direct or indirect interaction with regulators such as DNMT1 (Di Ruscio et al. 2013) or by directly binding to DNA (triplex) (Mondal et al. 2015) or other RNA-bind- ing proteins (Tichon et al. 2016). Analysis of cellular localization by fractionation followed by RNA-seq and in situ hybridization can indicate whether a given lncRNA may act in trans by quantify- ing its abundance in the nuclear soluble fraction as compared to cytoplasm. Although most lncRNAs in the nuclear soluble fraction may affect pathways associated with chromatin modification, ad- ditional experiments to globally understand their interaction part- ners will elucidate the molecular mechanism behind trans-acting lncRNAs (Li et al. 2017; Sridhar et al. 2017). Discussion (D) GTEx atlas across 54 tissues (N = 9662 samples) shows relatively high expression levels of RP11-398K22.12 in ct brain regions samples (highlighted). (E) Expression correlation for RP11-398K22.12 and KCNQ5 in eight out of 13 distinct brain regio hlighted in D. (F) Expression correlation for RP11-398K22.12 and CATG00000088862.1 in eight out of 13 distinct brain regions, as h ilowski et al. B A B A B C C D D E E F F Figure 5. RP11-398K22.12 down-regulates KCNQ5 and CATG00000088862.1 in cis. (A) Changes in expression levels of detectable genes in the same topologically associated domain (TAD) as RP11-398K22.12 based on Hi-C analysis. Both KCNQ5 and CATG00000088862.1 are down-reg- ulated (P < 0.05) upon the knockdown of RP11-398K22.12 by two independent ASOs in CAGE analysis (left) as further confirmed with RT-qPCR (right). (B) (Top) Representation of the chromatin conformation in the 4-Mb region proximal to the TAD containing RP11-398K22.12, followed by the locus gene annotation, CAGE, RNA-seq, and ATAC-seq data for native HDFs. (Bottom) Schematic diagram showing Hi-C predicted con- tacts of RP11-398K22.12 (blue) and KCNQ5 (gray) (25-kb resolution, frequency ≥5) in HDF cells. Red line indicates RP11-398K22.12 and KCNQ5 contact. (C) FISH image for RP11-398K22.12, suggesting proximal regulation. TUG1 FISH image (suggesting trans regulation) is included as a comparison; (bar = 10 µm). (D) GTEx atlas across 54 tissues (N = 9662 samples) shows relatively high expression levels of RP11-398K22.12 in 13 distinct brain regions samples (highlighted). (E) Expression correlation for RP11-398K22.12 and KCNQ5 in eight out of 13 distinct brain regions, as highlighted in D. (F) Expression correlation for RP11-398K22.12 and CATG00000088862.1 in eight out of 13 distinct brain regions, as high- lighted in D. could be vastly different across compartments (Johnsson et al. 2013). Since the majority of functional lncRNAs are reported to be inside the nucleus (Palazzo and Lee 2018; Sun et al. 2018), ASO-mediated knockdowns, which mainly target nuclear RNAs, in the nucleus, the siRNAs use the RNA-inducing silencing com- plex (RISC) mainly active in the cytoplasm. LncRNAs are known to function in specific subcellular compartments (Chen 2016) and their maturity, secondary structures, isoforms, and functions 1068 Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from Chromosome conformation capture (Hi-C) Hi-C libraries were prepared essentially as described previously (Lieberman-Aiden et al. 2009; Fraser et al. 2015a) with minor changes to improve the DNA yield of Hi-C products (Fraser et al. 2015b). For details, please refer to the Supplemental Methods. MKI67 staining upon lncRNA knockdown For the selected four lncRNA targets showing >25% growth inhibi- tion, we used two siRNAs and two ASOs with independent se- quences. The transfected cells were fixed by adding prechilled 70% ethanol and incubated at −20°C. The cells were washed with FACS buffer (2% FBS in PBS, 0.05% NaN3) twice. FITC-conju- gated MKI67 (20Raj1, eBioscience) was applied to the cells and subjected to flow cytometric analysis. Knockdown efficiency by siRNA was determined by real-time quantitative RT-PCR using the same three primer pairs as for ASO knockdown efficiency. For details, please refer to the Supplemental Methods. Acknowledgments We thank Linda Kostrencic, Hiroto Atsui, Emi Ito, Nobuyuki Takeda, Tsutomu Saito, Teruaki Kitakura, Yumi Hara, Machiko Kashiwagi, and Masaaki Furuno at RIKEN Yokohama for assistance in arranging collaboration agreements, ethics applications, com- putational infrastructure, and the FANTOM6 meetings. We also thank RIKEN GeNAS for generation and sequencing of the CAGE libraries and subsequent data processing. FANTOM6 was made possible by a Research Grant for RIKEN Center for Life Science Technology, Division of Genomic Technologies (CLST DGT) and RIKEN Center for Integrative Medical Sciences (IMS) from MEXT, Japan. I.V.K. and I.E.V. were supported by Russian Foundation for Basic Research (RFBR) 18-34-20024, B.B. is supported by the fel- lowship 2017FI_B00722 from the Secretaria d’Universitats i Recerca del Departament d’Empresa i Coneixement (Generalitat de Catalunya) and the European Social Fund (ESF), A. Favorov was supported by National Institutes of Health (NIH) P30 CA006973 and RFBR 17-00-00208, D.G. is supported by a “la Caixa”-Severo Ochoa pre-doctoral fellowship (LCF/BQ/SO15/ 52260001), E.L.I. and L.M. were supported by NIH National Cancer Institute Grant R01CA200859 and Department of Defense (DOD) award W81XWH-16-1-0739, M.K.-S. was support- ed by Versus Arthritis UK 20298, A.L. was supported by the Swedish Cancer Society, The Swedish Research Council, the Swedish Childhood Cancer fund, Radiumhemmets forsknigs- fonder; V.J.M. was supported by the Russian Academy of Sciences Project 0112-2019-0001; Y.A.M. was supported by Russian Science Foundation (RSF) grant 18-14-00240, A.S. was sup- ported by Novo Nordisk Foundation, Lundbeck Foundation, Danish Cancer Society, Carlsberg Foundation, Independent Research Fund Denmark, A.R.R.F. is currently supported by an Australian National Health and Medical Research Council Fellowship APP1154524, M.M.H. was supported by Natural Sciences and Engineering Research Council of Canada (RGPIN- 2015-3948), C.S. was supported by the Interuniversity Consortium for Biotechnology (CIB) from the Italian Ministry of Education, University and Research (MIUR) grant n.974, CMPT177780. J. Luginbühl was supported by Japan Society for the Promotion of Science (JSPS) Postdoctoral Fellowship for Foreign Researchers. C.J.C.P. was supported by RIKEN Special Post-Doctoral Research (SPDR) fellowship. Cap analysis of gene expression (CAGE) Four micrograms of purified RNA were used to generate libraries ac- cording to the nAnT-iCAGE protocol (Murata et al. 2014). For de- tails, please refer to the Supplemental Methods. References Ambasudhan R, Talantova M, Coleman R, Yuan X, Zhu S, Lipton SA, Ding S. 2011. Direct reprogramming of adult human fibroblasts to functional neurons under defined conditions. Cell Stem Cell 9: 113–118. doi:10 .1016/j.stem.2011.07.002 Bai J, Yao B, Wang L, Sun L, Chen T, Liu R, Yin G, Xu Q, Yang W. 2019. lncRNA A1BG-AS1 suppresses proliferation and invasion of hepatocel- lular carcinoma cells by targeting miR-216a-5p. J Cell Biochem 120: 10310–10322. doi:10.1002/jcb.28315 Ramilowski et al. Ramilowski et al. Wound closure assay The HDF cells were transfected with 20 nM ASO as described earlier in 12-well plates. The cells were replated at 24 h posttransfection into a 96-well ImageLock plate (Essen BioScience) at a density of 20,000 cells per well. At 24 h after seeding, cells form a spatially uniform monolayer with 95%–100% cell confluence. The cells were incubated with 5 µg/mL mitomycin C for 2 h to inhibit cell division. Then, medium was refreshed and a uniform scratch was created in each well by the WoundMaker (Essen BioScience). The closure of the wound was monitored by Incucyte live-cell imaging system (Essen Bioscience) every 2 h for 24 h. The RNA was harvest- ed after the assay for real-time quantitative RT-PCR. For details, please refer to the Supplemental Methods. Cell morphology quantification For each transfection, a representative phase-contrast image at a single time point was exported from the Incucyte time-series. These raw images were first transformed to probability maps of cells by pixel classification using ilastik (1.3.2) (Berg et al. 2019). The trained model was then applied to all images where the pre- dicted probability maps of cells (grayscale, 16 bits tiff format) were subsequently used for morphology quantification in CellProfiler (3.1.5) (Carpenter et al. 2006). For details, please refer to the Supplemental Methods. Gene models and lncRNA target selections The gene models used in this study were primarily based on the FANTOM CAGE-associated transcriptome (CAT) at permissive level as defined previously (Hon et al. 2017). From this merged as- sembly, there were ∼2000 lncRNAs robustly expressed in HDFs (TPM ≥1). However, we selected lncRNA knockdown targets in an unbiased manner to broadly cover various types of lncRNAs (TPM ≥0.2). Briefly, we first identified a list of the lncRNA genes expressed in HDFs, with RNA-seq expression at least 0.5 fragments per kilobase per million and CAGE expression at least 1 tag per mil- lion. Then, we manually inspected each lncRNA locus in the ZENBU genome browser for (1) its independence from neighbor- ing genes on the same strand (if any), (2) support from RNA-seq (for exons and splicing junctions) and CAGE data (for TSSs) of its transcript models, and (3) support from histone marks at TSSs for transcription initiation (H3K27ac) and along the gene body for elongation (H3K36me3), from the Roadmap Epigenomics Consortium (Roadmap Epigenomics Consortium et al. 2015). A representative transcript model, which best represents the RNA- seq signal, was manually chosen from each locus for design of an- tisense oligonucleotides. In total, 285 lncRNA loci were chosen for ASO suppression. Additional controls (NEAT1, protein coding genes) (Supplemental Table S1) were added, including MALAT1 p Phase-contrast images of transfected cells were captured every 3 h for 2 d with three fields per well by the Incucyte live-cell imag- ing system (Essen Bioscience). The confluence in each field was an- alyzed by the Incucyte software. The mean confluence of each well was taken along the timeline until the mean confluence of the NC_A control in the same plate reached 90%. The growth rate in each well was calculated as the slope of a linear regression. A nor- malized growth rate of each replicate was calculated as the growth rate divided by the mean growth rate of the six NC_A controls from the same plate. Negative growth rate was derived when cells shrink and/or detach. As these rates of cell depletion could not be normal- ized by the rate of growth, negative values were maintained to in- dicate severe growth inhibition. Student’s t-test was performed between the growth rate of the duplicated samples and the six NC_A controls, assuming equal variance. Genome Research www.genome.org 1069 Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from FANTOM6 pilot study analysis software for identifying and quantifying cell phenotypes. Genome Biol 7: R100. doi:10.1186/gb-2006-7-10-r100 Joung J, Engreitz JM, Konermann S, Abudayyeh OO, Verdine VK, Aguet F, Gootenberg JS, Sanjana NE, Wright JB, Fulco CP, et al. 2017. Genome- scale activation screen identifies a lncRNA locus regulating a gene neigh- bourhood. Nature 548: 343–346. doi:10.1038/nature23451 Carrieri C, Cimatti L, Biagioli M, Beugnet A, Zucchelli S, Fedele S, Pesce E, Ferrer I, Collavin L, Santoro C, et al. 2012. Long non-coding antisense RNA controls Uchl1 translation through an embedded SINEB2 repeat. Nature 491: 454–457. doi:10.1038/nature11508 Kalluri R. 2016. The biology and function of fibroblasts in cancer. Nat Rev Cancer 16: 582–598. doi:10.1038/nrc.2016.73 Chen L-L. 2016. Linking long noncoding RNA localization and function. Trends Biochem Sci 41: 761–772. doi:10.1016/j.tibs.2016.07.003 Kendall RT, Feghali-Bostwick CA. 2014. Fibroblasts in fibrosis: novel roles and mediators. Front Pharmacol 5: 123. doi:10.3389/fphar.2014.00123 Cho SW, Xu J, Sun R, Mumbach MR, Carter AC, Chen YG, Yost KE, Kim J, He J, Nevins SA, et al. 2018. Promoter of lncRNA gene PVT1 is a tumor-sup- pressor DNA boundary element. Cell 173: 1398–1412.e22. doi:10.1016/ j.cell.2018.03.068 p Kopp F, Mendell JT. 2018. Functional classification and experimental dis- section of long noncoding RNAs. Cell 172: 393–407. doi:10.1016/j .cell.2018.01.011 Li L, Chang HY. 2014. Physiological roles of long noncoding RNAs: insight from knockout mice. Trends Cell Biol 24: 594–602. doi:10.1016/j.tcb .2014.06.003 j Chu C, Zhang QC, da Rocha ST, Flynn RA, Bharadwaj M, Calabrese JM, Magnuson T, Heard E, Chang HY. 2015. Systematic discovery of Xist RNA binding proteins. Cell 161: 404–416. doi:10.1016/j.cell.2015 .03.025 Li B, Wang JH-C. 2011. Fibroblasts and myofibroblasts in wound healing: force generation and measurement. J Tissue Viability 20: 108–120. doi:10.1016/j.jtv.2009.11.004 Collado-Torres L, Nellore A, Kammers K, Ellis SE, Taub MA, Hansen KD, Jaffe AE, Langmead B, Leek JT. 2017. Reproducible RNA-seq analysis using re- count2. Nat Biotechnol 35: 319–321. doi:10.1038/nbt.3838 j j Li X, Zhou B, Chen L, Gou L-T, Li H, Fu X-D. 2017. GRID-seq reveals the global RNA–chromatin interactome. Nat Biotechnol 35: 940–950. doi:10.1038/nbt.3968 De Hoon M, Shin JW, Carninci P. 2015. Paradigm shifts in genomics through the FANTOM projects. Mamm Genome 26: 391–402. doi:10 .1007/s00335-015-9593-8 Lieberman-Aiden E, van Berkum NL, Imakaev M, Ragoczy T, Telling A, Amit I, Lajoie BR, Sabo PJ, Dorschner MO, Williams L, et al. 2009. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326: 289–293. FANTOM6 pilot study doi:10.1126/sci ence.1181369 Di Ruscio A, Ebralidze AK, Benoukraf T, Amabile G, Goff LA, Terragni J, Figueroa ME, De Figueiredo Pontes LL, Alberich-Jorda M, Zhang P, et al. 2013. DNMT1-interacting RNAs block gene-specific DNA methyl- ation. Nature 503: 371–376. doi:10.1038/nature12598 Liu SJ, Horlbeck MA, Cho SW, Birk HS, Malatesta M, He D, Attenello FJ, Villalta JE, Cho MY, Chen Y, et al. 2017. CRISPRi-based genome-scale identification of functional long noncoding RNA loci in human cells. Science 355: eaah7111. doi:10.1126/science.aah7111 Engreitz JM, Haines JE, Perez EM, Munson G, Chen J, Kane M, McDonel PE, Guttman M, Lander ES. 2016a. Local regulation of gene expression by lncRNA promoters, transcription and splicing. Nature 539: 452–455. doi:10.1038/nature20149 Science 355: eaah7111. doi:10.1126/science.aah7111 Liu Y, Cao Z, Wang Y, Guo Y, Xu P, Yuan P, Liu Z, He Y, Wei W. 2018. Genome-wide screening for functional long noncoding RNAs in human cells by Cas9 targeting of splice sites. Nat Biotechnol 36: 1203–1210. doi:10.1038/nbt.4283 Engreitz JM, Ollikainen N, Guttman M. 2016b. Long non-coding RNAs: spa- tial amplifiers that control nuclear structure and gene expression. Nat Rev Mol Cell Biol 17: 756–770. doi:10.1038/nrm.2016.126 The FANTOM Consortium, Suzuki H, Forrest ARR, Van Nimwegen E, Daub CO, Balwierz PJ, Irvine KM, Lassmann T, Ravasi T, Hasegawa Y, et al. 2009. The transcriptional network that controls growth arrest and dif- ferentiation in a human myeloid leukemia cell line. Nat Genet 41: 553–562. doi:10.1038/ng.375 Lu Z, Zhang QC, Lee B, Flynn RA, Smith MA, Robinson JT, Davidovich C, Gooding AR, Goodrich KJ, Mattick JS, et al. 2016. RNA duplex map in living cells reveals higher-order transcriptome structure. Cell 165: 1267–1279. doi:10.1016/j.cell.2016.04.028 Luo S, Lu JY, Liu L, Yin Y, Chen C, Han X, Wu B, Xu R, Liu W, Yan P, et al. 2016. Divergent lncRNAs regulate gene expression and lineage differen- tiation in pluripotent cells. Cell Stem Cell 18: 637–652. doi:10.1016/j .stem.2016.01.024 g Fraser J, Ferrai C, Chiariello AM, Schueler M, Rito T, Laudanno G, Barbieri M, Moore BL, Kraemer DCA, Aitken S, et al. 2015a. Hierarchical folding and reorganization of chromosomes are linked to transcriptional changes in cellular differentiation. Mol Syst Biol 11: 852. doi:10.15252/msb .20156492 Ma L, Cao J, Liu L, Du Q, Li Z, Zou D, Bajic VB, Zhang Z. 2019. LncBook: a curated knowledgebase of human long non-coding RNAs. Nucleic Acids Res 47: D128–D134. doi:10.1093/nar/gky960 Fraser J, Williamson I, Bickmore WA, Dostie J. 2015b. FANTOM6 pilot study An overview of ge- nome organization and how we got there: from FISH to Hi-C. Microbiol Mol Biol Rev 79: 347–372. doi:10.1128/MMBR.00006-15 Mondal T, Subhash S, Vaid R, Enroth S, Uday S, Reinius B, Mitra S, Mohammed A, James AR, Hoberg E, et al. 2015. MEG3 long noncoding RNA regulates the TGF-β pathway genes through formation of RNA- DNA triplex structures. Nat Commun 6: 7743. doi:10.1038/ ncomms8743 The GTEx Consortium. 2015. Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in hu- mans. Science 348: 648–660. doi:10.1126/science.1262110 Gupta RA, Shah N, Wang KC, Kim J, Horlings HM, Wong DJ, Tsai M-C, Hung T, Argani P, Rinn JL, et al. 2010. Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 464: 1071–1076. doi:10.1038/nature08975 Murata M, Nishiyori-Sueki H, Kojima-Ishiyama M, Carninci P, Hayashizaki Y, Itoh M. 2014. Detecting expressed genes using CAGE. Methods Mol Biol 1164: 67–85. doi:10.1007/978-1-4939-0805-9_7 Palazzo AF, Lee ES. 2018. Sequence determinants for nuclear retention and cytoplasmic export of mRNAs and lncRNAs. Front Genet 9: 440. doi:10 .3389/fgene.2018.00440 Guttman M, Rinn JL. 2012. Modular regulatory principles of large non-cod- ing RNAs. Nature 482: 339–346. doi:10.1038/nature10887 Guttman M, Donaghey J, Carey BW, Garber M, Grenier JK., Munson G, Young G, Lucas AB, Ach R, Bruhn L, et al. 2011. lincRNAs act in the cir- cuitry controlling pluripotency and differentiation. Nature 477: 295– 300. doi:10.1038/nature10398 g Proudfoot NJ. 2016. Transcriptional termination in mammals: stopping the RNA polymerase II juggernaut. Science 352: aad9926. doi:10.1126/sci ence.aad9926 Quek XC, Thomson DW, Maag JLV, Bartonicek N, Signal B, Clark MB, Gloss BS, Dinger ME. 2015. lncRNAdb v2.0: expanding the reference database for functional long noncoding RNAs. Nucleic Acids Res 43: D168–D173. doi:10.1093/nar/gku988 Hardwick SA, Bassett SD, Kaczorowski D, Blackburn J, Barton K, Bartonicek N, Carswell SL, Tilgner HU, Loy C, Halliday G, et al. 2019. Targeted, high-resolution RNA sequencing of non-coding genomic regions associ- ated with neuropsychiatric functions. Front Genet 10: 309. doi:10.3389/ fgene.2019.00309 g Quinn JJ, Chang HY. 2016. Unique features of long non-coding RNA bio- genesis and function. Nat Rev Genet 17: 47–62. doi:10.1038/nrg.2015.10 g Hon C-C, Ramilowski JA, Harshbarger J, Bertin N, Rackham OJL, Gough J, Denisenko E, Schmeier S, Poulsen TM, Severin J, et al. 2017. An atlas of human long non-coding RNAs with accurate 5′ ends. Nature 543: 199–204. doi:10.1038/nature21374 Ransohoff JD, Wei Y, Khavari PA. 2018. Data access All raw and processed sequencing data generated in this study have been submitted to the DNA Data Bank of Japan (DDBJ; https:// www.ddbj.nig.ac.jp/) under accession numbers DRA008311, DRA008312, DRA008436, and DRA008511 or can be accessed through the FANTOM6 project portal https://fantom.gsc.riken .jp/6/datafiles. The analysis results can be downloaded from https://fantom.gsc.riken.jp/6/suppl/Ramilowski_et_al_2020/data/ and interactively explored using our in-house portal https ://fantom.gsc.riken.jp/zenbu/reports/#FANTOM6. Berg S, Kutra D, Kroeger T, Straehle CN, Kausler BX, Haubold C, Schiegg M, Ales J, Beier T, Rudy M, et al. 2019. ilastik: interactive machine learning for (bio)image analysis. Nat Methods 16: 1226–1232. doi:10.1038/ s41592-019-0582-9 Carninci P, Kasukawa T, Katayama S, Gough J, Frith MC, Maeda N, Oyama R, Ravasi T, Lenhard B, Wells C, et al. 2005. The transcriptional land- scape of the mammalian genome. Science 309: 1559–1563. doi:10 .1126/science.1112014 Carpenter AE, Jones TR, Lamprecht MR, Clarke C, Kang I, Friman O, Guertin DA, Chang J, Lindquist RA, Moffat J, et al. 2006. CellProfiler: image 1070 Genome Research www.genome.org 1070 Genome Research www.genome.org Cold Spring Harbor Laboratory Pre on September 24, 2020 - Published by genome.cshlp.org Downloaded from FANTOM6 pilot study The functions and unique features of long intergenic non-coding RNA. Nat Rev Mol Cell Biol 19: 143–157. doi:10.1038/nrm.2017.104 Roadmap Epigenomics Consortium, Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, Heravi-Moussavi A, Kheradpour P, Zhang Z, Wang J, et al. 2015. Integrative analysis of 111 reference human epigenomes. Nature 518: 317–330. doi:10.1038/nature14248 Imada E-L, Sanchez DF, Collado-Torres L, Wilks C, Matam T, Dinalankara W, Stupnikov A, Lobo-Pereira F, Yip C-W, Yasuzawa K, et al. 2020. Recounting the FANTOM CAGE-Associated Transcriptome. Genome Res (this issue). doi:10.1101/gr.254656.119 Roux BT, Lindsay MA, Heward JA. 2017. Knockdown of nuclear-located en- hancer RNAs and long ncRNAs using locked nucleic acid GapmeRs. Methods Mol Biol 1468: 11–18. doi:10.1007/978-1-4939-4035-6_2 g Iyer MK, Niknafs YS, Malik R, Singhal U, Sahu A, Hosono Y, Barrette TR, Prensner JR, Evans JR, Zhao S, et al. 2015. The landscape of long noncod- ing RNAs in the human transcriptome. Nat Genet 47: 199–208. doi:10 .1038/ng.3192 Sokolova M, Turunen M, Mortusewicz O, Kivioja T, Herr P, Vähärautio A, Björklund M, Taipale M, Helleday T, Taipale J. 2017. Genome-wide screen of cell-cycle regulators in normal and tumor cells identifies a dif- ferential response to nucleosome depletion. Cell Cycle 16: 189–199. doi:10.1080/15384101.2016.1261765 Johnsson P, Ackley A, Vidarsdottir L, Lui W-O, Corcoran M, Grandér D, Morris KV. 2013. A pseudogene long-noncoding-RNA network regulates PTEN transcription and translation in human cells. Nat Struct Mol Biol 20: 440–446. doi:10.1038/nsmb.2516 Genome Research www.genome.org 1071 Ramilowski et al. Sridhar B, Rivas-Astroza M, Nguyen TC, Chen W, Yan Z, Cao X, Hebert L, Zhong S. 2017. Systematic mapping of RNA-chromatin interactions in vivo. Curr Biol 27: 602–609. doi:10.1016/j.cub.2017.01.011 Tichon A, Gil N, Lubelsky Y, Havkin Solomon T, Lemze D, Itzkovitz S, Stern- Ginossar N, Ulitsky I. 2016. A conserved abundant cytoplasmic long noncoding RNA modulates repression by Pumilio proteins in human cells. Nat Commun 7: 12209. doi:10.1038/ncomms12209 Stojic L, Lun ATL, Mangei J, Mascalchi P, Quarantotti V, Barr AR, Bakal C, Marioni JC, Gergely F, Odom DT. 2018. Specificity of RNAi, LNA and CRISPRi as loss-of-function methods in transcriptional analysis. Nucleic Acids Res 46: 5950–5966. doi:10.1093/nar/gky437 Ulitsky I. 2016. Evolution to the rescue: using comparative genomics to un- derstand long non-coding RNAs. Nature Reviews Genetics 17: 601–614. doi:10.1038/nrg.2016.85 Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al. 2005. Received July 12, 2019; accepted in revised form June 24, 2020. 1072 Genome Research www.genome.org Tichon A, Gil N, Lubelsky Y, Havkin Solomon T, Lemze D, Itzkovitz S, Stern- Ginossar N, Ulitsky I. 2016. A conserved abundant cytoplasmic long noncoding RNA modulates repression by Pumilio proteins in human cells. Nat Commun 7: 12209. doi:10.1038/ncomms12209 Ulitsky I. 2016. Evolution to the rescue: using comparative genomics to un- derstand long non-coding RNAs. Nature Reviews Genetics 17: 601–614. doi:10.1038/nrg.2016.85 Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org Downloaded from FANTOM6 pilot study Gene set en- richment analysis: a knowledge-based approach for interpreting ge- nome-wide expression profiles. Proc Natl Acad Sci 102: 15545–15550. doi:10.1073/pnas.0506580102 Volders P-J, Verheggen K, Menschaert G, Vandepoele K, Martens L, Vandesompele J, Mestdagh P. 2015. An update on LNCipedia: a data- base for annotated human lncRNA sequences. Nucleic Acids Res 43: D174–D180. doi:10.1093/nar/gku1060 Xue Z, Hennelly S, Doyle B, Gulati AA, Novikova IV, Sanbonmatsu KY, Boyer LA. 2016. A G-rich motif in the lncRNA Braveheart interacts with a zinc-finger transcription factor to specify the cardiovascular lin- eage. Mol Cell 64: 37–50. doi:10.1016/j.molcel.2016.08.010 p Sun Q, Hao Q, Prasanth KV. 2018. Nuclear long noncoding RNAs: key reg- ulators of gene expression. Trends Genet 34: 142–157. doi:10.1016/j.tig .2017.11.005 Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K, Yamanaka S. 2007. Induction of pluripotent stem cells from adult hu- man fibroblasts by defined factors. Cell 131: 861–872. doi:10.1016/j .cell.2007.11.019 Received July 12, 2019; accepted in revised form June 24, 2020. 1376 Genome Research www.genome.org Corrigenda Corrigenda Genome Research 28: 285–294 (2018) Genome Research 28: 285–294 (2018) Corrigendum: 3′ UTR lengthening as a novel mechanism in regulating cellular senescence Meng Chen, Guoliang Lyu, Miao Han, Hongbo Nie, Ting Shen, Wei Chen, Yichi Niu, Yifan Song, Xueping Li, Huan Li, Xinyu Chen, Ziyue Wang, Zheng Xia, Wei Li, Xiao-Li Tian, Chen Ding, Jun Gu, Yufang Zheng, Xinhua Liu, Jinfeng Hu, Gang Wei, Wei Tao, and Ting Ni The authors would like to correct Figure 3, panel J, in which the rightmost upper image of SA-β-gal stained 293T cells following short hairpin RNA (shRNA)-mediated knockdown of RRAS2 with sh769 (RRAS2-KD- sh769) was inadvertently, and due to a labeling error, taken from the same original source image presented in the middle upper panel, which shows increased SA-β-gal activity following RRAS2 knockdown by a different shRNA (sh646). This correction does not affect any of the conclusions of the article. The corrected image representative of RRAS2-KD-sh769 is provided below, and Figure 3 has been updated in the article online. The authors thank Ning Yuan Lee for bringing this error to their attention and apologize for any confusion this may have caused. Additionally, the authors have provided a revised Supplemental Figure S7 file in which the redundant succes- sive Supplemental figure files have been removed. This can be found in the Revised Supplemental Material online Figure 3. Panel J, rightmost upper image. Figure 3. Panel J, rightmost upper image. Figure 3. Panel J, rightmost upper image. Figure 3. Panel J, rightmost upper image. Figure 3. Panel J, rightmost upper image. The authors thank Ning Yuan Lee for bringing this error to their attention and apologize for any confusion this may have caused. The authors thank Ning Yuan Lee for bringing this error to their attention and apologize for any confusion this may have caused. The authors thank Ning Yuan Lee for bringing this error to their attention and apologize for any confusion this may have caused. Additionally, the authors have provided a revised Supplemental Figure S7 file in which the redundant succes- sive Supplemental figure files have been removed. This can be found in the Revised Supplemental Material online. doi: 10.1101/gr.270165.120 1376 Genome Research www.genome.org Corrigenda Genome Research 30: 1060–1072 (2020) Genome Research 1377 www.genome.org These updates are reflected in the revised manuscript online. unctional annotation of human long noncoding RNAs via molecula phenotyping Jordan A. Ramilowski, Chi Wai Yip, Saumya Agrawal, Jen-Chien Chang, Yari Ciani, Ivan V. Kulakovskiy, Mickaël Mendez, Jasmine Li Ching Ooi, John F. Ouyang, Nick Parkinson, Andreas Petri, Leonie Roos, Jessica Severin, Kayoko Yasuzawa, Imad Abugessaisa, Altuna Akalin, Ivan V. Antonov, Erik Arner, Alessandro Bonetti, Hidemasa Bono, Beatrice Borsari, Frank Brombacher, Christopher J.F. Cameron, Carlo Vittorio Cannistraci, Ryan Cardenas, Melissa Cardon, Howard Chang, Josée Dostie, Luca Ducoli, Alexander Favorov, Alexandre Fort, Diego Garrido, Noa Gil, Juliette Gimenez, Reto Guler, Lusy Handoko, Jayson Harshbarger, Akira Hasegawa, Yuki Hasegawa, Kosuke Hashimoto, Norihito Hayatsu, Peter Heutink, Tetsuro Hirose, Eddie L. Imada, Masayoshi Itoh, Bogumil Kaczkowski, Aditi Kanhere, Emily Kawabata, Hideya Kawaji, Tsugumi Kawashima, S. Thomas Kelly, Miki Kojima, Naoto Kondo, Haruhiko Koseki, Tsukasa Kouno, Anton Kratz, Mariola Kurowska-Stolarska, Andrew Tae Jun Kwon, Jeffrey Leek, Andreas Lennartsson, Marina Lizio, Fernando López-Redondo, Joachim Luginbühl, Shiori Maeda, Vsevolod J. Makeev, Luigi Marchionni, Yulia A. Medvedeva, Aki Minoda, Ferenc Müller, Manuel Muñoz-Aguirre, Mitsuyoshi Murata, Hiromi Nishiyori, Kazuhiro R. Nitta, Shuhei Noguchi, Yukihiko Noro, Ramil Nurtdinov, Yasushi Okazaki, Valerio Orlando, Denis Paquette, Callum J.C. Parr, Owen J.L. Rackham, Patrizia Rizzu, Diego Fernando Sánchez Martinez, Albin Sandelin, Pillay Sanjana, Colin A.M. Semple, Youtaro Shibayama, Divya M. Sivaraman, Takahiro Suzuki, Suzannah C. Szumowski, Michihira Tagami, Martin S. Taylor, Chikashi Terao, Malte Thodberg, Supat Thongjuea, Vidisha Tripathi, Igor Ulitsky, Roberto Verardo, Ilya E. Vorontsov, Chinatsu Yamamoto, Robert S. Young, J. Kenneth Baillie, Alistair R.R. Forrest, Roderic Guigó, Michael M. Hoffman, Chung Chau Hon, Takeya Kasukawa, Sakari Kauppinen, Juha Kere, Boris Lenhard, Claudio Schneider, Harukazu Suzuki, Ken Yagi, Michiel J.L. de Hoon, Jay W. Shin, and Piero Carninci The authors would like to correct the misspelling of an author’s name and the inadvertent omission of two affiliations for that author, which are as follows: Christopher J.F. Cameron, Department of Biochemistry, Rosalind and Morris Goodman Cancer Research Center, McGill University, Montréal, Québec H3G 1Y6, Canada and Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06510, USA. These updates are reflected in the revised manuscript online. doi: 10.1101/gr.270330.120 Genome Research www.genome.org Genome Research www.genome.org Cold Spring Harbor Laboratory Press on September 24, 2020 - Published by genome.cshlp.org oaded from Material Supplemental Related Content Genome Res. September , 2020 30: 1377-1 Jordan A. Ramilowski, Chi Wai Yip, Saumya Agrawal, et al. molecular phenotyping Corrigendum: Functional annotation of human long noncoding RNAs via References http://genome.cshlp.org/content/30/7/1060.full.html#related-urls Articles cited in: http://genome.cshlp.org/content/30/7/1060.full.html#ref-list-1 This article cites 58 articles, 9 of which can be accessed free at: Open Access Open Access option. Genome Research Freely available online through the License Commons Creative . http://creativecommons.org/licenses/by/4.0/ Commons License (Attribution 4.0 International), as described at , is available under a Creative Genome Research This article, published in Service Email Alerting click here. top right corner of the article or Receive free email alerts when new articles cite this article - sign up in the box at the http://genome.cshlp.org/subscriptions go to: Genome Research To subscribe to © 2020 Ramilowski et al.; Published by Cold Spring Harbor Laboratory Press
https://openalex.org/W3016922176	https://www.e3s-conferences.org/10.1051/e3sconf/202016101090/pdf	English	null	Features of using solid cellulose-containing domestic wastes for production of bioethanol	E3S web of conferences	2,020	cc-by	2,019	Features of using solid cellulose-containing domestic wastes for production of bioethanol T.A. Bolotnikova1, E.B. Aronova1, J.G. Bazarnova1, O.I. Bolotnikova1,2,, A.I. Ginak3 1Peter the Great St. Petersburg Polytechnic University, Polytechnicheskaya Str. 29, Saint Petersburg, 195251, Russian Federation 2Petrozavodsk State University, Lenina Str. 33, Petrozavodsk, 185910, Republic of Karelia, Russian Federation 3Saint-Petersburg State Institute of Technology, Moskovskiy Av. 26, Saint Petersburg, 190013, Russian Federation Abstract. In this paper, we consider the process of producing biogas with a high methane content when used as a co-substrate for fermentation of plant residues of microalgae. Microalgae Chlorella sorokiniana are a valuable source for obtaining valuable components such as lipids, pigments, proteins, chlorophyll and others. After the extraction of valuable components, residual biomass is formed, which requires further disposal. In this experiment, the digestion process is carried out using an inoculant — lyophilically dried activated sludge from sewage treatment plants in Hamburg in the amount of 450 ml and residual biomass of the microalga Chlorella sorokiniana in the amount of 2.1 g. The studies were carried out in the Anaerobes Test system AMPT-II system. Fermentation produces 205 ml of methane gas. 1 Introduction yeast to utilize acid wood hydrolysates are known [6]. Therefore, the purpose of this work was to optimize the parameters of paper and paperboard wastes acid hydrolysis and to evaluate the ethanol yield from these substrates with the help of the special strain of Sacch. cerevisiae "Omskie". The problem of household wastes disposal is becoming increasingly acute from year to year. Almost half of the volume is waste paper and cardboard. Just one family of four people on the average throws out about 100 kg of paper and packaging wastes per year. The period of the disintegration in natural conditions is from 3 months to a year [1]. The traditional burning of paper wastes is extremely harmful to the environment [2]. Therefore, the development of efficient recycling technologies for such materials is one of the aims for the state policy of developed countries in the field of ecology and waste management. © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). Corresponding author: olga-bolotnikova@rambler.ru DP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 es/by/4 0/) ICEPP-2020 E3S Web of Conferences 161, 010 (2020) 90 ICEPP-2020 E3S Web of Conferences 161, 010 (2020) 90 https://doi.org/10.1051/ conf/202016101090 e3s 2 Methods This ensured not only their crushing to garbage, but also the dispersion of contaminants to sizes that did not significantly affect the process of acid hydrolysis of cellulose. Table 1. Obtaining of the bottom product. Table 1. Obtaining of the bottom product. Characteristics Stage Atmospheric pressure, +95°C The dispersion 2 hours, H2SO4 (70%); 2 hours, H2SO4 (40%). Acid hydrolysis Milk of lime Neutralization +80°C, pressure at the top of the column 0.7 MPa Distillation vacuum 60 minutes to pH=4.8 Air purge Cooling to a temperature of +25°C, Cooling To a final concentration of H2SO4 of 10%, Dilution with water 25% NH4OH to pH=6.0. Neutralization Evaporation by 50% Filtration, evaporation It is known that the technological parameters of concentrated and diluted acid hydrolysis of wood differ significantly [10]. The undoubted advantages of the first method are the high efficiency of the breaking the β- glycosidic bonds and, the mild temperature treatment of plant materials. At the same time, significant amounts of acid, the duration and high consumption of gypsum to neutralize the sugar-containing substrate, prior to its microbiological conversion, reduce the commercial attractiveness of such hydrolysis for the disposal of household cellulose-containing waste [10]. The main advantage of dilute acid hydrolysis is the minimal temporary loss. However, the extreme temperature regime, as well as the impossibility of complete degradation of cellulose macromolecules, cast doubt on the environmental friendliness of this method [11]. Therefore, the destruction of ground samples of waste paper and cardboard to glucose-containing substrates was carried out by the combined method, mixing the elements of concentrated and diluted hydrolysis. By the end of 4 hours of such treatment, the fibrous component of municipal solid waste was 95% converted into a glucose-containing solution. The combined method greatly simplifies the technological scheme and, is characterized by a sufficiently high efficiency and, lower financial costs achieved by reducing corrosion of equipment and reducing the neutralization of acid hydrolysates. In addition, it allows to reduce the volume of water required for acid hydrolysis, and avoid using significant external pressure as a factor that increases the efficiency of the destruction β-1,4-glycosidic bonds in cellulose macromolecules. These features well illustrate the environmental safety of the developed method. The bottom product was cooled (+40°C), diluted with water to a concentration of reducing substances of 10-13 g/l, and 0.02% (NH4)2SO4 and 0.01% ((CaH2PO4)2×H2O+2CaSO4×2H2O) were added. This wort was fermented with a batch culture of Sacch yeast. 2 Methods For the experiment, paper and cardboard were taken from the house bin in quantities of 1 kg of each sample. These samples were dispersed in a hydraulic diluent to a homogeneous pulp with fragments of size: 150-200×4 mm (newsprint) and 100×5 mm (cardboard). The dispersion was carried out at atmospheric pressure and +95°C. H2SO4 was added to the resulting homogeneous mass to a final concentration of 70%. The mixture was stirred, incubated for 2 hours at +60-70°C. A part of the acid fraction was removed and water was added to a final concentration of H2SO4 of 40%. This solution has been incubated for 2 hours under the previous conditions (hydraulic module of 8.7). Using different microorganisms fermenting D- glucose as the main component of cellulose is the main topic of discussion. Bioethanol is widely used not only in various industries, but also as an additive to boiler fuel for internal combustion engines, which meets modern environmental requirements and has a positive economic effect [1]. Enzymatic hydrolysis of abundant paper and paperboard wastes to glucose-containing substrates has significant cost and requires special technologies [3]. Cheapening of this process can be achieved through the use of mineral acids [4]. Similar methods of cellulose destruction were widely used in the USSR in the hydrolysis industry. It was at that time when soft methods for wood hydrolysis were developed, which made possible to reduce the anthropogenic impact on the environment [5]. However, cellulose hydrolysates obtained in this way are unsuitable for ethanol conversion by bacteria, which metabolic activity depends on pH level of the medium [5]. At the same time, the special strains of Saccharomyces cerevisiae The preparation of glucose-containing substrates obtained by the acid method for ethanol conversion was carried out during sequential neutralization with milk of lime, distillation vacuum (+80°C, pressure at the top of the column 0.7 MPa), continuously purged with air for 60 minutes to pH=4.8. The glucose-containing mixture was cooled down to the temperature of +25°C, diluted with water to the final concentration of H2SO4 of 10%, neutralized with 25% NH4OH to pH=6.0. The neutralized solution was filtered and evaporated by 50% (Table 1). The concentration of ICEPP-2020 E3S Web of Conferences 161, 010 (2020) 90 https://doi.org/10.1051/ conf/202016101090 e3s reducing sugars in the bottom product was determined by Fehling's test. 2 Methods cerevisiae "Omsk" (Collection of industrial microorganisms, VNIIHydrolysis, St. Petersburg). Ethanol fermentation conditions were: 100 ml round- bottom Erlenmeyer flasks with 100 ml of wort on a thermostatic rotary shaker for 24 hours at +30-32°C. After 18-24 hours 10 ml of yeast suspension from these flasks were transferred to a 500 ml flasks containing 100 ml of the wort which were incubated under similar conditions. Yeast biomass was used for anaerobic fermentation in an amount of 10.0±0.5 g a.d.m. (absolutely dry matter)/l. After fermentation the yeast biomass was separated in a laboratory centrifuge in 5-10 minutes at 5000 rpm. Fractional distillation of ethanol-containing mash was carried out on a distillation unit with a deflector (200 mm). On epuration column ethanol raw was purified to separate the ether-aldehyde fraction. Ethanol concentration was determined in the distillate obtained after distillation of the test sample by gas chromatography under conditions [7]. The experimental results were statistically processed using Student's test at a significance level of 0.05 with the MS Excel computer program. It is widely known that yeast, in comparison with bacteria and mycelial fungi, is characterized by greater metabolic activity, acidophilicity and fermentative type of sugar catabolism [12]. Special strains Sacch. cerevisiae have already been constructed in the USSR for the disposal of acid wood hydrolysates and by- products of hydrolysis plants. These strains are traditionally used as ethanol producers [6]. The choice of the strain of Sacch. cerevisiae "Omskie" as a biocatalyst made it possible to prepare wort for fermentation using traditional industrial methods (neutralization, vacuum rectification, continuous air purging, neutralization and filtration) [13]. Towards the end of 24-hour anaerobic fermentation the concentration of RB in the wort decreased to 2.5%, and the yield of ethanol was 52 ml/kg (paper waste) and 34 ml/kg (cardboard waste). Theoretical calculations show that scaling the process discussed above will make possible to obtain up to 150 L of rectified ethanol per ton of absolutely dry cellulose-containing household waste, which corresponds to the standard for the hydrolysis wood industry [14]. 3 Results and Discussion Today, cellulose-containing waste is a valuable source of recyclable materials, which are widely used to obtain a number of national economic products [8]. However, household waste paper and paperboard of poor quality usually contain bitumen, wax, paraffin, glue and other impurities, clogging nets and cloth of paper machines, sticking to the surface of the drying cylinders [9]. Such contaminated household waste, unsuitable for re- extraction of cellulose, is subjected to acid degradation. Therefore, samples of paper and cardboard from the house garbage container were first subjected to thermomechanical processing (so-called cold method). 2 ICEPP-2020 E3S Web of Conferences 161, 010 (2020) 90 https://doi.org/10.1051/ conf/202016101090 e3s 4 Conclusion [13] B. Tsegaye, C. Balomajumder, P. Roy, Bulletin of the National Research Centre, 43, 51 (2019). During the experiment, optimal conditions were selected for the conversion of household cellulose-containing garbage into glucose-containing substrates. The fundamental possibility for the utilization of household cellulose-containing garbage to bioethanol was proved. This method greatly simplifies the typical technological process of acid hydrolysis of cellulose, reduces its toxic effect on the environment. The use of a special biocatalyst allows optimization of the financial costs of preparing acidic cellulose hydrolysates for ethanol bioconversion through the use of standard detoxification technology. The well-developed network of pulp and paper mill industry in the Russian Federation will become a good technological potential for processing a variety of household and industrial wastes into a target product of great economic importance. [14] I.I. Korol'kov, Perkolyacionnyj gidroliz rastitel'nogo syr'ya, M.: Lesnaya prom-t', 271 (1990) (In Russian). This work was supported by the grant No 18-44-100001 from Russian Foundation for Basic Research (RFBR). References [1] N. Ferronato, V. Torretta, Int. J. Environ. Res. Public Health, 16, 6 (2019). [2] N. Ferronato, E.C. Rada, M.A. Gorritty Portillo, L.I. Cioca, M. Ragazzi, V. Torretta, J. Environ Manage, 15 (2019). [3] G.P. Maitan-Alfenas, E.M. Visser, V.M. Guimarães, Cur. Opin. Food Scien., 1 (2015). [4] M.V. Kharina, O.N. Grigoreva, Proceedings of Universities, Applied Chemistry and Biotechnology, 20, 13 (2017) (in Russian). [5] O.I. Bolotnikova, N.P. Mikhailova, J.G. Bazarnova, E.B. Aronova, T.A. Bolotnikova, J.N. Akinina, Proceedings of Universities, Applied Chemistry and Biotechnology, 9, 4 (2019) (in Russian). [6] V.I. Kudryavcev, Katalog kul'tur mikroorganizmov, podderzhivaemyh v SSSR, M: Nauka, 1981, 244 (In Russian). [7] O. Bolotnikova, J. Bazarnova, E. Aronova, N. Mikhailova, T. Bolotnikova, Pu J. E3S Web of Conferences, 140, 02008 (2019). DOI: 10.1051/e3sconf/201914002008. [8] L. Simão1, D. Hotza, F. Raupp-Pereira1, J.A. Labrincha, O.R.K. Montedo, Cerâmica, 64, 371 (2018). DOI: 10.1590/0366-69132018643712414. [9] M. Yılmaz, T. Gümüşkaya, Europ. J. Accounting Auditing and Finanace Research, 3, 4 (2015). [10] O.I. Bolotnikova, N.P. Mikhailova, A.I. Ginak, Izvestiya Sankt-Peterburgskogo gosudarstvennogo tekhnologicheskogo instituta (tekhnicheskogo universiteta), 39, 89 (2017) (In Russian). [11] A. Nancib, N. Diboune, N. Nancib, J. Boudrant, J. Appl. Biotech. Bioeng., 4, 1 (2017). EISSN: 2572- 8466. [12] C.P. Kurtzman, J.W. Fell, T. Boekhout (eds.), The Yeasts A Taxonomic Study, 5th Edition, Elsevier, Amsterdam, Netherlands, 2354 (2011). DOI: 10.1016/B978-0-444-52149-1.00007-0. 3
https://openalex.org/W4288685594	https://journal.uokufa.edu.iq/index.php/ksc/article/download/9712/9309	Arabic	null	تأثير الوعي الاجتماعي للمستثمر على قرارات الاستثمار دراسة استطلاعية لآراء عينة من المستثمرين في سوق العراق للأوراق المالية	Mağallaẗ markaz dirāsāt al-Kūfaẗ	2,022	cc-by	10,719	المقدمة: في الوقت الحاضر تغيرت ظروف العمل في العالم بوتيرة سريعة جدًا تأثرت على أثرها عملية اتخاذ القرار في مختلف المجاالت. وتطلبت وتيرة النمو هذه أداءً عاليًا من المديرين وصناع القرار بسبب التغيير السريع في التكنولوجيا والمنافسة بين الشركات، أن وتيرة النمو هذه لها تأثير كبير على سوق االوراق المالية والمستثمرين العاملين في هذه األسواق، اذ يضطر المستثمرون أو صناع القرار المرتبطون بهذه األسواق المالية إلى تبني هذه التغييرات في األسواق المالية. ومن اجل الحصول على نتائج أفضل للقرارات، فانه ليس من الضروري أن يركز صانعوا القرار االستثمار ي في عملهم على الصعيد المادي فقط ولكن المطلوب منهم التركيز على العوامل االجتماعية ، وكيفية توظيفها بذكاء، وحتى تتمكن األسواق االوراق المالية من مواجهة الصعوبات والعقبات بسبب التغيرات التي توجهها والتي تؤثر على قرارات المستثمرين فيه فقد ظهرت الحاجة الى ايجاد حلول في هذا الجانب وذلك من خالل مقدرات الوعي االجتماعي. ان الوعي االجتماعي له تأ ثير في عملية اتخاذ القرارات االستثمارية، اذ كلما تمتع المستثمرون بوعي اجتماعي ،عال، كانت نتائجهم جيدة من حيث االرباح الناتجة عن عملياتهم االستثمارية وبالعكس بالنسبة للمستثمرين ذوي الوعي االجتماعي المنخفض واذ تكون نتائجهم من عملياتهم االستثمارية غير جيدة، و هذا بدورة ينعكس على االسواق المالية سواء كانت نتائج جيدة او غير جيدة، ومن هذا المنطلق يظهر دور واهمية الوعي االجتماعي للمستثمرين في قراراتهم االستثمارية، لذا كانت الحاجة ضرورية لمعرفة مدى في الوقت الحاضر تغيرت ظروف العمل في العالم بوتيرة سريعة جدًا تأثرت على أثرها عملية اتخاذ القرار في مختلف المجاالت. وتطلبت وتيرة النمو هذه أداءً عاليًا من المديرين وصناع القرار بسبب التغيير السريع في التكنولوجيا والمنافسة بين الشركات، أن وتيرة النمو هذه لها تأثير كبير على سوق االوراق المالية والمستثمرين العاملين في هذه األسواق، اذ يضطر المستثمرون أو صناع القرار المرتبطون بهذه األسواق المالية إلى تبني هذه التغييرات في األسواق المالية. ومن اجل الحصول على نتائج أفضل للقرارات، فانه ليس من الضروري أن يركز صانعوا القرار االستثمار ي في عملهم على الصعيد المادي فقط ولكن المطلوب منهم التركيز على العوامل االجتماعية ، وكيفية توظيفها بذكاء، وحتى تتمكن األسواق االوراق المالية من مواجهة الصعوبات والعقبات بسبب التغيرات التي توجهها والتي تؤثر على قرارات المستثمرين فيه فقد ظهرت الحاجة الى ايجاد حلول في هذا الجانب وذلك من خالل مقدرات الوعي االجتماعي. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثير الوعي االجتماعي للمستثمر على قرارات االستثمار دراسة استطالعية آلراء عينة من المستثمرين في سوق العراق لألوراق المالية أ.م.د. زينب هادي الشريفي الباحث حسين بديوي محمد كلية اإلدارة واالقتصاد/ جامعة الكوفة تأثير الوعي االجتماعي للمستثمر على قرارات االستثمار دراسة استطالعية آلراء عينة من المستثمرين في سوق العراق لألوراق المالية أ.م.د. زينب هادي الشريفي الباحث حسين بديوي محمد كلية اإلدارة واالقتصاد/ جامعة الكوفة المقدمة: ان الوعي االجتماعي له تأ ثير في عملية اتخاذ القرارات االستثمارية، اذ كلما تمتع المستثمرون بوعي اجتماعي ،عال، كانت نتائجهم جيدة من حيث االرباح الناتجة عن عملياتهم االستثمارية وبالعكس بالنسبة للمستثمرين ذوي الوعي االجتماعي المنخفض واذ تكون نتائجهم من عملياتهم االستثمارية غير جيدة، و هذا بدورة ينعكس على االسواق المالية سواء كانت نتائج جيدة او غير جيدة، ومن هذا المنطلق يظهر دور واهمية الوعي االجتماعي للمستثمرين في قراراتهم االستثمارية، لذا كانت الحاجة ضرورية لمعرفة مدى العدد65 ج1 حزيران2022 271 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 2 . تمكين المستثمرين على تحسين مستوى وعيهم االجتماعي ، وانعكاس ذلك في قراراتهم االستثمارية .المتخذة 2 . تمكين المستثمرين على تحسين مستوى وعيهم االجتماعي ، وانعكاس ذلك في قراراتهم االستثمارية .المتخذة 3 . أهمية المتغيرات المزمع دراستها وهما الوعي االجتماعي وقرارات االستثمار. 4 . تساعد الدراسة الحالية المستثمرين على اتخاذ القرارات ذات المساس باألخرين الذين يعملون داخل سوق االوراق المالية التي يَنعكس اثارها في قراراتهم المتخذة وفي أ.داء سوق االوراق المالية 3 - اه:داف الدراسة :تهدف الدراسة الحالية الى تحقيق األهداف االتية 1 . الكشف عن مدى توافر مقدرات الوعي االجتماعي .لدى المستثمرين في سوق العراق لألوراق المالية 2 . الكشف عن العالقات المتوقعة بين كل من الوعي االجتماعي وقرارات االستثمار، والنتائج التي تتولد عن هذه.العالقة 3 . أهمية المتغيرات المزمع دراستها وهما الوعي االجتماعي وقرارات االستثمار. 4 . تساعد الدراسة الحالية المستثمرين على اتخاذ القرارات ذات المساس باألخرين الذين يعملون داخل سوق االوراق المالية التي يَنعكس اثارها في قراراتهم المتخذة وفي أ.داء سوق االوراق المالية سوق االوراق المالية التي يَنعكس اثارها في قراراتهم المتخذة وفي أ.داء سوق االوراق المالية 3 - اه:داف الدراسة :تهدف الدراسة الحالية الى تحقيق األهداف االتية 1 . الكشف عن مدى توافر مقدرات الوعي االجتماعي .لدى المستثمرين في سوق العراق لألوراق المالية 2 . الكشف عن العالقات المتوقعة بين كل من الوعي االجتماعي وقرارات االستثمار، والنتائج التي تتولد عن هذه.العالقة 2 . الكشف عن العالقات المتوقعة بين كل من الوعي االجتماعي وقرارات االستثمار، والنتائج التي تتولد عن هذه.العالقة ا 3 . " التعرف على مدى تأثر قرارات االستثمار للمستثمرين بمستوى وعيهم االجتماعي. 4 - نموذج الفرضي للدراسة وفرضياتها :في ضوء مشكلة الدراسة وأهدافها تم بناء المخطط الفرضي للدراسة الذي يوضح عالقات االرتباط والتأثير بين متغيرات الدراسة والتي تتمثل بمتغيرين رئيسين هما الوعي االجتماعي بوصفة متغير مستقالً وقرارات االستثمار بوصفة متغيراً تابعاً، اذ تمثل حركة السهم ذي االتجاهين الى عالقة االرتباط بين الوعي االجتماعي وقرارات االستثمار ، اما حركة السهم ذي االتج ( الواحد فتشير الى تأثير المتغير المستقل الوعي االجتماعي )) في المتغير التابع ( قرارات االستثمار ( ويصور شكل1 ) المخطط فرضي للدراسة، وهو عبارة عن تفاعل حركي وسببي بين مكو نين فكريي :يمكن توضيحهما كما يأتي ( شكل رقم1) المخطط الفرضي لفرضيات للدراسة الوعي االجتماعي قرارات االستثمار 3 . " التعرف على مدى تأثر قرارات االستثمار للمستثمرين بمستوى وعيهم االجتماعي. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار اهمية العوامل االجتماعية لل مستثمرين وتأثيرها على قرارتهم وخاصة في ظل المتغيرات التي شهدها العالم .وتأثيرها على اداء االسواق االوراق المالية وتضمنت الدراسة اربعة مباحث خصص االول الى منهجية البحث والثاني الى الجانب النظري فيما تناول المبحث الثالث الجانب العملي واخيراً خصص المبحث .الرابع لالستنتاجات والتوصيات 1 - مشكلة الدراسة : ان ، ًصنع القرار عملية مستمرة ومتكررة يوميا على المرء أن يتخذ عددًا من القرارات على أساس يومي، بعض القرارات ذات طبيعة حاسمة وال تؤثر فقط على صانع القرار ولكن أيضً ا على اآلخرين. ان اتخاذ القرار العقالني ليس بالمهمة السهلة، و وفقًا لفرضية السوق الكفؤ ، يكون المستثمرون في السوق الكفؤ على دراية جيدة بح ، الة السوق ويتخذون قرارات عقالنية، ولكن وفقًا للتمويل السلوكي هناك عدد من التحيزات التي تشكل عقبة في عملية اتخاذ القرار ، و الوعي االجتماعي هي واحدة من تلك التي قد يفشل المستثمرون بسببها في اتخاذ قرار عقالني اذا لم يكن لهم وعي عال يمكن استغالله بالشكل الصح يح، حيث يواجه المستثمرون مشكلة رئيسية أثناء اتخاذ قرارات استثمارية تتمثل في عدم قدرتهم على إدارة وعيهم االجتماعي ، ومن خالل ما تقدم يمكن تأطير مشكلة الدراسة بالسؤال الرئيسي اآل تي: هل يؤثر الوعي االجتماعي للمستثمر على قراراته االستثمارية ؟ ومن هذا التساؤل ا:لرئيسي تتفرع التساؤالت التالية 1 . هل يمتلك المستثمرون مقدرات الوعي االجتماعي ؟ 2 . هل توجد عالقة بين الوعي االجتماعي و قرارات االستثمار ؟ 3 . ما اثر امتالك مقدرات الوعي االجتماعي على قراراتهم االستثمارية ؟ 4 . ما هو مستوى الوعي االجتماعي للعينة المبحوثة ؟ 2 - اهمية الدراسة : :تكمن أهمية الدراسة بما يأتي 1 . تعزيز قدرات المستثمرين في ممارسة الوعي االجتماعي .عند اتخاذهم قرارات االستثمار العدد65 ج1 حزيران2022 272 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار والجدول1 :) يوضح االتي " ( جدول1) مجتمع الدراسة من القطاعات المدرجة في سوق العراق لألوراق المالية ت اسم القطاع عدد الشركات المدرجة في القطاع 1 قطاع المصرفي 42 2 قطاع التأمين 5 3 قطاع االستثمار 6 4 قطاع الخدمات 10 5 القطاع الصناعي 21 6 قطاع الفنادق والسياحة 10 7 القطاع الزراعة 7 8 قطاع االتصاالت 2 المجموع 103 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 4 - نموذج الفرضي للدراسة وفرضياتها : في ضوء مشكلة الدراسة وأهدافها تم بناء المخطط الفرضي : للدراسة الذي يوضح عالقات االرتباط والتأثير بين متغيرات الدراسة والتي تتمثل بمتغيرين رئيسين هما الوعي االجتماعي بوصفة متغير مستقالً وقرارات االستثمار بوصفة متغيراً تابعاً، اذ تمثل حركة السهم ذي االتجاهين الى عالقة االرتباط بين الوعي االجتماعي وقرارات االستثمار ، اما حركة السهم ذي االتجاه ( الواحد فتشير الى تأثير المتغير المستقل الوعي االجتماعي ،)) في المتغير التابع ( قرارات االستثمار ( ويصور شكل1 ) المخطط فرضي للدراسة، وهو عبارة عن تفاعل حركي وسببي بين مكو نين فكريين :يمكن توضيحهما كما يأتي ( شكل رقم1) المخطط الفرضي لفرضيات للدراسة الوعي االجتماعي قرارات االستثمار 3 . " التعرف على مدى تأثر قرارات االستثمار للمستثمرين بمستوى وعيهم االجتماعي. 4 - نموذج الفرضي للدراسة وفرضياتها : في ضوء مشكلة الدراسة وأهدافها تم بناء المخطط الفرضي : للدراسة الذي يوضح عالقات االرتباط والتأثير بين متغيرات الدراسة والتي تتمثل بمتغيرين رئيسين هما الوعي االجتماعي بوصفة متغير مستقالً وقرارات االستثمار بوصفة متغيراً تابعاً، اذ تمثل حركة السهم ذي االتجاهين الى عالقة االرتباط بين الوعي االجتماعي وقرارات االستثمار ، اما حركة السهم ذي االتجاه ( الواحد فتشير الى تأثير المتغير المستقل الوعي االجتماعي ،)) في المتغير التابع ( قرارات االستثمار ( ويصور شكل1 ) المخطط فرضي للدراسة، وهو عبارة عن تفاعل حركي وسببي بين مكو نين فكريين قرارات االستثمار الوعي االجتماعي العدد65 ج1 حزيران2022 273 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المصدر: من اعداد الباحث تمممم صمممياغة فرضمممية الدراسمممة علمممى وفمممق مشمممكلة الدراسمممة وأهمممدافها ومخططهممما ويمكمممن اسمممتعرا فرضممميات : الدراسة فيما يأتي- الفرضية : االولى- " توجد عالقة ارتباط ذات داللة معنوية بين.الوعي االجتماعي وقرارات االستثمار الفرضية الثانية: - توجد عالقة تأثير ذات داللة معنوية ب ين.الوعي االجتماعي وقرارات االستثمار 5 - مجتمع الدراسة وعينتها :لإليفاء بمتطلبات الجانب التطبيقي للدراسة الحالية، وتحقيق أهدافها ًومساعيها، تم اختيار سوق العراق لألوراق المالية ليكون مجتمعا للدراسة ، وشملت كل القطاعات المدرجة ( في سوق العراق لألوراق المالية وعددها8 ( ) وكما هو موضحة في جدول1 )، أما عينة الدراسة فكانت هي المستثمر ين المسؤولين عن اتخاذ القرار في المؤسسات المالية الموجودة في سوق العراق لألوراق ( المالية. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المصدر: من اعداد الباحث تمممم صمممياغة فرضمممية الدراسمممة علمممى وفمممق مشمممكلة الدراسمممة وأهمممدافها ومخططهممما ويمكمممن اسمممتعرا فرضممميات : الدراسة فيما يأتي- الفرضية : االولى- " توجد عالقة ارتباط ذات داللة معنوية بين.الوعي االجتماعي وقرارات االستثمار الفرضية الثانية: - توجد عالقة تأثير ذات داللة معنوية ب ين.الوعي االجتماعي وقرارات االستثمار 5 - مجتمع الدراسة وعينتها :لإليفاء بمتطلبات الجانب التطبيقي للدراسة الحالية، وتحقيق أهدافها ًومساعيها، تم اختيار سوق العراق لألوراق المالية ليكون مجتمعا للدراسة ، وشملت كل القطاعات المدرجة ( في سوق العراق لألوراق المالية وعددها8 ( ) وكما هو موضحة في جدول1 )، أما عينة الدراسة فكانت هي المستثمر ين المسؤولين عن اتخاذ القرار في المؤسسات المالية الموجودة في سوق العراق لألوراق ( المالية. والجدول1 :) يوضح االتي " ا ( جدول1) مجتمع الدراسة من القطاعات المدرجة في سوق العراق لألوراق المالية ت اسم القطاع عدد الشركات المدرجة في القطاع 1 قطاع المصرفي 42 2 قطاع التأمين 5 3 قطاع االستثمار 6 4 قطاع الخدمات 10 5 القطاع الصناعي 21 6 قطاع الفنادق والسياحة 10 7 القطاع الزراعة 7 8 قطاع االتصاالت 2 المجموع 103 العدد65 ج1 حزيران2022 274 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .المصدر: من اعداد الباحث باالعتماد على النشرات والتقارير الخاصة بسوق العراق لألوراق المالية 6 - أداة الدراسة: طورت ألغرا الدراسة الحالية استمارة استبيان ، اعدت الستكمال الجوانب التطبيقية منها ،اذ تضمنت االستمارة جانبين، األول اشتمل على المعلومات التعريفية للمستثمرين، اذ اشتمل على ( 5 .) فقرات اما الجانب الثاني فقد ضم محورين، االول لقياس الوعي االجتماعي تم تفسيرها (ب10 ) فقرة، ( جرى االعتماد في اعدادها على دراسةSilvita, 2017 ) والثاني لقرارات االستثمار ، تم تفسيرها من ( خالل4 ) (متغيرات مستقلة مفسرة ب27) فقرة، ( جرى االعتماد في اعدادها على دراسةAburas & Ab Hamid, 2013 ) ( وتم استخدام مدرج تفضيلLikert) الخماسي الذي يتصف بدقة القياس. المبحث الثاني: اإلطار النظري للبحث: 1 - :مفهوم الوعي االجتماعي في األساس، يمكّن الوعي االجتماعي الفرد من مراقبة وتعديل االستراتيجية والتوجيه والعمل نحو تحقيق رؤية مشتركة. ويساعد الفرد على معرفة متى يجب التقدم واالستفادة من زخم المجموعة ومتى يجب ان يتراجع، ويشجع التفكير ومراجعة اوضاع الجماعة و معرفة االخطاء وتجنبها للتحقيق االهداف واألولويات(Ramsey, 2014; 87) . ان الوعي االجتماعي هو قدرة الفرد على التقاط مشاعر االخرين بدقة وفهم ما يحدث معهم. ويعني هذا غالبًا إدراك ما يفكر فيه اآلخرون ويشعرون به حتى لو لم تشعر بنفس الشعور. ومن السهل أن تنغمس في م شاعرك وتنسى التفكير في منظور الطرف اآلخر. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ويعرف الباحث الوعي االجتماعي بأنه " هو قدرة الفرد على فهم ما يشعر به اآلخرون م ن حوله ." وتحديد مشاعرهم والعمل على مساعدتهم في الظروف التي يحتاجون فيها للمساعدة 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, ويعرف الباحث الوعي االجتماعي بأنه " هو قدرة الفرد على فهم ما يشعر به اآلخرون م ن حوله ." وتحديد مشاعرهم والعمل على مساعدتهم في الظروف التي يحتاجون فيها للمساعدة ويعرف الباحث الوعي االجتماعي بأنه " هو قدرة الفرد على فهم ما يشعر به اآلخرون م ن حوله ." وتحديد مشاعرهم والعمل على مساعدتهم في الظروف التي يحتاجون فيها للمساعدة 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, 2011;48 ): ي م ي م 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, 2011;48 ): ي ي 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, 2011;48 ): 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, 2011;48 ): 1 . ( التعاطفempathy): وهو القدرة على معرفة ما يجول في اذهان االخرين، ومن ثم تحفيزهم على تقديم أداء اكثر مما هو متوقع او مطلوب منهم، يتمكن المدير على سبيل المثال من تحفيز االفراد لما .يمتلكه من قدرة في الوصول الى مستويات تفكيرهم ومخاطبتهم باللغة التي يفهمونها 1 . ( التعاطفempathy): وهو القدرة على معرفة ما يجول في اذهان االخرين، ومن ثم تحفيزهم على تقديم أداء اكثر مما هو متوقع او مطلوب منهم، يتمكن المدير على سبيل المثال من تحفيز االفراد لما .يمتلكه من قدرة في الوصول الى مستويات تفكيرهم ومخاطبتهم باللغة التي يفهمونها 2 . ( الوعي التنظيميOrganizational awareness .): هو الشعور باآلخرين واإلحساس بهم .وبسبب هذه القدرة على اإلحساس باألخرين يمكن معرفة كيفية التفاعل معهم 2 . ( الوعي التنظيميOrganizational awareness .): هو الشعور باآلخرين واإلحساس بهم .وبسبب هذه القدرة على اإلحساس باألخرين يمكن معرفة كيفية التفاعل معهم أإ 3 . ( توجه الخدمةService orientation ): القدرة واالستعداد على تقديم المساعدة لألخرين عند .الحاجة 3 - :خصائص الوعي االجتماعي يتميز الوعي االجتماعي بعدد من الخصائص وهي ،(القوس2018 : 8 - 9 ): اا 3 - :خصائص الوعي االجتماعي يتميز الوعي االجتماعي بعدد من الخصائص وهي ،(القوس2018 : 8 - 9 ): 1 . :مكتسب وهي التي يتم اكتسابها من البيئة المحيطة ،وسائل التواصل ، والعالقات االجتماعية بي .االفراد 1 . تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار يضمن الوعي االجتماعي للفرد الحفاظ على ا لتركيز واستيعاب المعلومات المهمة (Bradberry & Greaves, 2009; 35) . ويمكن تعريف الوعي االجتماعي وفقًا لجولمان بأنه "القدرة على فهم سلوك "اآلخرين ودوافعهم ، وتطوير التعاطف معهم; (Goleman, 2001; 35 ) (Fatum, 2008; 36 . كما يعرف بأنه " القدرة على إدراك وفهم ما قد يفكر فيه اآلخرون ويشعرون به بدقة حتى لو لم يتم مشاركة "هذه المشاعر(Hohlbein, 2015; 20) . العدد65 ج1 حزيران2022 275 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ويعرف الباحث الوعي االجتماعي بأنه " هو قدرة الفرد على فهم ما يشعر به اآلخرون م ن حوله ." وتحديد مشاعرهم والعمل على مساعدتهم في الظروف التي يحتاجون فيها للمساعدة 2 - :مقدرات الوعي االجتماعي يحتوي بعُد الوعي االجتماعي على ثالث مقدرات هي ( Goleman, 2011;48 ): 1 . ( التعاطفempathy): وهو القدرة على معرفة ما يجول في اذهان االخرين، ومن ثم تحفيزهم على تقديم أداء اكثر مما هو متوقع او مطلوب منهم، يتمكن المدير على سبيل المثال من تحفيز االفراد لما .يمتلكه من قدرة في الوصول الى مستويات تفكيرهم ومخاطبتهم باللغة التي يفهمونها 2 . ( الوعي التنظيميOrganizational awareness .): هو الشعور باآلخرين واإلحساس بهم .وبسبب هذه القدرة على اإلحساس باألخرين يمكن معرفة كيفية التفاعل معهم 3 . ( توجه الخدمةService orientation ): القدرة واالستعداد على تقديم المساعدة لألخرين عند .الحاجة اا تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار :مكتسب وهي التي يتم اكتسابها من البيئة المحيطة ،وسائل التواصل ، والعالقات االجتماعية بي .االفراد العدد65 ج1 حزيران2022 276 2 . .الشمولية: ان الوعي االجتماعي يرتبط بالوجود االجتماعي ككل 3 . :التنوع ان الوعي االجتماعي يختلف من فرد آلخر ومن مجتمع آلخر تبعاً الختالف المفاهيم المهيمنة على المجتمع واختالف فهمها و كذلك.تفسيرها 4 . .النسبية: الوعي االجتماعي يتميز بالتغير النسبي وعدم الثبات المطلق 5 . القابلية للتطور والتجديد: حسب تطور العالقات االجتماعية ووسائل التواصل التي يستخدمها اإلنسان في كل عصر، كما يشمل ذلك القابلية للتزييف .وتشويه 2 . .الشمولية: ان الوعي االجتماعي يرتبط بالوجود االجتماعي ككل 3 . :التنوع ان الوعي االجتماعي يختلف من فرد اآلخر ومن مجتمع آلخر تب على المجتمع واختالف فهمها و كذلك.تفسيرها 2 . .الشمولية: ان الوعي االجتماعي يرتبط بالوجود االجتماعي ككل 3 . :التنوع ان الوعي االجتماعي يختلف من فرد آلخر ومن مجتمع آلخر تبعاً الختالف المفاهيم المهيمنة على المجتمع واختالف فهمها و كذلك.تفسيرها 4 . .النسبية: الوعي االجتماعي يتميز بالتغير النسبي وعدم الثبات المطلق 5 . القابلية للتطور والتجديد: حسب تطور العالقات االجتماعية ووسائل التواصل التي يستخدمها اإلنسان في كل عصر، كما يشمل ذلك القابلية للتزييف .وتشويه العدد65 ج1 حزيران2022 276 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 6 . .التعقيد: ان الخصائص السابقة تدل على الطبيعة المعقدة للوعي االجتماعي 7 . فهم المجتمع: يكسب الفرد القدرة على تفسير وتحليل قضايا مجتمعه المعاصرة والوصول لحلول لها فهو وليد فهم أفراد المجتمع لثقافتهم وقيمهم .وتاريخهم وحاضرهم 8 . .ايجابي: في تعمله مع قضايا المجتمع ثانياً: قرارات االستثمار ا 1 - مفهوم قرارات االستثمار : حظي موضوع اتخاذ القرار باهتمام بالغ في العديد من نواحي العلوم كعلم الفلسفة، وعلم النفس، وعلم ادارة االعمال، وعلم االقتصاد، وغيرها من العلوم االخرى، وق د ظهرت العديد من النظريات في المجاالت النفسية حول عملية اتخاذ القرار في العام1954 اكدت هذه النظريات على أسباب اتخاذ القرار معللة هذه العملية بانها تحدث نتاج التحديات واالحتماالت التي يعيشها الفرد، وان اختالف المواقف يتطلب اختالفا في عملية اتخاذ القرار (ال ،منصور2015 : 60 .) أورد الباحثون العديد ،من التعريفات بشأنه، فتعرفه كل من (ماجدة2014 : 75 ،; حنان وسمية2019 : 12 ) بأنه " عملية اختيار البديل االستثماري الذي يحقق اعلى عائد من بين العديد من البدائل، وتسبق عملية االختيار دراسات الجدوى التي في ضوئها يجري اختيار البديل المالئم الذي يحقق اهداف المستثمر ويتفق مع ،طبيعته". ويرى (الجعارات2007) بان القرار االستثماري هو القرا ر الذي يترتب عليه تخصيص جزء من األموال الخاصة بالمؤسسة من اجل استثمارها بهدف الحصول على األرباح مستقبال، ويتعر هذا ،االستثمار للمخاطر وعدم التأكد وبدرجات حدة متفاوتة" (شهيرة و نزيهة2019 : 14 .) 2 - اهمية االستثمار: يسهم اإلنفاق االستثماري بشكل مباشر في النش اط االقتصادي ألن االستثمار هو العنصر األكثر تقلباً في الناتج المحلي اإلجمالي، ويلعب االستثمار دورًا حيويًا على المدى الطويل والنمو قصير المدى، حيث يربط الحاضر بالمستقبل، واالستثمار هو جزء من التخطيط المالي العام، فإذا كان لدينا بعض المدخرات ، فسنحاول االست( ثمار لتعظيم العائدAhmad et al, 2012; 680 .) ان اهمية ،االستثمار يمكن ان نلخصها بما يلي: ( العبودي2020 : 223 :) ي ( ب ما) 2 - اهمية االستثمار: يسهم اإلنفاق االستثماري بشكل مباشر في النش اط االقتصادي ألن االستثمار هو العنصر األكثر تقلباً في الناتج المحلي اإلجمالي، ويلعب االستثمار دورًا حيويًا على المدى الطويل والنمو قصير المدى، حيث يربط الحاضر بالمستقبل، واالستثمار هو جزء من التخطيط المالي العام، فإذا كان لدينا بعض المدخرات ، فسنحاول االست( ثمار لتعظيم العائدAhmad et al, 2012; 680 .) ان اهمية ،االستثمار يمكن ان نلخصها بما يلي: ( العبودي2020 : 223 :) العدد65 ج1 حزيران2022 277 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار • يؤدي االستثمار الى زيادة انتاجية المؤسسات، والذي يؤدي بدورة الى زيادة الدخل القومي وارتفاع .المستوى المعيشي للمواطنين • ، تأمين الخدمات للمستثمرين.والمواطنين • .تأمين الفرص العمل للعاملين والحد من ظاهرة البطالة • .تأمين كافة القدرات والتخصصات واإلمكانيات المالية واالدارية والفنية وكذلك العمال المهرة 3 - ( ابعاد قرارات االستثمار: بحسبAburas & Ab Hamid, 2013) فان هناك أربعة ابعاد للقرار ( االستثماري وكما الموضحة في الشكل2 :)، وفيما يلي توضيح لكل بعد منها ( الشكل رقم2 ) أبعاد قرارات االستثمار المصدر: من اعداد الباحث باالعتماد على السابقة الذكر قرارات االستثمار القدرة المالية الفائدة المدركة المنتجات وسمات الشركة المعرفة والخبرات السابقة • يؤدي االستثمار الى زيادة انتاجية المؤسسات، والذي يؤدي بدورة الى زيادة الدخل القومي وارتفاع .المستوى المعيشي للمواطنين • يؤدي االستثمار الى زيادة انتاجية المؤسسات، والذي يؤدي بدورة الى زيادة الدخل القومي وارتفاع .المستوى المعيشي للمواطنين • ، تأمين الخدمات للمستثمرين.والمواطنين • .تأمين الفرص العمل للعاملين والحد من ظاهرة البطالة • .تأمين كافة القدرات والتخصصات واإلمكانيات المالية واالدارية والفنية وكذلك العمال المهرة 3 - ( ابعاد قرارات االستثمار: بحسبAburas & Ab Hamid, 2013) فان هناك أربعة ابعاد للقرار ( االستثماري وكما الموضحة في الشكل2 :)، وفيما يلي توضيح لكل بعد منها رين ين ين و و • .تأمين الفرص العمل للعاملين والحد من ظاهرة البطالة • .تأمين كافة القدرات والتخصصات واإلمكانيات المالية واالدارية والفنية وكذلك العمال المهرة 3 - ( ابعاد قرارات االستثمار: بحسبAburas & Ab Hamid, 2013) فان هناك أربعة ابعاد للقرار ( االستثماري وكما الموضحة في الشكل2 :)، وفيما يلي توضيح لكل بعد منها ( الشكل رقم2 ) أبعاد قرارات االستثمار المصدر: من اعداد الباحث باالعتماد على السابقة الذكر قرارات االستثمار القدرة المالية الفائدة المدركة المنتجات وسمات الشركة المعرفة والخبرات السابقة العدد65 ج1 حزيران2022 278 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار أ- ( القدرة الماليةFinancial ability ): تعد القدرة المالية للمستثمر من أهم المتطلبات للمشاركة في عملية االستثمار في سوق األوراق المالية. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار اذ يمثل القرار االستثمار في البورصة الرغبة في االست حواذ على أسهم الشركة أو األسهم المعروضة للتداول، وفي أغلب األحيان ال يرتبط قرار االستثمار بالتكلفة التي يتحملها الفرد المستثمر، اذ يعتمد على رغبته في الحصول على االسهم والتي تعتمد على القدرة المالية للمستثمر لشراء االسهم عندها يتحول المستثمر الى الطلب ا لفعال، والطلب الفعال هو الرغبة في الحصول على األسهم التي ترتبط بالقدرة المالية، ومن ثم فإن الطلب الفعال يتكون من عنصرين هما ( الرغبة في االستحواذ على األسهم والقدرة المالية على شراء األسهمAburas & Ab Hamid, 2013; 71-72 .) أ- ( القدرة الماليةFinancial ability ): تعد القدرة المالية للمستثمر من أهم المتطلبات للمشاركة في عملية االستثمار في سوق األوراق المالية. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار أ- " ( مسممممتوى التحصمممميل العلمممممي: وتشممممير النتممممائج فممممي الجممممدول3 ) التكممممرارات ونسممممبتها لمسممممتوى التحصمممميل :العلمي ( جدول3 ): توزيع أفراد العينة حسب المؤهل العلمي المؤهل العلمي اعدادية دبلوم عالي بكالوريوس ماجستير دكتوراه التكرار 8 18 30 17 12 % النسبة 9.4 % 21.1 % 35.3 % 20 % 14.2 % المصدر : إعداد الباحث على وفق نتائج الدراسة ( ممممممن خمممممالل جمممممدول3 ) يتبمممممين ممممممن الدراسمممممة ان نسممممممبة افمممممراد العينمممممة المممممذين مسمممممتوى التحصممممميل العلمممممممي ( لمممممديهم (بكمممممالوريوس) هممممممي النسمممممبة األعلممممممى والتمممممي بلغممممممت35.3 % )، امممممما اقممممممل نسمممممبة فقممممممد كانمممممت عنممممممد تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( جدول2) : توزيع أفراد العينة حسب النوع االجتماعي والفئات العمرية النوع االجتماعي )الفئة العمرية (سنة ذكر انثى 30 سنة فأقل 31 - 40 41 - 50 51 - فأكثر التكرار 80 5 5 29 11 40 %النسبة 94.1 % 5.9 % 5.9 % 34.1 % 13 % 47 % المصمممممدر : إعداد الباحث على وفق نتائج الدراسة ( جدول2) : توزيع أفراد العينة حسب النوع االجتماعي والفئات العمرية النوع االجتماعي )الفئة العمرية (سنة ذكر انثى 30 سنة فأقل 31 - 40 41 - 50 51 - فأكثر التكرار 80 5 5 29 11 40 %النسبة 94.1 % 5.9 % 5.9 % 34.1 % 13 % 47 % المصمممممدر : إعداد الباحث على وفق نتائج الدراسة تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار اذ يمثل القرار االستثمار في البورصة الرغبة في االست حواذ على أسهم الشركة أو األسهم المعروضة للتداول، وفي أغلب األحيان ال يرتبط قرار االستثمار بالتكلفة التي يتحملها الفرد المستثمر، اذ يعتمد على رغبته في الحصول على االسهم والتي تعتمد على القدرة المالية للمستثمر لشراء االسهم عندها يتحول المستثمر الى الطلب ا لفعال، والطلب الفعال هو الرغبة في الحصول على األسهم التي ترتبط بالقدرة المالية، ومن ثم فإن الطلب الفعال يتكون من عنصرين هما ( الرغبة في االستحواذ على األسهم والقدرة المالية على شراء األسهمAburas & Ab Hamid, 2013; 71-72 .) ب- ( الفائدة المدركةPerceived usefulness : ) إ ن المستثمر عقالني ويحاول تعظيم فائدته من خالل اختيار أفضل بديل من بين البدائل المتاحة، وذلك من اجل تحقيق افضل عائد ممكن وتعظيم االرباح ، كما أن للمشاعر والرؤى التي يشار إليها على أنها تحيزات سلوكية لها تأثير على اتخاذ قرارات ( االستثمار، حيث تدفعه نحو البدائل التي تحقق له افضل فائدة ممكنهMahalakshmi & Anuradha, 2018; 1671 .) ان الفائدة المدركة " هي مدى اعتقاد المستثمر بأن استخدام االموال في ( " االستثمار وفي مشروع معين سوف يعود علية بفوائد وأرباحHamid et al, 2016; 646 .) جم- ( المنتجات وسمات الشركةProducts & Company Attributes :) أن العديد من المستثمرين يختارون األسهم بناءً على مجموعة من المعايير النوعية، وهذا يمثل تحديًا هائالً لمجتمع استثماري معتاد على التحليل والتواصل بشأن القيم النسبية لألوراق المالية، وتشمل هذه المعايير ال ،نوعية سمعة الشركة وحالة الشركة، والمشاعر حول منتجات الشركة ، اذ تعتبر من المحددات الرئيسة الختيار المستثمرين الشركة واالستثمار فيها، فكلما كانت سمعتها ومنتجاتها جيدة زادت رغبتهم في االستثمار في هذه الشركة ( Shiundu, 2009; 9 .) ويرى الباحث ان منتجات و سمات ا لشركة هي معلومات مهمة عن الشركة يسعى المستثمر للحصول عليها قبل اتخاذ القرار االستثماري، مثالً سمعتها وجودة منتجاتها وكفاءة ادارتها العدد65 ج1 حزيران2022 279 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ومقدار توزيع االرباح على اصحاب أالسهم الشركة و كذلك معدل نمو الشركة (األصول ، حقوق ، رأس .)المال ومقدار توزيع االرباح على اصحاب أالسهم الشركة و كذلك معدل نمو الشركة (األصول ، حقوق ، رأس .)المال د- المعرفة والخبرات( السابقةKnowledge & past experiences :) يظهر المستثمرون في اثناء عملية صنع القرار مشاعر مختلفة مثل القلق والخوف والسعادة والجشع وما إلى ذلك والتي تؤثر بدورها على قدرتهم على اتخاذ القرار ونوعية القرارات، كما أنهم يتخذون قرارات بناءً على خبرتهم السابقة ومعرف تهم وتفضيالتهم تجاه المنتجات المالية المختلفة، كل هذه العوامل ال بد أن تؤثر على قرار المستثمر إلى حد كبيرJohnsi & Sunitha, 2019; 22 ( ) ان المعرفة و الخبرة السابقة تؤثر على القرارات االستثمارية التي يتخذها المستثمرين في المستقبل، فمن المنطقي أنه عندما ينتج شيء إيجابي عن قرار ما ، فمن المرجح أن يتخذ المستثمرون القرار في المستقبل بطريقة مماثلة ، في ضوء موقف مشابه ( Qureshi et al, 2012; 283 ،) ويرى الباحث ان المعرفة والخبرة السابقة هي المعرفة التي حصل عليها المستثمر نتيجة الخبرة السابقة، حيث ان الخبرة السابقة ا لغير ناجحة تقلل من فشل المستثمر او التهاون في العمل، اما الخبرة السابقة الناجحة فإنها تعزز من قدرة المستثمر في النجاح في العملية .االستثمارية اوالً: وصف عينة الدراسة : سيتم في هذه الفقرة و صف عينة الدراسة وكما يأتي: .ا( النوع االجتماعي و الفئة العمرية ألفراد عينة الدراسة: النتائج في الجدول2 ) توضح التكرارات ونسبة النوع االجتماعي والفئات العمرية:" العدد65 ج1 حزيران2022 280 280 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( جدول2) : توزيع أفراد العينة حسب النوع االجتماعي والفئات العمرية النوع االجتماعي )الفئة العمرية (سنة ذكر انثى 30 سنة فأقل 31 - 40 41 - 50 51 - فأكثر التكرار 80 5 5 29 11 40 %النسبة 94.1 % 5.9 % 5.9 % 34.1 % 13 % 47 % المصمممممدر : إعداد الباحث على وفق نتائج الدراسة ممممن خمممالل جمممدول ( 2 %( ) يتمممممبين ان عينمممة الدراسمممة حسمممب الفئمممة االجتماعيمممة نسمممبة المممذكور بلغمممت94.1 ) ( وبلغممت نسممبة النسمماء5.9 % ) وهممذا يشممير الممى الفئممة األكبممر التممي تسممتثمر فممي سمموق االوراق الماليممة والتممي (تتمثل بالذكور، اما الفئة العمريمة فكانمت النسمب األعلمى عنمد الفئمة51 - ( فمأكثر) وبنسمبة47 % )، امما اقمل (نسبة فكانت عند الفئة العمرية30 (سنة فأقل) وبنسبة5.9 % )، وهذا يشير الى ان اغلب المستثمرين كمانوا .كبار السن. ( جدول3): توزيع أفراد العينة حسب المؤهل العلمي المؤهل العلمي اعدادية دبلوم عالي بكالوريوس ماجستير دكتوراه التكرار 8 18 30 17 12 % النسبة 9.4 % 21.1 % 35.3 % 20 % 14.2 % ال ا ا ل إ ا ال ا ال 281 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار إذ سمممميتم تشممخيص مسممتوى اسممتجابة العينممة المبحوثممة علممى محمماور االسممتبيان (المتغيممرات واالبعمماد الفرعيممة) ألجممراء عمليممة التحليممل والكشممف عممن طبيعممة توجهممات المسممتجيبين وردراكهممم للمتغيممرات المبحوثممة فممي المنظمممة قيممد ( من خالل جدول4 ) يتبين من الدراسة ان نسبة افراد العينة الذين سنوات خبرتهم ( 11 - 15) همي النسمبة ( األعلى والتي بلغت31.7 % )، اما اقل نسبة فقد كانت عنمد سمنوات الخ بمرة ( 1 - 5 ( ) اذ بلغمت5.9 % ، ) ويتبين من النسب أعاله الفئة األكبر من المستثمرين هم من سنوات خبرتهم ( 11 - 15 .) العدد65 ج1 حزيران2022 282 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار الدراسممة. مممما يتوجممب االعتممماد علممى مجموعممة مممن المؤشممرات الوصممفية الم تمثلممة بممم(الوسط الحسممابي) الممذي يوضح استجابة العينة للمتغيرات المبحوثة ومؤشر(االنحراف المعياري) الذي يبين مدى انحمراف القميم عمن وسطها الحسابي. فضالً عن مؤشر (األهمية النسبية) لكل بعد وفقرة تم قياس المتغيرات من خاللها، إذ تم ( اعتماد قيمة الوسط الفرضي وبقيمة3) أ ي أن الوسط الحسابي المتحقمق المذي يسماوي أو يفموق عمن قيممة )الوسمممط الفرضمممي تعمممد قيممممة مقبولمممة أي (توجمممد اسمممتجابة) وبخالفمممه تعمممد مرفوضمممة أي (ال توجمممد اسمممتجابة )ًباعتمممماد علمممى مقيممماس ليكمممرت خماسمممي الرتمممب (أتفمممق تمامممماً ، أتفمممق، محايمممد ، ال أتفمممق، ال أتفمممق تمامممما :وكاالتي أ : متغير الوعي االجتماعي : يتم قياس هذا المتغير من عدة فقرات :هي كاالتي ( تبين النتائج الموضحة في الجدول5 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لمتغير (الوعي االجتماعي)، إذ يتضح ان المتغير المذكور قد حقق ( استجابة مقبولة وبوسط حسابي3.82 ،) أي أعلى من قيمة الوسط الفرضي المعتمدة مما يدل على أن العينة المبحوثة لديها إدراك اال انه ليس بال مستوى المطلوب لمضمون فقرات المتغير ( المذكور، وبلغت قيمة االنحراف المعياري1.03 ( ) اما النسبة المئوية فقد بلغت0.76 )، أما على مستوى الفقرات التي تم قياس المتغير المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبولة (وبمعدالت متفاوتة نوعا ما ،إذ حققت الفقرة أنا قادر بشكل عام على فهم ما يشعر به المستثمرين من .حولي( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي4.30 ) في حين( حققت الفقرة أستطيع أن .أكون داعمًا عند إعطاء أخبار سيئة للمستثمرين لآلخرين ( ) أقل استجابة وبوسط حسابي3.12 .) تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( مسممممممممتوى التعلمممممممميم ( اعداديممممممممة) اذ بلغممممممممت9.4 % ) ، ويتبممممممممين مممممممممن النسممممممممب أعمممممممماله الفئممممممممة األكبممممممممر مممممممممن المستثمرين هم من حملة البكال وريوس". ب- سمممممنوات الخبمممممرة: وتشمممممير( النتمممممائج فمممممي الجمممممدول4 ) والتمممممي توضمممممح التكمممممرارات ونسمممممبنها للسمممممنوات الخدممممممة ( مسممممممممتوى التعلمممممممميم ( اعداديممممممممة) اذ بلغممممممممت9.4 % ) ، ويتبممممممممين مممممممممن النسممممممممب أعمممممممماله الفئممممممممة األكبممممممممر مممممممممن المستثمرين هم من حملة البكال وريوس". أا المستثمرين هم من حملة البكال وريوس". ب- سمممممنوات الخبمممممرة: وتشمممممير( النتمممممائج فمممممي الجمممممدول4 ) والتمممممي توضمممممح التكمممممرارات ونسمممممبنها للسمممممنوات الخدممممممة . للمستثمرين في سوق العراق لألوراق المالية" م ي ي ب- سمممممنوات الخبمممممرة: وتشمممممير( النتمممممائج فمممممي الجمممممدول4 ) والتمممممي توضمممممح التكمممممرارات ونسمممممبنها للسمممممنوات الخدممممممة . للمستثمرين في سوق العراق لألوراق المالية" سنوات الخدمة سنة فأقل 1 مم5 6 مم10 11 مم15 16 مم20 21 مم فأكثر التكرار 8 5 22 27 12 11 %النسبة 9.4 % 5.9 % 25.8 % 31.7 % 14.2 % 13 % المصدر : إعداد الباحث سنوات الخدمة ( من خالل جدول4 ) يتبين من الدراسة ان نسبة افراد العينة الذين سنوات خبرتهم ( 11 - 15) همي النسمبة ( األعلى والتي بلغت31.7 % )، اما اقل نسبة فقد كانت عنمد سمنوات الخ بمرة ( 1 - 5 ( ) اذ بلغمت5.9 % ، ) ويتبين من النسب أعاله الفئة األكبر من المستثمرين هم من سنوات خبرتهم ( 11 - 15 .) ثانيمماً: التحليممل االحصمممائي الوصممفي : تركمممز هممذه الفقمممرة علممى اختبمممار التحليممل االحصمممائي الوصمم في لتقمممديم ملخص عن استجابة العينة المبحوثة لمتغيرات الدراسة (الرئيسة واالبعاد الفرعية) ويتم ذلك باالعتماد على ،المؤشممرات اإلحصممائية الوصممفية، اذ يمثممل التحليممل الوصممفي جانبمماً مهممماً لوصممف وتحليممل البيانممات الكميممة كونه يعمل علمى وصمف سملوك (العينمة المبحوثمة) أي العينمة التم ي اعتممدت فمي جممع البيانمات، ممما يعطمي الباحممث تصممور وفهممم طبيعممة البيانممات التممي تممم تحليلهمما ويمكممن االعتممماد علممى الجممداول أو االشممكال البيانيممة ( التممممي غالبمممماً ممممما تممممدعم بمناقشممممات تفصمممميلية لتكممممون الصممممورة أوضممممح للقممممارTriola,2012:4 ). تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار الدراسممة. مممما يتوجممب االعتممماد علممى مجموعممة مممن المؤشممرات الوصممفية الم تمثلممة بممم(الوسط الحسممابي) الممذي يوضح استجابة العينة للمتغيرات المبحوثة ومؤشر(االنحراف المعياري) الذي يبين مدى انحمراف القميم عمن وسطها الحسابي. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار فضالً عن مؤشر (األهمية النسبية) لكل بعد وفقرة تم قياس المتغيرات من خاللها، إذ تم ( اعتماد قيمة الوسط الفرضي وبقيمة3) أ ي أن الوسط الحسابي المتحقمق المذي يسماوي أو يفموق عمن قيممة )الوسمممط الفرضمممي تعمممد قيممممة مقبولمممة أي (توجمممد اسمممتجابة) وبخالفمممه تعمممد مرفوضمممة أي (ال توجمممد اسمممتجابة )ًباعتمممماد علمممى مقيممماس ليكمممرت خماسمممي الرتمممب (أتفمممق تمامممماً ، أتفمممق، محايمممد ، ال أتفمممق، ال أتفمممق تمامممما :وكاالتي أ : متغير الوعي االجتماعي : يتم قياس هذا المتغير من عدة فقرات :هي كاالتي ( تبين النتائج الموضحة في الجدول5 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لمتغير (الوعي االجتماعي)، إذ يتضح ان المتغير المذكور قد حقق ( استجابة مقبولة وبوسط حسابي3.82 ،) أي أعلى من قيمة الوسط الفرضي المعتمدة مما يدل على أن العينة المبحوثة لديها إدراك اال انه ليس بال مستوى المطلوب لمضمون فقرات المتغير ( المذكور، وبلغت قيمة االنحراف المعياري1.03 ( ) اما النسبة المئوية فقد بلغت0.76 )، أما على مستوى الفقرات التي تم قياس المتغير المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبولة (وبمعدالت متفاوتة نوعا ما ،إذ حققت الفقرة أنا قادر بشكل عام على فهم ما يشعر به المستثمرين من .حولي( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي4.30 ) في حين( حققت الفقرة أستطيع أن .أكون داعمًا عند إعطاء أخبار سيئة للمستثمرين لآلخرين ( ) أقل استجابة وبوسط حسابي3.12 .) العدد65 ج1 حزيران2022 283 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار (الجدول5 ) قيم مؤشرات اإلحصاء الوصفي لمتغير الوعي االجتماعي ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 أفكر في تأثير قراراتي االستثمارية على .المستثمرون االخرين 4.035 0.944 0.807 2 2 أستطيع أن اعرف بسهولة إذا كان المستثمرون .من حولي قد أصبحوا منزعجين 3.506 1.130 0.701 8 3 استطيع ان اشعر عندما يتغير مزاج اي مستثمر .من حولي 3.471 1.064 0.694 9 4 أستطيع أن أكون داعمًا عند إعطاء أخبار سيئة .للمستثمرين لآلخرين 3.129 1.378 0.626 10 5 أنا قادر بشكل عام على فهم ما يشعر به .المستثمرون من حولي 4.306 0.772 0.861 1 6 يمكن ان يخبرني المستثمرون من حولي عما .يعتري نفوسهم 4.165 0.829 0.833 3 7 يزعجني رؤية المستثمرين اآلخرين منزعجين .نتيجة تعرضهم لخسائر مالية بسبب قراراتهم 3.918 1.049 0.784 5 8 لدي القدرة على معرفة متى ابدي برأيي الى .المستثمرين االخرين ومتى اوقف 3.871 1.044 0.774 6 9 .يهمني ما يحدث للمستثمرين لآلخرين 3.976 1.091 0.795 4 10 أفهم متى تتغير خطط وقرارات المستثمرين .االخرين 3.847 1.018 0.769 7 المعدل العام لبعد الوعي االجتماعي 3.822 1.032 0.764 العدد65 ج1 حزيران2022 284 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 :ب : متغير قرارات االستثمار: يتم قياس هذا المتغير من اربعة ابعاد فرعية هي كاالتي 1 : . تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار القدرة المالية ( تبين النتائج الموضحة في الجدول6 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (القدرة المالية)، إذ يتضح ان ( البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي3.64 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أن العينة المبحوثة لديها إدراك اال انه ليس بالمستوى المطلوب لمضمون فقرات ( البعد المذكور، وبلغت قيمة االنحراف المعياري1.00 ( ) اما النسبة المئوية فقد بلغت0.72 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبولة وبمعدال (ت متقاربة نوعا ما ،إذ حققت الفقرة.تتوافر لدي القدرة المالية على شراء األسهم ) أعلى استجابة ( بين الفقرات األخرى وبوسط حسابي3.78 ( ) في حين حققت الفقرة هناك سهوله في الحصول على األموال من المؤسسات المالية االخرى للعمالء المحتملين ) ( أقل استجابة وبوسط حسابي3.47 .) الجدول ( 6) قيم مؤشرات اإلحصاء الوصفي لبعد القدرة المالية الجدول () قيم مؤشرات اإلحصاء الوصفي لبعد القدرة المالية ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 .تتوافر لدي القدرة المالية على شراء األسهم 3.788 1.036 0.758 1 2 تؤثر سهولة حصولي على األموال المقترضة لشراء .األسهم على قراري االستثماري 3.671 0.905 0.734 2 3 هناك سهوله في الحصول على األموال من .المؤسسات المالية االخرى للعمالء المحتملين 3.471 1.064 0.694 3 المعدل العام لبعد القدرة المالية 3.643 1.002 0.729 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 العدد65 ج1 حزيران2022 285 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 2 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار الفائدة المدركة ( تبين النتائج الموضحة في الجدول76 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (الفائدة المدركة)، إذ يتضح ان ( البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي3.87 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أن العين ة المبحوثة لديها إدراك اال انه ليس بالمستوى المطلوب لمضمون فقرات ( البعد المذكور، وبلغت قيمة االنحراف المعياري0.95 ( ) اما النسبة المئوية فقد بلغت0.77 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبولة وبمعد (الت متفاوتة نوعا ما ،إذ حققت الفقرة الخدمات المالية واالقتصادية المقدمة في سوق االوراق المالي تكون مفيدة( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي4.22 ( ) في حين حققت الفقرة وجود عدد .كبير من العمالء في سوق األوراق المالية يؤثر على قراري االستثماري )أق ( ل استجابة وبوسط حسابي 3.43 .) العدد65 ج1 حزيران2022 286 الجدول ( 7) قيم مؤشرات اإلحصاء الوصفي لبعد الفائدة المدركة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 السياسات االقتصادية التي تضعها الحكومة بشأن العوائد .على االستثمارات تكون واضحة 4.071 0.973 0.814 5 2 الخدمات المالية واالقتصادية المقدمة في سوق .االوراق المالي تكون مفيدة 4.224 0.822 0.845 1 3 التأخر في اصدار االسهم يؤثر على قراري .االستثماري 4.106 0.637 0.821 3 4 أقوم بتنويع استثماراتي من اجل تجنب المخاطر .االستثمارية 4.059 0.696 0.812 6 5 اذ كانت القيمة االسمية لألسهم منخفضة نسبيًا تؤثر4.071 0.784 0.814 4 العدد65 ج1 حزيران2022 286 الجدول ( 7) قيم مؤشرات اإلحصاء الوصفي لبعد الفائدة المدركة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 السياسات االقتصادية التي تضعها الحكومة بشأن العوائد .على االستثمارات تكون واضحة 4.071 0.973 0.814 5 2 الخدمات المالية واالقتصادية المقدمة في سوق .االوراق المالي تكون مفيدة 4.224 0.822 0.845 1 3 التأخر في اصدار االسهم يؤثر على قراري .االستثماري 4.106 0.637 0.821 3 4 أقوم بتنويع استثماراتي من اجل تجنب المخاطر .االستثمارية 4.059 0.696 0.812 6 5 اذ كانت القيمة االسمية لألسهم منخفضة نسبيًا تؤثر4.071 0.784 0.814 4 العدد65 ج1 حزيران2022 العدد65 ج1 حزيران2022 286 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .على قراري االستثماري 6 .يشجعني استقرار السوق المالي على االستثمار 4.094 0.934 0.819 2 7 يؤثر وجود عدد كبير من المستشارين والمحللين الماليين في سوق االوراق المالية على قراري .االستثماري 3.647 0.972 0.729 8 8 وجود عدد كبير من العمالء في سوق األوراق .المالية يؤثر على قراري االستثماري 3.435 1.210 0.687 11 9 توصيات األصدقاء أو زمالء العمل تؤثر على .قراري االستثماري 3.553 1.160 0.711 9 10 تؤثر توصيات الوسيط الذي أتعامل معه على قراري .باالستثمار 3.541 1.419 0.708 10 11 يؤثر وجود سياسات اقتصادية واضحة من قبل .الحكومة على قراري باالستثمار 3.859 0.847 0.772 7 المعدل العام لبعد الفائدة المدركة 3.878 0.950 0.776 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .على قراري االستثماري 6 .يشجعني استقرار السوق المالي على االستثمار 4.094 0.934 0.819 2 7 يؤثر وجود عدد كبير من المستشارين والمحللين الماليين في سوق االوراق المالية على قراري .االستثماري 3.647 0.972 0.729 8 8 وجود عدد كبير من العمالء في سوق األوراق .المالية يؤثر على قراري االستثماري 3.435 1.210 0.687 11 9 توصيات األصدقاء أو زمالء العمل تؤثر على .قراري االستثماري 3.553 1.160 0.711 9 10 تؤثر توصيات الوسيط الذي أتعامل معه على قراري .باالستثمار 3.541 1.419 0.708 10 11 يؤثر وجود سياسات اقتصادية واضحة من قبل .الحكومة على قراري باالستثمار 3.859 0.847 0.772 7 المعدل العام لبعد الفائدة المدركة 3.878 0.950 0.776 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 3 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المنتجات وسمات الشركة ( تبين النتائج الموضحة في الجدول8 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (المنتجات ( وسمات الشركة)، إذ يتضح ان البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي4.00 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أن العينة المبحوثة لديها إدراك واضح لمضمون فقرات ( البعد المذكور، وبلغت قيمة االنحراف المعياري0.82 ( ) اما النسبة المئوية فقد بلغت0.80 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مرتفعة وبمعدالت متفاوتة نوعا م ا ،إذ حققت الفقرة (سمعة مالكي الشركة ومدراءها أمر بالغ األهمية بالنسبة لي ) تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 3 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المعرفة والخبرات السابقة ( تبين النتائج الموضحة في الجدول9 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم(الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (المعرفة والخبرات ( السابقة)، إذ يتضح ان البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي3.69 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أ ن العينة المبحوثة لديها إدراك اال أنه ليس بالمستوى المطلوب ( لمضمون فقرات البعد المذكور، وبلغت قيمة االنحراف المعياري1.07 ( ) اما النسبة المئوية فقد بلغت 0.73 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبول (ة وبمعدالت متقاربة نوعا ما ،إذ حققت الفقرة امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات األسهم( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي3.82 ) في حين ( حققت الفقرة لديّ من المعرفة والخبرات السابقة على ان اتوقع متى ترتفع اسعار االسهم في ا لمستقبل ) ( أقل استجابة وبوسط حسابي3.60 .) الجدول ( 9) قيم مؤشرات اإلحصاء الوصفي لبعد المعرفة والخبرات السابقة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات.األسهم 3.824 0.889 0.765 1 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .االستثماري 10 تلعب سمعة مساهمي الشركة دورًا حيويًا في .قراراتي االستثمارية 4.094 0.750 0.819 3 المعدل العام لبعد المنتجات وسمات الشركة 4.004 0.820 0.801 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 4 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المنتجات وسمات الشركة ( تبين النتائج الموضحة في الجدول8 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم( الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (المنتجات ( وسمات الشركة)، إذ يتضح ان البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي4.00 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أن العينة المبحوثة لديها إدراك واضح لمضمون فقرات ( البعد المذكور، وبلغت قيمة االنحراف المعياري0.82 ( ) اما النسبة المئوية فقد بلغت0.80 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مرتفعة وبمعدالت متفاوتة نوعا م ا ،إذ حققت الفقرة (سمعة مالكي الشركة ومدراءها أمر بالغ األهمية بالنسبة لي ) العدد65 ج1 حزيران2022 287 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( أعلى استجابة بين الفقرات األخرى وبوسط حسابي4.10 ( ) في حين حققت الفقرة المعلومات المطروحة حول عر أسهم الشركة تكون معلومة وشفافة للجميع ) ( أقل استجابة وبوسط حسابي3.75 .) (الجدول8 )قيم مؤشرات اإلحصاء الوصفي لبعد المنتجات وسمات الشركة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 كفاءة ادارة سوق االوراق المالي تدفعني الى .االستثمار في سوق االوراق المالي 4.047 0.925 0.809 5 2 الشركات المدرجة في سوق االوراق المالي عالية .األداء 3.835 0.998 0.767 9 3 المعلومات المطروحة حول عر أسهم الشركة .تكون معلومة وشفافة للجميع 3.753 1.011 0.751 10 4 يعتبر الهيكل اإلداري للشركة عامل مهمًا في .اتخاذ قراري االستثماري 4.082 0.658 0.816 2 5 سمعة مالكي الشركة و مدراءها أمر بالغ األهمية .بالنسبة لي 4.106 0.724 0.821 1 6 يتأثر قراري االستثماري بأرباح الشركات .المتوقعة 4.012 0.732 0.802 8 7 يؤثر األداء السابق ألسهم الشركة على قراراتي .االستثمارية 4.012 0.699 0.802 7 8 تلعب سمعة الشركة دورًا حيويًا في قراراتي .االستثمارية 4.071 0.842 0.814 4 9 تؤثر األخالق المدركة للشركة أيضً ا على قراري4.024 0.859 0.805 6 ( أعلى استجابة بين الفقرات األخرى وبوسط حسابي4.10 ( ) في حين حققت الفقرة المعلومات المطروحة حول عر أسهم الشركة تكون معلومة وشفافة للجميع ) ( أقل استجابة وبوسط حسابي3.75 .) (الجدول8 )قيم مؤشرات اإلحصاء الوصفي لبعد المنتجات وسمات الشركة العدد65 ج1 حزيران2022 288 288 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .االستثماري 10 تلعب سمعة مساهمي الشركة دورًا حيويًا في .قراراتي االستثمارية 4.094 0.750 0.819 3 المعدل العام لبعد المنتجات وسمات الشركة 4.004 0.820 0.801 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 4 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المعرفة والخبرات السابقة ( تبين النتائج الموضحة في الجدول9 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم(الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (المعرفة والخبرات ( السابقة)، إذ يتضح ان البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي3.69 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أ ن العينة المبحوثة لديها إدراك اال أنه ليس بالمستوى المطلوب ( لمضمون فقرات البعد المذكور، وبلغت قيمة االنحراف المعياري1.07 ( ) اما النسبة المئوية فقد بلغت 0.73 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبول (ة وبمعدالت متقاربة نوعا ما ،إذ حققت الفقرة امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات األسهم( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي3.82 ) في حين ( حققت الفقرة لديّ من المعرفة والخبرات السابقة على ان اتوقع متى ترتفع اسعار االسهم في ا لمستقبل ) ( أقل استجابة وبوسط حسابي3.60 .) الجدول ( 9) قيم مؤشرات اإلحصاء الوصفي لبعد المعرفة والخبرات السابقة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات.األسهم 3.824 0.889 0.765 1 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .االستثماري 10 تلعب سمعة مساهمي الشركة دورًا حيويًا في .قراراتي االستثمارية 4.094 0.750 0.819 3 المعدل العام لبعد المنتجات وسمات الشركة 4.004 0.820 0.801 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 4 :. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( مستوى داللة5 % ) وعلى فق هذه النتائج تقبل الفرضية التي مفادها توجد عالقة ارتباط ذات داللة .معنوية وموجبة بين الوعي االجتماعي وقرارات االستثمار على مستوى الدراسة الحالية رابع ًا : اختبار عالقات التأثير رابع ا : اختبار عالقات التأثير بعد ان تم اختبار فرضيات االرتباط وت بين ان هناك عالقة ارتباط للوعي االجتماعي وابعاده القرارات ( االستثمارية سوف يتم التعرف على مقدار تأثيرB ( ) ومعامل التفسير او التحديد2 R ) ومستوى المعنوية ( اي مقدار اخطاء القرار والتي يفرضها الباحث5 % ( ) فضال عن امكانية استخدام اختبار قيمةt ) ( المحسوبة ومقارنتها مع القيمة الجدولية البالغة1.96 ) لقبول الفرضية اذا كانت القيمة تزيد عن الجدولية تقبل الفرضية والعكس بالعكس ، ويفتر الباحث فرضية رئيسة واحدة ، و الفرضية الرئيسة تفيد بان الوعي االجتماعي له تأثير معنوي وا ،يجابي في القرارات االستثمارية و سيتم استعمال البرنامج االحصائي ( SmartPLS :) في استخراج نتائج التحليل وهي كاالتي تفيد هذه الفرضية بوجود تأثير للوعي االجتماعي معنوي وايجابي في قرارات االستثمار، وهذا يوضح ان المتغير التابع قرارات االستثمار دالة لل متغير المستقل الوعي االجتماعي ، اي ان اي زيادة تحدث في المتغير المستقل ستؤدي الى تغييرا مماثال في المتغير التابع، وسيتم استعمال البرنامج االحصائي ( SmartPLS V.3 ) في استخالص نتائج االنحدار البسيط. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار المعرفة والخبرات السابقة ( تبين النتائج الموضحة في الجدول9 ) قيم المؤشرات اإلحصائية الوصفية المتحققة والمتمثلة بم(الوسط الحسابي، االنحراف المعياري، النسبة المئوية) لبعد (المعرفة والخبرات ( السابقة)، إذ يتضح ان البعد المذكور قد حقق استجابة مقبولة وبوسط حسابي3.69 ) أي أعلى من قيمة الوسط الفرضي المعتمدة، مما يدل على أ ن العينة المبحوثة لديها إدراك اال أنه ليس بالمستوى المطلوب ( لمضمون فقرات البعد المذكور، وبلغت قيمة االنحراف المعياري1.07 ( ) اما النسبة المئوية فقد بلغت 0.73 )، أما على مستوى الفقرات التي تم قياس البعد المذكور من خاللها فقد كانت قيم األوساط الحسابية مقبول (ة وبمعدالت متقاربة نوعا ما ،إذ حققت الفقرة امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات األسهم( ) أعلى استجابة بين الفقرات األخرى وبوسط حسابي3.82 ) في حين ( حققت الفقرة لديّ من المعرفة والخبرات السابقة على ان اتوقع متى ترتفع اسعار االسهم في ا لمستقبل ) ( أقل استجابة وبوسط حسابي3.60 .) العدد65 ج1 حزيران2022 289 الجدول ( 9) قيم مؤشرات اإلحصاء الوصفي لبعد المعرفة والخبرات السابقة ت الفقرات الوسط الحسابي االنحراف المعياري النسبة المئوية الترتيب 1 امتلك من المعرفة والخبرات السابقة ما يمكنني من توقع تحركات.األسهم 3.824 0.889 0.765 1 2 لديّ القدرة على ان اتوقع المخاطر المالية في3.647 1.172 0.729 2 العدد65 ج1 حزيران2022 289 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .سوق االوراق المالية 3 لديّ من المعرفة والخبرات السابقة على ان اتوقع .متى ترتفع اسعار االسهم في المستقبل 3.600 1.177 0.720 3 المعدل العام لبعد المعرفة والخبرات السابقة 3.690 1.079 0.738 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 ًثالثا :اختبار عالقات االرتباط سيتم في هذه الفقرة اختبار عالقات االرتباط اذ سيتم التأكد بان الوعي االجتماعي مرتبطة ارتباط معنوي ( وطردي بقرارات االستثمار، اذ سيتم استخدام البرنامج االحصائيSmartpls ) في تقدير معامالت ( االرتباط اذ ان معامل االرتباط له نسبة تتراوح بين- + 1 ) ويفتر الباحث ان هناك عالقة ارتباط مقبولة معنويا ومرتبط المتغير المستقل بالتابع ارتباطا ايجاب( يا، ونتائج االختبار كما يعر الجدول10 ) الذي :يمثل مصفوفة معامالت االرتباط بين متغيرات وابعاد الدراسة- ( جدول10) مصفوفة معامالت االرتباط بين متغيرات وابعاد الدراسة الوعي االجتماعي قرارات االستثمار الوعي االجتماعي 1 قرارات االستثمار 0.711** 1 ( المصدر : من اعداد الباحث باالعتماد على البرنامجSmartPLS ) ( على وفق نتائج الجدول10) سيتم اختبار الفرضية :وهي كاالتي افتر الباحث بوجود عالقة ارتباط معنوية وايجابية بين الوعي االجتماعي وقرارات االستثمار، وحسب ( نتائج وجدول13 () تبين ان الوعي االجتماعي يرتبط ارتباط قدرة ب0.71 ) بقرارات االستثمار وهو ( طردي اي ان كلما تغير المتغير المستقل تغير المتغير التابع بمقدار0.71) وهو ارتباط معنوية عند تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار .سوق االوراق المالية 3 لديّ من المعرفة والخبرات السابقة على ان اتوقع .متى ترتفع اسعار االسهم في المستقبل 3.600 1.177 0.720 3 المعدل العام لبعد المعرفة والخبرات السابقة 3.690 1.079 0.738 المصدر : من إعداد الباحث باالعتماد على مخرجات برنامجSPSS V.24 اا ( المصدر : من اعداد الباحث باالعتماد على البرنامجSmartPLS ) ( على وفق نتائج الجدول10) سيتم اختبار الفرضية :وهي كاالتي افتر الباحث بوجود عالقة ارتباط معنوية وايجابية بين الوعي االجتماعي وقرارات االستثمار، وحسب ( نتائج وجدول13 () تبين ان الوعي االجتماعي يرتبط ارتباط قدرة ب0.71 ) بقرارات االستثمار وهو ( طردي اي ان كلما تغير المتغير المستقل تغير المتغير التابع بمقدار0.71) وهو ارتباط معنوية عند العدد65 ج1 حزيران2022 290 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار اذ يتم تقدير معامل التحديد والتعرف على جودة النموذج ومن ثم التعرف على مقدار التأث ير بين المتغيرات التي تم اختبارها والمعيار لقبول الفرضية ( هو مستوى خطا5 %) اذا سجل االختبار نسبة خطا ادنى ،منه تقبل الفرضية والعكس صحيح الجدول ( 11 :) يعر بعض احصاءات االختبار، وهي كاالتي- بعد ان تم اختبار فرضيات االرتباط وت بين ان هناك عالقة ارتباط للوعي االجتماعي وابعاده القرارات ( االستثمارية سوف يتم التعرف على مقدار تأثيرB ( ) ومعامل التفسير او التحديد2 R ) ومستوى المعنوية ( اي مقدار اخطاء القرار والتي يفرضها الباحث5 % ( ) فضال عن امكانية استخدام اختبار قيمةt ) ( المحسوبة ومقارنتها مع القيمة الجدولية البالغة1.96 ) لقبول الفرضية اذا كانت القيمة تزيد عن الجدولية تقبل الفرضية والعكس بالعكس ، ويفتر الباحث فرضية رئيسة واحدة ، و الفرضية الرئيسة تفيد بان الوعي االجتماعي له تأثير معنوي وا ،يجابي في القرارات االستثمارية و سيتم استعمال البرنامج االحصائي ( SmartPLS :) في استخراج نتائج التحليل وهي كاالتي تفيد هذه الفرضية بوجود تأثير للوعي االجتماعي معنوي وايجابي في قرارات االستثمار، وهذا يوضح ان المتغير التابع قرارات االستثمار دالة لل متغير المستقل الوعي االجتماعي ، اي ان اي زيادة تحدث في المتغير المستقل ستؤدي الى تغييرا مماثال في المتغير التابع، وسيتم استعمال البرنامج االحصائي ( SmartPLS V.3 ) في استخالص نتائج االنحدار البسيط. اذ يتم تقدير معامل التحديد والتعرف على جودة النموذج ومن ثم التعرف على مقدار التأث ير بين المتغيرات التي تم اختبارها والمعيار لقبول الفرضية ( هو مستوى خطا5 %) اذا سجل االختبار نسبة خطا ادنى ،منه تقبل الفرضية والعكس صحيح الجدول ( 11 :) يعر بعض احصاءات االختبار، وهي كاالتي- العدد65 ج1 حزيران2022 291 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ( جدول11 ) إحصاءات تأثير الوعي االجتماعي في قرارات االستثمار Original Sample (O) 2 R Standard Deviation (STDEV) T Statistics (\|O/STDEV\|) P Values X -> Y 0.238 0.5 2 0.099 2.399 0.017 ( المصدر : من اعداد الباحث باالستناد الى البرنامجSmartPLS ) ( المصدر : من اعداد الباحث باالستناد الى البرنامجSmartPLS ) ا ( علمممى وفمممق جمممدول11 ) تبمممين ان هنممماك جمممودة للنمممموذج حسمممب مممما فسمممره الممموعي االجتمممماعي ممممن قمممرارات ( االسممتثمار والتمممي بلغمممت0.52 ) وهممي تعنمممي ان المتغيمممر المسممتقل يوضمممح نسمممبة كبيممرة ممممن المتغيمممر التمممابع ،والمتبقي لم يدرس في هذه الدراسة و افتر الباحث بوجود عالقة تأثير معنوي وايجابي للوعي االجتماعي فمممي قمممر ارات االسمممتثمار، وحسمممب نتمممائج( جمممدول15 ()ان الممموعي االجتمممماعي لهممما تمممأثير قمممدر ب0.24 ) فمممي قمممرارات االسمممتثمار وهمممو ايجمممابي اي ان كلمممما زاد البعمممدُ المسمممتقل وحمممدة واحمممدة سممميزاد المتغيمممر التمممابع بمقمممدار ( 0.24 ( ) وهممو تممأثير معنويممة عنممد مسممتوى داللممة5 %) وعلممى فممق هممذه النتممائج تقبممل الفرضممية ا لتممي مفادهمما توجد عالقة تأثير ذات داللة معنوية وموجبة للوعي االجتماعي في قرارات االستثمار على مستوى الدراسة الحالية. المبحث الرابع: االستنتاجات والتوصيات:اا اوالً: االستنتاجات: اوالً: االستنتاجات: اا 1 . تشيممر نتائج التحليل الى صحة الفرضية االولى والتي تنص (على وجود عالقة ارتباط ذات داللة معنوية بين الوعي االجتماعي ،)وقرارات االستثمار مما يعني ان أي تعديل في مستوى الوعي االجتماعي في العينة المبحوثة سوف يسهم في تحسين وزيادة مستوى كفاءة قرارات االستثمار المتخذة في العينة .المبحوثة العدد65 ج1 حزيران2022 292 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 2 . تشيممر نتائ ج التحليل الى صحة الفرضية الثانية التي تنص (على وجود عالقة تأثير ذات داللة معنوية بين الوعي االجتماعي وقرارات االستثمار)، مما يعني ان أي تعديل في مستوى الوعي االجتماعي في العينة المبحوثة سوف يسهم في تحسين وزيادة مستوى كفاءة قرارات االستثمار المتخذة في العي نة .المبحوثة 3 . تبين من خالل نتائج التحليل االحصائي "وجود ضعف في االمكانيات التي تمكن المستثمر من ان يعرف ان المستثمرين من حولة قد اصبحو ا منزعجين.في العينة المبحوثة 4 . تبين من خالل نتائج التحليل االحصائي وجود ضعف في اإلمكانيات التي تمكن المستثمر بان يشعر عند تغير مزاج المستثمرين من حول ه .في العينة المبحوثة 5 . تبين من خالل نتائج التحليل االحصائي وجود ضعف فيج اإلمكانيات المستثمرين بان يكونوا داعمين عند اعطاء اخبار سيئة للمستثمرين االخرين في العينة المبحوثة مما يعني ان المنافسة بين المستثمرين على درجة عالية جدا.وقد ال يقدم البعض منهم الدعم الكافي لزمالئه من المستثمرين االخرين ثانياً: التوصيات: 3 . تبين من خالل نتائج التحليل االحصائي "وجود ضعف في االمكانيات التي تمكن المستثمر من ان يعرف ان المستثمرين من حولة قد اصبحو ا منزعجين.في العينة المبحوثة 4 . تبين من خالل نتائج التحليل االحصائي وجود ضعف في اإلمكانيات التي تمكن المستثمر بان يشعر 3 . تبين من خالل نتائج التحليل االحصائي "وجود ضعف في االمكانيات التي تمكن المستثمر من ان يعرف ان المستثمرين من حولة قد اصبحو ا منزعجين.في العينة المبحوثة 5 . تبين من خالل نتائج التحليل االحصائي وجود ضعف فيج اإلمكانيات المستثمرين بان يكونوا داعمين عند اعطاء اخبار سيئة للمستثمرين االخرين في العينة المبحوثة مما يعني ان المنافسة بين المستثمرين على درجة عالية جدا.وقد ال يقدم البعض منهم الدعم الكافي لزمالئه من المستثمرين االخرين ثانياً: التوصيات: 1 . حسممممممب فرضمممممممية االولمممممممى" التمممممممي تمممممممم اثباتهممممممما بوجممممممود ارتبممممممماط" بمممممممين الممممممموعي االجتمممممممماعي وقمممممممرارات االسممممممممتثمار، حتممممممممى نعممممممممزز" كفمممممممماءة قممممممممرارات االسممممممممتثمار المتخممممممممذة يجممممممممب ان نعممممممممزز مقممممممممدرات المممممممموعي االجتماعي". 2 . حسمممممممب فرضمممممممية الثانيمممممممة" التمممممممي تمممممممم اثباتهممممممما بوجمممممممود تمممممممأثير" بمممممممين الممممممموعي االجتمممممممماعي وقمممممممرارات االسمممممممتثمار ، حتمممممممى نعمممممممزز" كفممممممماءة قمممممممرارات االسمممممممتثمار المتخمممممممذة يجمممممممب ان نعمممممممزز مقمممممممدرات الممممممموعي االجتماعي". 3 . ضرورة تعزيز امكانيات المستثمرين حتى يتمكنوا من ان يعرف ان المستثمرين من حولة قد ا صبحوا منزعجين. 4 . ضرورة تعزيز قدرات المستثمرين حتى يتمكنوا من يعرفوا عند تغير مزاج المستثمرين من.حولهم العدد65 ج1 حزيران2022 العدد65 ج1 حزيران2022 العدد65 ج1 حزيران2022 293 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار ير ر و و ى ير ر زيز رور .االخرين :المراجع :ًاوال المراجع العربية: 1 . (، حنان، سعيني ،وسمية، شحان2019 ،)أ ثر تكلفة راس المال على اتخاذ القرار االستثماري في المؤسسة االقتصادية – ( دراسة حالة مؤسسة إيباش لألشغال العامة بأدرار خالل الفترة الممتدة من2014 - 2016 ) ، رسالة مقدمة لنيل درجة الماجستير، جامعة أحمد دراية– أدرار ، كلية العلوم االقتصادية، التجارية وعلوم الت.سيير 2 . (، شهيرة، شعاب ،و نزيهة، جدة2019 )، تحسين القرارات االستثمارية باستخدام جودة المعلومات المالية- دراسة حالة مصنع االسمنت بسور الغزالن– البويرة، رسالة ماجستير، جامعة اكلي محند اولحاج– البويرة ، كلية العلوم االقتصادية و التجارية و علوم التسيير، الجزا.ئر 3 . (، العبودي، وسن سعدون عجس2020 )، دور دراسة الجدوى االقتصادية في اتخاذ القرار االستثماري- حالة دراسية مشروع دواجن إلنتاج بيض المائدة نموذجا، مجلة اإلدارة واالقتصاد، المجلد9 ، العدد34 . 4 . ( ، القوس، سعود بن سهل2018 )، دور وسائل التواصل االجتماعي في تشكيل الوعي االجتماعي– دراسة ميدانية على عينة من الشباب السعودي بمنطة الريا ، مجلة جامعة الفيوم للعلوم التربوية والنفسية، العدد10 ، الجزء1 . 5 . (، ماجدة، ميسور2014 )، اثر قائمة تدفقات الخزينة على قرارات االستثمارية- دراسة حالة مؤسسة مطاحن الحضنة – بالمسيلة، رسالة.ماجستير، جامعة المسيلة، كلية العلوم االقتصادية و التجارية و علوم التسيير، الجزائر 6 . (، المنصور، زينة2015 )، الذكاء وعالقته بمهارة اتخاذ القرار- ،دراسة ميدانية على عينة من طالب جامعة دمشق رسالة مقدمة لنيل درجة الماجستير، جامعة دمشق، كلية التربية، قسم علم نف.س :ًثانيا المراجع:األجنبية 1 . (، حنان، سعيني ،وسمية، شحان2019 ،)أ ثر تكلفة راس المال على اتخاذ القرار االستثماري في المؤسسة االقتصادية – ( دراسة حالة مؤسسة إيباش لألشغال العامة بأدرار خالل الفترة الممتدة من2014 - 2016 ) ، رسالة مقدمة لنيل درجة الماجستير، جامعة أحمد دراية– أدرار ، كلية العلوم االقتصادية، التجارية وعلوم الت.سيير 2 . (، شهيرة، شعاب ،و نزيهة، جدة2019 )، تحسين القرارات االستثمارية باستخدام جودة المعلومات المالية- دراسة حالة مصنع االسمنت بسور الغزالن– البويرة، رسالة ماجستير، جامعة اكلي محند اولحاج– البويرة ، كلية العلوم االقتصادية و التجارية و علوم التسيير، الجزا.ئر الماجستير، جامعة أحمد دراية– أدرار ، كلية العلوم االقتصادية، التجارية وعلوم الت.سيير 2 . (، شهيرة، شعاب ،و نزيهة، جدة2019 )، تحسين القرارات االستثمارية باستخدام جودة المعلومات المالية- دراسة حالة مصنع االسمنت بسور الغزالن– البويرة، رسالة ماجستير، جامعة اكلي محند اولحاج– البويرة ، كلية العلوم االقتصادية و التجارية و علوم التسيير، الجزا.ئر التجارية و علوم التسيير، الجزا.ئر 3 . تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار (، العبودي، وسن سعدون عجس2020 )، دور دراسة الجدوى االقتصادية في اتخاذ القرار االستثماري- حالة دراسية مشروع دواجن إلنتاج بيض المائدة نموذجا، مجلة اإلدارة واالقتصاد، المجلد9 ، العدد34 . 4 . ( ، القوس، سعود بن سهل2018 )، دور وسائل التواصل االجتماعي في تشكيل الوعي االجتماعي– دراسة ميدانية على عينة من الشباب السعودي بمنطة الريا ، مجلة جامعة الفيوم للعلوم التربوية والنفسية، العدد10 ، الجزء1 . 3 . (، العبودي، وسن سعدون عجس2020 )، دور دراسة الجدوى االقتصادية في اتخاذ القرار االستثماري- حالة دراسية مشروع دواجن إلنتاج بيض المائدة نموذجا، مجلة اإلدارة واالقتصاد، المجلد9 ، العدد34 . مشروع دواجن إلنتاج بيض المائدة نموذجا، مجلة اإلدارة واالقتصاد، المجلد9 ، العدد34 . 4 . ( ، القوس، سعود بن سهل2018 )، دور وسائل التواصل االجتماعي في تشكيل الوعي االجتماعي– دراسة ميدانية على عينة من الشباب السعودي بمنطة الريا ، مجلة جامعة الفيوم للعلوم التربوية والنفسية، العدد10 ، الجزء1 . 5 . (، ماجدة، ميسور2014 )، اثر قائمة تدفقات الخزينة على قرارات االستثمارية- دراسة حالة مؤسسة مطاحن الحضنة – بالمسيلة، رسالة.ماجستير، جامعة المسيلة، كلية العلوم االقتصادية و التجارية و علوم التسيير، الجزائر 6 . (، المنصور، زينة2015 )، الذكاء وعالقته بمهارة اتخاذ القرار- ،دراسة ميدانية على عينة من طالب جامعة دمشق رسالة مقدمة لنيل درجة الماجستير، جامعة دمشق، كلية التربية، قسم علم نف.س :ًثانيا المراجع:األجنبية أ 1. Aburas, R. & Ab Hamid, N. Raihan ,(2013), A Construct Validity of Investment Decision in the Banking Sector in Libya (A Structural Equation Modeling Approach), Iosr Journal of Business and Management, , Vol (10), No (3). أ 1. Aburas, R. & Ab Hamid, N. Raihan ,(2013), A Construct Validity of Investment Decision in the Banking Sector in Libya (A Structural Equation Modeling Approach), Iosr Journal of Business and Management, , Vol (10), No (3). العدد65 ج1 حزيران2022 294 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 2. Ahmad, N. & Luqman, M. & Hayat, M. Farhat ,(2012), Importance of Investment for Economic Growth: Evidence From Pakistan, Interdisciplinary Journal Of Contemporary Research In Business, , Vol (4), No (4). 2. Ahmad, N. & Luqman, M. & Hayat, M. Farhat ,(2012), Importance of Investment for Economic Growth: Evidence From Pakistan, Interdisciplinary Journal Of Contemporary Research In Business, , Vol (4), No (4). 3. Bradberry , T & Greaves, J ,(2009), Emotional Intelligence 2.0., Talent Smart. 3. Bradberry , T & Greaves, J ,(2009), Emotional Intelligence 2.0., Talent Smart. 4. Fatum, B. A. تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار (2008), The Relationship Between Emotional Intelligence And Academic Achievement In Elementary - School Children, Thesis of Doctor of Education, The University of San Francisco. 5. Goleman, D, (2001), Emotional intelligence: Issues in paradigm building - The emotionally intelligent workplace, Library of Congress Cataloging-in-Publication Data , 1st ed. 6. Goleman, D, (2011), The Brain and Emotional Intelligence: New Insights; More Than Sound LLC: Florence, MA, USA. 7. Hamid, A. Abd & Abdul R. & Zaidi , F. & Abu Baka, A. & Abdullah, Wan S. Wong ,(2016), The Effects Of Perceived Usefulness And Perceived Ease Of Use On Continuance Intention To Use E-Government, Procedia Economics and Finance, , Vol (35). 8. Hohlbein, P. J. ,(2015), The Power Of Play In Developing Emotional Intelligence Impacting Leadership Success: A Study Of The Leadership Team In A Midwest Private, Liberal Arts University , Thesis of Doctor of Organizational Leadership, University of P di 7. Hamid, A. Abd & Abdul R. & Zaidi , F. & Abu Baka, A. & Abdullah, Wan S. Wong ,(2016), The Effects Of Perceived Usefulness And Perceived Ease Of Use On Continuance Intention To Use E-Government, Procedia Economics and Finance, , Vol (35). 7. Hamid, A. Abd & Abdul R. & Zaidi , F. & Abu Baka, A. & Abdullah, Wan S. Wong ,(2016), The Effects Of Perceived Usefulness And Perceived Ease Of Use On Continuance Intention To Use E-Government, Procedia Economics and Finance, , Vol (35). 8. Hohlbein, P. J. ,(2015), The Power Of Play In Developing Emotional Intelligence I ti L d hi S A St d Of Th L d hi T I A Mid t P i t 8. Hohlbein, P. J. ,(2015), The Power Of Play In Developing Emotional Intelligence Impacting Leadership Success: A Study Of The Leadership Team In A Midwest Private, Liberal Arts University , Thesis of Doctor of Organizational Leadership, University of Pepperdine. 9. Johnsi, S. & Sunitha, K. , (2019), Impact of Personality and Emotional Intelligence on Investor Behaviour , Sdmimd Journal of Management, , Vol (10), No (1). 10. Mahalakshmi, T.N. & Anuradha, N , (2018), Factors affecting Investment Decision making & Investment Performance among Individual Investors in India, International Journal of Pure and Applied Mathematics , Vol (118), No (18). 9. Johnsi, S. & Sunitha, K. , (2019), Impact of Personality and Emotional Intelligence on Investor Behaviour , Sdmimd Journal of Management, , Vol (10), No (1). تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 10 Mahalakshmi T N & Anuradha N (2018) Factors affecting Investment Decision العدد65 ج1 حزيران2022 295 تأثري الوعي االجتماعي للمستثمر على قرارات االستثمار 11. Qureshi, S. Ali & Rehman , K. ur & Hunjra, A. Imran ,(2012) Factors Affecting Investment Decision Making of Equity Fund Managers, Wulfenia Journal, , Vol (19), No (10). 12. Ramsey, R. (2014), Correlative Study of Emotional Intelligence and the Career Intentions of First-Year School of Business Students, Thesis of Doctor of Philosophy, Indiana Institute of Technology (Indiana Tech). 11. Qureshi, S. Ali & Rehman , K. ur & Hunjra, A. Imran ,(2012) Factors Affecting Investment Decision Making of Equity Fund Managers, Wulfenia Journal, , Vol (19), No (10). 12. Ramsey, R. (2014), Correlative Study of Emotional Intelligence and the Career Intentions of First-Year School of Business Students, Thesis of Doctor of Philosophy, Indiana Institute of Technology (Indiana Tech). 11. Qureshi, S. Ali & Rehman , K. ur & Hunjra, A. Imran ,(2012) Factors Affecting Investment Decision Making of Equity Fund Managers, Wulfenia Journal, , Vol (19), No (10). 12. Ramsey, R. (2014), Correlative Study of Emotional Intelligence and the Career Intentions of First-Year School of Business Students, Thesis of Doctor of Philosophy, Indiana Institute of Technology (Indiana Tech). 13. Shiundu, M. Vincent ,(2009), A Survey Of The Factors Influencing Investment Decisions: The Case Of Individual Investors At The Nse, Master thesis of Business Administration (MBA), University of Nairobi. 14. Triola, M. (2012), Elementary Statistics, 11th Edition, Pearson Education, United State. 14. Triola, M. (2012), Elementary Statistics, 11th Edition, Pearson Education, United State. العدد65 ج1 حزيران2022 296
https://openalex.org/W1591565644	https://www.fs.usda.gov/treesearch/pubs/download/10256.pdf	English	null	An interpretation of landscape structure from historic and present land cover data in the Eastern Upper Peninsula of Michigan.	null	1,997	public-domain	11,968	An Interpretatio Landscape Historic and Cover Data Upper Penins Michigan USDAUnitedStates Departmentof An Interpretation of Agriculture Forest Landscape Structure from Service North,Central Historic and Present Land ForestExperiment Station Cover Data in the Eastern GeneralTechnical Report NC.192 Upper Peninsula of _ Michigan Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Outwash Bedrock-controlled Lowland Sand Lake Plain USDAUnitedStates Departmentof An Interpretation of Agriculture Forest Landscape Structure from Service North,Central Historic and Present Land ForestExperiment Station Cover Data in the Eastern GeneralTechnical Report NC.192 Upper Peninsula of _ Michigan Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Outwash Bedrock-controlled Lowland Sand Lake Plain USDAUnitedStates Departmentof An Interpretation of Agriculture Forest Landscape Structure from Service North,Central Historic and Present Land ForestExperiment Station Cover Data in the Eastern GeneralTechnical Report NC.192 Upper Peninsula of _ Michigan Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Outwash Bedrock-controlled Lowland Sand Lake Plain North Central Forest Experiment Station Forest Service--U.S. Department of Agriculture 1992 Folwell Avenue St. Paul, Minnesota 55108 Manuscript approved for publication July 31, 1997 1997 Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Policy makers and planners dealing with broad scale issues must account for landscape and regional scale factors in their design. In light of contemporary broad-scale environmental issues, landscape pattern analysis has become a necessity for both resource managers and scientists. Questions such as ... "how will managing a 40-ha tract affect adjacent areas or regions?..." have promoted new approaches to society's dilemmas like air and water qual- ity, species and habitat losses, and increased carbon dioxide levels (King 1993). Maps of regional- and landscape-scale ecological units are available to, and now frequently used by, policy decision-makers and land managers to approach these sorts of issues. Regional and landscape ecological units based on biophysi- cal factors provide an ecologically based framework for land use decisions. However, for many of these ecological units, measurable descriptions of pattern, composition, and other CharacterisUcs are lacking. Measures that quantify and describe landscape structure (pattern and composition) improve knowledge Janet SIIb rnagel is with the Department of Horticulture and _dscape Architecture, Washington State University, Pullman, WA 99164. Jlquan Chen, Margaret Gale, and Kurt Pregitzer are with the School of Forestry and Wood Products, Michigan Technological University, Houghton, MI 49931. John Probst is with the Forestry Sciences Laboratory, USDA Forest Service, North Central Forest Experiment Station, Rhinelander, WI 5450 I-9128 Policy makers and planners dealing with broad scale issues must account for landscape and regional scale factors in their design. In light of contemporary broad-scale environmental issues, landscape pattern analysis has become a necessity for both resource managers and scientists. Questions such as ... "how will managing a 40-ha tract affect adjacent areas or regions?..." have promoted new approaches to society's dilemmas like air and water qual- ity, species and habitat losses, and increased carbon dioxide levels (King 1993). Maps of regional- and landscape-scale ecological units are available to, and now frequently used by, policy decision-makers and land managers to approach these sorts of issues. Regional and landscape ecological units based on biophysi- cal factors provide an ecologically based framework for land use decisions. However, for many of these ecological units, measurable descriptions of pattern, composition, and other CharacterisUcs are lacking. Measures that quantify and describe landscape structure (pattern and composition) improve knowledge of ecological units and strengthen the frame- work on which both finer resolution landscape decisions may be based. An Interpretation of Landscape Structure from Historic and Present Land Cover Data in the Eastern Upper Peninsula of Michigan An Interpretation of Landscape Structure from Historic and Present Land Cover Data in the Eastern Upper Peninsula of Michigan Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst An Interpretation of Landscape Structure from Historic and Present Land Cover Data in the Eastern Upper Peninsula of Michigan USDAUnitedStates Departmentof Agriculture Forest Service North,Central ForestExperiment Station GeneralTechnical Report NC.192 _ Report NC.192 Upper Peninsula of _ Michigan Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Outwash Bedrock-controlled Lowland Sand Lake Plain Morainal Origin Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Outwash Bedrock-controlled Lowland Sand Lake Plain Bedrock-controlled Lowland Sand Lake Plain Lowland Sand Lake Plain Morainal Origin A. Study Area 1. Geographic location: The study area is the Eastern Upper Peninsula of Michigan, comprising six counties: Alger, Chippewa, Delta, Luce, Mackinac, and Schoolcraft; and 1,795,524 ha _. The area is bordered by three Great _es--Huron, Michigan, and Supe- rior--and is characterized by lake-moderated climates and land types including hardwood forest on glacial moraines; pine barrens on outwash and beach ridges; coastal marshes; interior bogs, fens, and swamps; and llme- stone outcrops (Peterson 1986). The area lies completely within the Northern Great Lakes Ecological Section (map I), (Barley eta/. 1994, McNab and Avers 1994). The distinction of this study from important works like Frelich's and others' lies in its geographic size, its sequence of time, and its provisionalassessment units. We chose landtype association groups as the basis for comparison to control for broad physiographic patterns, while strengthening the knowledge base of these units and their value as planning elements (McNab 1996, Sims et al. 1996). Many previous landscape assessments have been based on administrative or arbitrary boundaries (e.g., 7.5' USGS quadrangle maps), rather than an ecological frameworlc Without such a framework, it is not possible to account for variation in the physical environment, 2. Physiographic setting: The Northern Great Lakes Section is a level to gently rolling lowland of fiat outwash or lacustrine plain with dune fields near the Great _es. Par- tially buried end moraines protrude from the lowlands roughly paranel to the lakes. A large portion of the area is covered by Pleistocene stratified drift, mostly outwash sand. Lacus- trine deposits occur between morainal ridges and are widespread in the Eastern Upper Peninsula. Pleistocene and Holocene sand dunes occur near the Great Lakes (Martin 1957). Silurian limestones and dolomites are locally exposed along _es Huron and Michi- gan, while Cambrian and Precambrian sand- stone outcrops occur along Lake Superior. The dominant geomorphic processes for this area are fluvial erosion, transport, and deposi- tion; lake-shore erosion and deposition; and minor dune construction (Albert 1995, McNab and Avers 1994). Janet Silbernagel, Jiquan Chen, Margaret R. Gale, Kurt S. Pregitzer, and John Probst Already there have been rigorous efforts linked with physical geography to describe land surface patterns (GoUey 1995). But, many of these studies, regardless of the investigator's perspective, have been of limited extent (Pastor and Broschart 1990, Simpson et al. 1994), low resoluUon (Dunn et aL 199 I, Hulshoff 1995, O'Neill et al. 1988), or of narrow temporal scale (Luque et aL 1994, Muller and Middleton 1994, Simpson et a/. 1994, Turner and Ruscher 1988). What is needed are works that can increase spatial and temporal extent while maintaining high resolution so that results can be applied to a larger area and provide a longer frame of reference, yet sun be meaning- ful for site level assessments. There has been significant interest in land- scapes surrounding the Great Lakes, where land use has had a particularly large influence on the lake ecosystems (e.g., Upper Great Lakes Biodiversity Committee; Lake Superior Binational; SOLEC, in press). Fluctuations in human population in the Lake States since European settlement have played a role in development of current forest patterns and have been the focus of several research pro- grams. For instance, Frelich (I 995) docu- mented the amount of "old forest" in historic and present forests of the Lake States, includ- ing amounts present in various land cover types. He defined "old forest" as both old growth (forests dominated by long-lived spe- cies beyond 120 years) and old-seral forests (dominated by short-lived species that are relaUvely old at age 80). The results showed that about 5.2 to 8.3 percent of the _e States forest is now old forest, compared with Janet SIIb rnagel is with the Department of Horticulture and _dscape Architecture, Washington State University, Pullman, WA 99164. Jlquan Chen, Margaret Gale, and Kurt Pregitzer are with the School of Forestry and Wood Products, Michigan Technological University, Houghton, MI 49931. John Probst is with the Forestry Sciences Laboratory, USDA Forest Service, North Central Forest Experiment Station, Rhinelander, WI 5450 I-9128 II. BACKGROUND an estimated 68 percent before European settlement. Frelich also found that in the largest single block of remaining unlogged forest in the Lake States, Minnesota's Bound- ary WaterS Canoe Area Wilderness, the spatial •pattern, age structure, and species composi- tion have been significantly altered since the 1800's. Still, Frelich's study was limited to assessment of "old forest," was coarse in resolution (errors of +/- 50 percent for forest areas 400 ha in size or less), and did not quantify spatial pattern to augment additional landscape analyses, an estimated 68 percent before European settlement. Frelich also found that in the largest single block of remaining unlogged forest in the Lake States, Minnesota's Bound- ary WaterS Canoe Area Wilderness, the spatial •pattern, age structure, and species composi- tion have been significantly altered since the 1800's. Still, Frelich's study was limited to assessment of "old forest," was coarse in resolution (errors of +/- 50 percent for forest areas 400 ha in size or less), and did not quantify spatial pattern to augment additional landscape analyses, _ One hectare equals 2.471 acres. The objective of this report was to: The objective of this report was to: Compile quantitative InformatWn, or landscape metrics, to supplement existing qua!itative descr_tlons of land type associations in the study area, Compare the distribution of land- scape measures among landtype association groups in hlstorlc and present landscapes, This study occurred in the context of a larger research project whereby historic and contemporary landscape variability was de- scribed and quantified, landscape change was assessed over a 150-year time period, cultural settlement patterns were assessed over a 3,000-year time period, and the landscape change was discussed in terms of the combi- nation of biophysical and cultural factors, This report addresses the first portion of the research: description and quantification of historic and contemporary landscape patterns, 3. Land ownership: Much of the Eastern Upper Peninsula is public land including the Hiawatha National Forest, Lake Superior State Forest, Seney National Wildlife Refuge, Pic- tured Rocks National Lakeshore, and Tahquamenon State Parl_ Other large land holdings belong to The Nature Conservancy and three major industrial landowners. To- gether these holdings make up about 12 percent of Michigan's land area in one of the most undeveloped portions of the Eastern United States. 2 B. Ecological Classification yet still provide enough information to identify differences in ecological potential or trends and to improve coordinated management among multiple land owners or regulatory agencies (Cleland pers. comm., ECOMAP 1993). . Because of the infinite complexity inherent in any ecosystem, land areas are often catego- rized into discrete units so that biophysical components can be measured and understood (Allen and Start 1982, I_jn and Udo de Haes 1994, Urban eta/. 1987). Unified functional ecosystems are divided and categorized into similar and dissimilar pieces at various scales, inthe interests of description and understand- ing (Rowe 1996, 1991). A hierarchical frame- work of ecological units was developed for the United States, from which .units for this study were based (ECOMAP 1993). The Framework was modeled after the multi-factor, multi-scale approach to understanding ecosystem struc- ture and function pioneered in Baden- Wurttemberg, Germany (Albert 1995, Barnes 1984, Spurr and Barnes 1980). At the broad- est scale of the framework, global ecosystems were identified, within which a series of finer scale ecoregions are nested (Barley et a/. 1994, Barley 1996). Next, Albert et aL (1986, 1995) mapped regional landscape ecosystems of Michigan, Minnesota, and Wisconsin to the subsection level, based on climatic and broad physiographic factors. The finest level units of their work provided the framework for subdivi- sion into local landscape ecosystem units characterized by topography, microclimate, soft, and vegetation (Albert et al. 1986). Through an interagency effort, a first approxi- marion of landtype associations (LTA's) were mapped across the Eastern Upper Peninsula to help coordinate natural resource steward- ship (map I). The LTA's were identified in accordance with concepts and methods de- scribed in The National Hierarchical Frame- work of Ecologlcal Units (ECOMAP 1993, Ban and Padley 1993, Uhlig and Jordan 1996). The delineation process is iterative, using both top-down (regionalization) and bottom-up (classification) information. At this stage, the delineations have been based primarily on intuitive and visual methods. However, bot- tom-up ecological classification and research such as in this report contribute to, refine, and validate the landscape boundary Ioca- tions. Similarly, macroscale research (e.g., regional climatic data) is underway that will provide quantitative top-down contributions. B. Ecological Classification Preliminary delineation criteria included terrain (major slope and elevation changes or escarpments, variability of slope) from 1:250,000 USGS topography maps; geology (bedrock type, surficial geology, major glacial features, paleoshorelines) from Anderton (1993), Mart/n (1957), and Farrand (1982); soil (communities of softs, drainage, texture, depth to bedrock, source and mode of deposi- tion) from county-level soft association maps; localized climate (lake effect wind-blown sand and wave action, major cold air drainages, etc.) from Santer (1977); large lakes or open water systems; watersheds (drainage pattern, water level, and fluctuation variability); and vegetation (LANDSAT 1984 image, pre-Euro- pean settlement vegetation from General Land Office surveyors notes, aerial photography); existing subsection maps; plot data; and personal experience (Jordan et oL, in prep.). A, Data Assembly We used only four of the eight groupings: bedrock-controlled, lowland sand lake plain, morainal origin, and outwash for analysis because we believed they were the most internally consistent and because they occurred repeatedly across the study area (map 2). 2. Present land cover: Land cover data derived from Landsat Thematic Mapper (TM) satellite imagery were used to measure present •landscape structure. The TM imagery was classified into land cover categories for the entire Upper Peninsula with 60-m resolution, Maclean Consultants, Ltd. of Houghton, Michigan, obtained Landsat Thematic Mapper (TM) imagery for the Upper Peninsula of Michigan from Eosat, Inc., in contract with the Michigan Department of NatUral Resources, to classify TM images into land cover categories (Maclean 1994). Images were selected to provide: (1) complete coverage of the Upper Peninsula; (2) minimal cloud cover; (3) cover- age no earlier than 1990; and (4) data from the correct time of the growing season for best species differentiation (june I through July •31). Classification accuracy was presented in the form of contingency tables that showed both errors of omission and errors of commis- sion (Maclean 1994). The result was a com- plete land cover classification for the entire Upper Peninsula with an average 90.2-percent correct (accuracy) within a 95-percent confi- dence interval, Land cover maps were compiled by the Michi- gan Natural Features Inventory on 1:24,000 USGS topographic maps. Section line informa- tion recorded by land surveyors in the early 1800's was plotted onto the USGS maps. Cover type boundaries between each section line were interpolated using elevation lines, surface geology maps, and other early vegetation maps; the resulting resolution was definable to 259 ha, although many smaller polygons were delineated. Cover classes were based on ex- panded Michigan Resource Information System (MIRIS) land cover codes. Small cover types not transected by the surveyors were probably not recorded. Bias in tree species selected by surveyors has been documented (Bourdo 1956) and was assumed in map production (MNFI 1994). Nevertheless, previous investigations have concluded that the GLO records could be used to reconstruct historic vegetation (Bourdo 1956, Curtis 1959). In fact, in a study by Price (1994) of composition, structure, and distur- bance regimes of historic forests in westem Chippewa County, Michigan (within the Eastern Upper Peninsula), tests found only moderate bias toward the 15- and 20-cm diameter classes in the mixed conifer/deciduous lowland forest type. A, Data Assembly 1, Landtype association: Landscape eco- systems, defined as "a cluster of interacting ecosystems repeated in similar form through- out," (Crow 1991; Forman and Godron 1986; Jordan et oL, in prep.) provide a spatial foot- print that can account for mesoscale physical environmental factors. They are a useful size for many analyses because they are commonly based on easily identifiable physiographic factors, they are wen-suited to scales of mod- em inventory technology (remote sensing, GIS, etc.), and they can be easily recognized from automobile or airplane (Smith and Carpenter 1996). Boundaries separating landscapes are distinct, as often seen in satellite images, due to geomorphology and human activity (Forman 1995). Within a relatively short time, land- scape ecosystems can be delineated and described across large areas (multiple States), For this study, LTA's were intuitively grouped into the following eight categories based on dominant physiographic simflariUes: bedrock- controlled, lowland sand lake plain, morainal origin, outwash, clay lake plain, ground mo- raine, transitional/complex, and inland lake (Gary et aL 1974). The groupings are compa- rable to Michigan's Land Types mapped by the 4 Pre-European settlement landscape data based on the GLO survey notes were used as the historical dataset of land cover for this study (Ewert, pers. comm.; Albert, pers. comm.; MNFI 1994). Surveyors of Michigan in the early 1800's took detailed notes on the location, species, and diameter of each tree used to mark section lines and comers. Their notes also detailed landscape features such as lakes, wetlands, fields, natural disturbances, trails, and settlements. For our study, advantages of these records were: (I) they are available in digital format for the entire Upper Peninsula of Michigan, (2) original field work was conducted according to a predetermined plan, and (3) the surveys consUtute a systematic aligned sample of the historic forest species and thus are useable for quantitative analysis (Bourdo 1956, Price 1994). Michigan Department of Natural Resources (Santer 1977). We used only four of the eight groupings: bedrock-controlled, lowland sand lake plain, morainal origin, and outwash for analysis because we believed they were the most internally consistent and because they occurred repeatedly across the study area (map 2). Michigan Department of Natural Resources (Santer 1977). A, Data Assembly Therefore, Price concluded that GLO records from western Chippewa County were largely free of bias toward any specific diameter classes or tree species. •3, Historic land cover: General Land Office •(GLO) survey records have been commonly used in historic vegetation and landscape assessment studies. Early uses of these records involved reconstruction of pre-Euro- pean settlement forests in the form of maps and compositional descriptions (Curtis 1959, Veatch 1959). The suitability of survey records for.quantitative analysis of historic vegetation was first documented by Bourdo (I 956). Subsequent studies by Lorimer (1977, 1980), Canham (1978), Canham and Loucks (1984), Noss (1985), and Whitney (1986) described the use of GLO notes and other historic records in characterizing historic forests and critiqued the various approaches. In a more recent application of historic data in the Lake States, Baker (1991) simulated effects of settlement and fire suppression by using a GIS-based spatial model and historical data on fire size and frequency, Each layer was compiled in digital format and georeferenced to Universal Transverse Mercator projection on a Unix platform ARC/INFO Geo- graphic Information System. A legend defining the cover types was included with both the 5 B. Data Analysis Class statistics and indices calculated for this report included class area (CA), N, MPS, landscape similarity index (LSIM), LPI, PD, PSCV, and mean shape index (MSI). Land- scape similarity index measures the proportion of a class area to the total landscape area. Mean shape index is a ratio of patch area to patch perimeter. Selected landscape and class calculations were tabulated and graphed by LTA group. Indices measuring several components of landscape pattern and composition together are a means to quantify structure and change (Hess 1994, Mflne 199 I, O'Neill et al. 1988). Indices and basic metrics (measures of pattern and composition) can be computed for three scales: patch, class, and landscape. Patch- scale metrics are primarily used as building blocks to calculate class and landscape metrics, and relate to an individual patch or stand of a relatively homogenous (consistent) cover type. Class statistics relate to individual land cover classes, or cover types, based on dominant overstory species within a landscape and include: number of patches, percent of a class in a landscape, mean patch size of the class, and average perimeter/area ratios (shape indices). Class metrics are known for their application in estimating levels of frag- mentation or in determining the degree to which the matrix (most connected) class is fragmented by another class type. Landscape statistics relate to the landscape or LTA group as a whole and include number of patches, total area, mean patch size, and patch size variability; diversity of classes, dominance of one class over others, and shape, or fractal indices. Landscape-scale metrics are used to examine landscape diversity and structure of an entire landscape, or to compare landscapes of different time periods or location (MacGarlgal and Marks 1993). There is a limitation to conclusions that may be drawn from comparing historic to present data due to methodological and resolution differences between the two. The objectives of this paper do not include comparisons over time, and inappropriate inferences should be avoided. Metrics can be categorized according to what they measure. Some metrics measure pattern, others measure composition. We used the FRAGSTATS Spatial Pattern Analysis Program (MacGarigal and Marks 1993) to quantify and describe patterns within and among LTA groups in the historic and present land covers. In this report, we focus on the following land- scape metrics: number of patches (N), total area (TA), mean patch size (MPS), patch size coefficient of variation (PSC_, largest patch index (LPI), patch density (PD), and Shannon's diversity index (SHDI), the combination of which we believed best represented aspects of landscape structure. Largest patch index indicates the contribution of the largest patch to total area. Patch size coefficient of variation provides a measure of relative variation. Patch density is the number of patches per 100 ha, and Shannon's diversity index indicates the amount of different class types in a landscape. Equations for each of the indices used are provided in Appendix A. historic and present coverages (tables 1 and 2). Because the two legends were not identi- cal, they had to be combined so that cover types in one coverage matched those in the other, or in some cases, a cover type was renamed (table 3). Of the total 24 cover types in the combined legend, I0 were selected for analysis based on the following criteria: the type was present in both time periods in at least two of the LTA groups; the types repre- sented a broad range of land cover classes indicative of the region and pertinent to the objectives-of this study (e.g., mixed conifer and cedar were selected, but white spruce was not); based on our familiarity with the two datasets, the type seemed to be relatively consistent between and within the two datasets (e.g., herbaceous shrub and openlands were inconsistent, particularly in the historic set); or the type was of particular importance to land use change questions (e.g., agriculture). B. Landscape Measures 1. Historic Landscape Metrics: Of the four LTA groups studied, the lowland group was largest in total area (509,511 ha), with 5,409 patches, and the outwash group was smallest (198,831 ha), with 1,640 patches (table 5). In the historic landscape, LPI was highest in the moraine group (10.280), with a mean patch size of 121 ha and lowest in the lowland group (2.757, MPS = 94 ha), although the other three groups (bedrock, lowland, and outwash) were relatively close (table 5, fig. I). Because LPI in the lowland group was lower than in the other groups, we can determine that the lowland MPS was less affected by a single large patch (table 5). The four groups had relatively close values for PD and SHDI. PD was highest in the lowland group (1.062) and lowest in the bedrock group (0.63 I). Outwash had the highest diversity historically (SHDI = 2.069), and lowland had the lowest (0.520). 2. Present landscape composition: The present coverage was substantially more detailed due to the higher resolution from the Thematic Mapper (TM)classification (map 3b. However, a few cover types were absent in the LTA groups considered (table 4). There were no oak, balsam fir, or hemlock types in any of the present datasets. In the original present legend, most conifer types were coded into two classifications: <70-percent but >40-percent canopy closure contributed by that species, and >70 percent canopy closure contributed by that species. In the LTA groups assessed, cedar occurred only in the >70-percent canopy closure class, except for the lowland group, where it occurred only in the <70-percent class. In other words, in the bedrock, mo- mine, and outwash groups, all cedar patches 12 I0 8 o_ m_ x 6 "o 4 0 I ,' I Bedrock Lowland Moraine Outwash LTA group _ Largest patch index __ Patch density _ Shannon's diversity index Figure 1.---Values for three landscape level _Ldlces: largest patch index (LPI),patch density (PD), and Shannon's Diversity Index (SHDI) calculated by LTA group on the historic land cover. 12 I0 8 o_ m_ x 6 "o 4 0 I ,' I Bedrock Lowland Moraine Outwash LTA group _ Largest patch index __ Patch density _ Shannon's diversity index Figure 1.---Values for three landscape level _Ldlces: largest patch index (LPI),patch density (PD), and Shannon's Diversity Index (SHDI) calculated by LTA group on the historic land cover. A. Landscape Composition 1. Historic Landscape Composition: No areas were mapped as "urban" in MNFI's transcription of the GLO maps as we have for the present cover; rather, a separate map layer was created by MNFI showing cultural settle- ments, primarily Indian villages. Non-veg- etated areas did not occur in the outwash group, and agriculture did not occur in any of the four groups. Herb/openland occurred only 7 7 in outwash, and shrub land did not occur in any of the groups studied. Oak cover type was absent in an LTA groups. Neither dry nor wet hardwood-conifer mixes were found in the historic datasets. Red pine was not found in the bedrock group, and white spruce was absent from the historic lowland landscape (map 3a, table 4). in outwash, and shrub land did not occur in any of the groups studied. Oak cover type was absent in an LTA groups. Neither dry nor wet hardwood-conifer mixes were found in the historic datasets. Red pine was not found in the bedrock group, and white spruce was absent from the historic lowland landscape (map 3a, table 4). in the present land cover were quite dense, and dominated by cedar. In the lowland group, however, cedar stands were less dense, or included other species. B. Landscape Measures o_ m_ x "o Shannon's diversity index Figure 1.---Values for three landscape level _Ldlces: largest patch index (LPI),patch density (PD), and Shannon's Diversity Index (SHDI) calculated by LTA group on the historic land cover. 8 2. Present Landscape Metrics" Similar results were obtained for the present land- scape, in which the lowland group had the greatest number of patches (63,994), and the remaining three groups all had approximately 25,000 patches. LPI was highest in the mo- raine group (5.63 I), but due to the higher resolution, MPS was only I 1 ha (fig. 2). Low- land had the lowest LPI again (I. 170, MPS = 8 ha). MPS for all four groups was between 8 and 11 ha. Present PD was highest in the lowland group (12.560) and lowest in the moraine group (9.056), although all four groups reflected high PD due to higher resolu- tion data. Diversity for all groups in the present landscape was low (0.578 - 0.615) (table 5, fig. 2). outwash had more mixed pine than other groups, with no one class being as dominant. The lowland group was historically more dominated by mixed conifer and wetlands (fig. 3a). In the present moraine landscape, northern hardwoods were most prevalent (LSIM). In present bedrock, northern hardwoods and mixed conifer were still important, but agricul- ture was as well, rather than cedar. The present outwash landscape was quite well distributed among several classes including northern hardwoods, mixed pine, mixed conifer, white pine, wetlands, and agriculture. The lowland group, presently, was weighted to mixed conifer and northern hardwoods, prima- rily, followed by wetlands and mixed pine (fig. 3b). C. Class Measures Northern hardwoods, especially, and mixed conifer to a lesser extent were the most preva- lent (LSIM) of the 10 classes analyzed in the historic moraine landscape. Historic bedrock had a relatively high amount of mixed conifer, northern hardwoods, and cedar, while historic Classes with high LPIhistorically in the too- raine group were northern hardwoods, mixed conifer, mixed pine, and cedar. The wetlands class had the highest LPI in historic lowland. In the bedrock and outwash groups, the classes were less driven by single large Figure 2._Values for three landscape level indices: largest patch index (LPI), patch denslhj (PD), and Shannon's Diversity Index (SHDI) calculated by LTA group on the present land cover. Figure 2._Values for three landscape level indices: largest patch index (LPI), patch denslhj (PD), and Shannon's Diversity Index (SHDI) calculated by LTA group on the present land cover. I0 0 :_,......................................................................................................................................................................................................................................................................................................................................................... 20-i. ................................................................................................................................................................................................................................................................................................................. i.i .................I1 lo i !i ,.,- -4- _....R..... t--- _;.......... i .... _.... Non-veg Agricul . N.hdwds Aspen Low hdwd Wetlands W.pine Mix pine Cedar Mix conifer Cover class _ Bedrock _ Lowland _ Moraine Outwash Figure 3a.--Landscape s/m//ar/ty index (LSIM) calculated by cover type for each LTA group using the historic landscape dataset (MNFI 1994). LSIM measures the proportion of one cover hjpe in area per the total landscape (LTA group) area. 0 :_,......................................................................................................................................................................................................................................................................................................................................................... 20-i. ................................................................................................................................................................................................................................................................................................................. i.i .................I1 lo i !i ,.,- -4- _....R..... t--- _;.......... i .... _.... Non-veg Agricul . N.hdwds Aspen Low hdwd Wetlands W.pine Mix pine Cedar Mix conifer Cover class _ Bedrock _ Lowland _ Moraine Outwash Figure 3a.--Landscape s/m//ar/ty index (LSIM) calculated by cover type for each LTA group using the historic landscape dataset (MNFI 1994). LSIM measures the proportion of one cover hjpe in area per the total landscape (LTA group) area. 60 [ ! ]I i i 504.................................................................................................................................................................... I 40J t , I 1 _ -. ! _,_ t t > 30-.J ! i" "..................................................................................................................................................................................................................................... \| 20-!.................................................................................... _ i Non-veg Agricul N'.I_dwds Aspen Lowhdwd Wetlands W.pine MIXpine Cedar Mixconifer Coverclass Figure 3b.--Landscape similarity index (LSIM) calculated by cover type for each LTA group using the present landscape dataset (Maclean 1994). LSIM measures the proport/on of one cover hjpe in area per the total landscape (LTA group) area. 60 [ ! ]I i i 504....................................................................................................................................................................... I 40J t , I 1 _ -. ! _,_ t t > 30-.J ! i" "..................................................................................................................................................................................................................................... \| 20-!.................................................................................... C. Class Measures Presently, the moraine and outwash groups had variable patterns of mixed and white pine and northern hardwoods. Howevel in present bedrock, northern hardwoods and agriculture were most variable, and in presenl lowlands, northern hardwoods had high PSC_ (table 7). Mean shape index (MSI) was fairly even in th_ historic landscape among landscape groups and cover classes, except that the non-veg- etated had much higher MSI values in all bu the outwash group (fig. 5a). Present shape indices were also quite even among groups and classes, with non-vegetated class gener- ally slightly higher. In the present landscap high MSI for non-vegetated areas was attrib uted to the inclusion of roadways within th_ spectral class, which have high perimeter tc area ratios, and therefore, high MSI. Asper was generally the lowest in MSI of the 10 classes, and all landscape groups reflected nearly the same MSI value in aspen (fig. 5b The greatest number of patches for any one class type was 6,076 patches of mixed pine in the present lowland landscape (tables 6 and 7). This was believed to be a reflection of the lowland sand ridge-swale complexes found within this LTA group, in which a mosaic Wetland/bog landscape is interspersed by a pattern of linear sand ridges vegetated with Jack pine and other seral species. Patch density indicates how fine the pattern of landscape is, or how much it is fragmented into many small patches. PD is measured as the number of patches per 100 ha. In all LTA groups, except lowlands, historically and presently, mixed conifer had the highest PD. In the lowland group, mixed pine, wetlands, and northern hardwoods had high PD (tables 6 and 7). Also, in the historic bedrock group, non-vegetated areas had very high PD, even though the mixed conifer class was highest (fig. 4a). PD was much lower in historic data due to the resolution differences and should not be compared to the present data. In the present landscape, all LTA groups had high mixed conifer PD, although in the lowland group, wetlands and mixed pine were also quite high. All four groups reflected slmilar moderate PD values in the agriculture, north- ern hardwoods, aspen, and mixed pine classes (fig. 4b). C. Class Measures _ i Non-veg Agricul N'.I_dwds Aspen Lowhdwd Wetlands W.pine MIXpine Cedar Mixconifer Coverclass Figure 3b.--Landscape similarity index (LSIM) calculated by cover type for each LTA group using the present landscape dataset (Maclean 1994). LSIM measures the proport/on of one cover hjpe in area per the total landscape (LTA group) area. Figure 3b.--Landscape similarity index (LSIM) calculated by cover type for each LTA group using the present landscape dataset (Maclean 1994). LSIM measures the proport/on of one cover hjpe in area per the total landscape (LTA group) area. 11 is weighted per unit area. In historic bedrock, mixed conifer, cedar, and northern hardwoods had the most variable pattern. Historic outwash was similar with mixed conifer, northern hardwoods, and mixed pine having the highest PSCV. But, in historic moraine, northern hardwoods and wetlands had the most variable size pattern, while wetlands and mixed pine had high PSCV in historic lowland,¢ (table 6). Presently, the moraine and outwash groups had variable patterns of mixed and white pine and northern hardwoods. Howevel in present bedrock, northern hardwoods and agriculture were most variable, and in presenl lowlands, northern hardwoods had high PSC_ (table 7). patches. In historic bedrock and outwash, classes with high LPI were northern hard- woods, mixed conifer, cedar, and mixed pine (table 6). In present bedrock and outwash, classes with high LPI included northern hardwoods, agriculture, mixed pine and white pine (table 7). Of all four groups in the present coverage, wetlands had the highest LPI in the lowland group, indicating larger interior wetland areas within the lowland group. Overall, the moraine group had the largest single patches among an cover types, but especially in northern hardwoods, which suggests a greater amount of core areas/ interior habitats in the moraine landscapes, In the bedrock and moraine landscapes, MPS was very high in the northern hardwoods class, but both groups also had high LPI and PSCV, which implies there might be a few very large patches and many sman patches in these groups, is weighted per unit area. In historic bedrock, mixed conifer, cedar, and northern hardwoods had the most variable pattern. Historic outwash was similar with mixed conifer, northern hardwoods, and mixed pine having the highest PSCV. But, in historic moraine, northern hardwoods and wetlands had the most variable size pattern, while wetlands and mixed pine had high PSCV in historic lowland,¢ (table 6). D. Assumptions About the Data and Inherent Bias or Limitations This project used three different geographl datasets, each developed by different met] As a result, each land cover layer had uni, limitations. The LTA coverage, discussed previously, was derived from intuitive mal procedures in which the boundaries repr_ sented intangible ecological transitions. result, neither map error nor accuracy c¢ be estimated for this layer. Nevertheless, a meaningful framework for landscape a., ments and planning, and one that can b, strengthened by studies of this type. Th units, like regional ecosystems, should t considered hypotheses for testing, devel, from ecological theory and knowledge of landscape. By using the units for studi, such as this, we can identify units that functionally different versus those that respond to landscape disturbances in s ways (Albert 1995). We refined our measure of landscape pattern by considering how regular or variable the pattern is. PSCV measures patch size start- dard deviation divided by mean patch size and 12 0.5 0.45 0.4 0.35 ® 0.3 >_0.25 0.2 0.15 0.1 O.05 _ I - - _..__- - o : : : g'-'- : _'- i :__- : ' _ I Non-veg Agrtcul N.hdwds Aspen Lowhdwd Wetlands W.plne MIXpine Cedar Mixconifer Coverclass _Bedrock _ Lowland _Moraine _Outwash Figure 4a.--Patch density, the number of patches per 100 ha, calculated by cover type as a function of LTA group using the historic landscape dataset (MNH 1994). 0.5 0.45 0.4 0.35 ® 0.3 >_0.25 0.2 0.15 0.1 O.05 _ I - - _..__- - o : : : g'-'- : _'- i :__- : ' _ I Non-veg Agrtcul N.hdwds Aspen Lowhdwd Wetlands W.plne MIXpine Cedar Mixconifer Coverclass ash ash Figure 4a.--Patch density, the number of patches per 100 ha, calculated by cover type as a function of LTA group using the historic landscape dataset (MNH 1994). 1.5 _=¢o > 1 r_i 0.5 o l I Non-veg Agricul N.hdwds Aspen Lowhdwd Wetlands W.pine Mixpine Cedar Mixconifer Coverclass _ Bedrock _ Lowland _ Moraine _ Outwash Figure 4b.---Patch density, the number of patches per 100 ha, calculated by cover type as a function of LTA group using the present landscape dataset (Maclean 1994). D. Assumptions About the Data and Inherent Bias or Limitations The transcribed digital data from MNFI, however, contained little attribute information, or legend about the cover types derived from the original line tree data, except the compositional information derived from this study, and its resolution was definable only to 1 square mile. These data limitations are not uncommon to landscape research, •however. Frelich's (1995) study of "old forest" changes in the Lake States over the past 150 _years relied on historic vegetation maps by Veatch (1928), who inductively mapped his- toric Vegetation from GLO notes and his knowledge of soil patterns at a broader scale than the maps by MNFI. hardwoods and mixed conifer were the most prevalent cover types of the 10 studied, his- toricaUy and currently. Northern hardwoods were especially prevalent in the moraine group, while the mixed conifer type was more prevalent in the bedrock group. Wetlands and mixed pine, in addition to northern hard- woods, were also prominent in the lowland group. In the outwash group, these types were also present but mixed and white pine were more prevalent. Largest single patches (LPI) were found to be redundant to LSIM, and therefore were not assessed to the same extent as other indices. However, in comparisons of mean patch sizes (MPS) of classes, LPI can be used to gage how much MPS is determined by a single large patch. LPI also may suggest areas with more interior habitat. The highest LPI values were found in northern hardwoods within the moraine group. Patch density was useful in comparing the scale of pattern within landscapes. The bedrock group had the lowest patch density, or coarsest (least patchy) pat- tern historically, yet in general, mixed conifer, which was a dominant class in the bedrock group, tended to have higher patch density than other cover types. Generally, cover types in the moraine group had the most variable (least regular) pattern, but no one cover class was consistently more variable than others. The present coverage was derived from differ- ent methodology: classification of sateUite imagery, which comprehensively recorded land cover in the total study area based on the spectral signature of vegetation types within the image area, rather than a sample. Resolu- tionwas as high as 60 m to an accuracy of 90.2 percent. However, some types, such as non-vegetated, had an accuracy as low as 28.6-percent correct. D. Assumptions About the Data and Inherent Bias or Limitations 1.5 _=¢o > 1 r_i 0.5 o l I Non-veg Agricul N.hdwds Aspen Lowhdwd Wetlands W.pine Mixpine Cedar Mixconifer Coverclass _ Bedrock _ Lowland _ Moraine _ Outwash Figure 4b.---Patch density, the number of patches per 100 ha, calculated by cover type as a function of LTA group using the present landscape dataset (Maclean 1994). Figure 4b.---Patch density, the number of patches per 100 ha, calculated by cover type as a function of LTA group using the present landscape dataset (Maclean 1994). 13 4 [ 3.5 -_......................................................................................................................... 3 ............ R ........................................................................................................................................................... _ t ti]' i l i >: 2!........ ;! ...................................................................................................i 1._! i,ii: !i i 1!...... i! ............................................................. i ........... I ............................i •-I!!i_i_:_ :i ! i Non-veg Agricul N.hdwds Aspen Low hdwd Wetlands W.pine Mix pine Cedar Mix conifer Cover class _Bedrock _ Lowland _Moraine _Outwash Figure 5a.---Meen shape index, a perimeter to area ratio, cal_J!_ted by cover type as a function of LTA group using the historic landscape dataset (MNFI 1994). Figure 5a.---Meen shape index, a perimeter to area ratio, cal_J!_ted by cover type as a function of LTA group using the historic landscape dataset (MNFI 1994). 4 3.5 'I" ............................................................................................................................................................................................................. ! ! o3 2.5 ....t....................................................................................................................................................................................................... I 2- i....................................................................................................................................................................................................... N a,__'N N 1.5 i ................................ ....... i .,, 1 Non-veg Agricul N hdwds Aspen Low hdwd Wetlands W.pme Mix pine Cedar Mix conifer Cover class Figure 5b.---Mean shape index, a perimeter to area ratio, calculated by cover type as a function of LTA group using the present landscape dataset (MNFI 1994). 4 3.5 'I" ............................................................................................................................................................................................................. ! ! o3 2.5 ....t........................................................................................................................................................................................................ I 2- i........................................................................................................................................................................................................ N a,__'N N 1.5 i ................................ ....... i .,, 1 Non-veg Agricul N hdwds Aspen Low hdwd Wetlands W.pme Mix pine Cedar Mix conifer Cover class Figure 5b.---Mean shape index, a perimeter to area ratio, calculated by cover type as a function of LTA group using the present landscape dataset (MNFI 1994). Figure 5b.---Mean shape index, a perimeter to area ratio, calculated by cover type as a function of LTA group using the present landscape dataset (MNFI 1994). 14 Sampling completeness of the land cover datasets was another concern. The historic coverage was based on the equivalent of a systematic aligned sampling strategy. The surveyors walked and recorded trees along the township and section survey lines (1-mile intervals). D. Assumptions About the Data and Inherent Bias or Limitations Furthermore, the present coverage, like the historic, lacked a compositional/structural legend for the cover classes. Neither historic or present datasets described, or were intended to describe, verti: cal structure. Nor did they provide detailed species composition information, beyond the compositional assessment included in this study, as would maps developed from plot samples and quantitative ecological classiflca- tion techniques, Shape indices, based on perimeter to area ratios, were relatively consistent across land- scape groups and cover classes, with the exception of the non-vegetated class. Non- vegetated areas had higher shape indices in both historic and present coverages. In the present coverage, this finding was attributed to inclusion of roadways, which have high perimeter/area values in the non-vegetated cover class. However, it is not clear what factors would have contributed to high shape indices in historic non-vegetated areas. V. CONCLUSIONS Diversity, based on Shannon's Diversity Index, was highest in the outwash group and lowest in the lowland group of the historic land- scapes. In present landscapes, diversity was very similar among the four groups; it was only slightly higher in the lowland group and lower in the moraine group than in others. This study has quantified landscape patterns •for four physiographically based landtype association groups across the Eastern Upper Peninsula of Michigan from historic (1840's) and current (1990-1992) datasets. Prevalence or dominance of cover classes was based on class area (CA) and landscape similarity index (LSIM), or class area weighted by total land- scape area. In the four groups studied-- bedrock-controlled, lowland sand lake plain, •morainal origin, and outwash--northern Conclusions about a specific cover type may be drawn by referring to tables 6 and 7. These tables should be a reference for land managers and planners making decisions based on 15 knowledge of actual landscape structure. The broad landscape data in this report may be used in conjunction with s/milar landscape studies in the Lake States or in other areas with similar biophysical landscape units to estimate patterns beyond northern Michigan. Albert, Dennis A.; Denton, Shirley R.; Barnes, Burton V. 1986. Regional landscape eco- systems of Michigan. Ann Arbor, MI: Uni- versity of Michigan, School of Natural Re- sources. 32 p. knowledge of actual landscape structure. The broad landscape data in this report may be used in conjunction with s/milar landscape studies in the Lake States or in other areas with similar biophysical landscape units to estimate patterns beyond northern Michigan. Allen, T.F.H.; Start, Thomas B. 1982. HIerar- ©hy: perspectives for ecological complex- ity. Chicago, IL: University of Chicago Press. 30 p. ACKNOWLEDGMENTS Seleen, and I_ Brittain greatly advanced the landtype association delineation program. AppreciaUon also goes to J. Johnson and j. Durbin, of Mead Paper, Inc., in Escanaba, Mich/gan, and to Q. Zhang, C. Herlevich, Q. Zhao, and M. Hyslop of Michigan Technological University, for GIS preparaUon, and Dr. N. Crookston, of the Intermountain Forest Experiment Station in Moscow, for computer support. Special •thanks to Dr. T. Crow of the Forestry Sciences Laboratory, North Central Forest Experiment Station, in Rhinelander, Wisconsin for his 'support and many suggestions on this project. Anderton, John A. 1993. Paleoshoreline geoarchaeology in the Northern Great Lakes: Hiawatha National Forest. Heritage Program Monograph No. 1. g gy Lakes: Hiawatha National Forest. Heritage Program Monograph No. 1. Barley, Robert G. 1996. Multi-scale ecosys- tern analysis. Environmental Monitoring & Assessment. 39:21-24. Barley, I_G.; Avers, P.E.; King, T.; McNab, W.H., eds. 1994. Ecoregions and subre- gions of the United States. Washington, DC: U.S. Department of Agriculture, Forest Service, ECOMAP Team. 1 map (1:7,500,000). Baker, William L. 1991. Effects of settlement and fire suppression on landscape strue- ture. Ecology. 73: 1879-1887. Ball, Janet; Padley, Eunice A. 1993. Draft delineation of landtype associations in the Eastern Upper Peninsula of Michigan. On file, Hiawatha National Forest. Barnes, Burton V. 1984. Forest ecosystem classification and mapping in Baden- Wurttemberg, West Germany. In: Bockheim, J.G., ed. Proceedings of the symposium on forest land classificaUon: experience, problems, perspectives; 1984 March; Madison, WI. NCR-102. 276 p. Bourdo, Eric A., Jr. 1956. A review of the General Land Office survey and of its use in quantitative studies of former forests. Ecology. 37(4): 754-768. ACKNOWLEDGMENTS Partial support and funding were provided by the Michigan Department of Natural Re- sources, the National Council for Air and Stream Improvement, The Nature Conser- Vancy, Michigan Technological University, the Hiawatha National Forest, and Washington State University. Recognition for development of spatial land cover datasets goes to Dr. A. Maclean and Maclean Consultants, Ltd., and to Dr. D. Albert of Michigan Natural Features Inventory. Dr. E. Padley and J. K. Jordan, with the Eastern Region of the Forest Service, and I_ L. Watson, initially developed the landtype associaUon mapping on Forest Service lands in this region, and collaborated on expansion of the map units across the study area. We also appreciate the vision of D. Howlett, Ecosystem Team Leader, to extend the landtype association mapping beyond the boundaries of the Hiawatha National Forest. A number of other people on the Hiawatha National Forest, especially W. F. Spinner, D. Evans, J. Schultz, D. Gerdes, K. Seleen, and I_ Brittain greatly advanced the landtype association delineation program. AppreciaUon also goes to J. Johnson and j. Durbin, of Mead Paper, Inc., in Escanaba, Mich/gan, and to Q. Zhang, C. Herlevich, Q. Zhao, and M. Hyslop of Michigan Technological University, for GIS preparaUon, and Dr. N. Crookston, of the Intermountain Forest Experiment Station in Moscow, for computer support. Special •thanks to Dr. T. Crow of the Forestry Sciences Laboratory, North Central Forest Experiment Station, in Rhinelander, Wisconsin for his 'support and many suggestions on this project. Partial support and funding were provided by the Michigan Department of Natural Re- sources, the National Council for Air and Stream Improvement, The Nature Conser- Vancy, Michigan Technological University, the Hiawatha National Forest, and Washington State University. Recognition for development of spatial land cover datasets goes to Dr. A. Maclean and Maclean Consultants, Ltd., and to Dr. D. Albert of Michigan Natural Features Inventory. Dr. E. Padley and J. K. Jordan, with the Eastern Region of the Forest Service, and I_ L. Watson, initially developed the landtype associaUon mapping on Forest Service lands in this region, and collaborated on expansion of the map units across the study area. We also appreciate the vision of D. Howlett, Ecosystem Team Leader, to extend the landtype association mapping beyond the boundaries of the Hiawatha National Forest. A number of other people on the Hiawatha National Forest, especially W. F. Spinner, D. Evans, J. Schultz, D. Gerdes, K. Hess, George. 1994. Pattern and error in landscape ecology: a commentary. Land- scape Ecology. 9(I)" 3-5. Crow, Thomas I_ 1991. Landscape ecology: the big picture approach to resource management. In: Decker, Daniel J., et a/., eds. Challenges in the conservation of bio- lOgical resources: a practitioner's guide, Boulder, CO: Westview Press: 55-65. Hulshoff, Maureen I_ 1995. Landscape indi- ces describing a Dutch landscape. Land- scape Ecology. 10(2): 101-111. Jordan, James K.; Padley, Eunice; Cleland, David T.; Gates, Joseph ,_; Hoppe, David J.; Kempf, Leonard S.; Leuelling, Barbara; Seleen, Kirstin L.; Shadis, David A.; Sflbernagel, Janet. In prep. Landtype asso- ciations--odgin, concepts, mapping, and application in the Lake States National Forests. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. Curtis, JohnT. 1959. Vegetation of Wiscon- sin. Madison, WI: University of Wisconsin Press. 657 p. ' Dunn, Christopher P.; Sharpe, David M.; ) Guntenspergen, Glenn I_; Stearns, Forrest; Yang, Zhao. 1991. Methods for analyzing temporal changes in landscape pattern. In: Turner, Monica G.; Gardener, Robert H., eds. Quantitative methods in landscape ecology. Ecol. Stud. 82. New York, NY: Springer-Verlag: 173-198. King, ,aLW. 1993. Consideration of scale and hierarchy. In: Woodley, S.; Roy, J.; Francis, G., eds. Ecological integrity and the manage- ment of ecosystems. Delray Beach, FL: St. Lucie Press. 220 p. ECOMAP. 1993. National hierarchical frame- work of ecological units. Washington, DC: U,S. Department of Agriculture, Forest Service. KliJn, Frans; Udo de Haes, Helias _ 1994. A hierarchical approach to ecosystems and its implications for ecological land classi- flcation. Landscape Ecology. 9(2)" 89-104. Farrand, W.I_ 1982. Quaternary geology of Michigan. Ann Arbor, MI: State of Michigan . Department of Natural Resources, Geological Survey, University of Michigan. 2 maps Michigan. Ann Arbor, MI: State of Michigan . Department of Natural Resources, Geological Survey, University of Michigan. 2 maps Li, Habin; Reynolds, James F. 1993. A new contagion index to quantify spatial pat- terns of landscapes. Landscape Ecology. 8(3)" 155-162. Forman, Richard T.T. 1995. Some general principles of landscape and regional ecology. Landscape Ecology. 10(3): 133-142. Li, Habin; Reynolds, James F. 1994. A shnula- tion experiment to quantify spatial het- erogeneity in categorical maps. Ecology. 75(8): 2446-2455. Forman, Richard T.T.; Godron, Michael. 1986. Landscape ecology. New York, NY: John Wiley and Sons, Inc. 619 p. Frelich, Lee E. 1995. Old forest in the Lake States today and before European settle- ment. Natural Areas Journal. 15: 157-167. Lorimer, Craig G. 1977. The presettlement forest and natural disturbance cycle of northeastern Maine. Ecology. 58: 139-148. Gary, Margaret; McAfee, Robert, Jr.; Wolf, Carol L., eds. 1974. Glossary of geology, Washington, DC: American Geological Insti- tute. 805 p. Canham_ Charles D.; Loucks, Orie L. 1984. Catastrophic windthrow in the pre-settle- merit forests of Wisconsin. Ecology. 65(3)" 803-809. Canham_ Charles D.; Loucks, Orie L. 1984. Catastrophic windthrow in the pre-settle- merit forests of Wisconsin. Ecology. 65(3)" 803-809. Hess, George. 1994. Pattern and error in landscape ecology: a commentary. Land- scape Ecology. 9(I)" 3-5. LITERATURE CITED Albert, Dennis ,_ 1995. Regional landscape ecosystems of Michigan, Minnesota, and Wisconsin: a working map and classiflca- tion. Gen. Tech. Rep. NC-178. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 250 p. Canham, Charles D. 1978. Catastrophic windthrow in the hemlock-hardwood forest of Wisconsin. Madison, WI: Unlver- sity of Wisconsin. 94 p. M.S. thesis. 16 GoUey, Frank B. 1995. Reaching a landmark. Landscape Ecology. 10(1): 3-4. Canham_ Charles D.; Loucks, Orie L. 1984. Catastrophic windthrow in the pre-settle- merit forests of Wisconsin. Ecology. 65(3)" 803-809. Hess, George. 1994. Pattern and error in landscape ecology: a commentary. Land- scape Ecology. 9(I)" 3-5. Frelich, Lee E.; Lorlmer, Craig G. 1991. Natu- ral disturbance regimes in hemlock- hardwood forests of the upper Great Lakes region. Ecological Monographs. 61(2): 145- 164. Lorimer, Craig G. 1980. The use of land survey records in estimating presettle- ment fire frequency. In: Stokes, MA.; Dieterich, J.H., eds. Proceedings of the Fire history workshop; 1980 October 20-24; Tucson, AZ. Gen. Tech. Rep. GTR-RM-81. Fort Collins, CO: U.S. Department of Agricul- ture, Forest Service, Rocky Mountain Forest and Range Experiment Station. 142 p. Gary, Margaret; McAfee, Robert, Jr.; Wolf, Carol L., eds. 1974. Glossary of geology, Washington, DC: American Geological Insti- tute. 805 p. 17 17 Lor/mer, Craig C. 1984. Methodological considerations in the analysis of forest disturbance history. Canadian Journal of Forest Research. 15: 200-213. Muller, Michael I_; Middleton, John. 1994. A Markov model of land-use change dynam- ics in the Niagara Region, Ontario, Canada. Landscape Ecology. 9(2): 151-157. Noss, Reed F. 1985. On characterizing presettlement vegetation: how and why. Natural Areas Journal. 5: 5-19. Luque, Sandra S.; Lathrop, Richard G.; Bognar, John A. 1994. Temporal and spa- tial changes in an area of the New Jersey Pine Barrens landscape. Landscape Ecol- ogy. 9(4}: 287-300. O_leill, I_V.; Drummel, J.l_; Gardner, I_H.; Sugthara, G.; Jackson, B.; DeAngelis, D.L.; Mflne, B.T.; Turner; M.G.; Zygmunt, B.; Chr/stensen, S.W.; Dale, V.H.; Graham, I_L. 1988. Indices of landscape pattern. Land- scape Ecology. 1(3}: 153-162. MacGarlgal, Key/n; Marks, Barbara J. 1993. FRAGSTATS: spatial analysis program for quantifying landscape structure. Corvallis, O1_ Oregon State University. 62 p. Maclean, Gordon, A. 1994. Inventory of deer habitat in Michigan's Upper Peninsula utilizing Landsat thematic mapper data. Final report for Michigan Department of Natural Resources, _sing, MI. Pastor, John; Broschart, Michael. 1990. The spatial pattern of a northern conifer- hardwood landscape. Landscape Ecology. 4(1): 55-68. Peterson, W.L. 1986. Late Wisconsinan glacial history of Northeastern Wisconsin and Western Upper Michigan. USGS Bull. 1652. Washington, DC: U.S. Geological Survey. United States Government Printing Office. 14 p. Martin, Helen M. 1957. Map of the surface formations of the Northern Peninsula of Michigan. Publ. 49. Michigan Department of Conservation, Geological Survey Division. 1 •map. Price, David L. 1994. An ecological study of the composition, structure and distur- bance regimes of the pre-European settle- ment forests of western Chippewa County, Michigan. East Lansing, MI: Michi- gan State University, Department of For- estry. 214 p. M.S. thesis. McNab, W. Henry. 1996. Santer, Richard _ 1977. Michigan: heart of the Great Lakes. Dubuque, IA: Kendall/ Hunt Publishing Co. 364 p. Hess, George. 1994. Pattern and error in landscape ecology: a commentary. Land- scape Ecology. 9(I)" 3-5. Classification of •local- and landscape-scale ecological types in the Southern Appalachian Moun- tains. Env/ronmental Monitoring & Assess- ment. 39:215-229. McNab, W. Henry; Avers, Peter E., eds. 1994. Ecological subregions of the United States: section descriptions. WO-WSA-5. Prepared in cooperation with R_gional Compilers and the ECOMAP Team of the Forest Service. Rowe, J. Start. 1991. Forests as landscape ecosystems: implications for their regionalization and classification. In: Symposium: Ecological land classification: applications to identify the productive poten- tial of southern forests; 1991 January 7-9; Charlotte, NC. Ashevflle, NC: U.S. Depart- ment of Agriculture, Forest Service, South- eastern Forest Experiment Station. 149 p. Mflne, Bruce T. 1991. Lessons from applying fractal models to landscape patterns. In: .Turner, Monica G.; Gardener, Robert H., eds. Quantitative methods in landscape ecology, Ecol. Stud. 82. New York, NY: Springer- Verlag: 199-235. R_we, J. Stan. 1996. Land classification and ecosystem classification. Environmental Monitoring & Assessment. 39:11-20. MNFI. 1994. Pre-European settlement land- scape of Michigan. Lansing, MI: Michigan Natural Features Inventory. Nature Conservancy. 1993. IAmdscape con- servation--the next big step. News from the Michigan Chapter of The Nature Conser- vancy, Autumn. MNFI. 1994. Pre-European settlement land- scape of Michigan. Lansing, MI: Michigan Natural Features Inventory. Santer, Richard _ 1977. Michigan: heart of the Great Lakes. Dubuque, IA: Kendall/ Hunt Publishing Co. 364 p. Nature Conservancy. 1993. IAmdscape con- servation--the next big step. News from the Michigan Chapter of The Nature Conser- vancy, Autumn. 18 Uhlig, Peter W.C.; Jordan, James K. 1996. A spatial hierarchical framework for the co- management of ecosystems in Canada and the United States for the Upper Great Lakes Region. Environmental Monitoring & Assessment. 39: 59-73. Simpson, John W.; Boerner, Ralph E.J.; DeMers, Michael N.; Berns, Leslie IL; Artigas, Francisco J.; Silva, Alejandra. 1994. Forty-eight years of landscape change on two contiguous Ohio landscapes. Land- scape Ecology. 9(4): 261-270. Urban, Dean L.; O'Neill, Robert V.; Shugart, Herman H., Jr. 1987. Landscape ecology, a hierarchical perspective can help scien- tists understand spatial patterns. Bio- science. 37: 119-127. Slms, Richard ,_; Corns, Ian G.W.; Kllnka, Karel. 1996. Introduction--global to local: ecological land classification. Environmen- tal Monitoring & Assessment. 39: 1-10. Smith, Marie-Louise; Carpenter, Constance. i996. Application of the USDA Forest Service national hierarchical framework of ecological units at the sub-regional , level: the New England-New York example, Environmental Monitoring & Assessment. 39: 187-i98. USGS. 1990. In: Anderson, James I_, et aL, eds. Turner, Monica G.; Ruscher, C. Lynn. 1988. Changes in landscape patterns in Georgia, USlL Landscape Ecology. 1(4): 241-251. Spurr, S.H.; Barnes, Burton V., eds. 1980. Forest ecology. 3d ed. New York, NY: John Wfley and Sons. 687 p. Hess, George. 1994. Pattern and error in landscape ecology: a commentary. Land- scape Ecology. 9(I)" 3-5. A land use and land cover classiflca- tion system for use with remote sensor data. Reston, V_: U.S. Geological Survey. Veatch, J.O. 1928. Reconstruction of forest cover based on soil maps. Quarterly BulIe- tin of the Michigan Agricultural Experiment Station. I0:116-126. SOLEC. In Press. Impacts of changing land use. In: Thorp, Steve; Rivers, Ray; Pebbles, Victoria, eds. Proceedings of the 1996 State of the Lakes ecosystem conference; 1996 November; Windsor, Ontario. EPA 905-D-96- 00 l e. Windsor, Ontario: Environment Canada and U.S. Environmental Protection Agency. Veatch, J.O. 1959. Presettlement forests of Michigan. East Lansing, MI: Department of Natural Resources, Michigan State Univer- sity. Map. Whitney, Gordon G. 1986. Relation of Michigan's presettlement pine forests to substrate and disturbance history. Ecol- ogy. 67: 1548-1559. Whitney, Gordon G. 1986. Relation of Michigan's presettlement pine forests to substrate and disturbance history. Ecol- ogy. 67: 1548-1559. Spurr, S.H.; Barnes, Burton V., eds. 1980. Forest ecology. 3d ed. New York, NY: John Wfley and Sons. 687 p. 19 APPENDIX A: Formulae used to calculate landscape, class, and relative indices Landscape Indices: 1. Number of Patches = N(lJ) 2. Total Area = SUM a(iJ) = A (ha) 3. Mean Patch Size (MPS) = SUM a(iJ)/N(ij) 4. _est Patch Index (LPI) = max a(iJ)/A * I00 5. Patch Density (PD) = n(iJ)/A * I0000 * I00 (# per I00 ha) 6. Diversity {SI-IDI)= -SUM (P(i) * log P(i)) Class Indices: 1. Number of Patches = N(J) 2. Class Area = SUM a(J)= CA (ha) 3. Mean Patch Size (MPS) = SUM a(J)/N(i) 4. Landscape Simfliarity Index (LSIM) = SUM a(j)/A * I00, or CA/A * I00 5. Largest Patch Index (LPI) = max a(J)/CA * 100 6. Patch Density (PD) = n_])/CA * I0000 * I00 (# per I00 ha) 7. Patch Size Coefficient of Variation (PSC_ = patch size standard deviation / mean patch size 8. Mean Shape Index (MSI) = SUM (p(J)/2sqrt(pieaU))) / N(J) a(iJ} = area of patch iJ p(iJ) = perimeter of patch lJ P(i} = proportion of patch type I = a(iJ)/A Landscape Indices: 1. Number of Patches = N(lJ) 2. Total Area = SUM a(iJ) = A (ha) 3. Mean Patch Size (MPS) = SUM a(iJ)/N(ij) 4. _est Patch Index (LPI) = max a(iJ)/A I00 5. Patch Density (PD) = n(iJ)/A * I0000 * I00 (# per I00 ha) 6. Diversity {SI-IDI)= -SUM (P(i) * log P(i)) Class Indices: 1. Number of Patches = N(J) 2. Class Area = SUM a(J)= CA (ha) 3. Mean Patch Size (MPS) = SUM a(J)/N(i) 4. Landscape Simfliarity Index (LSIM) = SUM a(j)/A * I00, or CA/A * I00 5. Largest Patch Index (LPI) = max a(J)/CA * 100 6. Patch Density (PD) = n_])/CA * I0000 * I00 (# per I00 ha) 7. Patch Size Coefficient of Variation (PSC_ = patch size standard deviation / mean patch size 8. Mean Shape Index (MSI) = SUM (p(J)/2sqrt(pieaU))) / N(J) a(iJ} = area of patch iJ p(iJ) = perimeter of patch lJ P(i} = proportion of patch type I = a(iJ)/A Landscape Indices: Landscape Indices: 1. Number of Patches = N(lJ) 2O Table l.--IJst of cover classes and numeral codes for historic land cover map developed by MN_ (1994) GIX) Surveyors' notes of 1840"s; legend is based on expanded Michigan Resowce Informatk)n System (MIRIS) legend of land use cover classes Category • 1st order 2nd order 3rd order 4th order Land cover class description Palustrlne 62 Non-forestedWetland , .. 622 Emergentmarshmeadowprairie 6221 Eemergentmarsh , 6222 Great lakes marsh 6223 Interdunalwetland " 6224 Wet meadow 6225 Inlandsalt marsh 6226 Llakeplainprairie 6227 Inlandwet prairie 6228 Intermittentwetland 623 Mudflats 6231 Marlflats 612 Shrub-dominatedwetland 6121 Bog 6122 Alder/willow/Bogbirchthicket 6123 Bottonbush/DogwoodNVi!low swamp 6124 Patternedpeatland 6125 Muskeg 91 LandscapeComplex 911 Wooded dune/swalecomplex 4 ForestedWetlands 41 Hardwood/Conifer- Hardwoodspredominating 414 Lowlandhardwood(broadleaf) 4141 Ash 4142 Elm 4143 Silver/Redmaple 4144 Cottonwood 4145 Balsampoplar 4146 Aspen 4147 White birch 4148 blackwillow 42 Hardwood/Conifer-Coniferspredominating 423 Lowlandconifer 4231 Cedar 4232 blackspruce 4233 Tamarack 4234 Balsamfir/Whitespruce Palustrlne 62 Non-forestedWetland , .. (table I continued on next page) (table I continued on next page) Formulae used to calculate landscape, class, and relative indices Landscape Indices: 1. Number of Patches = N(lJ) 2. Total Area = SUM a(iJ) = A (ha) 3. Mean Patch Size (MPS) = SUM a(iJ)/N(ij) 4. _est Patch Index (LPI) = max a(iJ)/A I00 5. Patch Density (PD) = n(iJ)/A * I0000 * I00 (# per I00 ha) 6. Diversity {SI-IDI)= -SUM (P(i) * log P(i)) Class Indices: 1. Number of Patches = N(J) 2. Class Area = SUM a(J)= CA (ha) 3. Mean Patch Size (MPS) = SUM a(J)/N(i) 4. Landscape Simfliarity Index (LSIM) = SUM a(j)/A * I00, or CA/A * I00 5. Largest Patch Index (LPI) = max a(J)/CA * 100 6. Patch Density (PD) = n_])/CA * I0000 * I00 (# per I00 ha) 7. Patch Size Coefficient of Variation (PSC_ = patch size standard deviation / mean patch size 8. Mean Shape Index (MSI) = SUM (p(J)/2sqrt(pieaU))) / N(J) a(iJ} = area of patch iJ p(iJ) = perimeter of patch lJ P(i} = proportion of patch type I = a(iJ)/A 622 Emergentmarshmeadowprairie 6221 Eemergentmarsh , 6222 Great lakes marsh 6223 Interdunalwetland " 6224 Wet meadow 6225 Inlandsalt marsh 6226 Llakeplainprairie 6227 Inlandwet prairie 6228 Intermittentwetland 623 Mudflats 6231 Marlflats 612 Shrub-dominatedwetland 6121 Bog 6122 Alder/willow/Bogbirchthicket 6123 Bottonbush/DogwoodNVi!low swamp 6124 Patternedpeatland 6125 Muskeg 91 LandscapeComplex 911 Wooded dune/swalecomplex 4 ForestedWetlands 41 Hardwood/Conifer- Hardwoodspredominating 414 Lowlandhardwood(broadleaf) 4141 Ash 4142 Elm 4143 Silver/Redmaple 4144 Cottonwood 4145 Balsampoplar 4146 Aspen 4147 White birch 4148 blackwillow 42 Hardwood/Conifer-Coniferspredominating 423 Lowlandconifer 4231 Cedar 4232 blackspruce 4233 Tamarack 4234 Balsamfir/Whitespruce 4235 Balsamfir 4236 Jackpine , 4237 Hemlock 4238 White pine LacUstrine and Rlverine 51 MajorRiver 52 Lake or Pond 53 Great Lakes (table I continued on next page) 21 et order 2nd order 3rd order 4th order Land cover class description _. Formulae used to calculate landscape, class, and relative indices Landscape Indices: 1. Number of Patches = N(lJ) 2. Total Area = SUM a(iJ) = A (ha) 3. Mean Patch Size (MPS) = SUM a(iJ)/N(ij) 4. _est Patch Index (LPI) = max a(iJ)/A I00 5. Patch Density (PD) = n(iJ)/A * I0000 * I00 (# per I00 ha) 6. Diversity {SI-IDI)= -SUM (P(i) * log P(i)) Class Indices: 1. Number of Patches = N(J) 2. Class Area = SUM a(J)= CA (ha) 3. Mean Patch Size (MPS) = SUM a(J)/N(i) 4. Landscape Simfliarity Index (LSIM) = SUM a(j)/A * I00, or CA/A * I00 5. Largest Patch Index (LPI) = max a(J)/CA * 100 6. Patch Density (PD) = n_])/CA * I0000 * I00 (# per I00 ha) 7. Patch Size Coefficient of Variation (PSC_ = patch size standard deviation / mean patch size 8. Mean Shape Index (MSI) = SUM (p(J)/2sqrt(pieaU))) / N(J) a(iJ} = area of patch iJ p(iJ) = perimeter of patch lJ P(i} = proportion of patch type I = a(iJ)/A Terrestrial 3 Non-Forested(grassland- savannas) 31 Herbaceous- UplandGrassland 32 Shrub- ShrubSavanna 33 Tree Savanna 331 Lakeplainoak opening 332 Oak barrens 333 Pinebarrens 334 Oak/pinebarrens 335 Buroak savanna • 336 Oakopening 4 Forested 411 Northernhardwoods 4111 Sugarmaple, Beech 4119 Beech,Hemlock 412 Centralhardwoods 4121 Beech,Sugar maple, Basswood,Redoak 4122 Whiteoak, Hickory,Blackoak 4123 Blackoak, Whiteoak 4124 Pin/Blackoak 413 Aspen/Whitebirch 421 Pine 4211 Whitepine 4212 Red pine 4213 Jackpine 4215 Red pine/Jackpine 4216 Red pine/Whitepine 4217 Whitepine/Whiteoak 4218 Red pine/Oak 4219 White pinelBeechlMaple 422 Other uplandconifer 4221 White spruce 4223 FirlSprucelCedar 4226 Hemlock 4227 Hemlock/Whitepine • 4228 Hemlock/Sugarmaple 4229 HemloddYellow birch Open, LltUe/NoVegetation 72 Beach,Riverbank 73 Open SandDune 74 ExposedBedrock 741 Alvar 742 Bedrockglade 743 Sinkhole 744 Limestoneledge/Outcrop 745 Sandstoneledge/Outcrop 746 Igneous-met.ledgelOutcrop (table I continued on next p (table I continued on next page) 22 (table I conttnue_ Category 1st order 2nd order 3rd order 4th order Land cover class description Natural - DistUrbances/ Cultural Feature 92 Wildfire 93 Windthrow 94 Beaver Flooding 95 Great LakesLevelChange 96 CulturalFeatures 97 Jack PineThicket Wetland Grade I Intact D Degraded E Eliminated M Manipulated (table I conttnue_ 2nd order 3rd order 4th order Land cover class description DistUrbances/ Cultural Feature 92 Wildfire 93 Windthrow 94 Beaver Flooding 95 Great LakesLevelChange 96 CulturalFeatures 97 Jack PineThicket Wetland Grade I Intact D Degraded E Eliminated M Manipulated 23 Table 2.DLIst of cover classes and numer_al codes for present land cover map (Maclean 1994) through classifi- cation of 1991 Thematic Mapper satellite imagery Table 2.DLIst of cover classes and numer_al codes for present land cover map (Maclean 1994) through classifi- cation of 1991 Thematic Mapper satellite imagery Cover class Numerical code Category Cover type name Cover class Numerical code Non-Coniferous Urban 1 Non-vegetative •2 Agricultural-cropland 3 Herbaceousopenland 4 Shrubland 5 Northernhardwood 6 Oak 7 Aspen/birch 8 Lowlandhardwoods 9 Dry hardwood/conifermix 10 Wet hardwood/conifermix 11 Wetlands 12 Water 13 Coniferous Pines Red pine 14 Jackpine 15 White pine 16 Other(mixedpine) 17 Tamarack 18 Hemlock <70% crownclosure 19 >70% crownclosure 20 Blackspruce <70% crownclosure 21 >70% crownclosure 22 White spruce <70% crownclosure 23 >70% crownclosure 24 Balsamfir <70% crownclosure 25 >70% crownclosure 26 White cedar <70% crownclosure 27 >70% crownclosure 28 Mixedconifer <70% crownclosure 29 >70% crownclosure 30 Non-Coniferous 24 of 24 Cross-llsted cover classes used to compare historic and present land cover maps Table 3.--List of 24 Cross-llsted cover classes used to compare historic and present land cover maps MNFI Code" TM code2 Class name MNFI Code _ TMcode: Class name not applicable 1 Urban 6221 12 72 2 Non-vegetated 6222 12 73 2 6223 12 74 2 6224 12 741 2 6227 12 742 2 6228 12 743 2 6231 12 744 2 51 13 Water 745 2 52 13 746 2 54 13 notapplicable 3 Agriculutralcropland 4212 14 Red pine 31 4 Herbaceousopenland 4211 15 Jackpine 32 5 Shrubland 4238 15 4111 6 Northernhardwoods 333 16 White pine 4119 6 334 16 332 7 Oak 4213 16 4121 7 4236 16 4122 7 4215 17 Mixedpine 4123 7 4216 17 4124 7 4217 17 413 8 Aspen/Birch 4218 17 •41 9 Lowlandhardwoods 4219 17 414 9 4233 18 Tamarack 4141 9 4226 19 Hemlock 4142 9 4227 19 4143 9 4228 19 4144 9 4229 19 4145 9 4237 19 4146 9 4232 21 Blackspruce 4147 9 4242 21 4148 9 4221 23 White spruce notapplicable 10 Dry hardwood-conifermix 42 25 Balsamfir 42 11 Wet hardwood-conifermix 4234 25 6121 12 Wetlands 4235 25 6122 12 4231 27 Whitecedar .6124 12 423 29 6125 12 4223 29 Mixed conifer • . tural Features Inventory (MNFI 1994) code from historic cover class legend (table I). _Michlgan Natural Features Inventory (MNFI 1994) code from historic cover class legend (table I). ='[hematlc Mapper code from present cover class legend (Maclean 1994, table 2). tural Features Inventory (MNFI 1994) code from historic cover class legend (table I). apper code from present cover class legend (Maclean 1994, table 2). Formulae used to calculate landscape, class, and relative indices Landscape Indices: 1. Number of Patches = N(lJ) 2. Total Area = SUM a(iJ) = A (ha) 3. Mean Patch Size (MPS) = SUM a(iJ)/N(ij) 4. _est Patch Index (LPI) = max a(iJ)/A I00 5. Patch Density (PD) = n(iJ)/A * I0000 * I00 (# per I00 ha) 6. Diversity {SI-IDI)= -SUM (P(i) * log P(i)) Class Indices: 1. Number of Patches = N(J) 2. Class Area = SUM a(J)= CA (ha) 3. Mean Patch Size (MPS) = SUM a(J)/N(i) 4. Landscape Simfliarity Index (LSIM) = SUM a(j)/A * I00, or CA/A * I00 5. Largest Patch Index (LPI) = max a(J)/CA * 100 6. Patch Density (PD) = n_])/CA * I0000 * I00 (# per I00 ha) 7. Patch Size Coefficient of Variation (PSC_ = patch size standard deviation / mean patch size 8. Mean Shape Index (MSI) = SUM (p(J)/2sqrt(pie*aU))) / N(J) a(iJ} = area of patch iJ p(iJ) = perimeter of patch lJ P(i} = proportion of patch type I = a(iJ)/A _Michlgan Natural Features Inventory (MNFI 1994) code from historic cover class legend (table I). ='[hematlc Mapper code from present cover class legend (Maclean 1994, table 2). 25 g_'U.S. GOVERNMENT PRINTING OFFICE: 1997-557-052 Pdnted on recydable paper. KEY WORDS: Landtype association framework, landscape composi- tion, landscape pattern. 29 80 _lB_ 41 Pdnted on recydable paper. Sflbernagel, Janet; Chen, Jiquan; Gale, Margaret R.; Pregitzer, Kurt S.; Probst, John. 1997. An interpretation of landscape structure from historic and present land cover data in the Eastern Upper Peninsula of Michi- gan. Gen. Tech. Rep. NC-192. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 30 p. Compares historic and present landscape structure among four landtype association groups in Upper Michigan. Provides an example of a landtype association framework for assessing landscape composi- •tion and pattern. KEY WORDS: Landtype association framework, landscape composi- tion, landscape pattern. Sflbernagel, Janet; Chen, Jiquan; Gale, Margaret R.; Pregitzer, Kurt S.; Probst, John. Sflbernagel, Janet; Chen, Jiquan; Gale, Margaret R.; Pregitzer, Kurt S.; Probst, John. 1997. An interpretation of landscape structure from historic and present land cover data in the Eastern Upper Peninsula of Michi- gan. Gen. Tech. Rep. NC-192. St. Paul, MN: U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. 30 p. Compares historic and present landscape structure among four landtype association groups in Upper Michigan. Provides an example of a landtype association framework for assessing landscape composi- •tion and pattern. KEY WORDS: Landtype association framework, landscape composi- tion, landscape pattern. 30 p. Compares historic and present landscape structure among four landtype association groups in Upper Michigan. Provides an example of a landtype association framework for assessing landscape composi- •tion and pattern. Our job at the North Central Forest Experiment Station is discovering and creating new knowledge and technology in the field of natural resources and conveying this information to the people who can use it. As a new generation of forests emerges in our region, managers are confronted with two unique challenges: (1) Dealing with the great diversity in composition, quality, and ownership of the forests, and (2) Reconciling the conflicting demands of the people who use them. Helping the forest manager meet these challenges while protecting the environment is what research at North Central is all about. NORTHCENTRAL_ EHPERII11ENT STATION
https://openalex.org/W4285509676	https://ojs.lib.unideb.hu/competitio/article/download/8541/7775	Hungarian	null	digitalizáció módszertani kihívásai az üzleti képzésben	Competitio	2,021	cc-by	8,050	Competitio 2020. XIX. évfolyam doi: 10.21845/comp/2020/1–2/5 Competitio 2020. XIX. évfolyam doi: 10.21845/comp/2020/1–2/5 A digitalizáció módszertani kihívásai az üzleti képzésben Digitalizáció az üzleti képzésben KISPÁL-VITAI ZSUZSANNA egyetemi tanár Pécsi Tudományegyetem Közgazdaságtudományi Kar, Vezetés- és Szervezéstudományi Intézet vitai.zsuzsanna@ktk.pte.hu KISPÁL-VITAI ZSUZSANNA egyetemi tanár Pécsi Tudományegyetem Közgazdaságtudományi Kar, Vezetés- és Szervezéstudományi Intézet vitai.zsuzsanna@ktk.pte.hu Az egyetemi oktatás a 21. században alapvetően megváltozott környezetben kell, hogy helytálljon. A megszerzendő ismeretek az információtechnológia fejlődésével megszaporodni látszanak. A munkaerőpiac igényeinek megváltozása is megköve- teli az egyetemi tananyagok folyamatos fejlesztését és hatékonyabb átadását. A cikk ebben a környezetben a technológia hatását vizsgálja az oktatás hatékonyságára, bemutat és értékel néhány oktatásmódszertani megközelítést. A tanulmány követ- keztetései arra mutatnak rá, hogy az információ megszerzése sokkal könnyebbé vált, mint akár 10 évvel ezelőtt volt, az ismereteket azonban még mindig ugyanúgy kell rögzíteni, a tanulás neurobiológiai folyamata alapvetően ugyanaz, mint 100 évvel ezelőtt. JEL kódok: A20, A22, A23 JEL kódok: A20, A22, A23 Kulcsszavak: digitalizáció, oktatásmódszertan, a tanulás pszichológiája 1. Bevezetés „Nem vezet királyi út a geometriához” – Eukleidésznek tulajdonítják ezt az idéze- tet (van der Waerden–Taisbak 2019), de sokszor hallható a tudás megszerzésével kapcsolatban is. A tudás megszerzésének könnyebbé tétele évszázados törekvés, a tanulási folyamat segítése, azaz a tanítás pedig egyidős az emberiséggel. Minden korszaknak megvolt a saját oktatási módszertana, amely módszertan együtt fejlő- dött a tudás tartalmával. A 21. század sem más, a tanítás ebben a korban is azért létezik, hogy a tanuló el tudja sajátítani azokat az ismereteket, amelyekre szüksége van, amelyek segítik az előre haladását az életben. Az egyetemi oktatásnak is követnie kell a környezete által teremtett igényeket, és a 21. század ezt egyrészt megkönnyíti, másrészt igen nehézzé teszi. Jelen tanul- mány egyik célja, hogy bemutasson az üzleti oktatásban használt néhány oktatási 1 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN netika és a biológia szerepét a viselkedés kialakulásában. (Hucyznski–Buchanan 2007, McLeod 2017a) módszert, és mérlegelje az értéküket. Üzleti oktatáson értjük mindazon képzési formákat, ahol a hallgató a végzés után az üzleti élet valamely területén szakirá- nyú diplomával el tud helyezkedni. Itt a képzés szintje a felsőfokú szakképzéstől a mesterképzésig terjedhet. A tanulmány nem foglalkozik a doktori képzéssel, mivel az ott használatos módszertan elemzése külön értékelést igényel. Keresi viszont arra a kérdésre is a választ, hogy azok az oktatási módszerek, amelyeket az utóbbi húsz évben elkezdtek használni az üzleti oktatásban, mennyire viszik előre a tanu- lási folyamatot, mennyire segítik a hallgatót az ismeretek elsajátításában. Hatéko- nyabb lesz-e a tanulás az új módszerekkel, és ha igen, akkor milyen folyamatokon keresztül? A vizsgálandó hipotézis így szól: Az üzleti képzésben használt modern oktatási módszerek könnyebbé és hatékonyabbá teszik az ismeretek megszerzését, és megkönnyítik a tanulási folyamatot. A behaviorista elméletek szigorúan viselkedésközpontú megközelítése után az 1950-es években alakult ki a kognitív pszichológia nevű irányzat, mely az elmét, mint információ feldolgozó eszközt kutatja. Itt a hangsúly azokon a tényezőkön van, amelyek az előző megközelítésben nem lényegesek, (mert nem megfigyelhe- tők) azokon a mechanizmusokon, amelyek az inger/információ és az arra adott válasz/kimenet között vannak. Kutatják az észlelés, a memória, a figyelem és a problémamegoldás mechanizmusait (McLeod 2015). A harmadik tanuláselméleti megközelítés a konstruktivizmus. Itt az az alap- feltételezés, hogy az emberek aktívan hozzák létre a tudást, a tanuló nem passzív befogadó. A tudás a tanár és a diák együttműködése által jön létre, ahol a keletke- zett ismeretanyagot befolyásolják a létrehozó személy tapasztalatai és az őt ért új élmények. 1. Bevezetés A tanulás a tapasztalatok alapján felépített mentális modelleken keresz- tül történik (Berg–Simonson 2016, McLeod 2017b). A tanulmány felépítése a következő: rövid tanuláselméleti bevezető után meg- vizsgálunk az üzleti oktatásra ható néhány lényeges tényezőt a 21. században, majd áttekintjük, milyen módszertant használnak általában az oktatásra az üzleti képzésekben is. Az egyre inkább elfogadott digitális módszerek elemzése után kö- vetkeztetéseket vonunk le a kutatási kérdésre és a hipotézisre vonatkozóan. g Mindegyik felfogás alapvetően befolyásolja azt, hogy mit gondolunk a tanulá- si folyamatról, hogy hogyan kívánjuk elősegíteni, azaz milyen módszertant kívá- nunk használni. A behaviorizmusról gondolhatnánk, hogy elavult, mivel a 21. szá- zadban igen nagy hangsúlyt helyezünk a gondolkodásra, és a tudatos viselkedésre, de mégsem tekinthetjük elfeledettnek. Amikor a hallgatókat értékeljük, amikor jutalmat/büntetést adunk a munkáért, amikor rendszeres munkára kényszerítjük őket a rendszeres feladatadással és értékeléssel, akkor egyértelműen a viselkedés következményeit használjuk a hallgató irányítására. A kognitív pszichológia ered- ményeit használjuk, amikor önreflexiós naplót iratunk egy projekt végén, vagy amikor nyílt forráshasználatot engedélyezünk egy vizsgán. Itt arra ösztönözzük a diákot, hogy ne csak használja a memóriáját, és mechanikusan számoljon be a tanultakról, hanem kritikus gondolkodásra, értékelésre, esetleg új következtetés levonására késztetjük. Az ilyen feladatok átvezetnek a konstruktivista oktatáshoz, ahol a tudás a tanár és diák együttműködése következtében jön létre, ahol a diák felfedez, létrehoz, kérdéseket tesz fel, és ezáltal tanul. A tanulásnak nincsen vége soha, tartalmazza a továbbfejlődés és az állandó ismeretszerzés igényét is. 2. Az oktatás alapja – a tanulás pszichológiája Mielőtt az oktatás módszertanáról beszélnénk, feltétlenül meg kell határoznunk azt a folyamatot, amelynek érdekében használjuk az oktatást: a tanulást. Lényeges a tanár számára, hogy értse, milyen pszichológiai folyamatokon keresztül történik a tudás megszerzése, mert ha ezzel nincsen tisztában, akkor nem képes megfele- lő, a hallgató számára alkalmas módszert választani. Pszichológiai meghatározása szerint a tanulás az a folyamat, amikor a tapasztalataink alapján tudást szerzünk, és ez állandó változást okoz a viselkedésünkben (Huczynsky–Buchanan 2007). Arról, hogy hogyan történik a folyamat, számos elmélet született, a módszertan szempontjából itt csak három csoportot említünk. Ez a három elmélettípus szoros kapcsolatban áll a pszichológia egyes iskoláival és a filozófia egyik ágával. ii A tanuláselméletek egyik igen nagy hatású csoportját képzik a behaviorista elméletek, ahol a viselkedést kizárólag az inger és az arra adott válasz határozza meg. Ezek az elméletek nem foglalkoznak a gondolati folyamatokkal, kizárólag a megfigyelhető és mérhető ingereket és válaszokat vizsgálják. Feltételezéseik közé tartozik, hogy minden viselkedési formát a környezetből tanulunk meg. Két alap- típusuk van: a metodológiai és a radikális behaviorizmus. A metodológiai beha- viorizmus azt feltételezi, hogy az ember agya születéskor „üres”, és mindent az emberi fejlődés során tanulunk meg. A radikális behaviorizmus elfogadja, hogy az élőlények veleszületett viselkedési formákkal rendelkeznek, tehát belátja a ge- 3. A 21. század összetett tanulási környezete Ahhoz, hogy a megfelelő módszertant ki tudjuk választani, először meg kell ha- tároznunk az oktatás célját, környezetét és azt a tanuláselméleti megközelítést, amelynek keretében dolgozni fogunk. Jelen tanulmány az üzleti felsőoktatást vizs- gálja, és néhány itt használatos módszertani alkalmazást értékel. A lista nem kíván teljes lenni, a hangsúly a számítógépen keresztüli oktatáson van, amelynek jelen- tősége az elmúlt években nagymértékben felértékelődött. A 2020 márciusában a 2 3 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN Érdekes, hogy a digitális képességek megszerzésének hangsúlyozása mellett egyéb képességek meglétét is elengedhetetlennek tekinti, például a magas szintű kogni- tív és társadalmi-érzelmi képességekét (OECD 2019).i világban elterjed koronavírus-járvány pedig megkerülhetetlenné tette az online oktatást. Először az oktatásra általánosan jellemző tulajdonságokat elemezzük, ezeken belül vizsgáljuk az üzleti felsőoktatást. A „skill”-ek elég környezetspecifikusak, és noha önállóan is elsajátíthatók, erre kidolgozott oktatásmódszertannal gyorsabban fejleszthetők (Bates 2019). Gyak- ran felmerül a kérdés, hogy a felsőoktatás mennyire szolgálja a piac igényeit, és sokan képviselik azt az álláspontot, hogy jobban ki kellene szolgálni a vállalati szükségleteket (például Chamorro–Premuzic–Frankievicz 2019). Ha azt tekintjük, hogy az oktatás feladata a kiművelt emberfők képzése és olyan munkavállalók ne- velése, akik képesek azokat a problémákat is megoldani, amelyek még nem is lé- teznek, akkor nem feltétlenül kellene kizárólag a piac igényeit figyelembe venni az oktatásnál. Az viszont elfogadott tény, hogy csakis teoretikus tudással, a fent em- lített képességek nélkül igen nehéz vagy lehetetlen elhelyezkedni (Fallow–Steven 2000). Annak eldöntése, hogy mire és hogyan képezzük a hallgatót, befolyásolni fogja a használt oktatásmódszertant is. 3.2. Az információtechnológia fejlődése A 20. század vége és a 21. század eleje az információtechnológia olyan fejlődését hozta el, amely az élet minden területét befolyásolja. A Gartner Inc. globális ku- tató és tanácsadó cég szerint az információtechnológiára való költés 2020-ban, globálisan várhatóan el fogja érni a 3,9 trillió dollárt (Gartner 2020). A korona- vírus-járvány előtt is elfogadott tény és jó üzlet volt az oktatástechnológiába való befektetés. Bainbridge 2016-ban írta a TechCrunch online folyóiratban, hogy az „edtech az új fintech”, vagyis jó pénzügyi műveletnek ítélte meg az oktatástechno- lógiába való befektetést. 2011-től 2018-ig az ilyen befektetések száma az USA-ban folyamatosan emelkedett, 2019-ban 1,45 milliárd dollár volt (Wan 2019). A munkáltató igényei is változnak, egyre inkább követelnek a tényszerű tu- dáson kívül „skill”-eket, azaz „képességeket, amelyek lehetővé teszik, hogy bizo- nyos tevékenységeket jól végezzen” a dolgozó (Mathis et al. 2017). Ilyen képes- ségek például: a csoportban való munka képessége, időmenedzsment, technikai képességek, kreativitás, a hiteles kommunikáció, gyors tanulás és alkalmazkodás a nehéz helyzetekhez (What are… 2019). Wagner és a Change Leadership Group a Harvard University-n 2010-ben 7 „skill”-csoportot azonosított, melyekre a hall- gatóknak szükségük van azért, hogy a 21. században sikeresen élhessenek és dol- gozhassanak. Ezek: 3.1. A munkaerőpiac igényeinek változása Általában a tanulás egyik fontos célja olyan hasznos ismeretek megszerzése, me- lyek képessé teszik a tanulót/hallgatót arra, hogy munkát végezzen. Az oktatási intézmények a munkaerőpiac igényeit is ki kell szolgálják, olyan képzett végzősö- ket kell kibocsátaniuk, akik azonnal hasznos munkát tudnak végezni, és értéket tudnak teremteni a vállalatnak. Már a 2000-es években is világossá vált, hogy a munkaerőpiac alapjaiban változik meg. Az élet minden területén megnöveke- dett az olyan munkaerő iránti igény, amelyet „tudásmunkásnak” neveznek. Az ilyen emberek elsősorban az iskolarendszerű oktatásban megszerzett tudásukat használják a munkahelyeken, hogy termékeket és szolgáltatásokat hozzanak létre, alkalmazzanak és fejlesszenek. Az ő munkájuk elsősorban intellektuális termé- szetű, a foglalkoztatási elithez tartoznak, a munkáltatójuk számára jelentős érték létrehozására képesek (Wilkinson et al. 2017). Még olyan területeken is megjelent a tudásintenzív oktatás igénye, amelyek eddig inkább csak manuális készségeket igényeltek (például az autószerelés és a mezőgazdasági gépszerelés, ahol informa- tikai ismeretekre is szükség van) (Bates 2019). 3.3. A felsőoktatási piac változásai – 110 millió hallgatójuk van – ebben Kína nincs benne – 2500 kurzust ajánlottak a szolgáltatók – 11 online diplomát adó degree programot – 170 microcredential-t (egy bizonyos „skill”-t fejlesztő rövid online kurzus [Blazevic 2018]) – 170 microcredential-t (egy bizonyos „skill”-t fejlesztő rövid online kurzus A milleniumi generációt a mítoszok alapján az különbözteti meg a többitől, hogy készség szinten használja a digitális eszközöket. Azt is mondják róluk, hogy képesek a „multitasking”-ra, egyszerre több dolgot is megfelelően elvégezni. Ezek az állítások azonban tudományos kutatások alapján, szisztematikusan megvizs- gálva szintén csak mítoszok. 3.4. A hallgatókra jellemző tulajdonságok a 21. században, „városi legendák” és a valóság Tanulmányok sokasága áll rendelkezésre, melyek igen részletesen tárgyalják a 21. század fiataljainak értékrendjét és jellemző tulajdonságaikat. A kutatások eredmé- nye bekerült a tankönyvekbe is (például Robbins–Judge 2017 190. oldal). Magyar- országon is többen foglalkoztak ezzel a témával, részletesen például Tari Annamá- ria több könyvben is. Itt kizárólag az oktatás szempontjából vizsgáljuk ezt a témát. g Kirschner (2017) több cikkben is kifejtette, hogy a digitális bennszülött nem létező fogalom, Prensky nem tudományos kutatás alapján teremtette ezt a fogal- mat, hanem az alapján, amit maga körül látott. A kutatások azonban nem iga- zolták Prensky feltételezéseit, hogy a milleniumi generáció készség szinten képes használni a számítógép adta lehetőségeket. Az 1984 után született fiatalok egy limitált és igen egyszerű hányadát tudják használni a digitális lehetőségeknek. Ezek a kész tartalmak és a szórakozás digitális formáinak használata, valamint a kapcsolattartás az interneten keresztül. Ezek nem kreatív tevékenységek, és az, hogy a fiatalok el tudták sajátítani őket, nem a veleszületett ügyességüket illuszt- rálja, hanem a programozás és a digitális technológiák felhasználóbarát mivoltát. Oktatás szempontjából ez annyit is jelent, hogy egyrészt a tanárok nem „digitális bevándorlók”, akik a diákjaik fejlettebb tudását nem tudják sem kihasználni, sem bővíteni, másrészt hiba volna készpénznek venni a diákok digitális képességeit (Kirschner 2017). Az úgynevezett milleniumi generációra jellemző értékek a munkára vonat- kozóan a Robbins & Judge Organizational Behavior könyvben az önbizalom, a pénzügyi sikerek értékelése, a saját magukhoz való lojalitás és a barátság fontos- sága. Tari Annamária az azonnali kívánság kielégítésének vágyát, a „most rög- tön” attitűdöt, az állandó online jelenlétet, a türelmetlenséget és a kapkodást emeli ki. (Széles–Horváth 2016) Jellemzően Marc Prensky 2001-ben az On the Horizon című oktatási folyóiratban megjelent cikke után terjedt el, hogy a mai hallgatók megváltoztak, és az oktatási rendszer már nem képes őket a régi módszerekkel tanítani. i Prensky híres cikkében azt állítja, hogy a 21. század fiataljai rendelkeznek olyan képességekkel, amelyek lehetővé teszik számukra a digitális technológiák használatát, azért, mert ilyen környezetben nőttek fel. Az ilyen fiatalok a digitális bennszülöttek, és akik 1984 előtt születtek, azok még nem képesek erre, ők a digi- tális bevándorlók. Prensky nem kutatások alapján állította ezt, hanem megfigyelte a környezetében élő fiatalokat, akik több digitális eszközt használtak, mint előde- ik. Ő azt feltételezte, hogy a használat mögött megértés és tudás is van. A többféle tevékenység egyidejű végzésének képessége, azaz a „multitasking” a modern agyi vizsgálati technikák használatával (funkcionális mágneses rezonan- cia) kiderült, hogy szintén nem létező fogalom. 3.3. A felsőoktatási piac változásai A 20. század végének politikai és gazdasági változásai együtt az információtechno- lógia fejlődésével nemcsak az információ, hanem az emberek szabad áramlását is lehetővé tették a 2020-as világjárványig. Ezzel nemcsak információcsere valósult meg, hanem lehetővé vált a külföldön való tanulás és tanítás, a tapasztalatszerzés és a megszerzett ismeretek beépítése az oktatásba. A szabad mozgás következtében a piac kitágult, és az egyes országok egyetemeinek nemcsak hazai, hanem külföldi versenytársai is keletkeztek. A megfelelő erőforrásokkal, pénzzel, nyelvtudással és rugalmassággal rendelkező hallgatók és oktatók elvileg bárhol, de gyakorlatilag is igen sok helyen tudtak rövidebb-hosszabb ideig tanulni és tanítani. Ezt a vál- tozást más szintre helyezte a Massive Open Online kurzusok megjelenése 2008- ban, amikor Stephen Downes és George Siemens elindították a Connectivism and – Kritikai gondolkodás és problémamegoldó készség – Együttműködési képességek és vezetési képessége – Tettrekészség és alkalmazkodó készség – Kezdeményezés és vállalkozó szellem – Hatékony kommunikáció a beszélt és az írott kommunikáció terén – Az információ megszerzésének és elemzésének a képessége – Kíváncsiság és képzelőerő Wagner 2010, idézi Scott 2015. Az OECD a 2019-es Skills Outlook 2019: Thriving in a Digital World című kiad- ványában a technológiai haladás által létrehozott új „skill”-igényeket részletezi. 5 4 A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN KISPÁL-VITAI ZSUZSANNA Connectivity Knowledge című kurzusukat. 25 hallgató járt erre a University of Manitoba-n, további 2300 vett részt rajta online. Innentől kezdve az ilyen típusú kurzusok csak szaporodtak (A Brief… 2020). Shah (2019) a Class Central webol- dalon a következő adatokat közölte a MOOC-okról 2019-ben: Connectivity Knowledge című kurzusukat. 25 hallgató járt erre a University of Manitoba-n, további 2300 vett részt rajta online. Innentől kezdve az ilyen típusú kurzusok csak szaporodtak (A Brief… 2020). Shah (2019) a Class Central webol- dalon a következő adatokat közölte a MOOC-okról 2019-ben: hely iránt és a munkahelyváltási szándék voltak. Ezekben a változókban az egyes generációknak eltérő attitűdöket tulajdonítanak. A kutatásnak az volt az eredmé- nye, hogy nincsen érdemi különbség az idősebb (tradicionális és baby-boomer) és a fiatalabb (X és milleniumi) generációk munkához kapcsolódó értékrendjében, ami pedig van, az valószínűleg nem a generációhoz való tartozással magyarázha- tó. Az IBM 2014 nyarán több szervezet 1784 munkatársának értékrendjét mérte fel 12 országban és 6 iparági területen. A kapott eredmények megmutatták, hogy a milleniumi generáció ugyanolyan vagy nagyon hasonló prioritásokat határozott meg a munkahelyén, mint a baby boomer vagy az X (Baird 2015). 3.4. A hallgatókra jellemző tulajdonságok a 21. században, „városi legendák” és a valóság Aki egyszerre több dolgot csinál, az tulajdonképpen igen gyorsan képes váltani a feladatok között. Ha ezek mély kognitív funkciókat érintenek, akkor azok a funkciók sérülni fognak, azaz a haté- konyságuk sokkal gyengébb lesz (American Psychological Association 2006). Ezek a kiadványok azonban csak egy véleményt képviselnek, van egy kevéssé meghallgatott másik oldal is, ahol egészen más módszerekkel vizsgálják a fiata- lok generációit. Constanza és munkatársai 2012-ben publikálták egy metaanalí- zis eredményeit, amelyben 20 kutatást kvantitatív módszerekkel vizsgáltak meg (Constanza et al. 2012). A kutatás célja az volt, hogy kvantitatív módszerekkel elemezzék a munkahelyi attitűdökben előforduló generációs különbségeket. A vizsgált változók a munkával való megelégedettség, az elkötelezettség a munka- Igen gyakori vélemény, gyakran hangoztatott tényként kezelt nézet az is, hogy a mai hallgató képes önállóan kutatni, hogy a projekt alapú tanulás hatékonyabb, mint a hagyományos, tanár által vezetett tanulás (Bell 2010). A minimális tanári irányítással végzett tanulásnak igen sok támogatója van, és nem is új módszerként áll rendelkezésre. A problémaalapú tanulást először a ’60-as években, Kanadá- ban vezették be egy orvosi egyetemen (Barrows 1996). Igen hasonló a felhasznált módszerek sora, például a kutatás alapú (inquiry learning), a felfedezésen alapuló 7 6 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN (discovery learning), a kísérletes tanulás (experiential learning). Mindegyikben azok a feltételezések működnek, hogy a hallgatók valódi problémák megoldása által jobban tanulnak, komplex tudást képesek elsajátítani, mivel az információ- ban gazdag környezetben a saját megoldásaikat kell kifejleszteniük. A tudást ak- kor lehet jól elsajátítani, ha a tanult tudományág folyamatait tanulmányozzák, és gyakorlati tapasztalatot szereznek. A tanár minimális segítséget nyújt a diáknak ebben a munkában, hogy ne zavarja meg a tanulás önálló, természetes folyamatát. Ez a felfogás teljesen figyelmen kívül hagyja az emberi gondolkodás és a memória működését. A tanulás célja, hogy az ismeretek a rövid távú munkamemóriából a hosszú távú memóriába kerüljenek. A minimális irányítással történő kutatás a rö- vid távú memória használatával történik, és az információ nem kerül át a hosszú távú memóriába – tehát igazi tanulás nem történik. A munkamemória kapacitása igen korlátozott, a hosszú távú memória kapacitását jelen pillanatban korlátlan- nak tekintik (Szabó 2000, Kirschner 2006).i a földet összezsugorító utazási szokásai egyszerre megszűntek, az utazás, mint te- vékenység ellehetetlenült gyakorlatilag egyik napról a másikra. Ezzel megszűnt az oktatói és hallgatói mobilitás is. 4. Oktatásmódszertan, rövid áttekintés Az oktatás módszertana igen sokrétű lehet, sokféle megoldást kínál a tanároknak a diákok képzésére. A következőkben Lada (2009) felsorolása alapján két szem- pontból vizsgáljuk a módszereket: egyrészt hogy mennyire oktató vagy hallgató- és diákközpontú. Az oktatásmódszertani filozófiákban igen gyakran tárgyalt kér- dés, hogy mennyire igényli a használt módszer a hallgató aktív közreműködését, vagy mennyire tekinti a diákot passzív befogadónak (Freeman 2014, Lada 2009). A másik kérdés, hogy a megváltozott oktatási/tanulási környezetben mennyire használhatók, mennyire tekinthetjük őket elavultnak. Az a feltételezés, hogy a 21. század fiataljainak nincs szüksége irányításra a tanulásban, hogy nincs szükség az ismeretek (gyakran igen hosszadalmas és fá- rasztó) rögzítésére, nem veszi figyelembe a kognitív folyamatok felépítését, vala- mint a tanulás pszichológiáját. Lehet, hogy a projekt nagyon jól fog sikerülni, és a diáknak örömet fog okozni a részvétel, de a tovább vitt tudásanyag nem biztos, hogy versenyképes lesz az irányított oktatás, problémamegoldás, tesztelés, ismét- lés útján szerzett ismeretekkel szemben. A módszerek listája: – előadás – magyarázat gy – elbeszélés – elbeszélés 3.4. A hallgatókra jellemző tulajdonságok a 21. században, „városi legendák” és a valóság Igaz, hogy az ismeretanyag tartalma átmegy a szá- mítógépen keresztül is, de a más kultúrában való túlélés, az emberek megismerése olyan környezetben, ahol tétje van a közös munkának, a mély megértése egy má- sik életmódnak, gondolkodásmódnak egyik pillanatról a másikra semmivé lett. A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN KISPÁL-VITAI ZSUZSANNA Freeman és munkatársai egy 2013-as kutatásban megvizsgálták az előadás haté- konyságát a természettudományok, a mérnöki tudományok és a matematika ok- tatásában az aktív tanítási módszerekkel összehasonlítva. Itt egyértelmű volt, hogy a hallgatók bukási aránya az aktív tanulást használva lényegesen alacsonyabb volt. Aktív oktatási módszer például a problémamegoldás, esettanulmány-használat, csoportban való tanulás (Active learning… 2020). Freeman és munkatársai egy 2013-as kutatásban megvizsgálták az előadás haté- konyságát a természettudományok, a mérnöki tudományok és a matematika ok- tatásában az aktív tanítási módszerekkel összehasonlítva. Itt egyértelmű volt, hogy a hallgatók bukási aránya az aktív tanulást használva lényegesen alacsonyabb volt. Aktív oktatási módszer például a problémamegoldás, esettanulmány-használat, csoportban való tanulás (Active learning… 2020). – a tudás átadása – tények megértése – a tudás alkalmazása – attitűd létrejötte – a tudás átadása – tények megértése – a tudás alkalmazása – attitűd létrejötte – a tudás átadása – tények megértése – attitűd létrejötte Ha Bloom taxonómiájával hasonlítjuk össze ezeket a szempontokat, akkor belát- ható, hogy szoros kapcsolatban állnak a Bloom-féle oktatási célokkal. Az első há- rom a kognitív célokkal, az attitűd pedig az affektív területekkel (Armstrong 2017). p ( g ) Nem vitatva a kutatások eredményeit, az előadást mégsem kellene teljesen kihagyni a modern oktatásmódszertani eszközök közül. Igen hatékony lehet, ha nagyobb közönségnek kell gyorsan ismereteket átadni, különösen ha ezek az is- meretek újak, és nincsen belőlük tankönyv sem. Az előadást ki lehet egészíteni 21. századi módszerekkel, videók vetítésével, kérdések feltételével a hallgatók szá- mára, ami már azonnal aktív módszertan használatát jelenti. Ez a módszer nem feltétlenül passzív, ha a hallgató jegyzetel, mivel a jegyzetelés nem az összes el- hangzott mondat leírása, hanem a lényeges ismeretek rögzítése kell, hogy legyen, tehát gondolkodást és elemzést igényel. Alapképzésben, amikor teljesen kezdő- ket tanít az oktató, valamilyen módon át kell adni a kezdeti ismereteket, erre az előadás alkalmasabb, mint ha kiadjuk a tankönyv megfelelő fejezetét elolvasásra, és utána kérdezünk belőle. Egy előadás lehet rendkívül érdekes, figyelemfelkeltő, ha a tanár felkészül és hatékonyan képes nyilvános kommunikációra. Kizárólagos használatra valóban nem alkalmas, hiszen a megszerzett ismeretek gyakorlása el- ismerten növeli és mélyíti a hallgató tudását, de teljesen eltávolítani a módszerek közül nem lehet. f A különböző kritériumok alapján az esettanulmányok nagyon jól szerepeltek minden téren, kivéve a tudás átadása kritériumon. A módszer feltételezi, hogy a hallgató képes önállóan elsajátítani az eset megoldásához szükséges alapvető el- méleti ismereteket. A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN Azt, hogy mennyire könnyű vagy nehéz a felsorolt módszerek- kel való munka mind a tanár, mind a diák számára, nem értékelték. A felsorolt módszerek már átvezetnek az aktív, hallgatóközpontú oktatási módszerek területére. Aktív és hallgatóközpontú módszereknek tekinthetők: Aktív és hallgatóközpontú módszereknek tekinthetők: – tanulói kiselőadás – kérdezés ita – kérdezés – vita – vita – projekt – kooperatív oktatási módszerek – szimuláció, szerepjáték, játék Az összes többi módszer valamilyen formában aktívnak ítélhető, de ezek is csak akkor hatásosak, ha megfelelő feltételek között használjuk őket. Például a tanulói kiselőadást az üzleti képzésben rendszeresen használjuk a kommuniká- ciós készségek fejlesztésére. Nem szabad elfelejteni, hogy a tanulói kiselőadás is előadás. Igaz, hogy aki előad, elvben sokkal mélyebben készült fel az órára, de aki hallgatja, az ugyanazt kapja, mintha a tanártól hallgatna előadást. – házi feladat A módszerek listájához az üzleti képzésben még hozzáadható az esettanulmányo- kon keresztüli oktatás, a „blended learning”, ahol offline és online módszerek ke- verékét használják. Ezek szintén a hallgató tevékeny közreműködését feltételezik, valamint a „flipped classroom”, a tükrözött osztályterem. Ezt a technikát először 1998-ban javasolták, majd 2000-ben egy bevezető közgazdaságtan órán használ- ták, hogy segítsék a hallgatókat, mert a különböző tanulási stílusok elmélete értel- mében a diákok más és más módszert használva tanulnak eredményesen (Brame 2013). Az elmélet szerint vannak hallás után tanuló, olvasással tanuló vagy éppen gyakorlás útján hatékonyabban tanuló diákok. (A legújabb neuropszichológiai kutatások alapján azonban kiderült, hogy a tanulási stílus nem létező konstruk- ció [Kirschner 2017].) Magyarországon is használták a tükrözött osztálytermet, a Budapesti Corvinus Egyetemen 2015-ben és az ELTE-n is. (Tóth 2015, A felsőok- tatási… 2017) A módszertani felsorolásban a kérdezés, vita és szemléltetés mind használható egy olyan előadás keretében, ahol az oktató képes munkát befektetni a technikák elsajátításába, és készség szinten hajlandó használni azokat. A módszerek ered- ményessége függ a tantárgy jellegétől, a hallgatók aktivitásától és a légkörtől is. A tantermi légkör meghatározó ilyen esetben, ha akár a hallgatók, akár a tanár nem érzi magát biztonságban, ha nem a kölcsönös ismeretszerzésen van a hangsúly, akkor ezek a módszerek sikertelenek lesznek. Az üzleti képzésben gyakran használunk projekteket, szimulációkat, szerepjá- tékokat és esettanulmányokat. De Juan Vigaray és munkatársai (2010) hatékony- ság szempontjából értékeltek oktatási módszereket az üzleti képzésben, köztük ezeket is. A hatékonysági szempontok a következők voltak: A tükrözött osztályterem a „feje tetejére” állítja a Bloom-féle oktatási taxonó- miát. 3.5. A globális egészségügyi helyzet váratlan és hirtelen megváltozása – tanulói kiselőadás – megbeszélés A 2020 tavaszán elkezdődő koronavírus-világjárvány földrengésszerűen rázta meg a felsőoktatást az egész világon (Govindarajan–Srivastava 2020). Olyan eddig ter- mészetesnek és megdönthetetlennek tartott szokásokat és oktatási gyakorlatokat kérdőjelezett meg, amire soha senki nem mert gondolni. Az oktatás szerkezetét, a tudás átadásának módját, a technológia szerepének hirtelen legfontosabbá válá- sát minden felsőoktatási szereplőnek újra kellett gondolnia. Itt nemcsak a nagyon gyors digitális átállás okozott nehezen megoldható gondokat, hanem az oktatás hagyományainak kényszerű felrúgása is, hiszen sokszor a személyes megjelenés helyett egy képernyőnek kellett órát tartani azzal a tudattal, hogy lehet, hogy mö- götte a hallgató csak bejelentkezett az órára, de nincs is ott. A vizsgák megszer- vezése más egy olyan környezetben, ahol nincs mód alapos és megbízható elle- nőrzésre. Együtt kellett élni mindenütt a világon a kétséggel, hogy mennyit ér az oktató (és a hallgató) munkája, volt-e értelme azoknak a hősies erőfeszítéseknek, amelyeket a felsőoktatási intézmények oktatói és dolgozói megtettek a hallgatóik érdekében. A felmerült kétségek mellett a kényszerű alkalmazkodás tapasztalatok halmazát hozta létre, amelyeknek elemzése, értékelése és hasznosítása még sok időbe fog telni. A teljességre való törekvés nélkül csak egy példa: a globalizációnak – strukturálás – kérdezés – kérdezés – a tanulói munka értékelése – vita – vita – vita – szemléltetés – projekt – projekt – kooperatív oktatási módszerek – szimuláció, szerepjáték, játék – házi feladat A lista élén álló módszerek közül az első három igen sok támadásnak volt kitéve, különösen a 21. század elején. Az előadás, magyarázat és elbeszélés mind mono- logikus módszerek, ahol a tanár beszél, a hallgatók pedig – elvben – figyelemmel kísérik, jegyzetelnek, és ha nem értenek valamit, kérdeznek. Ezeket a módszereket passzívnak ítélik meg, például Lada (2009) is, mivel a hallgatónak követnie kell a tanár monológját, a tanár tempójában, és „nincs módjában visszalapozni” (Lada 2009), mint egy könyvben, esetleg újra meghallgatni az nem értett ismereteket. 9 8 A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN Itt a hagyományos módon az osztályteremben történő olvasás és jegyzetelés, az „emlékezés” és a „megértés” otthon történik, az alkalmazás és a kreatív tevé- kenységek, amelyek a piramis tetején találhatók, az iskolában (Armstrong 2017). 10 11 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN a programok képesek érzékelni a környezetüket, képesek tanulni és cselekedni ugyanúgy, mint az ember, annak függvényében, hogy mit érzékelnek maguk kö- rül, és milyen céllal programozták őket (AACSB 2020). Érdemi tanulás akkor történik, ha a hallgató elfogadja a helyzetet, és hajlandó passzív olvasáson kívül gondolkodni, azaz alkotó módon részt venni a folyamat- ban. Az alapelv mindegyik felsorolt módszernél (nemcsak a tükrözött osztályte- remnél) a hallgató aktív kutató munkája, a gondolkodási képesség és a kíváncsi- ság feltételezése, a törekvés a tudás minél magasabb színvonalú megszerzésére. A módszerek feltételezik azt is, hogy a gyakorlati feladatok arra késztetik a diákot, hogy utánanézzen az elméleti háttérnek, és a sok feladat megoldásán keresztül egyszerre szerez magasabb szintű elméleti és gyakorlati tudást. Az oktatásban a mesterséges intelligencia az alkalmazkodó tanulásban játszhat nagy szerepet, a tanulni képes gépek személyre tudják szabni az oktatást, jobban ki tudják szolgálni a hallgató speciális igényeit. A gépek által gyűjtött adatokat fel lehet használni arra, hogy jobban megismerjék az egyes hallgatók és évfolyamok igényeit. Már létezik több ilyen alkalmazás, amely segíti a tanulókat, például a Duolingo nevű applikáció, amelyen keresztül 35 nyelvet lehet tanulni. Az appliká- ció játékos formában személyre szabja a tanulást a használó igényeinek megfele- lően (Preanandam 2018). Noha nagyon csábítónak tűnik a klasszikus előadásalapú oktatás felfrissítése ezekkel a technikákkal, igen sok feltételnek kell teljesülnie a hatékony megvaló- sításhoz. A hallgatói motivációt nemcsak felébreszteni kell, hanem a nehézségek esetén fenntartani is. A csoportmunka nehézségeinek kezelése állandó oktatói probléma, értékelése pedig még nehezebb. Nehéz azt kezelni, hogy mély és átfogó kutatás helyett a még éppen kielégítő megoldásokat keresik a diákok, hogy a jó kommunikáció eltakarja az esetleges tényhiányokat, hogy nem gondolják, hogy az elméleti tudás szükséges a feladatok megoldásához, mert van elég kész megol- dás, amit csak le kell tölteni az Internetről. Itt áttérünk az utóbbi években egyre elterjedtebb módszerek használatára és a digitalizáció hatására az üzleti oktatásra. Ide tartozik a „push research” fogalom is, amikor a számítógép megérti azt a kontextust, amiben a kutató dolgozik, és proaktívan küld olyan információt, ami- re szükség van, sokszor még az előtt, hogy azt a kutató keresné. 5. Oktatási módszerek a digitális környezetben Az információtechnológia, a számítógépek az élet mindent területén való hasz- nálata megváltoztatta, a jövőben pedig még jobban meg fogja változtatni az okta- tást. Az üzleti oktatásban az AACSB (Association to Advance Collegiate Schools of Business) globális non-profit intézmény innovációs bizottsága megvizsgálta a kreatívan romboló (disruptive technology) technológia lehetséges hatásait, az ál- tala biztosított lehetőségeket és a vele kapcsolatos problémákat (AACSB Research Reports 2020). A Georgia Institute of Technology sikeresen használt mesterséges intelligen- ciát hallgatói segítségként, Jill Watsonnak nevezték el a tanársegédet. Érdekesség, hogy a hallgatók nagy része észre sem vette, hogy nem emberi lénnyel kommu- nikál, pedig a Watson név árulkodó lehetett volna (az IBM Watsonjára utalva). A 300 hallgató a mester szintű AI-ről szemeszterenként 10.000 üzenetet posztolt az üzenőfalra, ennyit emberi tanársegéd képtelen lett volna megválaszolni. Az AI a gyakran ismétlődő kérdésekre az esetek 97%-ában helyesen válaszolt, így felsza- badította az emberek számára az időt, hogy az ismétlődő unalmas munka helyett értelmes munkát végezzenek (Neelakantan 2020). p Kutatást végeztek a mesterséges intelligencia, a blockchain, a virtuális és kiter- jesztett valóság (augmented reality), valamint a mobil és mikrotanulás területén. Nemcsak a számítógépek egyszerű használata jelenti a digitális oktatás techno- lógiáját, hanem az olyan technológiák is, mint a mesterséges intelligencia vagy a virtuális valóság használata. A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN A gyakorló jogá- szok számára már elérhető olyan szoftver, amely figyeli a vállalati kommunikációt a vállalati partnerekkel, és a szövegek elemzése alapján előre képes figyelmeztetni a céget a valószínűsíthető perek lehetőségére. A Google már beépítette a Google asszisztensbe ezt a jellegzetességet (Heller 2017, Hello HR… 2020). Az oktatásban az ilyen technológiák használata rendkívül könnyűvé teszi az információ megta- lálását, olyan információkét is egyben, amelyek egészen biztosan kapcsolódnak a hallgató/oktató igényeihez. Meggyorsítja és leegyszerűsíti az egyébként igen idő- igényes információkeresést, főleg az értékelését, hogy mennyire illeszkedik ahhoz, amit tanulni vagy tanítani akarnak. A projektek, szakdolgozatok, diplomamunkák megírásához szükséges kutatás lerövidíthető, a hallgatói motiváció könnyebben fenntartható. Az információkereső algoritmus alkalmazkodik az ember igényei- hez. 5.3. Virtuális és kiterjesztett valóság A virtuális valóságot az etikus döntések oktatására is felhasználják. A Univers- ity of St. Thomas – Opus College of Business vezetői döntési dilemmáknak teszik ki a hallgatóikat virtuális valóság használatával. Okostelefonra letölthető appli- káció, valamint egy viselhető szemüveg segítségével egy bizonyos problémában négyféle döntési kimenet közül választhat a hallgató, aki teljesen átéli a helyzet komolyságát és a döntése következményeit is (Innovations that… 2020). A virtuális valóság egyik definíciója: „Olyan három dimenziós, számítógép ál- tal létrehozott környezet, amelyet egy (élő) személy felfedezhet, és amellyel ez a személy interakcióba léphet. Az (élő) személy vagy részévé válik a virtuális kör- nyezetnek, vagy elmerülhet ebben a világban, és amíg ott van, képes lesz tárgyak mozgatására vagy cselekedetek véghezvitelére.” (What is… 2017) A felhasználó egy fejre illeszthető készüléken keresztül lép be a virtuális világba. A technológia változatai a kiterjesztett valóság, ahol a valóság bizonyos részeire helyezik rá a vir- tuális tartalmat, tehát nem teljesen „művi módon megteremtett” a világ, hanem a valóság egyes részeit kiegészíti digitális adatokkal, mozgó elemekkel, hogy fejles�- sze a felhasználói élményt. A „kevert valóság” (mixed reality) esetén a valós, létező elemek együtt léteznek a virtuálisan megteremtettekkel. A kevert valóságban a felhasználó látja a „valóságot”, ezzel együtt a virtuális elemeket is, és egyszerre le- het őket kezelni (AACSB 2018). Ennek a technológiának nem is olyan új a képzés- ben való felhasználása, hiszen a pilótákat már régen képzik szimulátorokban, de használják a katonaságnál, a tűzoltóságnál is. Az üzleti életben a Walmart eladói tréningen alkalmazzák, de a General Motors is használ Google Glass-t azért, hogy a gyárakban a dolgozók azonnal visszajelzést kapjanak a munkájukról. 5.1. Artificial Intelligence – mesterséges intelligencia g A mesterséges intelligencia használatának hátránya lehet, ha az algoritmus hi- bás, és nem azt adja, amit kérünk tőle. Az adaptív tanuláshoz szükséges eszközök és szoftverek igen költséges beruházást igényelnek, és a használatukra is meg kell tanítani a velük dolgozókat (AACSBa 2020). A mesterséges intelligencia olyan komputer program, amely az emberi gondolko- dási folyamatokat próbálja lemásolni (Rouse 2020). Olyan „gép”, amely tanulásra, fejlődésre, önálló problémamegoldásra képes. A programozása három képességre koncentrál: a tanulásra, érvelésre és a saját maga okozta hibák kijavítására. Ezek 13 12 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN 5.2. Blockchain – Megnövelheti a hallgatói motivációt a játékossága által, és kifejezetten alkalmas a mai világban az olyan hallgatók oktatására, akik képtelenek hosszabb ideig figyelni. – Megnövelheti a hallgatói motivációt a játékossága által, és kifejezetten alkalmas a mai világban az olyan hallgatók oktatására, akik képtelenek hosszabb ideig figyelni. A blockchain olyan technológia, amely egy nyitott főkönyvnek felel meg, és képes két személy közötti tranzakciókat hitelt érdemlően és megváltoztathatatlanul rög- zíteni. Ez a technológia van a bitcoin mögött, de a felsőoktatásban is használható például a bizonyítványok, transcriptek, fokozatok megbízható rögzítésére. Gya- korlatilag lehetetlenné teszi az okirathamisítást, mert a technológia ezt azonnal képes felfedni. Pozitív oldalról pedig sokkal részletesebben rögzíthető benne a hallgató profilja, mindent lehet benne tárolni, bizonyítható és nyomon követhető minden kurzus, amit egy hallgató valaha is tanult. Hátránya ennek is van, költség- gel rendelkezik, és noha hamisítani nem lehet, az általa szolgáltatott bizonyítvány minőségéről nem tud információt adni. Elképzelhető, hogy a jövőben használni fogják, de itt is számítani fog a kibocsátó személye (AACSB 2020, McArthur 2018). i – Aktív tanulást igényel. – A hallgatót azonnal értékelni lehet, és azonnali visszajelzést lehet adni neki a munkájáról (AACSBa 2018). Az Egyesült Államokban használt ilyet a Massachusetts Institute of Technology, Sloan School of Management, amikor a Sandy hurrikán miatt a hallgatók nem jutottak be az egyetemre, és virtuális platformon keresztül tanították őket. A Stan- ford University Graduate School of Business online programot indított, ahol a hallgatók virtuálisan vehetnek részt az oktatásban avatárok használatával. A hallgatók ezt a módszert jobban szerették, mint a videókonferencián keresztüli együttműködést, és természetesebben használták az egymással való közös mun- kában is (AACSBa 2018). 5.4. Mobil tanulási lehetőségek és viselhető (wearable) technológiák A tanulás könnyebbé tételének egy ígéretes lehetősége a mobileszközök csoportja. Okostelefonok, tabletek, a Google Glass egyre jobban terjednek az oktatásban. A mobileszközök szolgálhatnak kiegészítésként a hagyományos oktatáshoz, de bizo- nyos fejlődő országokban az internethez való kapcsolódás kizárólagos eszköze a mobiltelefon és a tablet, mint például Nepálban (AACSBb 2018). Ázsiában, Afri- kában és Latin-Amerikában érik el a felhasználók a világhálót főleg mobileszkö- zökön keresztül. A legnépszerűbbek a nyelvtanuló, játékos alkalmazások, mint a Duolingo, vagy a gyerekeknek készített alkalmazások és a „brain trainer”-ek, a memória javító mobil szoftverek. A Verified Market Research weboldal szerint a globális mobil tanulási piac 2019-ben 27,32 milliárd dollárra tehető, és 2027-re elérheti a 280,08 milliárd dollárt (Global Mobile Learning… 2018). Ezek a technológiák az üzleti oktatásban is nagyon kiterjesztik a lehetőségeket. Csak néhány példa: y – Olyan környezetet is be lehet mutatni, amit egyébként nem lehetne,i – Olyan környezetet is be lehet mutatni, amit egyébként nem lehetne, mert veszélyes, drága vagy egyszerűen fizikai képtelenség. A mobileszközöket hasznosító tanulás nem feltételezi, hogy a felhasználó hos�- szú ideig foglalkozik a tananyaggal. Kis rövid „falatnyi” (bite sized) információt közöl. A mobil-first technológia eleve kis képernyőre tervezi az információt és keveset. A „micro-learning” esetén is nagyon hasznos ez a technológia. Itt sokkal kevesebb információt közölnek, és a felhasználó bármikor eléri, rövid időinter- i – A komplex fogalmak megértését segíti azzal, hogy a fogalmat az „életben” mutatja meg. j g – A kreativitást nagyban elősegíti a virtuális világ manipulálhatósága. 15 14 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN vallumban is képes tanulni (például utazás vagy sorban állás közben) (AACSBb 2018). i ner 2006-os cikkében is hangsúlyozzák a munkamemória szerepét a tanulásban. A rövid távú, azaz a munkamemória (working memory) csak igen kevés ideig képes csekély információt tárolni. A hosszú távú memória kapacitása határtalan, és életünk végéig tart (Szabó 2000). A hosszú távú memóriából a rövid távú me- móriába visszahívható az információ, ezt hívják irányított gondolkodásnak, és ez az a folyamat, ami lehetővé teszi magát a normális életet. A rövid távú memóri- ából az új információ 30 másodperc alatt eltűnik. Ha a hosszú távú memóriából hívjuk vissza az információt, ez a korlátozás nincsen, akkor az ismeret megmarad. E folyamat nélkül képtelenek lennénk egészen egyszerű feladatok megoldására is, mert ha nem emlékszünk rá, nem tudjuk megtenni. A hosszú távú memóriának alapvető szerepe van a tanulásban. 5.4. Mobil tanulási lehetőségek és viselhető (wearable) technológiák Amikor tanulunk/tanítunk, a hosszú távú me- móriában okozunk változást, ha ez nem történik meg, tanulás sincsen (Kirschner 2006). A mobil tanulás támogatja a játékos tanulási formákat (gamification). Az in- ternetes felületeken és az okostelefonokon való játékok tulajdonságai köszönnek vissza a játékos szoftverekben. A tanulókat ugyanúgy motiválják, mint a játékok- ban, egyre magasabb szintekre érhetnek el, és jutalmakat gyűjthetnek. A sok po- zitív feedback segít abban, hogy a tanuló visszatérjen további pozitív élményekért, és így tanuljon. Ezek az alkalmazások könnyebbé, élvezetesebbé próbálják tenni a tanulást, és próbálják a memorizálás nehézségeit feledtetni. Ezekkel a technikák- kal azt feltételezik, hogy a tanuló motivációja tovább megmarad, ha a tanulás a játékkal asszociálódik. j A mobil tanuláshoz kapcsolódhat az adaptív tankönyvek rendszere, ahol a ta- nuló igényéhez alkalmazkodik a szoftver. A hagyományos tankönyvek helyett a szoftver mesterséges intelligenciája felméri a tanuló igényét, és olyan problémákat ad neki, amelyeket képes megoldani, gyakoroltat és igény szerint lépteti tovább. Ezzel a technológiával lehetővé válik bizonyos diszciplínákban, hogy a tantermek- ben a különböző képességű és igényű hallgatók személyre szabott oktatást kap- janak. Az oktatás hatékonysága így jelentősen megnövelhető. Lényeges azonban, hogy emberi közreműködés nélkül a szoftver nem működik, mert a tanulóknak szüksége lehet többlet magyarázatra, és olyat a gép egyelőre nem tud adni (Ore- mus 2015). Akkor hatékony az oktatás, ha a hallgató hosszú távú memóriájába kerül a szükséges információ, ami kizárólag irányított tanulás és az ismeret rögzítése út- ján történhet. Ha a modern módszereket tekintjük, akkor a flipped classroom is csak akkor tesz eleget ennek a követelménynek, ha a diák rögzíti a tananyag el- méleti részét, és ennek segítségével oldja meg a gyakorlati feladatokat. Itt komoly gond lehet, ha meg sem érti a feladott elméletet, mert nem volt előadás és ma- gyarázat. Nem egyértelmű a flipped classroom értékelése, például Ryan és Reid (2016) általános kémiát tanuló egyetemi hallgatóknál azt találta, hogy a vizsga- eredmények azonosak voltak a flipped classroomot használó és a kontrollcsoport- ban. A leggyengébb hallgatóknál volt csak szignifikáns eredményjavulás a flipped classroomot használva a kontrollcsoporthoz képest. Ők is hangsúlyozták, hogy a lehetséges jobb eredmény eléréséhez szükséges a hallgató és az oktató megnöveke- dett interakciója. Ez pedig ismét csak az irányított oktatás jelentőségét bizonyítja. 6. A digitális oktatás és a tanulás – következtetések A fentiekből belátható, hogy a tanulást és tanítást a 21. században igen sok hasz- nos eszköz, technológia és új kutatási ismeret segíti. Csábító lenne a feltételezés, hogy akkor változtassuk meg a tanítás technológiáját, hagyatkozzunk kizárólag az új módszerekre az oktatás megtervezésénél. Nem hagyhatjuk azonban figyelmen kívül a tanulás pszichológiáját sem, azt, hogy hogyan tud a hallgató ismereteket rögzíteni. Az emberi agy működéséről a digitális technológiák segítségével már sokkal világosabb képe van a kutatóknak, mint a 20. században. A fejlett képal- kotó módszerek képesek segíteni a folyamatok magyarázatában. A neurokémia kutatja az emlékezésben részt vevő vegyületek szerepét. A folyamatok, amelyeket vizsgálnak, nem változtak évszázadok óta. Az ismeretek rögzítésének agyi és bio- kémiai folyamata ugyanaz maradt, az ismeretek megszerzésének folyamata válto- zott meg alapvetően. A kognitív pszichológia részletesen elemzi a rövid és hosszú távú memória közötti kapcsolatot, és az ismeretek rögzítésének folyamatát. A fej- lett tanulásmódszertanok ezeket figyelembe veszik. Az utóbbi évek legsikeresebb Massive Open Online kurzusa a Learning How to learn volt, ahol a tanulás techni- káját tanították neuropszichológiai alapokon (Learning…. 2020). Itt is és Kirsch- Az összes digitális eszköz akkor hatékony, ha az általuk közvetített ismeretek átkerülnek a hosszú távú memóriába. Ez pedig csakis a hagyományos rögzítési folyamatok útján történhet. Hiába szép és érdekes a virtuális üzleti szimuláció, ha nincs hozzá ismeret, nincs mihez kötni. Az esettanulmányok elemzése is csak akkor hatékony, ha az irányított ismeretszerzés (előadás, feladatmegoldás) után próbáljuk a valós problémát elemeztetni a hallgatóval. Nem lehet a magyarázatot, a hallgató irányítását, az olvasást és a házi feladatot kiváltani egyéni kutatással. Az egyéni kutatás olyan terhelés alá veszi a rövid távú memóriát, hogy a hallgató nem lesz képes sem kiválogatni a szükséges tananyagot, sem megtanulni (Kirschner 2006). A digitális eszközök nagyon élvezetessé és könnyűvé teszik az ismeretekhez való hozzájutást. A szűretlen információ mindenhol való jelenléte nagyon nehéz- zé teszi a nem szakértő számára annak az eldöntését, hogy az valódi-e egyáltalán, és ha valódi, mennyire helytálló. Amikor a hallgatók önállóan dolgoznak projekte- ken, amikor prezentációkat készítenek elméleti alapok nélkül, egyrészt túlterhelik 17 16 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN A BRIEF HISTORY OF MOOCS (2020) letöltve: https://www.mcgill.ca/maut/cur- rent-issues/moocs/history 2020. 05. 19. A BRIEF HISTORY OF MOOCS (2020) letöltve: https://www.mcgill.ca/maut/cur- rent-issues/moocs/history 2020. 05. 19. a rövid távú memóriájukat, másrészt nem képesek szakértőként szűrni az adatok valódiságát, különösen akkor, ha kezdők. 6. A digitális oktatás és a tanulás – következtetések A játékos gyakorlatok, az érdekes vide- ók, a versenyek, esettanulmányok, mind csak illusztrációk, de a megtanulandó tartalmak még mindig a könyvekben (vagy digitális megfelelőjükben) találhatók, és a hagyományos módon, néha nagyon nehezen, sok hiba és bukás árán tanul- hatók meg. Ez nem jelenti azt, hogy vissza kell térni a tábla-kréta-előadás-vizsga módszeréhez, hiszen ezzel ma már nem versenyképes az oktató. De nem szabad elfelejteni, hogy a sok modern eszköz minek az érdekét szolgálja: a tanulásét, a jövő hatékony üzletembereinek a képzését. y ACTIVE LEARNING (2020) letöltve: https://cei.umn.edu/active-learning 2020. 04. 27. AMERICAN PSYCHOLOGICAL ASSOCIATION (2006, March 20): Multitas- king: Switching costs - subtle “switching” costs cut efficiency, raise risk. fi letöltve: http://www.apa.org/research/action/multitask.aspx. 2020.05.18. ARMSTRONG, P. (2017) Bloom’s taxonomy. Vanderbilt University Center for Tea- ching letöltve: https://cft.vanderbilt.edu/guides-sub-pages/blooms-taxonomy/ BAINBRIDGE D (2016) “Edtech is the next fintech” letöltve: https://techcrunch y y ching letöltve: https://cft.vanderbilt.edu/guides-sub-pages/blooms-taxonomy/ BAINBRIDGE, D. (2016) “Edtech is the next fintech.” letöltve: https://techcrunch. com/2016/08/13/edtech-is-the-next-fintech/ 2020. 04.30. t BAINBRIDGE, D. (2016) “Edtech is the next fintech.” letöltve: https://techcrunch. com/2016/08/13/edtech-is-the-next-fintech/ 2020. 04.30. „Nem vezet királyi út a geometriához” – ezzel kezdtük a tanulmányt. Azért ma már ez egy kicsit megváltozott, mert a tanuláshoz a 21. században használt esz- közök igen élvezetes utat is jelenthetnek. Az út valóban lehet élvezetes az elején, de maga az ismeretek rögzítése, az ismétlések, a gyakorlások, a sok teszt ma sem kellemesebb, mint száz évvel ezelőtt. A hipotézis, miszerint a modern oktatási módszertan könnyebbé és hatékonyabbá teszi az ismeretek megszerzését, és meg- könnyíti a tanulási folyamatot, csak részben fogadható el. A modern módszerek tehetik könnyebbé és hatékonyabbá az ismeretek megszerzését, de a tanulás pszi- chológiai folyamatát nem képesek megkönnyíteni. A megszerzett biztos tudás ál- tal adott önbizalom, stabil értékrend és magasabb színvonalú életminőség viszont megéri a sokszor keserves küzdelmet. i BATES, T. (2019) Teaching in a Digital Age Guidelines for designing teaching and learning. letöltve: https://opentextbc.ca/teachinginadigitalage/ 2020. 04. 28. BAIRD, C.H. (2015) Myths, exaggerations and uncomfortable truths. letöltve: htt- ps://www.ibm.com/downloads/cas/Q3ZVGRLP 2020. 04. 28. BARROWS, H. S. (1996) Problem-based learning in medicine and beyond: A brief overview. New Directions for Teaching and Learning. (68): 3–12. doi:10.1002/ tl.37219966804. BELL, S. (2010) Project-Based Learning for the 21st Century: Skills for the Future, The Clearing House, 83: 39–43, Routledge DOI: 10.1080/00098650903505415 h BERG, G.A., SIMONSON, M.(2018) Distance learning letöltve: https://www.bri- tannica.com/topic/distance-learning#ref1146520 2020.08.18.h Irodalomjegyzék BLAZEVIC, O. (2018) What Are Micro Credentials and How Can They Be- nefit You? letöltve: https://www.training.com.au/ed/how-micro-credenti- als-can-benefit-you/?utm_referrer=https%3A%2F%2Fwww.google.hu%2F 2020.08.18. AACSB Research Reports (2020) letöltve: https://www.aacsb.edu/publications/ researchreports 2020.04.16. BRAME, C. (2013) Flipping the classroom. Vanderbilt University Center for Tea- ching. letöltve: http://cft.vanderbilt.edu/guides-sub-pages/flipping-the-class- room/ 2020.05.19. AACSBa (2020) Technologies With Potential to Transform Business and Business Education: Artificial Intelligence. letöltve: https://www.aacsb.edu/publications/ researchreports/artificialintelligence 2020. 05.15. Chamorro-Premuzic, T. & Frankiewicz, B. (2019) Does Higher Education Still Pre- pare People for Jobs? i AACSBb (2020) Technologies With Potential to Transform Business and Business Education: Blockchain. letöltve: https://www.aacsb.edu/publications/research- reports/blockchain 2020. 05.15 letöltve: https://hbr.org/2019/01/does-higher-education-still-prepare-people-for- jobs 2020.04. 28. AACSBa (2018) Technologies With Potential to Transform Business and Business Education: Virtual and Augmented Reality. letöltve: https://www.aacsb.edu/ publications/researchreports/virtual-and-augmented-reality 2020. 05.15. COSTANZA, D. P., BADGER, J. M., FRASER, R. L., SEVERT, J. B., & GADE, P. A. (2012) Generational differences in work-related attitudes: A meta-analysis. Jour- nal of Business and Psychology, 27 (4), 375–394. https://doi.org/10.1007/ s10869-012-9259-4 AACSBb (2018) Technologies With Potential to Transform Business and Business Education:Mobile and Micro-Learning. letöltve: https://www.aacsb.edu/-/me- dia/aacsb/publications/research-reports/mobile_and_micro_learning_brief. ashx?la=en&hash=0B3788FA5096BBA862587A10C4A9685A6CF55CF4 2020. 06.09. DE JUAN VIGARAY, M. D., LÓPEZ, J.J., PERIS, J.E., YÁÑEZ, E., MARTINEZ, C., CUEVAS, J., POSADAS, J.A., VALLÉS, M.L. (2010) Teaching Methods: Stu- dy and Results in Several Modules of Business Studies. Proceedings of EDU- 19 18 KISPÁL-VITAI ZSUZSANNA A DIGITALIZÁCIÓ MÓDSZERTANI KIHÍVÁSAI AZ ÜZLETI KÉPZÉSBEN LADA, L. (2009) Oktatási módszerek letöltve: https://ofi.oh.gov.hu/tudastar/prob- lemak-kerdesek/oktatasi-modszerek 2020. 04. 29. LEARN10 Conference. 5th-7th July 2010, Barcelona, Spain. letöltve: https:// core.ac.uk/download/pdf/16369578.pdf 2020. 04. 27. LEARN10 Conference. 5th-7th July 2010, Barcelona, Spain. letöltve: https:// core.ac.uk/download/pdf/16369578.pdf 2020. 04. 27. LEARNING HOW TO LEARN: POWERFUL MENTAL TOOLS TO HELP YOU MASTER TOUGH SUBJECTS (2020) letöltve: https://www.coursera.org/ learn/learning-how-to-learn 2020.05.18. FALLOW, S. AND STEVEN, C. (2000) Integrating Key Skills in Higher Educa- tion: Employability, Transferable Skills and Learning for Life. London UK/ Sterling VA: Kogan Page/Stylus letöltve: https://books.google.ca/books?h- l=en&lr=&id=ldBTAQAAQBAJ&oi=fnd&pg=PP1&dq=Books+2012:+s- kills+development+&ots=ynzUemTGgy&sig=BbIrHGgmA3GJ0rzn9PJw7A- onV3E#v=onepage&q&f=false 2020.04. 28. g MATHIS, R.L., JACKSON, J.H., VALENTINE, S.R., MEGLICH, P.A. (2017) Hu- man Resource Management. Cengage learning 15. kiadás MCARTHUR, D.M. (2018) Will blockchains revolutionize education? letöltve: htt- ps://er.educause.edu/articles/2018/5/will-blockchains-revolutionize-educa- tion. 2020.05.10 FREEMAN, S., EDDY, S. L., MCDONOUGH, M., SMITH, M.K., OKOROAFOR, N., JORDT.H. (2014) Active learning increases student performance in science, engineering, and mathematics. in: PNAS, 111. kötet, 23. szám, 8410–8415. olda- lak doi: https://www.pnas.org/content/111/23/8410 MCLEOD, S. A. (2015) Cognitive psychology. Simply Psychology. letöltve: https:// www.simplypsychology.org/cognitive.html 2020.04.28. y y gy g g MCLEOD, S. (2017a) Behaviorist approach. letöltve: https://www.simplypsycho- logy.org/behaviorism.html 2020. 04.28. Irodalomjegyzék GARTNER (2020) Gartner Says Global IT Spending to Reach $3.9 Trillion in 2020 letöltve: https://www.gartner.com/en/newsroom/press-releases/2020-01-15- gartner-says-global-it-spending-to-reach-3point9-trillion-in-2020 2020.08.18. MCLEOD, S. (2017b) Constructivism as a theory for teaching and learning letöltve: https://www.simplypsychology.org/constructivism.html 2020.04.28.h GLOBAL MOBILE LEARNING MARKET SIZE BY END USER, BY APPLICA- TION, BY SOFTWARE SOLUTION, BY GEOGRAPHIC SCOPE AND FO- RECAST (2018) letöltve: https://www.verifiedmarketresearch.com/product/ global-mobile-learning-market-size-and-forecast-to-2025/ 2020. 06.09.t NEELAKANTAN, S. (2020) Successful AI Examples in Higher Education That Can Inspire Our Future letöltve: https://edtechmagazine.com/higher/artic- le/2020/01/successful-ai-examples-higher-education-can-inspire-our-future 2020.08.18. GOVINDARAJAN, V & SRIVASTAVA, A. (2020) What the Shift to Virtual Lear- ning Could Mean for the Future of Highered. letöltve: https://hbr.org/2020/03/ what-the-shift-to-virtual-learning-could-mean-for-the-future-of-higher-ed 2020.05.27. NUMBER OF SMARTPHONE USERS WORLDWIDE FROM 2016 TO 2021 (2020) letöltve: https://www.statista.com/statistics/330695/number-of-smart- phone-users-worldwide/ 2020.04.28. NUMBER OF INTERNET USERS WORLDWIDE FROM 2005 TO 2019 (2020) letöltve: https://www.statista.com/statistics/273018/number-of-inter- net-users-worldwide/ 2020.04.28.h HELLER, J. (2017) Push research: how ai is fundamentally changing the way we re- search the law. letöltve: https://abovethelaw.com/2017/11/push-research-how- ai-is-fundamentally-changing-the-way-we-research-the-law/ 2020.04.30. OECD (2019): OECD Skills Outlook 2019: Thriving in a Digital World. OECD Publishing, Paris https://doi.org/10.1787/df80bc12-en. HELLO HR AND RISK & COMPLIANCE MANGERS (2020) letöltve: https://int- raspexion.com/ 2020.05.10. ( ) l h k OREMUS, W. (2015) No More Pencils, No More Books. letöltve: http://www.slate. com/articles/technology/technology/2015/10/adaptive_learning_software_ is_replacing_textbooks_and_upending_american.html?via=gdpr-consent 2020.06.09. HUCZYNSKI, A.A. & BUCHANAN, D.A (2007) Organizational Behaviour 6. ki- adás FT. Prentice Hall, Pearson Education INNOVATIONS THAT INSPIRE (2020) letöltve: https://www.aacsb.edu/about/ advocacy-and-awareness/member-challenges/innova%ADtions-that-ins- pire/2018/university-of-st-thomas 2020.08.18. PERANANDAM, C. (2018) AI Helps Duolingo Personalize Language Learning. letöltve: https://www.wired.com/brandlab/2018/12/ai-helps-duolingo-perso- nalize-language-learning/ 2020. 04. 17. KIRSCHNER P.A., SWELLER, J., CLARK, R.E. (2006) Why Minimal Guidance During Instruction Does Not Work: An Analysis of the Failure of Constructivist, Discovery, Problem-Based, Experiential, and Inquiry-Based Teaching. Educa- tional Psychologist, 41. kötet. 2. szám, 75-86. oldalak https://doi.org/10.1207/ s15326985ep4102_1h g g g ROBBINS, S.P. & JUDGE, T. A. (2017) Organizational Behavior. Pearson educa- tion, 17. kiadás RYAN, M.D., REID, S.A. (2016) Impact of the Flipped Classroom on Student Performance and Retention: A Parallel Controlled Study in General Chemist- ry Journal of Chemical Education 93, 13-23. https://doi.org/10.1021/acs. jchemed.5b00717 y g s15326985ep4102_1 KIRSCHNER, P. A. & Bruyckere, P.D. (2017) The myths of the digital native and the multitasker. Teaching and Teacher Education 67 kötet, 135-142. oldalak. https://doi.org/10.1016/j.tate.2017.06.001 21 20 KISPÁL-VITAI ZSUZSANNA ROUSE, M. (2020) Artificial intelligence. letöltve: https://searchenterpriseai.tech- target.com/definition/AI-Artificial-Intelligence 2020.04.27. PRENSKY, M. (2001) Digital Natives, Digital Immigrants. On the Horizon (MCB University Press, 9. kötet, 5. szám. letöltve: https://www.marcprensky.com/ writing/Prensky%20-%20Digital%20Natives,%20Digital%20Immigrants%20 -%20Part1.pdfh SHAH, D. (2019) By The Numbers: MOOCS in 2019. letöltve: https://www.class- central.com/report/mooc-stats-2019/ 2020. 04.26. SIMONSON, M., BERG, G.A. (2017) Distance learning. letöltve: https://www.bri- tannica.com/topic/distance-learning SZABÓ, I. (2000) Klinikai neuropszichológia. oktatási segédanyag, Pécs É Á SZÉLES-HORVÁTH, a.(2016) „Mindent akarok, és rögtön, ez a jelszó” – Tari An- namária könyvét ajánljuk (1. rész) letöltve: https://pszichoforyou.hu/genera- cok-online/ 2020.05.18. VAN DER WAERDEN, B.L., TAISBAK, C.M. (2019) Euclid, Greek matematician. letöltve: https://www.britannica.com/biography/Euclid-Greek-mathematician 2020.04.27h WAGNER, T. (2010) Overcoming The Global Achievement Gap (online). Camb- ridge, Mass., Harvard University in: SCOTT, C.L. (2015) The Futures of lear- ning 2: what kind of learning for the 21st century? letöltve: https://unesdoc. unesco.org/ark:/48223/pf0000242996 2020. 04. 16. WAN, T. (2019) Us edtech investments peak again with $1.45 billion raised in 2018. letöltve: https://www.edsurge.com/news/2019-01-15-us-edtech-investments- peak-again-with-1-45-billion-raised-in-2018 WHAT ARE THE TOP 10 SKILLS THAT’LL GET YOU A JOB WHEN YOU GRADUATE? letöltve: https://targetjobs.co.uk/careers-advice/career-plan- ning/273051-the-top-10-skills-thatll-get-you-a-job-when-you-graduate 2020. 04. 28. WHAT IS VIRTUAL REALITY? (2020) letöltve: https://www.vrs.org.uk/vir- tual-reality/what-is-virtual-reality.html 2020. 05. 15. WILKINSON, A., REDMAN, T., DUNDON, T. (2017) Contemporary Human Re- source Management, Text and Cases. Pearson education 5. kiadás 22 22
https://openalex.org/W2570609296	https://bmjopen.bmj.com/content/bmjopen/7/1/e012818.full.pdf	English	null	Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study	BMJ open	2,017	cc-by	9,459	Strengths and limitations of this study Objectives: To examine adherence to serum creatinine and potassium monitoring and discontinuation guidelines following initiation of treatment with ACE inhibitors (ACEI) or angiotensin receptor blockers (ARBs); and whether high-risk patients are monitored. ▪This is the largest monitoring study until now, examining both adherence to creatinine and potassium monitoring and discontinuation guidelines following initiation of ACE inhibitors or angiotensin receptor blockers in UK primary care, and whether patients are monitored in accordance with their individual risk profile. Design: A general practice-based cohort study using electronic health records from the UK Clinical Practice Research Datalink and Hospital Episode Statistics. ▪Use of the UK Clinical Practice Research Datalink and Hospital Episode Statistics ensured that the study was population-based and not restricted to specific demographic, hospital or insurance groups. Setting: UK primary care, 2004–2014. Setting: UK primary care, 2004–2014. Subjects: 223 814 new ACEI/ARB users. ▸Prepublication history and additional material is available. To view please visit the journal (http://dx.doi.org/ 10.1136/bmjopen-2016- 012818). Main outcome measures: Proportion of patients with renal function monitoring before and after ACEI/ ARB initiation; creatinine increase ≥30% or potassium levels >6 mmol/L at first follow-up monitoring; and treatment discontinuation after such changes. Using logistic regression models, we also examined patient characteristics associated with these biochemical changes, and with follow-up monitoring within the guideline recommendation of 2 weeks after treatment initiation. ▪Blood tests performed in hospital systems were not recorded in the Clinical Practice Research Datalink, but the results were consistent for patients with no recent hospital admissions. on October 23, 2024 by guest. Protected by copyright. //bmjopen.bmj.com/ on October 23, 2024 by guest. Protected by copyright. //bmjopen.bmj.com/ on October 23, 2024 by guest. Protected by copyright. //bmjopen.bmj.com/ ▪If the recording of creatinine levels was not missing completely at random, the associations between patient characteristics and creatinine increase may have been underestimated. ▪If the recording of creatinine levels was not missing completely at random, the associations between patient characteristics and creatinine increase may have been underestimated. Received 27 May 2016 Revised 5 August 2016 Accepted 12 October 2016 Results: 10% of patients had neither baseline nor follow-up monitoring of creatinine within 12 months before and 2 months after initiation of an ACEI/ARB, 28% had monitoring only at baseline, 15% only at follow-up, and 47% both at baseline and follow-up. The median period between the most recent baseline monitoring and drug initiation was 40 days (IQR 12–125 days). Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study Morten Schmidt,1,2,3 Kathryn E Mansfield,1 Krishnan Bhaskaran,1 Dorothea Nitsch,1 Henrik Toft Sørensen,2 Liam Smeeth,1 Laurie A Tomlinson1 To cite: Schmidt M, Mansfield KE, Bhaskaran K, et al. Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016- 012818 st. Protected by copyright. st. Protected by copyright. st. Protected by copyright. To cite: Schmidt M, Mansfield KE, Bhaskaran K, et al. Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016- 012818 on October 23, 2024 by g http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from on October 23, 2024 by g http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from on October 23, 2024 by g http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 20 BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 20 BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 20 on October 23, 2024 by guest. Protec http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from on October 23, 2024 by guest. Protec http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from on October 23, 2024 by guest. Protec http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Received 27 May 2016 Revised 5 August 2016 Accepted 12 October 2016 Open Access Open Access Research Strengths and limitations of this study 34% of patients had baseline creatinine monitoring within 1 month before initiating therapy, but <10% also had the guideline-recommended follow- up test recorded within 2 weeks. Among patients experiencing a creatinine increase ≥30% (n=567, 1.2%) or potassium level >6 mmol/L (n=191, 0.4%), 80% continued treatment. Although patients with prior myocardial infarction, hypertension or baseline potassium >5 mmol/L were at high risk of ≥30% increase in creatinine after ACEI/ARB initiation, there was no evidence that they were more frequently monitored. Conclusions: Only one-tenth of patients initiating ACEI/ARB therapy receive the guideline-recommended creatinine monitoring. Moreover, the vast majority of the patients fulfilling postinitiation discontinuation criteria for creatinine and potassium increases continue on treatment. 3, 2024 by guest. Protected by copyright. 3, 2024 by guest. Protected by copyright. 3, 2024 by guest. Protected by copyright. Laboratory data y All creatinine test results were extracted from the general practice records of the study population, using creatinine-speciﬁc codes in CPRD. Cross-reference was then made to creatinine test results identiﬁed from a broad Read code search. Any irrelevant codes were excluded. Renal function testing in the UK includes cre- atinine and potassium, so it can be inferred that testing frequency is similar to creatinine for potassium. When we conducted analyses related to potassium levels, we repeated the procedure used to identify creatinine levels for potassium test results. INTRODUCTION For numbered affiliations see end of article. Renin angiotensin system blockade using ACE inhibitors (ACEI) and angiotensin receptor blockers (ARBs) is a mainstay in treatment of hypertension,1 heart failure,2 diabetic microalbuminuria or proteinuric Correspondence to Dr Morten Schmidt; morten.schmidt@clin.au.dk Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 1 Open Access BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2 renal diseases,3 and after myocardial infarction.4 However, some patients experience a sudden decline in kidney function when initiating these drugs, presumably due to antagonism of the angiotensin II-mediated effer- ent arteriolar constriction or impaired kidney excretion of potassium.5 6 treatment initiation,2–4 and for patients with myocardial infarction at least annually thereafter.4 A baseline assess- ment and follow-up test within 2 weeks is also recom- mended by the UK Renal Association,12 as well as the frequently used online web resource General Practice (GP) Notebook.13 GP Notebook additionally recom- mends monitoring 1, 3, 6 and 12 months after the ﬁrst follow-up test.13 NICE recommends not to initiate ACEI/ARBs in patients with a baseline potassium level >5 mmol/L and to discontinue therapy if potassium rises above 6 mmol/L. The potential impact on kidney function should be evaluated by comparing preinitiation and postinitiation levels of serum creatinine and potassium.7 Discontinuation is recommended if the rise in creatinine exceeds 30% above baseline or if hyperkalaemia develops.8 It is unclear whether these recommendations are routinely followed in clinical practice.9 on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Patient characteristics We obtained information for all patients on age, sex, cal- endar period of ACEI/ARB initiation (2004–2008 and 2010–2014), socioeconomic status (quintiles of the 2004 index of multiple deprivation scores), lifestyle factors (smoking, alcohol intake and body mass index), baseline potassium level (≤5 or >5 mmol/L), CKD, cardiovascu- lar comorbidities (heart failure, myocardial infarction, hypertension, peripheral arterial disease and arrhyth- mia) and diabetes.15 We used algorithms for smoking status, alcohol intake and body mass index based on the most recent records in the CPRD before ACEI/ARB ini- tiation.16 17 As measures of baseline creatinine and potassium levels, we used the single most recent meas- urement within 12 months before the ﬁrst ACEI/ARB METHODS Data sources We used the UK’s Clinical Practice Research Datalink (CPRD) linked to hospital record data from the Hospital Episode Statistics (HES) database. The CPRD database contains primary care electronic health record data from 7% of the UK population (∼15 million patient lives, with ∼8 million currently under follow-up).10 Patients included in the CPRD are largely representative of the UK population in terms of age, sex and ethnicity.10 11 Information recorded in the database includes demographics such as sex and year of birth, the location of the general practice, medical diagnoses (based on ‘Read’ codes), drug prescriptions and a range of routine laboratory test results. HES records cover all hospital admissions for patients covered by the National Health Service (NHS) who receive treatment either from English NHS trusts or independent providers.10 11 Fifty-eight per cent of general practices included in the CPRD have agreed to HES linkage.10 We obtained linked data on socioeconomic status (index of multiple deprivation) based on area of residence. ACEI/ARB initiators We identiﬁed a cohort of all HES-linked CPRD patients aged ≥18 years, who initiated ACEI/ARB treatment between 1 January 2004 and 31 March 2014. We did not include earlier calendar periods, as laboratory data before 2004 were incomplete due to interface problems between laboratory reporting software and GP practice management software.14 Also, creatinine testing was incentivised in 2004 with the introduction of the dia- betes Quality and Outcomes Framework (QOF) and further in 2006 with the CKD QOF.14 To rule out any potential inﬂuence of incomplete data around 2004, we also examined the most recent 5-year calendar period separately in a sensitivity analysis. New users were deﬁned as persons with at least 1 year of continuous registration in the CPRD before their ﬁrst recorded ACEI/ARB prescription. y p A few studies have compared baseline and follow-up monitoring results,9 but large studies using contempor- ary data with reference to current guidelines are lacking, and it is unknown whether patients’ individual risk of renal impairment inﬂuences their likelihood of being monitored.9 We therefore examined adherence to cre- atinine and potassium monitoring and treatment discon- tinuation guidelines following ACEI/ARB initiation in UK primary care, and whether patients are monitored in accordance with their individual risk proﬁle. on October 23, 2024 by guest. P http://bmjopen.bmj.com/ 9 January 2017. Downloaded from Statistical analysis We described ACEI/ARB users according to patient characteristics, both overall and according to creatinine monitoring status (no baseline or follow-up monitoring, baseline only, follow-up only, and both baseline and follow-up monitoring). Baseline monitoring was deﬁned as a test performed on the date of drug initiation or within either 12 months before (wide interval) or 1 month before initiation (more ideal interval assumed to be driven by planned ACEI/ARB initiation). To accord with the postinitiation monitoring interval recommended from previous trial data, we considered only follow-up monitoring within the ﬁrst 2 months after drug initiation.8 Finally, we ﬁtted a logistic regression model to identify patient characteristics associated with a severe decline in renal function (creatinine increase ≥30% or potassium level >6 mmol/L) and compared these characteristics with those associated with receiving postinitiation follow-up monitoring within 2 weeks. The model included age, sex, CKD stage, cardiovascular comorbid- ities, diabetes and baseline potassium level (>5 vs ≤5 mmol/L). In three additional model-based sensitivity analyses, we repeated the analyses (1) excluding patients with a recent hospitalisation (as deﬁned above); (2) omitting baseline potassium from the model to examine the extent of potential overﬁtting when both baseline potassium and CKD stage were kept in the model; and (3) also adjusting additionally for ethnicity. on October 23, 2024 by guest. Prot http://bmjopen.bmj.com/ anuary 2017. Downloaded from g We calculated the proportion of persons in the total cohort of new users who had baseline and follow-up monitoring (within 1, 3 and 12 months before drug initi- ation and within 2 weeks, 1 month and 2 months after initiation). We then computed the proportion of persons with both baseline and initial follow-up monitor- ing within the guideline-recommended interval of 2 weeks following drug initiation. All analyses were performed using the STATA 14 statis- tical software package. g g We repeated the analyses for continuing users, in order to examine adherence to the stricter guideline recommendations for ongoing monitoring (ie, monitor- ing within 1, 3, 6 and 12 months after the ﬁrst retest).13 Continuation was deﬁned as ACEI/ARB use beyond 30 days following the monitoring date, that is, when the end date of the ﬁrst continuous course of therapy was after the date of the ﬁrst monitoring date plus 30 days (to allow for stockpiling). Patient involvement We used the subcohort of patients with both baseline and follow-up monitoring to calculate the proportion of patients with creatinine increases ≥30% or potassium levels >6 mmol/L at the ﬁrst follow-up monitoring within 2 months after initiation, as well as the proportion of patients continuing treatment despite these contrain- dications for use. The study included no patient involvement. Open Access within 1 month before or after their ﬁrst ACEI/ARB pre- scription, in order to account for drug initiation and any subsequent renal function tests occurring in the hospital and therefore not captured in the CPRD; (4) focusing on speciﬁc patient subgroups (heart failure, myocardial infarction, hypertension, CKD (eGFR<60 mL/min/ 1.73 m2), peripheral arterial disease and diabetes); and (5) deﬁning drug use continuation as ACEI/ARB use beyond 90 days (instead of 30 days) after the ﬁrst retest date. within 1 month before or after their ﬁrst ACEI/ARB pre- scription, in order to account for drug initiation and any subsequent renal function tests occurring in the hospital and therefore not captured in the CPRD; (4) focusing on speciﬁc patient subgroups (heart failure, myocardial infarction, hypertension, CKD (eGFR<60 mL/min/ 1.73 m2), peripheral arterial disease and diabetes); and (5) deﬁning drug use continuation as ACEI/ARB use beyond 90 days (instead of 30 days) after the ﬁrst retest date. prescription. We calculated the estimated glomerular ﬁl- tration rate (eGFR) level from the most recent creatin- ine measurement and CKD stage from the CKD Epidemiology Collaboration (CKD-EPI) equation.18 Cardiovascular comorbidities and diabetes were identi- ﬁed from both the CPRD and HES based on diagnoses recorded prior to ACEI/ARB initiation. The code lists for all variables are provided in the online supplementary appendix. on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ rst published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Statistical analysis The end date of each pre- scription was calculated by adding the prescription dur- ation (total number of tablets prescribed divided by the speciﬁed number of tablets per day) to the prescription date. In identifying continuous courses of therapy, we allowed for a 30-day gap between the end date of one prescription and the start of the next consecutive prescription. Monitoring guidelines Consistent with other international guidelines, the National Institute for Health and Care Excellence (NICE) recommends baseline testing of creatinine when initiating ACEI/ARB therapy in patients with hyperten- sion,1 heart failure,2 myocardial infarction4 or chronic kidney disease (CKD).3 The time interval for baseline testing is not further speciﬁed.1–4 Among patients with heart failure, myocardial infarction and CKD, NICE recommends follow-up monitoring within 2 weeks of est. Protected by copyright. 2 Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Table 1 Characteristics of patients initiating ACE inhibitors or ARBs in the UK primary care during 2004–2014, by monitoring groups ACEI/ARB therapy despite baseline potassium above 5 mmol/L. on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from The median number of days between base- line monitoring and ﬁrst prescription date was 40 days Among all patients initiating ACEI/ARB therapy, the proportion of patients receiving creatinine testing before initiation was 76% within 12 months of treatment initi- Table 1 Characteristics of patients initiating ACE inhibitors or ARBs in the UK primary care during 2004–2014, by monitoring groups Serum creatinine monitoring* Total No baseline or follow-up tests Baseline test only Follow-up test only Baseline and follow-up tests Total number 21 411 (100) 63 359 (100) 33 185 (100) 105 859 (100) 223 814 (100) Female sex 8882 (41) 27 722 (44) 14 570 (44) 49 109 (46) 100 283 (45) Age (years) <50 5019 (23) 13 697 (22) 8732 (26) 19 910 (19) 47 358 (21) 50–59 5485 (26) 15 135 (24) 9115 (27) 24 866 (23) 54 601 (24) 60–69 4863 (23) 15 586 (25) 7776 (23) 27 790 (26) 56 015 (25) 70–79 3579 (17) 12 193 (19) 5066 (15) 22 152 (21) 42 990 (19) 80+ 2465 (12) 6748 (11) 2496 (8) 11 141 (11) 22 850 (10) Calendar period 2004–2008 14 814 (69) 40 667 (64) 19 808 (60) 60 902 (58) 136 191 (61) 2009–2014 6597 (31) 22 692 (36) 13 377 (40) 44 957 (42) 87 623 (39) SES quintiles 1 (low) 5153 (24) 15 290 (24) 8533 (26) 25 577 (24) 54 553 (24) 2 4725 (22) 14 331 (23) 7887 (24) 24 851 (23) 51 794 (23) 3 4341 (20) 13 028 (21) 6890 (21) 22 629 (21) 46 888 (21) 4 4254 (20) 12 140 (19) 5931 (18) 19 318 (18) 41 643 (19) 5 (high) 2925 (14) 8508 (13) 3898 (12) 13 359 (13) 28 690 (13) Missing 13 (0) 62 (0) 46 (0) 125 (0) 246 (0) Smoking status Never 7860 (37) 22 496 (36) 12 229 (37) 36 895 (35) 79 480 (36) Ever 13 433 (63) 40 797 (64) 20 915 (63) 68 939 (65) 144 084 (64) Missing 118 (1) 66 (0) 41 (0) 25 (0) 250 (0) Alcohol intake No use 2556 (12) 7819 (12) 3409 (10) 11 088 (10) 24 872 (11) Current 15 495 (72) 47 322 (75) 25 656 (77) 82 870 (78) 171 343 (77) Former 1328 (6) 4499 (7) 1933 (6) 7490 (7) 15 250 (7) Missing 2032 (9) 3719 (6) 2187 (7) 4411 (4) 12 349 (6) BMI groups Underweight 282 (1) 700 (1) 304 (1) 1008 (1) 2294 (1) Healthy weight 5666 (26) 15 406 (24) 8089 (24) 24 972 (24) 54 133 (24) Overweight 7677 (36) 23 755 (37) 12 484 (38) 40 556 (38) 84 472 (38) Obesity 6009 (28) 20 660 (33) 10 527 (32) 35 887 (34) 73 083 (33) Missing 1777 (8) 2838 (4) 1781 (5) 3436 (3) 9832 (4) CKD (eGFR)† Stage ≤2 (≥60) 10 326 (48) 53 773 (85) 19 470 (59) 87 484 (83) 171 053 (76) Stage 3a (45–59) 1137 (5) 7382 (12) 1766 (5) 13 913 (13) 24 198 (11) Stage 3b (30–44) 217 (1) 1885 (3) 265 (1) 3854 (4) 6221 (3) Stage 4 (15–29) 24 (0) 319 (1) 29 (0) 608 (1) 980 (0) Not measured 9707 (45) 0 (0) 11 655 (35) 0 (0) 21 362 (10) CV comorbidities‡ Heart failure 1568 (7) 3270 (5) 1386 (4) 4583 (4) 10 807 (5) Myocardial infarction 3881 (18) 4653 (7) 3203 (10) 4620 (4) 16 357 (7) Hypertension 13 023 (61) 44 273 (70) 24 195 (73) 80 946 (76) 162 437 (73) Peripheral arterial disease 471 (2) 1590 (3) 523 (2) 2547 (2) 5131 (2) Arrhythmia 2057 (10) 4973 (8) 2000 (6) 7123 (7) 16 153 (7) Diabetes mellitus 1399 (7) 13 586 (21) 1992 (6) 21 548 (20) 38 525 (17) Monitoring groups based on baseline (within 12 months before) and follow-up (within 2 months after) serum creatinine monitoring. Serum creatinine monitoring before and after ACEI/ARB initiation We identiﬁed 223 814 new users of ACEI/ARB. We com- pared these patients in four groups: 21 411 (10%) had no baseline or follow-up creatinine tests within 12 months before and 2 months after treatment initi- ation, 63 359 (28%) had only a baseline test, 33 185 (15%) had only follow-up tests, and 105 859 (47%) had both baseline and follow-up tests (table 1). Median age varied only slightly between the groups (60, 62, 59 and 63 years, respectively) and there were no substantial dif- ferences in socioeconomic status, lifestyle factors or per- ipheral arterial disease. Compared with patients with neither preinitiation nor postinitiation monitoring, patients with both were more likely to have diagnosed hypertension (76% vs 61%) and diabetes (20% vs 7%), but less likely to have diagnosed heart failure (4% vs 7%), myocardial infarction (4% vs 18%) and arrhythmia (7% vs 10%). Among patients with baseline monitoring, 83% did not have CKD, 13% stage 3a, 3% stage 3b, 0.5% stage 4 CKD. In the same population, 7% started er 23, 2024 by guest. Protected by copyright. est. Protected by copyright. In sensitivity analyses, we repeated the analyses (1) extending the follow-up window for the ﬁrst follow-up monitoring from 2 to 3 weeks to account for minor delays; (2) including only the most recent calendar period (2009–2014) to account for temporal changes in data completeness and quality of care; (3) excluding patients with a hospital admission or discharge date 3 Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 Open Access on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from †Calculated from most recent creatinine measurement within 12 months before the first prescription date. ‡Diagnosis ever registered before ACE/ARB initiation in CRPD or HES. ARB, angiotensin receptor blocker; BMI, body mass index; CKD, chronic kidney disease; CV, cardiovascular; eGFR, estimated glomerular filtration rate; HES, Hospital Episode Statistics; SES, socioeconomic status. guest. Protected by copyright. ACEI/ARB therapy despite baseline potassium above 5 mmol/L. The median number of days between base- line monitoring and ﬁrst prescription date was 40 days (IQR 12–125 days). Among all patients initiating ACEI/ARB therapy, the proportion of patients receiving creatinine testing before initiation was 76% within 12 months of treatment initi- ation, declining to 34% within 1 month before initiation Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 4 Open Access BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 20 on October http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from on October 23, 2024 by guest. Protecte http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from calendar period (eTables 1–2) and to patients with heart failure, myocardial infarction, hypertension, peripheral arterial disease, diabetes or no recent hospitalisation (eTable 3). Only patients with CKD received a slightly higher degree of monitoring (13%) within 2 weeks fol- lowing treatment initiation (eTable 3). The proportion with follow-up testing after treatment initiation was also unchanged when results were stratiﬁed by date of ACEI/ ARB initiation in 2-year intervals (eTable 4). Table 2 Prevalence of baseline and follow-up serum creatinine monitoring among patients initiating ACE inhibitors or angiotensin receptor blockers, 2004–2014 Serum creatinine, ≥1 test Total number n=223 814 (100%) Baseline testing ≤12 months before 169 218 (76%) ≤3 months before 115 348 (52%) ≤1 month before 75 476 (34%) Follow-up testing ≤2 weeks after 65 090 (29%) ≤1 month after 114 244 (51%) ≤2 months after 139 044 (62%) Table 2 Prevalence of baseline and follow-up serum creatinine monitoring among patients initiating ACE inhibitors or angiotensin receptor blockers, 2004–2014 on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Serum creatinine and potassium changes after ACEI/ARB initiation Among patients receiving the recommended renal func- tion monitoring, 567 (1.2%) experienced a creatinine increase ≥30% and 191 (0.4%) a potassium level >6 mmol/L at their ﬁrst follow-up test within 2 months of treatment initiation (1.4% experienced the increase in creatinine and/or potassium) (table 4). Among these patients, 80% continued treatment beyond 30 days fol- lowing the monitoring date (table 4). The sensitivity ana- lysis showed that 65% of patients with a creatinine increase ≥30% and 60% of those with a potassium level >6 mmol/L also continued treatment beyond 90 days after the monitoring date (eTable 5). The results remained consistent for longer baseline monitoring intervals (eTable 5). (table 2). The proportion with follow-up testing after treatment initiation was 29% within 2 weeks, increasing to 62% within 2 months. Among ACEI/ARB initiators who had a baseline test within 12 months, 21% also had a follow-up test within 2 weeks after starting treatment (table 3). However, among patients undergoing testing within 1 month prior to treatment initiation, only 9% had also the recommended follow-up test within 2 weeks of treatment start. When we extended the follow-up window to 3 weeks, this proportion increased to only 14% (table 3). Among patients continuing treatment, only 1% had follow-up measurements at 1, 3, 6 and 12 months after the ﬁrst retest, in compliance with the strictest rec- ommendation (eTable 1). These results were unchanged when the analysis was restricted to the most recent Patients at high risk for creatinine increases ≥30% Patients at high risk for creatinine increases ≥30% When we examined patient characteristics associated with a creatinine increase ≥30% and adjusted for the Table 3 Prevalence of baseline and follow-up serum creatinine monitoring among patients initiating ACE inhibitors or angiotensin receptor blockers according to clinical guideline recommendations Clinical guidelines All initiators n=223 814 (100%) NICE heart failure NICE MI NICE/UKRA hypertension NICE CKD GP Notebook Wide baseline interval (≤12-months) Ideal baseline interval (≤1 month) Baseline testing x x x x x 169 218 (76%) 75 476 (34%) +Follow-up test ≤2 weeks x NA x x x 46 486 (21%) 19 679 (9%) +Follow-up test ≤3 weeks† 70 792 (32%) 30 451 (14%) Follow-up test among those with baseline measurements. †Sensitivity analysis illustrating the importance of 2-week vs 3-week cut-off interval in follow-up test intervals. CKD, chronic kidney disease; GP, general practice; MI, myocardial infarction; NA, not applicable; NICE, National Institute for Health and Care Excellence; UKRA, United Kingdom Renal Association. Table 3 Prevalence of baseline and follow-up serum creatinine monitoring among patients initiatin angiotensin receptor blockers according to clinical guideline recommendations Table 4 Proportion of new users of ACE inhibitors or angiotensin receptor blockers who continue or discontinue treatment according to guideline recommended cut-off levels of serum creatinine and potassium at follow-up testing Continuation† Discontinuation† Total Total number, % 42 942 (93.1) 3178 (6.9) 46 120 (100) Serum creatinine increase ≥30%, n (%) 462 (81.5) 105 (18.5) 567 (100) Serum potassium >6 mmol/L, n (%) 150 (78.5) 41 (21.5) 191 (100) Calculated from the most recent measurements within 1 month before and 2 months after drug initiation. †A patient was considered a continuous user when the end date of the first continuous course of therapy was larger than the date of the first follow-up monitoring +30 days (to allow for stockpiling and irregular use). Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 Schmidt M, et al. BMJ Open 2017;7:e012818. ) ) ) ) ) ) ) ) measur ts (n= diabete iphera http m Patients at high risk for creatinine increases ≥30% doi:10.1136/bmjopen-2016-012818 5 ation of m increa d ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) ths after shed as s after hed a m creatinine increase ≥30% and follow-up monitoring within 2 weeks following initiation o weeks Serum creatinine increase ≥30% Serum potassium incr sted† Age-adjusted and sex-adjusted Fully adjusted† Age-adjusted and sex-adjusted to 1.09) 1.39 (1.26 to 1.53) 1.63 (1.47 to 1.80) 0.87 (0.66 to 1.16) ence) 1.00 (reference) 1.00 (reference) 1.00 (reference) to 1.02) 0.88 (0.74 to 1.05) 0.86 (0.72 to 1.03) 1.29 (0.79 to 2.11) to 1.09) 1.03 (0.88 to 1.21) 1.00 (0.85 to 1.19) 1.35 (0.84 to 2.17) to 1.23) 1.49 (1.27 to 1.74) 1.36 (1.15 to 1.61) 1.65 (1.02 to 2.66) to 1.23) 2.72 (2.32 to 3.20) 2.02 (1.68 to 2.44) 2.75 (1.67 to 4.53) ence) 1.00 (reference) 1.00 (reference) 1.00 (reference) to 1.04) 0.62 (0.53 to 0.73) 0.60 (0.51 to 0.70) 2.48 (1.66 to 3.71) to 1.08) 1.01 (0.82 to 1.24) 0.88 (0.71 to 1.09) 7.51 (4.75 to 11.9) to 1.66) 2.16 (1.52 to 3.05) 1.72 (1.18 to 2.51) 24.0 (13.5 to 42.6) to 1.23) 4.00 (3.49 to 4.58) 2.93 (2.51 to 3.42) 2.90 (1.90 to 4.42) to 0.82) 2.33 (1.98 to 2.74) 1.57 (1.32 to 1.87) 2.12 (1.33 to 3.39) to 1.11) 0.62 (0.56 to 0.68) 1.58 (1.36 to 1.84) 0.60 (0.45 to 0.80) to 1.20) 2.10 (1.70 to 2.60) 1.87 (1.50 to 2.33) 2.14 (1.18 to 3.86) to 1.01) 2.37 (2.07 to 2.71) 0.77 (0.69 to 0.86) 1.41 (0.90 to 2.21) to 0.96) 1.09 (0.97 to 1.22) 1.04 (0.92 to 1.18) 0.97 (0.69 to 1.36) to 1.09) 1.04 (0.86 to 1.25) 0.97 (0.80 to 1.17) 8.22 (6.14 to 11.0) seline measurements within 12 months before and first follow-up measurement within 2 months afte urements (n=105 859). ythmia, diabetes and calendar period of prescription start. AD, peripheral arterial disease. http://bmjo ed as 10.1136/bmjopen-2016-012818 on 9 January 2017. Patients at high risk for creatinine increases ≥30% Downloaded from % CIs) creatin usted usted 04 to 1 erence 94 to 1 02 to 1 4 to 1 4 to 1 erence 96 to 1 93 to 1 21 to 1 09 to 1 75 to 0 97 to 1 01 to 1 03 to 1 90 to 0 00 to 1 ce betw th base e, MI, h um; MI ober 2 09 to 1 75 to 0 97 to 1 01 to 1 03 to 1 90 to 0 00 to 1 ce betw th base e, MI, h um; MI ober 2 Open Access other characteristics in a multivariable analysis (table 5), we found an increased OR for women (1.6-fold increased), for age above 70 years (at least 1.3-fold increased), for CKD stage 4 (1.6-fold increased), heart failure (2.9-fold increased), peripheral arterial disease (1.9-fold increased), myocardial infarction (1.6-fold increased) and hypertension (1.6-fold increased). with prior myocardial infarction, hypertension or a high baseline potassium level were at higher risk of sudden decline in kidney function after ACEI/ARB initiation, there was no evidence that these patient groups were monitored more frequently while initiating the drugs. Patients at high risk for potassium >6 mmol/L on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ d as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Baseline potassium level and CKD stage, but not age and sex, were associated with potassium levels >6 mmol/ L after ACEI/ARB initiation. Thus, the OR was seven- fold increased for baseline potassium >5 mmol/L, twofold increased for CKD stage 3a, ﬁvefold increased for stage 3b, and 11-fold increased for stage 4 (table 5). Among cardiovascular comorbidities, heart failure was associated with the strongest OR of a potassium level >6 mmol/L (2.22, 95% CI 1.38 to 3.58). g p p g p Over the time course of this study, multiple factors have impacted on the prescribing of ACEI/ARB and measurement of renal function in primary care, for example, the introduction of the relevant NICE guide- lines, and QOF reimbursement for testing in certain subgroups. We also did not have information about clini- cal initiatives such as heart failure nurses and ACEI/ ARB stopping rules (‘sick-day rules’). While our main results provide summary measures over a 10-year period, sensitivity analyses conﬁrm that despite these changes, the proportion receiving the guideline suggested that biochemical monitoring does not vary during the study period. We did not have access to blood tests performed in hospital systems, which may have been reported to GPs, but not recorded in CPRD. However, restricting the analysis to patients with no recent hospital admissions who were most likely to have had renal function mea- sured and acted on in secondary care had little effect on our ﬁndings. We did not examine testing during initi- ation of dual blockade with ACEI and ARB as this com- bination is now used very infrequently for patients with severe comorbidities who are likely to be monitored in secondary care. Although some patients may also have been seen in outpatient specialty clinics, it is common practice for specialists to ask GPs to initiate new drugs such as ACEI/ARBs, with local biochemical monitoring, limiting misclassiﬁcation. Strengths and limitations Several issues should be considered when interpreting our study results. Its large sample size increased preci- sion. Use of the CPRD ensured that the study was general practice-based and not restricted to speciﬁc demographic, hospital or insurance groups. Monitoring high-risk patients Some characteristics associated with increased odds of having ≥30% rise in creatinine were also associated with a greater likelihood of having a follow-up test within 2 weeks following drug initiation. These included older age: persons aged 70 years or above compared with ≤50 years (1.18, 95% CI 1.13 to 1.23 for 70–79 years and 1.17, 95% CI 1.11 to 1.23 for 80+ years), CKD stage 4 compared with no CKD (1.41, 95% CI 1.20 to 1.66), heart failure (1.16, 95% CI 1.08 to 1.23) and peripheral arterial disease (1.11, 95% CI 1.02 to 1.20). However, other characteristics associated with increased odds of having ≥30% rise in creatinine were not associated with a greater likelihood of having a follow-up test within 2 weeks following drug initiation: there was no substan- tially increased OR (>10%) associated with female sex (1.07, 95% CI 1.04 to 1.09), prior history of myocardial infarction (0.77, 95% CI 0.72 to 0.82), hypertension (1.05, 95% CI 1.00 to 1.11) or baseline potassium >5 mmol/L (1.04, 95% CI 0.99 to 1.09). When we excluded patients with a recent hospital admission, the reduced OR for myocardial infarction was no longer observed (0.93, 95% CI 0.80 to 1.08) (eTable 6). Finally, the results remained consistent when we omitted adjust- ment for baseline potassium (data not shown) and when we adjusted additionally for ethnicity (eTable 6). Consistent with ﬁndings from other studies,19 we found that ∼50% of all ACEI/ARB initiators were moni- tored both before and after treatment start. If GPs are retesting renal function in patients at higher risk of sub- stantial biochemical changes, we may have overestimated the proportion of patients with high potassium levels or creatinine increases compared with the untested lower- risk general population. 23, 2024 by guest. Protected by copyright. DISCUSSION g p p GP system software is used for issuing prescriptions, ensuring the accuracy of prescription data. However, it cannot be inferred that all patients actually redeemed their prescription at the pharmacy and started medica- tion on the same day that it was prescribed.18 20 Similarly, the estimated coverage of prescriptions may not be completely accurate due to such factors as stock- piling and irregular use. We also do not know whether GPs contacted patients with elevated laboratory results to advise them to stop taking the medication prior to the end of their prescriptions. However, 80% of patients Only one-tenth of patients initiating ACEI/ARBs in UK primary care appear to receive the guideline- recommended creatinine monitoring. One in 15 patients started ACEI/ARBs despite baseline potassium above the recommended level, which was also shown to be a strong predictor for severe postinitiation hyperkalaemia. Among monitored patients, a creatinine increase ≥30% or a potassium level >6 mmol/L occurred in almost 1.5% of patients, and most did not discontinue therapy despite guideline recommendations to stop. Although patients est. Protected by copyright. 7 Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 Open Access who developed creatinine increase ≥30% after ACEI/ ARB initiation were still issued a subsequent ACEI/ARB prescription. 2005 and 2009, found that 19% had both a baseline (within 90 days) and follow-up measurement (within 2 weeks) of initiation. Within this cohort, 1.7% had both a creatinine increase of ≥30% and potassium level ≥5.6 mmol/L. on October 23, 2024 by guest. Protected by copyright. http://bmjopen.bmj.com/ n: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from We aimed to detect discontinuation related closely in time to the ﬁrst follow-up monitoring and hence most likely resulting from an elevated creatinine or potas- sium result. We therefore deﬁned continuation as ACEI/ARB use beyond 30 days (the median prescrip- tion duration) after the monitoring date. Extending the deﬁnition of continuous use beyond 90 days reduced the risk of misclassifying patients as continuing treatment when they had in fact stopped. However, extending the duration also increased the risk of identi- fying discontinuation due to other reasons than creatin- ine/potassium increase, for example, death or cough. Comparison with other studies To the best of our knowledge, this is the largest study conducted until now on adherence to monitoring and discontinuation guidelines after ACEI/ARB initiation. Only one previous study19 examined monitoring accord- ing to guideline-recommended intervals (<14 days). All others have used longer intervals (eg, 30 days26 or 6 months27 28), which make interpretations and implica- tions for clinical practice less clear. Poor adherence to monitoring guidelines after ACEI/ARB initiation is not restricted to the UK,19 28 29 but has also been reported in the USA,30–32 Canada33 and the Netherlands.26 34 Owing to our sample size, we were able to show that the lack of monitoring occurred in all patient groups with an indication for ACEI/ARB therapy. ation with diabetes or increasing age, as could have been expected due to diabetic nephropathy or age- dependent hyporeninaemic hypoaldosteronism.6 uest. Protected by copyright. A recent Dutch study, including 3353 patients initiat- ing ACEI/ARBs between 2005 and 2011, found that 19% had creatinine measured within 30 days and 66% within 1 year.26 Creatinine increases above 30% occurred in 1.6% of patients, and among these 70% did not dis- continue treatment.26 A Scottish study of 4056 patients with type 2 diabetes, prescribed an ACEI/ARB between DISCUSSION Determinants of increases in creatinine levels after ACEI/ARB initiation are less well understood than for hyperkalaemia, but increasing age is a consistently reported factor.19 Advanced CKD and a range of cardio- vascular comorbidities (mostly associated with athero- sclerosis) were also important determinants in our patient cohort. Consistent with previous studies, we found that the risk of hyperkalaemia was associated with CKD (most likely due to the impaired ability of the cor- tical collecting tubule to secrete potassium), heart failure (most likely due to the decreased delivery of sodium to the distal nephron), and high pretreatment potassium levels.6 8 19 37 We did not observe an associ- DISCUSSION Diagnoses recorded in the CPRD generally have been found to have adequate validity for research pur- poses,21 22 particularly in the domains assessed by the QOF.23 24 The magnitude of the risk of severe renal impairment, as measured by creatinine increase in these observa- tional studies, was consistent with our ﬁndings, but substantially higher than reported in clinical trials (eg, 0.2% in the Ongoing Telmisartan Alone and in Combination with Ramipril Global Endpoint Trial (ONTARGET)).35 It is not clear from the literature how often harm occurs around the time of initiation, when the risk of nephrotoxicity is thought to be greatest.8 If physicians are to understand why follow-up monitoring within 2 weeks of treatment start matters, the short-term risks need to be clariﬁed. Until now, most studies have reported only on cumulative risk over entire courses of treatment, such as the 1.1% 2-year risk for potassium of >6 mmol/L in the Studies of Left Ventricular Dysfunction (SOLVD) trials of patients with heart failure.36 In contrast to clinical trial reviews, reporting a 0.2% (3/1818) risk of potassium >6 mmol/L, we found a 0.4% risk of hyperkalaemia already at the time of ﬁrst retesting after ACEI initiation. In the logistic regression analysis to estimate factors associated with creatinine increase ≥30%, we excluded patients without pre and post measurements (complete case analysis). If the recording of creatinine levels was not missing completely at random, the associations between patient characteristics and creatinine increase may have been underestimated.25 While this assumption could not be tested directly, examination of baseline characteristics revealed no major differences in age, sex, socioeconomic status, and lifestyle between patients with and without premonitoring and postmonitoring. Furthermore, the results were consistent for each indi- vidual patient group examined. Patients with no testing before or after treatment initiation (including those with potentially haemolysed samples) only accounted for 10% of all ACEI/ARB initiators. g Extending the previous literature, our results support that advanced age, advanced CKD and heart failure, but not sex, increase the likelihood of being moni- tored.19 26 30 Consistent with some,26 30 but not all, previ- ous studies,28 we found no association for diabetes. However, these previous studies reporting an association for diabetes focused on monitoring within broader inter- vals (eg, 6 months),28 where patients with diabetes, irre- spective of ACEI/ARB initiation, were likely to receive blood testing owing to the diabetes QOF programme. Generalisability, implications and conclusions The majority of patients initiating treatment with ACEI/ ARBs experience only minor changes in renal function. However, substantial increases in creatinine levels after ACEI/ARB initiation may not be as rare as previously suggested, reinforcing the need for adherence to clinical guidelines for both pre-initiating and post-initiating monitoring. Moreover, the postinitiation creatinine increase and potassium levels used in this study are widely recognised cut-off levels, making the results inter- nationally applicable. The comparison with the previous literature also conﬁrms that the lack of systematic moni- toring is not exclusive to the UK. Of particular concern was that even when appropriate monitoring was per- formed, severe renal impairment only rarely led to treat- ment discontinuation. Individual patient counselling may also be helpful to ensure that those at highest risk are closely monitored. More work is needed to deter- mine the prognostic importance of the changes in renal function that we have observed. The majority of patients initiating treatment with ACEI/ ARBs experience only minor changes in renal function. However, substantial increases in creatinine levels after ACEI/ARB initiation may not be as rare as previously suggested, reinforcing the need for adherence to clinical guidelines for both pre-initiating and post-initiating monitoring. Moreover, the postinitiation creatinine increase and potassium levels used in this study are widely recognised cut-off levels, making the results inter- nationally applicable. The comparison with the previous literature also conﬁrms that the lack of systematic moni- toring is not exclusive to the UK. Of particular concern was that even when appropriate monitoring was per- formed, severe renal impairment only rarely led to treat- ment discontinuation. Individual patient counselling may also be helpful to ensure that those at highest risk are closely monitored. More work is needed to deter- mine the prognostic importance of the changes in renal function that we have observed. 2. National Institute for Health and Clinical Excellence (NICE). Chronic heart failure in adults: management. 2010. http://pathways.nice.org. uk/pathways/chronic-heart-failure (accessed 1 Apr 2016). 3. National Institute for Health and Clinical Excellence (NICE). Management of chronic kidney disease. 2014. http://pathways.nice. org.uk/pathways/chronic-kidney-disease#path=view%3A/pathways/ chronic-kidney-disease/management-of-chronic-kidney-disease. xml&content=view-node%3Anodes-blood-pressure-control-and- antihypertensive-treatment (accessed 1 Apr 2016. yp ( p 4. National Institute for Health and Clinical Excellence (NICE). Myocardial infarction: secondary prevention. 2013. http://pathways. nice.org.uk/pathways/myocardial-infarction-secondary-prevention (accessed 1 Apl 2016). 5. Lesogor A, Cohn JN, Latini R, et al. Twitter Follow Kathryn Mansfield @AnimaSophia Provenance and peer review Not commissioned; externally peer reviewed. Data sharing statement No additional data are available. Data sharing statement No additional data are available. Open Access This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http:// creativecommons.org/licenses/by/4.0/ Twitter Follow Kathryn Mansfield @AnimaSophia Twitter Follow Kathryn Mansfield @AnimaSophia recommendations for timing and frequency of monitor- ing over time,6 consensus-based (rather than evidence- based) monitoring guidelines, and a lack of guidelines tailored to particular high-risk patients, such as those with CKD and heart failure. Although we found that follow-up monitoring correlated well with the risk of renal impairment after ACEI/ARB initiation for most patient groups, it was not observed for patients with myocardial infarction or preinitiation high potassium. The second explanation may be patient non-adherence to ordered tests. This is particularly salient in UK primary care where blood samples may be taken in phlebotomy clinics that the patient has to visit rather than the GP practice. Patients may ﬁnd it burdensome to have blood tests, and GPs have no direct economic incentives to ensure that they are done. A third barrier is lack of evi- dence of the clinical importance of monitoring and its cost- effectiveness. ACEI/ARB-induced renal impairment is rare in clinical trials, even among patients with multiple risk factors for atherosclerotic renal artery stenosis.8 38 Trial results may therefore have led to a general percep- tion that the rarity of renal impairment obviates the need for close monitoring. However, as observed in our data, the risks in real-world practice may be somewhat higher and non-negligible. In addition, previous research has shown that potassium monitoring in high- risk patients with CKD and diabetes may reduce serious hyperkalaemia-associated adverse events.39 Still, the extent to which an initial creatinine increase ≥30% translates into adverse long-term outcomes in real-world patients remains to be clariﬁed in future studies. recommendations for timing and frequency of monitor- ing over time,6 consensus-based (rather than evidence- based) monitoring guidelines, and a lack of guidelines tailored to particular high-risk patients, such as those with CKD and heart failure. Although we found that follow-up monitoring correlated well with the risk of renal impairment after ACEI/ARB initiation for most patient groups, it was not observed for patients with myocardial infarction or preinitiation high potassium. The second explanation may be patient non-adherence to ordered tests. This is particularly salient in UK primary care where blood samples may be taken in phlebotomy clinics that the patient has to visit rather than the GP practice. Patients may ﬁnd it burdensome to have blood tests, and GPs have no direct economic incentives to ensure that they are done. REFERENCES 1. National Institute for Health and Clinical Excellence (NICE). Hypertension in adults: diagnosis and management. 2011. https:// www.nice.org.uk/guidance/cg127/chapter/1-Guidance#initiating-and- monitoring-antihypertensive-drug-treatment-including-blood- pressure-targets-2 (accessed 1 Apr 2016). Clinical relevance Several possible explanations exist for the divergence between the clinical guideline recommendations and the observed monitoring and response patterns in clinical practice. The ﬁrst is clinician non-adherence to ordering tests. This may be due to inconsistent 8 Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 Open Access Twitter Follow Kathryn Mansfield @AnimaSophia A third barrier is lack of evi- dence of the clinical importance of monitoring and its cost- effectiveness. ACEI/ARB-induced renal impairment is rare in clinical trials, even among patients with multiple risk factors for atherosclerotic renal artery stenosis.8 38 Trial results may therefore have led to a general percep- tion that the rarity of renal impairment obviates the need for close monitoring. However, as observed in our data, the risks in real-world practice may be somewhat higher and non-negligible. In addition, previous research has shown that potassium monitoring in high- risk patients with CKD and diabetes may reduce serious hyperkalaemia-associated adverse events.39 Still, the extent to which an initial creatinine increase ≥30% translates into adverse long-term outcomes in real-world patients remains to be clariﬁed in future studies. Contributors LAT conceived the study idea and acquired data permissions. MS, KEM and LAT designed the study. MS and KEM performed data management and established the cohort. MS, KEM and LAT reviewed the literature. The analyses were carried out by MS. All authors participated in the discussion and interpretation of the results. MS organised the writing and wrote the initial drafts. All authors critically revised the manuscript for intellectual content and approved the final version. MS is the guarantor. Funding MS was supported by the A.P. Møller Foundation for the Advancement of Medical Science, Snedkermester Sophus Jacobsen & Hustru Astrid Jacobsens Fond, and Christian og Ottilia Brorsons Rejselegat for yngre videnskabsmænd og–kvinder. HTS was supported by the Program for Clinical Research Infrastructure (PROCRIN) established by the Lundbeck Foundation and the Novo Nordisk Foundation. KB holds a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (grant number 107731/Z/15/Z). LAT and KEM are funded by a Wellcome Trust intermediate clinical fellowship to LAT (101143/Z/13/Z). http://bmjopen.bmj.co s 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Disclaimer None of these funding sources had a role in the design, conduct, analysis or reporting of the study. Competing interests None declared. Ethics approval London School of Hygiene and Tropical Medicine Ethics Committee (Approval number 6536) and the Independent Scientific Advisory Committee (ISAC) for Medicines and Healthcare Products Regulatory Agency (Approval number 16_025). Ethics approval London School of Hygiene and Tropical Medicine Ethics Committee (Approval number 6536) and the Independent Scientific Advisory Committee (ISAC) for Medicines and Healthcare Products Regulatory Agency (Approval number 16_025). Provenance and peer review Not commissioned; externally peer reviewed. on October 23, 2024 by guest. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from J Am Geriatr Soc 2005;53:2165–9. ( ) p ; 18. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604–12. g 19. Mathieson L, Severn A, Guthrie B. Monitoring and adverse events in relation to ACE inhibitor/angiotensin receptor blocker initiation in people with diabetes in general practice: a population database study. Scott Med J 2013;58:69–76. 33. McAlister FA, Tu K, Majumdar SR, et al. Laboratory testing in newly treated elderly hypertensive patients without co-morbidities: a population-based cohort study. Open Med 2007;1:e60–7. 34. Bootsma JE, Warlé-van Herwaarden MF, Verbeek AL, et al. Adherence to biochemical monitoring recommendations in patients starting with renin angiotensin system inhibitors: a retrospective cohort study in the Netherlands. Drug Saf 2011; 34:605–14. y 20. Tomlinson LA, Riding AM, Payne RA, et al. The accuracy of diagnostic coding for acute kidney injury in England—a single centre study. BMC Nephrol 2013;14:58. y p ; 21. Herrett E, Thomas SL, Schoonen WM, et al. Validation and validity of diagnoses in the General Practice Research Database: a systematic review. Br J Clin Pharmacol 2010;69:4–14. 35. Yusuf S, Teo KK, Pogue J, et al., ONTARGET Investigators. Telmisartan, ramipril, or both in patients at high risk for vascular events. N Engl J Med 2008;358:1547–59. on October 23, 2024 by guest. Protec http://bmjopen.bmj.com/ January 2017. Downloaded from y ; 22. Khan NF, Harrison SE, Rose PW. Validity of diagnostic coding within the General Practice Research Database: a systematic review. Br J Gen Pract 2010;60:e128–36. 36. de Denus S, Tardif JC, White M, et al. Quantification of the risk and predictors of hyperkalemia in patients with left ventricular dysfunction: a retrospective analysis of the Studies of Left Ventricular Dysfunction (SOLVD) trials. Am Heart J 2006;152:705–12. 23. Doran T, Kontopantelis E, Valderas JM, et al. Effect of financial incentives on incentivised and non-incentivised clinical activities: longitudinal analysis of data from the UK Quality and Outcomes Framework. BMJ 2011;342:d3590. 37. Desai AS, Swedberg K, McMurray JJV, et al. Incidence and predictors of hyperkalemia in patients with heart failure: an analysis of the CHARM Program. J Am Coll Cardiol 2007;50:1959–66. 24. Barbour SJ, Schachter M, Er L, et al. A systematic review of ethnic differences in the rate of renal progression in CKD patients. Nephrol Dial Transplant 2010;25:2422–30. 38. Pitt B, Segal R, Martinez FA, et al. on October 23, 2024 by guest. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from 11. Health & Social Care Information Centre. Hospital Episode Statistics. http://www.hscic.gov.uk/hes (accessed 1 May 2016). 27. McDowell SE, Coleman JJ, Evans SJW, et al. Laboratory monitoring and adverse patient outcomes with antihypertensive therapy in primary care. Pharmacoepidemiol Drug Saf 2010;19:482–9. p g ( y ) 12. UK Renal Association. Hypertension. http://www.renal.org/ information-resources/the-uk-eckd-guide/hypertension#sthash. QNqnBcyh.dpbs (accessed 1 Apr 2016). p y p g 28. Coleman JJ, McDowell SE, Evans SJW, et al. Oversight: a retrospective study of biochemical monitoring in patients beginning antihypertensive drug treatment in primary care. Br J Clin Pharmacol 2010;70:109–17. q y p ( p ) 13. GP Notebook. General Practice Notebook—a UK medical reference. http://www.gpnotebook.co.uk (accessed 1 Apr 2016). p gp ( p ) 14. McDonald HI. The epidemiology of infections among older people with diabetes mellitus and chronic kidney disease. London School of Hygiene and Tropical Medicine. 2015. 29. Kalra PA, Kumwenda M, MacDowall P, et al. Questionnaire study and audit of use of angiotensin converting enzyme inhibitor and monitoring in general practice: the need for guidelines to prevent renal failure. BMJ 1999;318:234–7. yg p 15. McDonald HI, Thomas SL, Millett ERC, et al. CKD and the risk of acute, community-acquired infections among older people with diabetes mellitus: a retrospective cohort study using electronic health records. Am J Kidney Dis 2015;66:60–8. 30. Raebel MA, McClure DL, Chan KA, et al. Laboratory evaluation of potassium and creatinine among ambulatory patients prescribed spironolactone: are we monitoring for hyperkalemia? Ann Pharmacother 2007;41:193–200. y 16. Bhaskaran K, Douglas I, Forbes H, et al. Body-mass index and risk of 22 specific cancers: a population-based cohort study of 5.24 million UK adults. Lancet 2014;384:755–65. 31. Hurley JS, Roberts M, Solberg LI, et al. Brief report: laboratory safety monitoring of chronic medications in ambulatory care settings. J Gen Intern Med 2005;20:331–3. ; 17. Bhaskaran K, Forbes HJ, Douglas I, et al. Representativeness and optimal use of body mass index (BMI) in the UK Clinical Practice Research Datalink (CPRD). BMJ Open 2013;3:e003389. 18. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate Ann Intern Med 2009;150:604 12 ; 17. Bhaskaran K, Forbes HJ, Douglas I, et al. Representativeness and optimal use of body mass index (BMI) in the UK Clinical Practice Research Datalink (CPRD). BMJ Open 2013;3:e003389. 32. Simon SR, Andrade SE, Ellis JL, et al. Baseline laboratory monitoring of cardiovascular medications in elderly health maintenance organization enrollees. Generalisability, implications and conclusions Interaction between baseline and early worsening of renal function and efficacy of renin-angiotensin- aldosterone system blockade in patients with heart failure: insights from the Val-HeFT study. Eur J Heart Fail 2014;15:1236–44. 5. Lesogor A, Cohn JN, Latini R, et al. Interaction between baseline and early worsening of renal function and efficacy of renin-angiotensin- aldosterone system blockade in patients with heart failure: insights from the Val-HeFT study. Eur J Heart Fail 2014;15:1236–44. y 6. Raebel MA. Hyperkalemia associated with use of angiotensin- converting enzyme inhibitors and angiotensin receptor blockers. Cardiovasc Ther 2012;30:e156–66. 6. Raebel MA. Hyperkalemia associated with use of angiotensin- converting enzyme inhibitors and angiotensin receptor blockers. Cardiovasc Ther 2012;30:e156–66. 7. McDowell SE, Thomas SK, Coleman JJ, et al. A practical guide to monitoring for adverse drug reactions during antihypertensive drug therapy. J R Soc Med 2013;106:87–95. 8. Bakris GL, Weir MR. Angiotensin-converting enzyme inhibitor- associated elevations in serum creatinine: is this a cause for concern? Arch Intern Med 2000;160:685–93. Author affiliations 1Department of Non-Communicable Disease Epidemiology, London School of Hygiene and Tropical Medicine, London, UK 2Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark Author affiliations 1Department of Non-Communicable Disease Epidemiology, London School of Hygiene and Tropical Medicine, London, UK 2Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark ; 9. McDowell SE, Ferner RE. Biochemical monitoring of patients treated with antihypertensive therapy for adverse drug reactions. Drug Saf 2011;34:1049–59. ; 10. Herrett E, Gallagher AM, Bhaskaran K, et al. Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015;44:827–36. ; 10. Herrett E, Gallagher AM, Bhaskaran K, et al. Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015;44:827–36. 3Department of Internal Medicine, Regional Hospital of Randers, Denmark Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 9 Open Access Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 on October 23, 2024 by guest. http://bmjopen.bmj.com/ BMJ Open: first published as 10.1136/bmjopen-2016-012818 on 9 January 2017. Downloaded from Randomised trial of losartan versus captopril in patients over 65 with heart failure (Evaluation of Losartan in the Elderly Study, ELITE). Lancet 1997;349: 747–52. p 25. White IR, Carlin JB. Bias and efficiency of multiple imputation compared with complete-case analysis for missing covariate values. Stat Med 2010;29:2920–31. 26. van Blijderveen JC, Straus SM, de Ridder MA, et al. Adherence to renal function monitoring guidelines in patients starting antihypertensive therapy with diuretics and RAAS inhibitors: a retrospective cohort study. Drug Saf 2014;37:369–77. 39. Raebel MA, Ross C, Xu S, et al. Diabetes and drug-associated hyperkalemia: effect of potassium monitoring. J Gen Intern Med 2010;25:326–33. 10 Schmidt M, et al. BMJ Open 2017;7:e012818. doi:10.1136/bmjopen-2016-012818 Open Access Miscellaneous Correction: Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin– angiotensin system blockade: a UK general practice-based cohort study Schmidt M, Mansfield KE, Bhaskaran K, et al. Adherence to guidelines for creatinine and potassium monitoring and discontinuation following renin–angiotensin system blockade: a UK general practice-based cohort study. BMJ Open 2017;7:e012818. doi: 10.1136/ bmjopen-2016-012818 In the 3rd column of Table 5 the heading ‘Serum potassium increase ≥30%*’ should read ‘Serum potassium increase >6 mmol/L’. Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/ © Article author(s) (or their employer(s) unless otherwise stated in the text of the article) 2017. All rights reserved. No commercial use is permitted unless otherwise expressly granted. BMJ Open 2017;7:e012818corr1. doi:10.1136/bmjopen-2016-012818corr1 BMJ Open 2017;7:e012818corr1. doi:10.1136/bmjopen-2016-012818corr1 1
https://openalex.org/W3132711208	https://air.unimi.it/bitstream/2434/972108/2/2.%20Da%20Silva%20Linge%20et%20al.%202021.pdf	English	null	Multi-Locus Genome-Wide Association Studies Reveal Fruit Quality Hotspots in Peach Genome	Frontiers in plant science	2,021	cc-by	16,647	ORIGINAL RESEARCH published: 25 February 2021 doi: 10.3389/fpls.2021.644799 ORIGINAL RESEARCH published: 25 February 2021 doi: 10.3389/fpls.2021.644799 ORIGINAL RESEARCH published: 25 February 2021 doi: 10.3389/fpls.2021.644799 Multi-Locus Genome-Wide Association Studies Reveal Fruit Quality Hotspots in Peach Genome Cassia da Silva Linge1, Lichun Cai1,2, Wanfang Fu1, John Clark3, Margaret Worthington3, Zena Rawandoozi4, David H. Byrne4 and Ksenija Gasic1* 1 Department of Plant and Environmental Sciences, Clemson University, Clemson, SC, United States, 2 Department of Horticulture, Michigan State University, East Lansing, MI, United States, 3 Department of Horticulture, University of Arkansas, Fayetteville, AR, United States, 4 Department of Horticultural Sciences, Texas A&M University, College Station, TX, United States Peach is one of the most important fruit crops in the world, with the global annual production about 24.6 million tons. The United States is the fourth-largest producer after China, Spain, and Italy. Peach consumption has decreased over the last decade, most likely due to inconsistent quality of the fruit on the market. Thus, marker-assisted selection for fruit quality traits is highly desired in fresh market peach breeding programs and one of the major goals of the RosBREED project. The ability to use DNA information to select for desirable traits would enable peach breeders to efﬁciently plan crosses and select seedlings with desired quality traits early in the selection process before fruiting. Therefore, we assembled a multi-locus genome wide association study (GWAS) of 620 individuals from three public fresh market peach breeding programs (Arkansas, Texas, and South Carolina). The material was genotyped using 9K SNP array and the traits were phenotyped for three phenological (bloom date, ripening date, and days after bloom) and 11 fruit quality-related traits (blush, fruit diameter, fruit weight, adherence, fruit ﬁrmness, redness around pit, fruit texture, pit weight, soluble solid concentration, titratable acidity, and pH) over three seasons (2010, 2011, and 2012). Multi-locus association analyses, carried out using mrMLM 4.0 and FarmCPU R packages, revealed a total of 967 and 180 quantitative trait nucleotides (QTNs), respectively. Among the 88 consistently reliable QTNs detected using multiple multi-locus GWAS methods and/or at least two seasons, 44 were detected for the ﬁrst time. Fruit quality hotspots were identiﬁed on chromosomes 1, 3, 4, 5, 6, and 8. Out of 566 candidate genes detected in the genomic regions harboring the QTN clusters, 435 were functionally annotated. Gene enrichment analyses revealed 68 different gene ontology (GO) terms associated with fruit quality traits. Data reported here advance our understanding of genetic mechanisms underlying important fruit quality traits and further support the development of DNA tools for breeding. Keywords: FarmCPU, mrMLM 4.0, candidate gene analyses, SNP array, RosBREED, QTN Correspondence: Correspondence: Ksenija Gasic kgasic@clemson.edu *Correspondence: Ksenija Gasic kgasic@clemson.edu Specialty section: This article was submitted to Plant Development and EvoDevo, a section of the journal Frontiers in Plant Science Received: 21 December 2020 Accepted: 04 February 2021 Published: 25 February 2021 Edited by: Jia-Long Yao, The New Zealand Institute for Plant and Food Research Ltd, New Zealand Reviewed by: Satish Kumar, The New Zealand Institute for Plant and Food Research Ltd, New Zealand Ana M. Casas, Aula Dei Experimental Station (EEAD), Spain Reviewed by: Satish Kumar, The New Zealand Institute for Plant and Food Research Ltd, New Zealand Ana M. Casas, Aula Dei Experimental Station (EEAD), Spain Aula Dei Experimental Station (EEAD), Spain INTRODUCTION The mrMLM 4.0 R package comprises the mrMLM (Wang et al., 2016), FASTmrMLM (Tamba and Zhang, 2018), FASTmrEMMA (Wen et al., 2017), ISIS EM-BLASSO (Tamba et al., 2017), pLARmEB (Zhang et al., 2017), and pKWmEB (Ren et al., 2018) two-step multi-locus GWAS methods. First, various algorithms are used to select all potentially associated markers. Second, these selected markers are put in one model, in which all the eﬀects are obtained by empirical Bayes, and all the non-zero eﬀects are further identiﬁed by likelihood ratio test for true Quantitative Trait Nucleotides (QTNs) (Zhang et al., 2020). Recent advances in next-generation high-throughput sequencing and genotyping techniques, such as development of the 9K peach SNP array by the International Peach SNP Consortium (IPSC) (Verde et al., 2012), allow use of DNA information to develop tools for facilitating breeding eﬀorts (Lambert et al., 2016; da Silva Linge et al., 2018). Understanding the genetic mechanisms that control a speciﬁc trait would enable peach breeders to eﬃciently apply marker-assisted breeding (MAB) through the development of DNA diagnostic tools, and consequently select seedlings with desired quality traits early in the selection process before the characters can be evaluated in the ﬁeld (Abdelghafar et al., 2020). The multi-locus model Fixed and random model Circulating Probability Uniﬁcation (FarmCPU) uses the associated markers as covariates in a ﬁxed-eﬀect model (FEM) and optimization on the associated covariate markers in a random eﬀect model (REM). FarmCPU adopts REML optimization to replace the criterion that the variance explained by kinship is near zero, which can only be arbitrarily determined. FarmCPU also adopted a binning approach from super to select pseudo QTNs. The whole genome is equally divided into bins, and only one signiﬁcant marker with the smallest P-value from each bin is selected as the candidate pseudo QTN. These candidate pseudo QTNs are determined by a REM. The candidate pseudo QTNs are ﬁrst ranked by P-value. Then, the best combinations between the diﬀerent bins and the number of candidate pseudo QTNs are determined by REM. Finally, the two types of models (FEM and REM) are performed iteratively until no change occurs in the selection of pseudo QTNs (Huang et al., 2018). Thus, FarmCPU decreases the computer time required, provides reliable results by eﬃciently removing the confounding between the population structure and Kinship, avoiding model over-ﬁtting, and controlling for false positives (Liu et al., 2016). INTRODUCTION GWAS methods utilizing mixed linear models (MLM), which take into account multiple levels of relatedness, have become standard methodology (Yu et al., 2006). Signiﬁcant marker-trait association based on the single-locus models, such as the general linear model (GLM) and MLM, were reported for several traits such as fruit pubescence, fruit shape, stone adhesion-ﬂesh texture, fruit ﬂesh color, non-melting/melting ﬂesh, fruit weight, titratable acidity, soluble solid concentration, leaf gland type, ﬂower type, bloom date, fruit development period, maturity date, ripening index, and total sugars (Micheletti et al., 2015; Cao et al., 2016, 2019; Elsadr et al., 2019; Font I Forcada et al., 2019). Single- locus models test one locus at a time and fail to match the true genetic model of complex traits that are controlled by numerous loci simultaneously (Xu et al., 2018). Thus, major improvements in GWAS statistical methodology have occurred, and multi- locus GWAS methods considering the information of all loci simultaneously have been developed (Wang et al., 2016). Peach [Prunus persica (L.) Batsch] is a diploid species, with a short juvenile period (2–4 years), relatively simple genome (∼230 Mbp), and one of the best genetically characterized deciduous trees (Verde et al., 2013). Peach is the third most cultivated temperate tree fruit in the world, after apple and pear, with a world production of approximately 24.6 million tons (Food and Agricultural Organization of the United Nations (FAOSTAT), 2018). Despite the high production, peach consumption has declined over the past decades. In the United States, peach per capita consumption decreased to 1.3 kg per year compared to ∼3 kg per year in the 1980s (Minas et al., 2018). Inconsistent and low fruit quality is recognized as the major limiting factor for consumer acceptance and, consequently, the low rates of peach consumption (Cirilli et al., 2016). Peach breeders have always selected for fruit quality with respect to size, color and ﬁrmness, as well as tried to expand harvest season (Laurens et al., 2018). Recently, more emphasis is on other traits such as internal quality and postharvest traits (Elsadr et al., 2019). Recently, six multi-locus GWAS approaches were integrated into an R package, named mrMLM (Zhang et al., 2020). Citation: da Silva Linge C, Cai L, Fu W, Clark J, Worthington M, Rawandoozi Z, Byrne DH and Gasic K (2021) Multi-Locus Genome-Wide Association Studies Reveal Fruit Quality Hotspots in Peach Genome. Front. Plant Sci. 12:644799. doi: 10.3389/fpls.2021.644799 Keywords: FarmCPU, mrMLM 4.0, candidate gene analyses, SNP array, RosBREED, QTN February 2021 \| Volume 12 \| Article 644799 1 Frontiers in Plant Science \| www.frontiersin.org Fruit Quality Hotspots in Peach Genome da Silva Linge et al. Frontiers in Plant Science \| www.frontiersin.org Phenotypic Data where: object represents a model ﬁtted with the mmer function; data represents the dataset used to ﬁt the model provided in the object argument; gTerm is a character vector specifying the genetic terms ﬁtted in the model; and eTerm is a character vector specifying the environment term ﬁtted in the model. For the level from the eTerm (environment) the heritability is calculated as: Phenotypic data were recorded over three seasons (2010–2012) at each fresh market peach program. Bloom data (BD; Julian days) were visually assessed in the ﬁeld and recorded for each tree when 60–80% of the blossoms were open. Ripening date (RD; Julian days) was determined when 20% of fruits were at commercial harvest by visually inspecting the presence of a few soft fruits in the ﬁeld for maturity two times per week. Days after bloom (DAB; Julian days) was calculated as the number of days between the date of full bloom and ripening date. 1 −(PEV/(md ∗Vg)) "PEV" is the predicted error variance for the genotype, "md" is the mean value from the diagonal of the relationship (genomic) matrix "G" and where "Vg" refers to the genotype variance. The model included in the h2.fun was: Approximately 20 fruits were harvested for phenotyping. A ﬁve ﬁrm fruit sample was selected for the following traits evaluations: Blush (0–5 scale, 0 = none, and 3 = 40–60%, 5 > 90% red blush on fruit surface) subjective scales were used as described by Frett et al. (2014). Fruit diameter (FDIA; mm) was evaluated with a millimeter caliper, while fruit weight (FW; g) was measured as the average weight of the ﬁve selected peaches. Flesh adherence (ADH) was evaluated using 1 - 4 scale where 1 = Freestone; 2 = Semi-freestone; 3 = Semi-clingstone; and 4 = Clingstone. Fruit ﬁrmness (FF; N) was measured using an electronic fruit texture analyzer (FTA) ﬁtted with an 8-mm diameter tip (GÜSS Fruit Texture Analyzer; GÜSS Manufacturing (Pty) Ltd., Strand, South Africa). All readings were recorded as kilogram-force (kgf) and then converted to Newton (N) by multiplying the reading by 9.807. Redness around Pit (RP) was measured following the scale mix < −mmer(Trait ∼Year, random = ∼vs ds (Year) , Selection, Gu = K + vs ds Local , rcov = ∼vs ds (Year) , units , data = Trait) mix < −mmer(Trait ∼Year, where "K" refers to the genomic relationship matrix. h2.fun(object, data, gTerm, eTerm) h2.fun(object, data, gTerm, eTerm) h2.fun(object, data, gTerm, eTerm) INTRODUCTION The link between the genetic markers and a particular trait could be determined using diﬀerent approaches. Quantitative trait loci analysis (QTL mapping) and genome- wide association studies (GWAS) are widely used for dissection of complex genetic traits (Meneses and Orellana, 2013). In peach, several linkage maps have been used in QTL discovery of key fruit quality traits such as fruit size, diameter, ﬁrmness, acidity, soluble solid concentration, individual sugars, maturity date, pubescence, blush, fruit texture, and phytochemical compounds (Eduardo et al., 2011; Martínez-García et al., 2013; Pirona et al., 2013; Frett et al., 2014; Vendramin et al., 2014; da Silva Linge et al., 2015; Zeballos et al., 2016; Ciacciulli et al., 2018; Nuñez- Lillo et al., 2019; Abdelghafar et al., 2020). These maps were typically developed for mapping particular traits in a speciﬁc parental background with limited recombination events and genetic diversity. Alternatively, GWAS has the advantage of increasing the recombination events and consequently mapping resolution with a signiﬁcant reduction of the research time (Zhu et al., 2008). However, false positives due to population structure or kinship among genotypes, or false negatives due to removal of rare alleles that are involved in natural variation are some of the weaknesses of GWAS (Brachi et al., 2011). To deal with this problem, The objective of this study was to identify signiﬁcant marker- trait association for 14 agronomic traits, using the multi-locus GWAS methods in mrMLM 4.0 and FarmCPU in a U.S. peach diversity germplasm panel of 620 individuals, managed by three public fresh market peach breeding programs at University of Arkansas System Division of Agriculture, Texas A&M University and Clemson University. February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 2 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. Plant Material, DNA Isolation, Quantiﬁcation, and Genotyping For biochemical traits, a composite sample of one approximately 2 cm wide longitudinal slice from each of the ﬁve fruits was used to extract juice with a juicer for the measurement of soluble solid concentration (SSC) using a digital refractometer, pH with a pH meter and titratable acidity (TA) using an automatic titrator (DL 22 Food and Beverage analyzer, Mettler Toledo, Columbus, OH, United States). TA was obtained by the titration of solution of 6 g of the peach juice diluted with 50 mL of distilled water to pH 8.2 with 0.1N NaOH and expressed as milliequivalents of malic acid. The following equation was used to calculate titratable acidity (the milliequivalent factor used corresponded to malic acid, 0.067): The material used in this study represents the U.S. peach breeding germplasm assembled under the RosBREED project (Iezzoni et al., 2010, 2020; Peace et al., 2014). A total of 72 cultivars/advanced selections and 548 individuals from three public fresh market peach breeding programs: University of Arkansas System Division of Agriculture (AR), Clemson University (SC), and Texas A&M University (TX), were chosen to eﬀectively represent alleles currently found within North American fresh market peach breeding germplasm (Supplementary Table 1). Peach DNA was extracted from young leaves using the E-Z 96 Tissue DNA Kit (Omega Bio-Tek, Inc., Norcross, GA, United States). DNA was quantitated with the QuantiT PicoGreen Assay (Invitrogen, Carlsbad, CA, United States), using the Victor multi-plate reader (Perkin Elmer Inc., San Jose, CA, United States). The ﬁnal DNA concentrations were adjusted to a minimum of 50 ng/µL and submitted to the Research Technology Support Facility at Michigan State University (East Lansing, MI, United States). Titratable acidity (%) = Titratable acidity (%) = MATERIALS AND METHODS 1 = red; 0 = no red. The fruit texture (FT) was evaluated using the scale 1 = melting; 2 = non-melting. Pit weight (PW; g) was measured as the average weight of the ﬁve selected pits. Descriptive Analysis, Genetic Diversity, and Population Structure The descriptive analysis and the correlations between the traits were performed using the software Past (Hammer et al., 2001). The genetic diversity analysis was performed using the GenAlEx software (Peakall and Smouse, 2012). The narrow sense heritability was calculated using the R package Sommer (Covarrubias-Pazaran, 2016) using the h2.fun: Samples were genotyped with the IPSC peach 9K SNP array v1 (Verde et al., 2012). The SNP data curation was performed using the workﬂow for high-resolution genetic marker data described in Vanderzande et al. (2019). After the SNP data curation, a total of 4005 SNPs distributed over the eight peach chromosomes remained and were used in the multi-locus GWAS (Supplementary Table 2). Genome-Wide Association Study Genome-Wide Association Study To validate and increase the accuracy of the multi-locus GWAS results, we used mrMLM 4.0 (Zhang et al., 2020) and FarmCPU (Liu et al., 2016). The six multi-locus GWAS methods (mrMLM, FASTmrMLM, FASTmrEMMA, pLARmEB, pKWmEB, and ISIS EM-BLASSO) from mrMLM 4.0 R package were used. The SNP data were converted to character, as described in the user manual, the population structure was the Q matrix obtained from fastSTRUCTURE and the kinship matrix was calculated by mrMLM 4.0. All parameters in GWAS were set at default values. The signiﬁcantly associated SNPs were determined by the critical threshold of LOD score ≥3 as described in previous studies (Tamba et al., 2017). Concerning FarmCPU, the SNP data were converted to numerical using the R package GAPIT (Lipka et al., 2012). Principle component analysis (PCA) was conducted using TASSEL 5.0, and the ﬁrst three components were incorporated as covariates in the GWAS model. Bonferroni-corrected P-value threshold was set at p < 0.01. Highly signiﬁcant (P < 0.01) correlations were observed between the traits (Supplementary Table 3). The highest correlation was observed between the FDIA_2011 and FW_2011 (0.92). As expected, a signiﬁcant negative correlation was detected between the pH and TA (−0.65 and −0.64 in 2011 and 2012, respectively). The traits DAB and RD revealed a signiﬁcant positive correlation in the years analyzed (0.82 and 0.87). Concerning the correlation between years, BD showed the highest correlation (0.99), followed by RD (0.92). We considered a QTN reliable when: QTNs repeatedly detected in at least four methods and/or two seasons using the mrMLM 4.0; QTN consistently detected in two seasons using FarmCPU; QTNs detected in at least three methods in mrMLM 4.0 and also identiﬁed in the FarmCPU approach. These QTNs were named as “qtn” + trait name abbreviation + scaﬀold + detected QTL order on chromosome. g y The narrow sense heritability (h2) was estimated for all 14 traits (Supplementary Table 4). High average values of h2 (> 0.6) were observed for TA (0.87), RD (0.83), BD (0.77), ADH (0.77), FT (0.76), DAB (0.74), FDIA (0.73), SSC (0.72), Blush (0.71), FW (0.70), PW (0.70), RP (0.69), pH (0.69), and FF (0.68). Phenotypic Data Population structure, multidimensional scaling (MDS) and Bayesian clustering were performed with fastSTRUCTURE (Raj et al., 2014). The MDS was performed using TASSEL (Bradbury et al., 2007). The MDS results were plotted with the R package “scatterplot3D” (Ligges and Mächler, 2002). The fastSTRUCTURE was run with a “simple prior” option February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 3 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. and remaining default parameters. The number of populations (K), ranging from 1 to 20, and the most probable number of populations was chosen for running the built-in script for multiple choices of K. The admixture proportions of each genotype, estimated by fastSTRUCTURE, were visualized using DISTRUCT plots (Rosenberg, 2004). Accessions were assigned to a speciﬁc subpopulation when the estimated membership coeﬃcients (Q) were above 0.80. 2011). Second, we compared the position of the already reported candidate genes responsible for regulating BD, RP (Cs locus), Blush, RD, pH, and TA (D locus) with the QTL hotspot regions detected in this study. 1www.rosaceae.org Phenotypic Data Linkage disequilibrium (LD) was measured by correlation coeﬃcients (r2) for all pairs of SNPs. The LD decay were calculated using PopLDdecay (Zhang et al., 2018a) with the following parameters: -MaxDist 3000 kb -MAF 0.05. Linkage disequilibrium (LD) was measured by correlation coeﬃcients (r2) for all pairs of SNPs. The LD decay were calculated using PopLDdecay (Zhang et al., 2018a) with the following parameters: -MaxDist 3000 kb -MAF 0.05. Six hundred twenty individuals from the three fresh market public peach breeding programs were evaluated for 14 diﬀerent fruit quality traits over 3 years (2010–2012) (Table 1). The BD, DAB, FDIA, and TA traits were analyzed in two seasons (2011 and 2012), with the BD, DAB, and TA exhibiting the highest mean values in 2011 and FDIA in 2012. The RD, FW, FF, PW, SSC, and pH, as well as the categorical traits Blush, ADH, RP, and FT were evaluated in three seasons (2010–2012). The RD, FW, and pH varied from 111 to 237 Julian days, 30.2 – 351.4 g and 2.8 – 5.1 respectively, with the highest mean values observed in 2010 (197.6 Julian days, 119.4 g and 3.9, respectively). The FF ﬂuctuated from 0.9 N to 106.4N with the highest mean value measured in 2011 (21.1 N). The traits PW and SSC ranged from 2.5 to 12.2 and 7.2 – 26.8, respectively, with the highest mean values in 2012 (6.8 g and 13.2). Candidate Genes The candidate gene analysis was performed using two strategies. First, the candidate gene analyses were performed within the haploblock regions in which a QTN was detected with at least three methods in mrMLM 4.0 and with FarmCPU. Haploblock regions encompassing the associated SNPs were determined in PLINK 1.9 (Chang et al., 2015) using the ﬂag “blocks” restricted to 500 kb. From the Prunus persica Whole Genome v2.0 Assembly & Annotation v2.1 (Verde et al., 2017) in Genome Database for Rosaceae1 (Jung et al., 2018), a systematic search was conducted to compile the predicted candidate genes associated with the quality traits. The candidate genes were further analyzed for GO (gene ontology) enrichment using GOseq 1.42.0 R package (Young et al., 2010). The GO terms were considered signiﬁcantly enriched or depleted at FDR < 0.05. The enriched GO terms were visualized using REVIGO semantic similarities (Supek et al., The observed mean heterozygosity (Ho) per individual was 0.35, ranging from 0.06 in “St John” to 0.68 in “Elberta.” The mean Ho per SNP was 0.36 ranging from 0.07 in SNP_IGA_598267 (scaﬀold 5) to 0.97 in SNP_IGA_573558 (chromosome 5). The chromosome 6 was the least heterozygous (Ho = 0.281) while the chromosome 1 revealed the highest heterozygosis (Ho = 0.385). The mean expected heterozygosity (He) was 0.370, ranging from 0.08 (SNP_IGA_624226; scaﬀold 6) to 0.500. The mean average inbreeding coeﬃcient (F = (Ho-He) /Ho) was 0.05, ranging from −0.949 to 0.595. The population structure was analyzed with Multidimensional scaling (MDS) and fastSTRUCTURE. The MDS revealed two main groups, in which the second group could be divided in two clusters. The ﬁrst group comprised the individuals from the TX breeding program, while the second group grouped the individuals from the AR and SC breeding programs February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 4 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. TABLE 1 \| Descriptive analysis of 14 phenotypic traits observed in 620 individuals from three U.S. public fresh market peach breeding programs (Univ. of Arkansas, Texas A&M and Clemson Univ.) over three seasons (2010–2012). s of 14 phenotypic traits observed in 620 individuals from three U.S. public fresh market peach breeding programs (Univ. of Arkansa ) over three seasons (2010–2012). TABLE 1 \| Descriptive analysis of 14 phenotypic traits observed in 620 individuals from three U.S. public fresh market peach breeding programs (Univ. Candidate Genes of Arkansas, Texas A&M and Clemson Univ.) over three seasons (2010–2012). Trait Year Min Max Mean SE SD BD 2011/2012 37/36 80/78 65.1/59.31 1.14/0.51 16.19/10.93 RD 2010/2011/2012 158/111/125 230/237/230 197.6/189.1/171.3 1.2/1.1.02 14.9/23.5/22.2 DAB 2011/2012 67/78 196/159 120.8/111.3 1.52/0.82 19.98/17.24 Blush 2010/2011/2012 0/0/0 5/5/5 2.8/3.0/2.7 0.1/0.05/0.04 1.3/1.0.93 FDIA 2011/2012 36.3/42.0 85.2/80.6 55.2/60.8 0.42/0.37 7.47/6.66 FW 2010/2011/2012 49.0/33.4/30.2 214.7/351.4/289.8 119.4/103.1/117.1 2.9/1.9/1.9 34.9/40.7/42.2 ADH 2010/2011/2012 1/1/1 4/4/4 2.2/2.5454 0.1/0.07/0.06 1.4/1.4/1.4 FF 2010/2011/2012 1.4/2.1/0.9 54.6/67.9/106.4 16.2/21.1/17.6 0.9/0.6/0.7 10.5/12.0/14.8 RP 2010/2011/2012 0/0/0 1/1/1 0.8/0.8/0.6 0.04/0.02/0.02 0.4/0.4/0.5 FT 2010/2011/2012 1/1/1 2/2/2 1.1/1.3/1.3 0.02/0.03/0.03 0.3/0.7/0.7 PW 2010/2011/2012 2.6/2.5/2.8 10.7/12.2/11.6 6.0/5.9/6.8 0.1/0.1/0.1 1.8/1.9/1.6 SSC 2010/2011/2012 7.5/7.2/7.6 17.9/22.2/26.8 11.7/13.1/13.2 0.2/0.1/0.1 2.1/2.9/3.3 TA 2011/2012 0.2/0.01 1.7/1.8 0.7/0.6 0.01/0.01 0.3/0.3 pH 2010/2011/2012 3.4/2.9/2.8 5.0/4.8/5.1 3.9/3.8/3.8 0.02/0.02/0.03 0.3/0.3/0.4 BD, bloom date (Julian days); RD, ripening date (Julian days); DAB, days after bloom (Julian days); blush (0–5 scale); FDIA, fruit diameter (mm); FW, fruit weight (g); ADH, adherence (1–4); FF, fruit ﬁrmness (N); RP, Redness around pit (0–1 scale); FT, fruit texture (1–2 scale); PW, pit weight (g); SSC, soluble solid concentration (◦Brix); TA, titratable acidity (% malic acid); SE, mean standard error; SD, standard deviation. BD, bloom date (Julian days); RD, ripening date (Julian days); DAB, days after bloom (Julian days); blush (0–5 scale); FDIA, fruit diameter (mm); FW, fruit weight (g); ADH, adherence (1–4); FF, fruit ﬁrmness (N); RP, Redness around pit (0–1 scale); FT, fruit texture (1–2 scale); PW, pit weight (g); SSC, soluble solid concentration (◦Brix); TA, titratable acidity (% malic acid); SE, mean standard error; SD, standard deviation. Candidate Genes and TA; BD and FT; BD and FF; BD and Blush; DAB and Blush; and Blush, FT, and PW; FDIA and FW; FDIA and ADH; RD and SSC; and PW and TA), 2 (RD and PW; RD and pH; FF and FT; and FF and SSC), 3 (RD and DAB; DAB and SSC; Blush and RP; and FDIA and FW, ADH and SSC), 4 (BD and RD; RD and DAB; RD, DAB, FF, RP and FT; RD, DAB, FDIA, ADH, and RP; RD, DAB and Blush; RD, DAB and ADH; RD and FF; RD and ADH, RD, DAB and SSC; RD and SSC; RD, ADH and RP; RD, FF and SSC; RD, DAB, FDIA and RP; DAB and FF; DAB and FF; DAB, FF and RP; DAB, FT and PW, FDIA and FW; FDIA and RP; FDIA and ADH; FW, FF, ADH and RP; FF, ADH and FT; FDIA, FW, FF and RP; FF and FT; ADH and FT; FF, ADH and pH, ADH and PW, ADH and RP; and ADH and FT), 5 (BD and TA; RD and SSC; DAB and Blush, FDIA and FW; FW and SSC; FT, SSC, and TA; FT and PW; SSC and TA; FT and TA; SSC and pH; and TA and pH), 6 (RD and DAB; RD and FW; FDIA, FW, FT and SSC; FDIA and ADH; and FW and PW) and 8 (BD and pH; BD and SSC; BD and RD; BD and FF; RD, FDIA and FF; RD, DAB, PW and SSC; and FT and SSC). (Supplementary Figure 1). Population structure analysis with fastSTRUCTURE suggested a number of K between 2 and 19. However, the population stratiﬁcation for K = 3 showed clear diﬀerences between groups based mainly on the pedigree information of the individuals belonging to each group (Supplementary Figure 1). The ﬁrst group accounted for the individuals related to ‘Tropic Beauty’, ‘TX2293_3’, ‘TX2B136’, ‘TXW1293_1’, and ‘TXW1490_1’. The second group comprised individuals linked to ‘A_663’, ‘A_760’, and ‘Bolinha’, while the third group contained individuals linked either to ‘Clayton’ and/or ‘O’Henry’. (Supplementary Figure 1). Population structure analysis with fastSTRUCTURE suggested a number of K between 2 and 19. However, the population stratiﬁcation for K = 3 showed clear diﬀerences between groups based mainly on the pedigree information of the individuals belonging to each group (Supplementary Figure 1). Candidate Genes The ﬁrst group accounted for the individuals related to ‘Tropic Beauty’, ‘TX2293_3’, ‘TX2B136’, ‘TXW1293_1’, and ‘TXW1490_1’. The second group comprised individuals linked to ‘A_663’, ‘A_760’, and ‘Bolinha’, while the third group contained individuals linked either to ‘Clayton’ and/or ‘O’Henry’. The LD decayed with increase of physical distance between SNPs in all groups (Supplementary Figure 2). Considering the admixed individuals, the average of r2 was 0.16. The physical distance over which LD decayed to half of its maximum value was around 540 kb. Diﬀerent patterns of LD decays were observed in the three diﬀerent groups. Group 3 revealed the highest average of r2 (0.32) and the longest physical distances in which LD decayed to half of its maximum value (1,620 kb), while group 2 showed shortest distance (480 kb). In the group 1, the LD decayed of its maximum value of r2 in ∼540 kb. The multi-locus model FarmCPU revealed a total of 180 QTNs (Supplementary Table 6). The highest number of QTNs were detected on chromosome 4 (33), while the smallest was observed on chromosome 7 (6). Consistently associated SNPs over at least two seasons were identiﬁed for BD, RD, ADH, RP, and SSC. In addition, SNPs associated with more than one trait were detected on chromosomes 1 (DAB and RD), 3 (Blush and RP; FW and RD), 4 (Blush and RD; RD and SSC; FF and SSC), 5 (TA and pH), and 8 (BD and RD; FT and SSC). Multi-Locus Genome-Wide Association y GWAS using the six multi-locus methods in the R package mrMLM 4.0 revealed a total of 967 QTNs associated with 14 traits (Supplementary Table 5). The highest number of associated SNPs was observed on chromosome 4 (99) and the lowest in chromosome 7 (23). Signiﬁcant QTNs detected in at least four methods in the same season, were detected for almost all traits except TA. In addition, consistently associated QTNs identiﬁed in at least two seasons were detected for BD, RD, DAB, ADH, RP, SSC, Blush, FF, FT, TA, and pH. Furthermore, SNPs associated with more than one trait were identiﬁed on chromosome 1 (BD To ensure reliable results, further analyses included only QTNs that met the following conditions: QTNs detected in at least four methods in mrMLM and/or detected in at least two seasons using mrMLM (Table 2 and Figure 1); QTNs detected in two seasons using FarmCPU (Table 3 and Figure 1); and QTNs detected in at least three methods in mrMLM 4.0 and also identiﬁed in the FarmCPU (Table 4 and Figure 1). February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 5 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. TABLE 2 \| Signiﬁcant associations between SNP markers and quality traits detected in at least three methods of mrMLM 4.0 and/or two seasons. Multi-Locus Genome-Wide Association QTN Trait Methoda SNP Chromosome Position (bp) LOD r2 (%) qtnBD_1.1 BD 1–4 SNP_IGA_131557 1 45022954 11.7 – 24.2 13.1 – 64.9 qtnBD_1.2 BD 1–4 SNP_IGA_126857 1 46125525 5.5 – 14.8 3.6 – 37.4 qtnBD_1.3 BD 1–4; 6 SNP_IGA_128189 1 45753343 3.7 – 11.1 1.3 – 6.3 qtnBD_1.4 BD 1–4 SNP_IGA_134730 1 43578596 4.9 – 9.4 0.4 – 2.4 qtnBD_1.5 BD 1,2,4,7 SNP_IGA_119391 1 40620294 4.0 – 5.9 1.4 – 8.5 qtnBD_1.6 BD 5; 7 SNP_IGA_84580 1 25541717 3.2 – 5.6 5.6 – 10.9 qtnBD_4.1 BD 2–5 SNP_IGA_440662 4 16306919 4.3 – 10.4 1.5 – 19.9 qtnBD_7.1 BD 1–3;5 SNP_IGA_779594 7 15842240 3.0 – 11.5 0.5 – 11.0 qtnBD_7.2 BD 4;6–7 SNP_IGA_759649 7 10525885 3.1– 4.6 1.1 – 28.7 qtnRD_4.1 RD 1–7 SNP_IGA_415301 4 12523245 3.5 – 12.8 0.8 – 37.2 qtnRD_4.2 RD 1–7 SNP_IGA_410398 4 10696489 4.3 – 43.0 5.7 – 36.3 qtnRD_4.3 RD 1–7 SNP_IGA_411637 4 10981971 5.7 –46.5 11.7 – 54.0 qtnRD_4.4 RD 1–4;6–7 SNP_IGA_386222 4 4045426 5.0 – 11.6 3.4 – 9.0 qtnRD_4.5 RD 1–4 SNP_IGA_417666 4 13091850 6.2 – 12.3 1.5 – 4.9 qtnRD_4.6 RD 1–3;7 SNP_IGA_410336 4 10676008 5.3 – 5.4 2.6 – 10.1 qtnRD_4.7 RD 1;6–7 SNP_IGA_410794 4 10890653 5.0 15.4 19.2 – 29.8 qtnRD_6.1 RD 1–2; 4–6 SNP_IGA_632033 6 8774913 5.1 – 10.1 1.6 – 6.6 qtnDAB_4.1 DAB 1–7 SNP_IGA_410398 4 10696489 5.2 – 38.4 7.9 – 28.4 qtnDAB_4.2 DAB 1–7 SNP_IGA_411637 4 10981971 5.4 – 56.6 16.9 – 56.1 qtnDAB_4.3 DAB 1–3; 5 SNP_IGA_403613 4 9052116 3.6 – 14.1 2.0 – 5.6 qtnDAB_5.1 DAB 1–3; 6 SNP_IGA_602331 5 16550893 5.7 – 10.0 13.0 – 20.3 qtnBlush_1.1 Blush 1–6 SNP_IGA_88046 1 26896332 3.0 – 15.0 3.7 – 24.4 qtnBlush_1.2 Blush 1–2; 7 SNP_IGA_7992 1 2518043 4.0 – 4.8 3.0 – 11.1 qtnBlush_3.1 Blush 1–2; 4–6 SNP_IGA_349831 3 20473077 3.1 – 5.2 1.5 – 8.0 qtnBlush_3.2 Blush 3; 6–7 SNP_IGA_341962 3 18179421 4.7 – 11.3 1.2 – 11.4 qtnBlush_4.1 Blush 2–4; 6 SNP_IGA_397470 4 6624729 3.8 – 5.7 12.1 – 26.1 qtnBlush_5.1 Blush 2–4; 6 SNP_IGA_602331 5 16550893 3.3 – 8.1 1.4 – 6.90 qtnFDIA_7.1 FDIA 1–2; 4; 6 SNP_IGA_726818 7 207697 3.0 – 5.1 4.8 – 7.9 qtnFW_1.1 FW 1–6 SNP_IGA_1129 1 209701 3.5 – 8.6 4.3 – 8.5 qtnFW_1.2 FW 1–2; 4–5 SNP_IGA_89193 1 27244316 4.7 – 10.4 8.5 – 19.4 qtnFW_2.1 FW 1–4; 6 SNP_IGA_275189 2 22195492 3.4 – 7.3 2.2 – 8.9 qtnFW_3.1 FW 1–2; 4; 6 SNP_IGA_298935 3 3989094 6.7 – 8.0 5.8 – 7.7 qtnFW_4.1 FW 1–2; 4; 6 SNP_IGA_404442 4 9321093 3.7 – 4.9 9.8 – 15.7 qtnFW_4.2 FW 1–6 SNP_IGA_439186 4 15742278 6.0 – 10.4 4.2 – 13.4 qtnFW_6.1 FW 1–6 SNP_IGA_652492 6 13508541 3.1 – 5.3 2.6 – 13.8 qtnFW_6.2 FW 1–2; 4–6 SNP_IGA_699516 6 29491714 4.7 – 6.5 5.6 – 7.7 qtnFF_1.1 FF 1–2; 4–5 SNP_IGA_126158 1 46430951 3.3 – 4.1 8.6 –16.4 qtnFF_4.1 FF 2–3; 5–6 SNP_IGA_379393 4 1391180 3.0 – 6.9 2.5 – 16.5 qtnFF_4.2 FF 2–5; 7 SNP_IGA_379856 4 1477791 3.3 – 7.9 1.7 – 4.0 qtnFF_4.3 FF 2–6 SNP_IGA_411161 4 10922075 4.3 – 11.2 2.2 – 9.8 qtnADH_4.1 ADH 1–6 SNP_IGA_450629 4 18235458 8.8 – 20.3 12.9 – 69.2 qtnADH_4.2 ADH 1–7 SNP_IGA_467302 4 19028425 6.3 – 69.0 5.3 – 30.4 qtnADH_4.3 ADH 1–2; 4–6 SNP_IGA_441749 4 16584598 4.1 – 12.4 2.5 –33.1 qtnADH_4.4 ADH 1–7 SNP_IGA_411147 4 10921604 3.8 – 13.0 2.2 – 29.6 qtnADH_4.5 ADH 1; 7 SNP_IGA_410398 4 10696489 4.6 – 7.2 2.8 – 6.5 qtnADH_4.6 ADH 1; 4; 6–7 SNP_IGA_387584 4 4601159 3.1 – 6.9 3.2 – 6.2 qtnADH_4.7 ADH 4; 6–7 SNP_IGA_410165 4 10641209 4.0 – 5.8 1.9 – 3.3 qtnADH_6.1 ADH 1–2; 4; 6 snp_6_13059650 6 13073956 4.1 – 8.5 4.7 – 10.2 qtnRP_3.1 RP 4; 6–7 SNP_IGA_341962 3 18179421 12.0 - 12.1 5.9 – 23.6 qtnRP_3.2 RP 1; 5; 7 SNP_IGA_343288 3 18666687 4.8 – 9.5 23.9 – 55.2 qtnRP_4.1 RP 1–7 SNP_IGA_410398 4 10696489 4.4 – 15.2 6.2 – 19.7 qtnRP_4.2 RP 1–2; 4; 6–7 SNP_IGA_411147 4 10921604 3.5 – 9.9 8.8 - 21.0 qtnRP_4.3 RP 4; 6–7 SNP_IGA_408223 4 10107085 3.3 – 4.2 1.2 – 4.1 (Continued) February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 6 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. Multi-Locus Genome-Wide Association PW, pit weight; SSC, soluble solid concentration; TA, titratable acidity. aMethods: 1 = mrMLM; 2 = FASTmrMLM; 3 = FASTmrEMMA; 4 = pLARmEB; 5 = pKWmEB; 6 = ISIS EM–BLASSO; 7 = detected in at least two seasons. , p g ; , ; , y aMethods: 1 = mrMLM; 2 = FASTmrMLM; 3 = FASTmrEMMA; 4 = pLARmEB; 5 = pKWmEB; 6 = ISIS EM–BLASSO; 7 = detected in at least two seasons. Multi-Locus Genome-Wide Association TABLE 2 \| Continued QTN Trait Methoda SNP Chromosome Position (bp) LOD r2 (%) qtnRP_6.1 RP 1–4; 6 SNP_IGA_698951 6 29242212 4.2 – 9.3 6.0 - 8.5 qtnFT_2.1 FT 3–4; 7 SNP_IGA_197236 2 6650921 3.3 – 7.2 2.1 – 4.9 qtnFT_2.2 FT 5–7 SNP_IGA_198691 2 6753400 4.9 – 7.5 5.2 – 16.4 qtnFT_4.1 FT 1–4 SNP_IGA_374610 4 994306 6.4 – 9.4 4.8 – 11.1 qtnFT_5.1 FT 1–3; 5 SNP_IGA_545448 5 850261 5.1 – 30.7 15.3 – 33.8 qtnFT_5.2 FT 1–2; 5–6 SNP_IGA_559057 5 3731800 7.9 –11.2 9.9 –16.7 qtnFT_5.3 FT 4; 6–7 SNP_IGA_553456 5 2477309 3.9 – 7.5 10.1 – 22.7 qtnFT_7.1 FT 1–7 SNP_IGA_769572 7 12248919 3.3 – 10.4 7.6 – 16.0 qtnFT_8.1 FT 1–4; 6 SNP_IGA_821894 8 5071328 3.5 – 6.8 3.4 – 6.4 qtnFT_8.2 FT 1–2; 4–5; 7 SNP_IGA_866041 8 15002010 3.7 – 4.2 0.9 – 4.2 qtnPW_6.1 PW 1–4; 6 SNP_IGA_680747 6 24132839 3.6 – 7.7 2.8 –5.7 qtnPW_6.2 PW 1–4; 6 SNP_IGA_697517 6 28795793 5.2 – 7.6 6.6 – 17.0 qtnPW_6.3 PW 1–2; 4–6 SNP_IGA_684085 6 25090090 3.7 – 9.1 8.4 – 28.6 qtnPW_8.1 PW 2; 4–6 SNP_IGA_879528 8 19234898 3.0 – 4.5 4.5 – 13.2 qtnSSC_1.1 SSC 1–4; 6 SNP_IGA_58626 1 17538855 12.6 – 25.0 1.6 – 17.4 qtnSSC_4.1 SSC 1–4 SNP_IGA_397710 4 6694626 11.8 – 15.0 12.3 – 16.4 qtnSSC_4.2 SSC 1–2; 4–6 SNP_IGA_426994 4 14898353 3.2 – 7.6 2.4 – 11.8 qtnSSC_4.3 SSC 1–2; 4–7 SNP_IGA_411161 4 10922075 4.2 – 13.4 0.8 – 7.5 qtnSSC_5.1 SSC 1–5 SNP_IGA_552247 5 2240224 8.7 – 31.4 5.3 – 24.0 qtnSSC_5.2 SSC 1–2; 4;6 SNP_IGA_544961 5 698215 4.2 – 7.1 3.3 – 8.4 qtnSSC_5.3 SSC 1–7 SNP_IGA_595786 5 13019899 3.2 – 7.4 1.1 – 5.6 qtnSSC_6.1 SSC 1–2; 4;6 SNP_IGA_673205 6 21277895 6.7 – 9.5 4.3 – 12.3 qtnTA_5.1 TA 1–2; 7 SNP_IGA_547830 5 1342919 3.1 – 20.4 1.0–38.7 qtnpH_2.1 pH 1–4; 6 SNP_IGA_288845 2 29376788 4.9 – 11.2 6.9 – 14.6 qtnpH_5.1 pH 1–2; 4–7 SNP_IGA_544428 5 557504 7.1 – 29.51 30.5 – 62.7 BD, bloom date; RD, ripening date; DAB, days after bloom; FF, fruit diameter; FW, fruit weight; FF, fruit ﬁrmness; ADH, adherence; RP, redness around pit; FT, fruit texture; PW, pit weight; SSC, soluble solid concentration; TA, titratable acidity. aMethods: 1 = mrMLM; 2 = FASTmrMLM; 3 = FASTmrEMMA; 4 = pLARmEB; 5 = pKWmEB; 6 = ISIS EM–BLASSO; 7 = detected in at least two seasons. BD, bloom date; RD, ripening date; ADH, adherence; RP, red in pit; SSC, soluble solid concentration. QTNs Detected With Four Methods and/or Two Seasons Using mrMLM R Package detected on chromosome 4 were detected in at least two seasons, except qtnRD_4.5. Three QTNs on chromosome 4 (qtnDAB_4.1, qtnDAB_4.2 and qtnDAB_4.3) were associated with DAB. The greatest phenotypic variation was explained by qtnDAB_4.2 (16.9– 56.1%). The qtnDAB_5.1 on chromosome 5 explained 13.0 to 20.3% of the phenotypic variance observed (LOD scores 5.7– 10.0). The qtnDAB_4.1 and qtnDAB_4.2 on chromosome 4 were detected in two seasons. The mrMLM revealed nine reliable QTNs distributed on 3 chromosomes signiﬁcantly associated with BD (Table 2 and Figure 1). The qtnBD_1.1, qtnBD_1.2, qtnBD_1.3, qtnBD_1.4, qtnBD_1.5, and qtnBD_1.6, on chromosome 1, explained 13.1–64.9, 3.6–37.4, 1.3–6.3, 0.4–2.4, 1.4–8.5, and 5.6–10.9% of total phenotypic variation, respectively. The qtnBD_4.1 located on chromosome 4, demonstrated LOD score of 4.3–10.4 and explained 1.5–19.8% of phenotypic variation. The reliable QTNs identiﬁed on chromosome 7 (qtnBD_7.1 and qtnBD_7.2) explained 0.5–11.1 and 1.1–28.7% of total phenotypic variation, respectively. In addition, the qtnBD_1.5, qtnBD_1.6 and qtnBD_7.2 were detected in two seasons. The reliable QTNs associated with blush were distributed on chromosomes 1 (qtnBlush_1.1 and qtnBlush_1.2), 3 (qtnBlush_3.1 and qtnBlush_3.2), 4 (qtnBlush_4.1), and 5 (qtnBlush_5.1). The qtnBlush_3.2 was identiﬁed in three seasons, while the qtnBlush_1.2 was identiﬁed in two. The qtnBlush_4.1 explained the highest phenotypic variation (12.1–26.1%). The qtnFDIA_7.1 associated with FDIA was located on chromosome 7, revealed LOD scores ranging from 3.0 to 5.1 and explained 4.8–7.9% of the phenotypic variance observed. Reliable QTNs associated with RD were detected on chromosomes 4 (qtnRD_4.1, qtnRD_4.2, qtnRD_4.3, qtnRD_4.4, qtnRD_4.5, qtnRD_4.6 and qtnRD_4.7) and 6 (qtnRD_6.1). The qtnRD_4.3 accounted for the highest phenotypic variation (11.7–54.0%) in comparison to the other QTNs identiﬁed on chromosome 4 (Table 2 and Figure 1). The qtnRD_6.1, on chromosome 6 demonstrated LD scores ranging from 5.1 to 10.1 and explained 1.6–6.6% of phenotypic variation. Almost all QTNs Concerning FW, QTNs were detected on chromosomes 1 (qtnFW_1.1 and qtnFW_1.2), 2 (qtnFW_2.1), 3 (qtnFW_3.1), 4 (qtnFW_4.1 and qtnFW_4.2), and 6 (qtnFW_6.1 and qtnFW_6.2). The qtnFW_1.1, qtnFW_4.2 and qtnFW_6.1 were detected with the six methods of the mrMLM R package. The greatest phenotypic variation was explained by qtnFW_1.2 (8.5–19.4%), followed by qtnFW_4.1 (9.8–15.7%). February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 7 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. FIGURE 1 \| Reliable QTNs detected using different multi-locus GWAS methods and/or at least two seasons. BD, bloom date; RD, ripening date; ADH, adherence; RP, red in pit; SSC, soluble solid concentration. QTNs Detected With Four Methods and/or Two Seasons Using mrMLM R Package Genetic distance scale in physical position (Mbp) is placed at left margin. FIGURE 1 \| Reliable QTNs detected using different multi-locus GWAS methods and/or at least two seasons. Genetic distance scale in physical position (Mbp) is placed at left margin. FIGURE 1 \| Reliable QTNs detected using different multi-locus GWAS methods and/or at least two seasons. Genetic distance scale in physical position (Mbp) is l d t l ft i TABLE 3 \| Signiﬁcant associations between SNP markers and fruit quality traits consistently detected in two seasons using FarmCPU. QTN Trait/year SNP Chromosome Position (bp) P-value qtnBD_4.2 BD_2011 SNP_IGA_370194 4 607467 1.37E-07 BD_2012 SNP_IGA_370194 4 607467 7.74E-07 qtnRD_4.7 RD_2010 SNP_IGA_410794 4 10890653 2.58E-16 RD_2011 SNP_IGA_410794 4 10890653 3.9E-39 RD_2012 SNP_IGA_410794 4 10890653 1.47E-55 qtnRD_4.8 RD_2011 SNP_IGA_415382 4 12546297 1.2E-07 RD_2012 SNP_IGA_415382 4 12546297 7.72E-11 qtnRD_8.1 RD_2011 SNP_IGA_803461 8 2534033 3.24E-13 RD_2012 SNP_IGA_803461 8 2534033 1.59E-06 qtnADH_4.8 ADH_2010 SNP_IGA_409379 4 10389254 1.01E-06 ADH_2011 SNP_IGA_409379 4 10389254 9.65E-09 qtnADH_4.5 ADH_2011 SNP_IGA_410398 4 10696489 8.03E-19 ADH_2012 SNP_IGA_410398 4 10696489 5.49E-09 qtnADH_4.2 ADH_2011 SNP_IGA_467302 4 19028425 8.31E-14 ADH_2012 SNP_IGA_467302 4 19028425 4.8E-10 qtnRP_3.1 RP_2010 SNP_IGA_341962 3 18179421 6.64E-11 RP_2011 SNP_IGA_341962 3 18179421 2.93E-16 qtnSSC_5.3 SSC_2011 SNP_IGA_595786 5 13019899 2.87E-08 SSC_2012 SNP_IGA_595786 5 13019899 1.89E-06 qtnSSC_6.1 SSC_2011 SNP_IGA_673205 6 21277895 1.73E-06 SSC_2012 SNP_IGA_673205 6 21277895 2.42E-06 BD bloom date RD ipening date ADH adhe ence RP ed in pit SSC sol ble solid concent ation February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 8 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. TABLE 4 \| Signiﬁcant associations between SNP markers and quality traits commonly detected using at least three mrMLM 4.0 GWAS methods and FarmCPU. QTNs Detected With Four Methods and/or Two Seasons Using mrMLM R Package QTN Trait SNP Chromosome Position (bp) qtnBD_1.1 BD SNP_IGA_131557 1 45022954 qtnBD_1.2 BD SNP_IGA_126857 1 46125525 qtnBD_1.3 BD SNP_IGA_128189 1 45753343 qtnBD_4.3 BD SNP_IGA_420316 4 13813285 qtnRD_1.1 RD SNP_IGA_99110 1 30864365 qtnRD_4.1 RD SNP_IGA_415301 4 12523245 qtnRD_4.2 RD SNP_IGA_410398 4 10696489 qtnRD_4.5 RD SNP_IGA_417666 4 13091850 qtnDAB_4.2 DAB SNP_IGA_411637 4 10981971 qtnBlush_3.3 Blush SNP_IGA_330725 3 16198112 qtnBlush_5.1 Blush SNP_IGA_602331 5 16550893 qtnFDIA_1.1 FDIA SNP_IGA_104819 1 34325189 qtnFW_6.1 FW SNP_IGA_652492 6 13508541 qtnFF_4.1 FF SNP_IGA_379393 4 1391180 qtnFF_4.3 FF SNP_IGA_411161 4 10922075 qtnADH_4.1 ADH SNP_IGA_450629 4 18235458 qtnADH_4.2 ADH SNP_IGA_467302 4 19028425 qtnADH_4.3 ADH SNP_IGA_441749 4 16584598 qtnRP_4.2 RP SNP_IGA_411147 4 10921604 qtnRP_6.1 RP SNP_IGA_698951 6 29242212 qtnRP_8.1 RP SNP_IGA_804739 8 2702428 qtnFT_5.1 FT SNP_IGA_545448 5 850261 qtnFT_5.4 FT SNP_IGA_543474 5 329318 qtnSSC_1.1 SSC SNP_IGA_58626 1 17538855 qtnSSC_4.1 SSC SNP_IGA_397710 4 6694626 qtnSSC_4.2 SSC SNP_IGA_426994 4 14898353 qtnSSC_4.3 SSC SNP_IGA_411161 4 10922075 qtnSSC_5.1 SSC SNP_IGA_552247 5 2240224 qtnSSC_5.3 SSC SNP_IGA_595786 5 13019899 qtnpH_2.1 pH SNP_IGA_288845 2 29376788 qtnpH_5.1 pH SNP_IGA_544428 5 557504 BD, bloom date; RD, ripening date; DAB, days after bloom; FDIA, fruit diameter; FW, fruit weight; FF, fruit ﬁrmness; ADH, adherence; RP, redness around pit; FT, fruit texture; SSC, soluble solid concentration. TABLE 4 \| Signiﬁcant associations between SNP markers and quality traits commonly detected using at least three mrMLM 4.0 GWAS methods and FarmCPU. was detected in three seasons, while the qtnRP_3.1, qtnRP_3.2, qtnRP_4.2 and qtnRP_4.3 were detected in two seasons. The highest phenotypic variation (23.9–55.2%) was explained by the qtnRP_3.2 located on the chromosome 3. The reliable QTNs associated with FT were distributed on chromosome 2 (qtnFT_2.1 and qtnFT_2.2), 4 (qtnFT_4.1), 5 (qtnFT_5.1, qtnFT_5.2 and qtnFT_5.3), 7 (qtnFT_7.1) and 8 (qtnFT_8.1 and qtnFT_8.2). The qtnFT_2.1, qtnFT_2.2, qtnFT_5.3, qtnFT_7.1 and qtnFT_8.2 were detected in two seasons. The qtnFT_5.1 and qtnFT_5.3 exhibited the highest phenotypic variation 15.3–33.8 and 10.1–22.7%, respectively. Quantitative trait nucleotides associated with PW were identiﬁed on chromosomes 6 (qtnPW_6.1, qtnPW_6.2 and qtnPW_6.3) and 8 (qtnPW_8.1). The qtnPW_6.2 and qtnPW_6.3 with LOD scores of 5.2–7.6 and 3.7–9.1 explained 6.6–17.0 and 8.4–28.6% of the phenotypic variation, respectively. For SSC, 8 reliable QTNs were detected on chromosomes 1 (qtnSSC_1.1), 4 (qtnSSC_4.1, qtnSSC_4.2 and qtnSSC_4.3), 5 (qtnSSC_5.1, qtnSSC_5.2 and qtnSSC_5.3), and 6 (qtnSSC_6.1), with the qtnSSC_4.3 and qtnSSC_5.3 detected in two seasons. The qtnSSC_4.1 and qtnSSC_5.1 accounted for the highest phenotypic variation (12.3–16.4 and 5.3–24.0%). The qtnTA_5.1 on chromosome 5 was associated with TA and detected in two seasons. QTNs Detected in Two Seasons Using FarmCPU The qtnBD_4.2 was consistently associated with BD in the two seasons and revealed a p values of 1.37E-07 and 7.74E-07, respectively (Table 3 and Figure 1). y The qtnRD_4.7 located on chromosome 4 (10.9 Mbp) was signiﬁcantly associated with RD in all three seasons. In addition, the qtnRD_4.8 and qtnRD_8.1 were associated with RD in 2011 and 2012 and were located at 12.5 Mbp (chromosome 4) and 2.5 Mbp (chromosome 8), respectively. BD, bloom date; RD, ripening date; DAB, days after bloom; FDIA, fruit diameter; FW, fruit weight; FF, fruit ﬁrmness; ADH, adherence; RP, redness around pit; FT, fruit texture; SSC, soluble solid concentration. Quantitative trait nucleotides associated with FF were located on chromosomes 1 (qtnFF_1.1) and 4 (qtnFF_4.1, qtnFF_4.2 and qtnFF_4.3). The qtnFF_4.2 was identiﬁed in two seasons and explained 1.7–4.0% of the phenotypic variation with the qtnFF_1.1 explaining the highest phenotypic variation for FF (8.6–16.4%). A reliable qtnADH_4.8 on chromosome 4 (10.4 Mb) was consistently associated with ADH in 2010 and 2011, while qtnADH_4.5 (10.69) and qtnADH_4.2 (19.02 Mbp) were associated with ADH in 2011 and 2012. For RP, the qtnRP_3.1 located at 18.17 Mbp on chromosome 3 was detected in two seasons with p values of 6.64E-11 and 2.93E- 16, respectively. For ADH, seven QTNs on chromosome 4 (qtnADH_4.1, qtnADH_4.2, qtnADH_4.3, qtnADH_4.4, qtnADH_4.5, qtnADH_4.6 and qtnADH_4.7), and one on chromosome 6 (qtnADH_6.1) were identiﬁed. The qtnADH_4.2 was detected in three seasons, while qtnADH_4.4, qtnADH_4.5, qtnADH_4.6 and qtnADH_4.7 were detected in two seasons. The qtnADH_4.1 was identiﬁed with the six methods of mrMLM and explained the highest phenotypic variance (12.9–69.2%). Consistent, signiﬁcant associations with the SSC were detected in two seasons with QTNs on chromosomes 5 (qtnSSC_5.3) and 6 (qtnSSC_6.1). Consistent, signiﬁcant associations with the SSC were detected in two seasons with QTNs on chromosomes 5 (qtnSSC_5.3) and 6 (qtnSSC_6.1). Consistent, signiﬁcant associations with the SSC were detected in two seasons with QTNs on chromosomes 5 (qtnSSC_5.3) and 6 (qtnSSC_6.1). QTNs Detected With Four Methods and/or Two Seasons Using mrMLM R Package The LOD score varied from 3.1 to 20.4 and explained 1.0–38.0% of the phenotypic variation. One reliable QTN associated with pH was identiﬁed on chromosome 2 (qtnpH_2.1) and chromosome 5 (qtnpH_5.1). The qtnpH_5.1 explained the highest phenotypic variation (30.5– 62.7%) and was detected in two seasons. Frontiers in Plant Science \| www.frontiersin.org QTNs Commonly Detected With Both mrMLM and FarmCPU Trait Hap Scaffold Start (bp) End (bp) Associated SNPs SSC 1_1 1 17400346 17538855 SNP_IGA_58626 RD 1_2 1 30644296 31160594 SNP_IGA_99110 FDIA 1_3 1 34122904 34404100 SNP_IGA_104819 BD 1_4 1 44904968 45237616 SNP_IGA_131557 BD 1_5 1 45753343 45821173 SNP_IGA_128189 BD 1_6 1 46012310 46430951 SNP_IGA_126857 pH 2_1 2 29241773 29376788 SNP_IGA_288845 Blush 3_1 3 16195795 16236799 SNP_IGA_330725 FF 4_1 4 1382161 1413701 SNP_IGA_379393 SSC 4_2 4 6688718 6712809 SNP_IGA_397710 RD 4_3 4 10676008 10760085 SNP_IGA_410398 RP 4_4 4 10760086 10981971 SNP_IGA_411147 FF 4_4 4 10760086 10981971 SNP_IGA_411161 SSC 4_4 4 10760086 10981971 SNP_IGA_411161 DAB 4_4 4 10760086 10981971 SNP_IGA_411637 RD 4_5 4 12429145 12523245 SNP_IGA_415301 RD 4_6 4 13078233 13108512 SNP_IGA_417666 BD 4_7 4 13561808 14018643 SNP_IGA_420316 SSC 4_8 4 14735598 15182577 SNP_IGA_426994 ADH 4_9 4 16511312 16674024 SNP_IGA_441749 ADH 4_10 4 18140428 18235458 SNP_IGA_450629 ADH 4_11 4 18719887 19206580 SNP_IGA_467302 FT 5_1 5 329318 481015 SNP_IGA_543474 pH 5_2 5 521865 821356 SNP_IGA_544428 FT 5_3 5 850261 882334 SNP_IGA_545448 SSC 5_4 5 2086499 2242971 SNP_IGA_552247 SSC 5_5 5 13014155 13019899 SNP_IGA_595786 Blush 5_6 5 16550893 16702088 SNP_IGA_602331 FW 6_1 6 13235506 13565811 SNP_IGA_652492 RP 6_2 6 29231386 29714220 SNP_IGA_698951 RP 8_1 8 2369263 2838462 SNP_IGA_804739 BD, bloom date; RD, ripening date; DAB, days after bloom; FDIA, fruit diameter; FW, fruit weight; FF, fruit ﬁrmness; ADH, adherence; RP, red in pit; FT, fruit texture; SSC, soluble solid concentration. 28.8–29.5 Mbp, QTNs involving PW, RP, and FW were detected. Furthermore, on top of chromosome 8 (2.5–5.1 Mbp), a hotspot with reliable QTNs associated with RD, RP, and FT was observed. DISCUSSION We have analyzed peach germplasm containing 620 individuals from three U.S. public fresh market breeding programs [University of Arkansas System Division of Agriculture (AR), Clemson University (SC) and Texas A&M University (TX)] for 14 traits over three seasons (2010, 2011, and 2012). Phenotypic variation was observed between individuals and seasons, and the mean values for BD, RD, FW, and SSC were lower than those reported in the Spanish and European germplasm (Hernández Mora et al., 2017; Font I Forcada et al., 2019).However, average TABLE 5 \| Haploblock regions encompassing SNPs markers signiﬁcantly associated with fruit quality traits in peach. Hotspots in Peach Genome The reliable QTNs revealed fruit quality hotspots in the peach genome (Figure 1). On chromosome 1, three reliable QTNs associated with BD, Blush and FW in the interval of 25.5– 27.2 Mbp were identiﬁed. In addition, at the bottom of the same chromosome (43.6–46.4 Mbp), we also observed QTNs associated with BD and FF. On chromosome 3, a hotspot involving the quality traits Blush and RP was observed in the region located at 18.2–20.2 Mbp. The majority of the reliable QTNs detected were located on chromosome 4 (0.6–19.0 Mbp), especially concentrated in the genomic region located at 9.0–12.5 Mbp with QTNs associated with DAB, FW, RP, ADH, RD, FF, and SSC. A hotspot was also observed on the top of chromosome 5 (0.3–3.7 Mbp) with signiﬁcant signals associated with FT, pH, SSC, and TA. In the genomic region on chromosome 6 spanning Candidate Genes A total of 566 candidate genes (CG) were detected within the haploblock regions for the signiﬁcantly associated QTNs (Supplementary Table 7), from which 93 CG were detected in the regions for BD, 89 for RD, 29 for DAB, 22 for Blush, 39 for FDIA, 24 for FW, 26 for ADH, 31 for FF, 148 for RP, 12 for FT and 90 for SSC. The gene ontology (GO) annotations were retrieved for 435 CG. The GO enrichment analysis revealed 68 GO terms in all three GO aspects, biological process, molecular function, and cellular component. Twenty-six GO terms (78 genes) were described as biological processes, 32 GO terms (108 genes) with the molecular function, and 10 GO terms (36 genes) with the cellular component (Supplementary Table 8). The GO term cluster representatives were joined into “superclusters” of terms loosely related to cellulose microﬁbril organization, THO complex part of transcription export complex and sulfotransferase activity in the biological process, cellular component and molecular function, respectively (Figure 2). 28.8–29.5 Mbp, QTNs involving PW, RP, and FW were detected. Furthermore, on top of chromosome 8 (2.5–5.1 Mbp), a hotspot with reliable QTNs associated with RD, RP, and FT was observed. 28.8–29.5 Mbp, QTNs involving PW, RP, and FW were detected. Furthermore, on top of chromosome 8 (2.5–5.1 Mbp), a hotspot with reliable QTNs associated with RD, RP, and FT was observed. QTNs Commonly Detected With Both mrMLM and FarmCPU Quantitative trait nucleotides on chromosome 3 (qtnRP_3.1 and qtnRP_3.2), 4 (qtnRP_4.1, qtnRP_4.2 and qtnRP_4.3), and 6 (qtnRP_6.1) were signiﬁcantly associated with RP. The qtnRP_4.1 A total of 31 QTNs were consistently detected with at least three methods in mrMLM and also with the FarmCPU (Table 4). Six QTNs associated with BD, RD, FDIA, and SSC were February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 9 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. identiﬁed on chromosome 1 (qtnBD_1.1, qtnBD_1.2, qtnBD_1.3, qtnRD_1.1, qtnFDIA_1.1, and qtnSSC_1.1). Only one QTN (qtnpH_2.1) associated with pH, located at 29.4 Mpb on chromosome 2 was detected using both approaches. A QTN associated with blush was detected on chromosome 3 (16.2 Mbp). Most of the QTNs were observed on chromosome 4. A total of twelve QTNs associated with BD (qtnBD_4.2), RD (qtnRD_4.1, qtnRD_4.2 and qtnRD_4.5), DAB (qtnDAB_4.2), FF (qtnFF_4.1 and qtnFF_4.3), ADH (qtnADH_4.1, qtnADH_4.2 and qtnADH_4.3), RP (qtnRP_4.2) and SSC (qtnSSC_4.1 and qtnSSC_4.2) and a QTN cluster of FF and SSC (qtnFF_4.3 and qtnSSC_4.3) were identiﬁed. QTNs for blush (qtnBlush_5.1), FT (qtnFT_5.1 and qtnFT_5.4), SSC (qtnSSC_5.1 and qtnSSC_5.3) and pH (qtnpH_5.1) were identiﬁed on chromosome 5 located at 16.55, 0.85, 0.32, 2.24, 13.01 and 0.55 Mbp, respectively. Two QTNs associated with FW (qtnFW_6.1) and RP (qtnRP_6.1) were present on chromosome 6. Lastly, the QTN qtnRP_8.1, associated with the RP, was present on chromosome 8 (2.7 Mbp). Candidate Genes Genomic regions encompassing the QTNs detected using the mrMLM 4.0 methods and FarmCPU revealed a total of twenty- eight haploblocks located at scaﬀolds 1, 2, 3, 4, 5, 6, and 8 (Table 5). A total of 566 candidate genes (CG) were detected within the haploblock regions for the signiﬁcantly associated QTNs (Supplementary Table 7), from which 93 CG were detected in the regions for BD, 89 for RD, 29 for DAB, 22 for Blush, 39 for FDIA, 24 for FW, 26 for ADH, 31 for FF, 148 for RP, 12 for FT and 90 for SSC. The gene ontology (GO) annotations were retrieved for 435 CG. The GO enrichment analysis revealed 68 GO terms in all three GO aspects, biological process, molecular function, and cellular component. Twenty-six GO terms (78 genes) were described as biological processes, 32 GO terms (108 genes) with the molecular function, and 10 GO terms (36 genes) with the cellular component (Supplementary Table 8). QTNs Commonly Detected With Both mrMLM and FarmCPU The GO term cluster representatives were joined into “superclusters” of terms loosely related to cellulose microﬁbril organization, THO complex part of transcription export complex and sulfotransferase activity in the biological process, cellular component and molecular function, respectively (Figure 2). Hotspots in Peach Genome The reliable QTNs revealed fruit quality hotspots in the peach genome (Figure 1). On chromosome 1, three reliable QTNs associated with BD, Blush and FW in the interval of 25.5– 27.2 Mbp were identiﬁed. In addition, at the bottom of the same chromosome (43.6–46.4 Mbp), we also observed QTNs associated with BD and FF. On chromosome 3, a hotspot involving the quality traits Blush and RP was observed in the region located at 18.2–20.2 Mbp. The majority of the reliable QTNs detected were located on chromosome 4 (0.6–19.0 Mbp), especially concentrated in the genomic region located at 9.0–12.5 Mbp with QTNs associated with DAB, FW, RP, ADH, RD, FF, and SSC. A hotspot was also observed on the top of chromosome TABLE 5 \| Haploblock regions encompassing SNPs markers signiﬁcantly associated with fruit quality traits in peach. DISCUSSION We have analyzed peach germplasm containing 620 individuals from three U.S. public fresh market breeding programs [University of Arkansas System Division of Agriculture (AR), Clemson University (SC) and Texas A&M University (TX)] for 14 traits over three seasons (2010, 2011, and 2012). Phenotypic variation was observed between individuals and seasons, and the mean values for BD, RD, FW, and SSC were lower than those reported in the Spanish and European germplasm (Hernández Mora et al., 2017; Font I Forcada et al., 2019).However, average values for RD and DAB observed in our study were in agreement with the values reported in the University of Guelph’s peach Frontiers in Plant Science \| www.frontiersin.org February 2021 \| Volume 12 \| Article 644799 10 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. FIGURE 2 \| TreeMap of the GO term cluster representatives joined into “superclusters” of biological processes (A), cellular component (B), and molecular functions (C) visualized with different colors. Each rectangle is a single cluster representative. GO term cluster representatives joined into “superclusters” of biological processes (A), cellular component (B), and molecular functions olors. Each rectangle is a single cluster representative. FIGURE 2 \| TreeMap of the GO term cluster representatives joined into “superclusters” of biological processes (A), cellular component (B), and molecular functions (C) visualized with different colors. Each rectangle is a single cluster representative. GURE 2 \| TreeMap of the GO term cluster representatives joined into “superclusters” of biological processes (A), cellular compone C) visualized with different colors. Each rectangle is a single cluster representative. term cluster representatives joined into “superclusters” of biological processes (A), cellular component (B), and molecular functions s. Each rectangle is a single cluster representative. February 2021 \| Volume 12 \| Article 644799 11 Frontiers in Plant Science \| www.frontiersin.org Fruit Quality Hotspots in Peach Genome da Silva Linge et al. conservative for detecting many important loci. Thus, multi- locus GWAS methods have been recommended to overcome the problem of stringent correction (Zhang et al., 2020). In this study, we have successfully performed a genome-wide association study using six multi-locus GWAS methods (mrMLM, FASTmrMLM, FASTmrEMMA, pLARmEB, pKWmEB, and ISIS EM-BLASSO) comprised in mrMLM 4.0 and FarmCPU R packages.The mrMLM 4.0 adopts the critical probability value or log of odds (LOD), a less stringent signiﬁcance threshold while FarmCPU requires Bonferroni correction to detect QTNs. DISCUSSION Nevertheless, the second group could further be separated in two clusters in which the ﬁrst cluster grouped the individuals linked to ‘A_663’, ‘A_760’, and ‘Bolinha’, and the second cluster contained individuals linked to ‘Clayton’ and/or ‘O’Henry’. The population structure indicated by fastSTRUCTURE, between K = 2 and 19, supported MDS clustering, as the K = 3 reﬂected grouping based on the pedigree background, and number of the breeding programs in the panel. g g p g p The population structure inﬂuences LD patterns within the genome (Thurow et al., 2019). The LD detected in this study decayed much slower in comparison with the observed by Thurow et al. (2019) and faster than the observed by Micheletti et al. (2015). The diﬀerence could be explained by the genetic material analyzed and the methods used for analyzing the LD decay. In addition, a slower decay of LD is expected in selﬁng materials. The LD decay over distance also determines the number of markers required to cover the genome. Considering the LD decay of our dataset (540 Kb), approximately 421 SNPs covering the total peach genome (227.4 Mb) should be suﬃcient to perform the GWAS. However, domestication regions containing key genes require more SNPs due to the faster LD decay (Cao et al., 2016). QTL mapping and association studies focused on ripening date have been widely reported in peach (Eduardo et al., 2011; Pirona et al., 2013; Fresnedo-Ramírez et al., 2015; Hernández Mora et al., 2017; Elsadr et al., 2019; Font I Forcada et al., 2019; Nuñez-Lillo et al., 2019; Rawandoozi et al., 2020a). The qtnRD_4.2, qtnRD_4.3, qtnRD_4.6 and qtnRD_4.7 overlapped with the major RD QTLs reported in the chromosome 4 located at approximately 10.6 and 11.1 Mbp (Pirona et al., 2013; Elsadr et al., 2019). In addition, the qtnRD_4.1, qtnRD_4.5 and qtnRD_4.8 were located in the same genetic interval (11.2 - 14.1 Mbp) of the RD QTL reported by Hernández Mora et al. (2017) using the pedigree-based QTL mapping in the European peach germplasm. Beside chromosome 4, we also detected reliable QTNs for RD on chromosome 1 (qtnRD_1.1), 6 (qtnRD_6.1) and 8 (qtnRD_8.1). DISCUSSION The multi-locus methods detected 967 and 180 QTNs using mrMLM 4.0 and FarmCPU, respectively, allowing the identiﬁcation of important regions in the peach genome that control fruit quality traits. Furthermore, consistently reliable QTNs (88) for all traits were detected using diﬀerent multi-locus GWAS methods and/or at least two seasons (Tables 2–4). Half of the reliable QTNs (44) detected have already been reported using diﬀerent progenies, germplasm and approaches. However, to our knowledge, the other 44 reliable QTNs controlling fruit quality traits in peach have not been previously described. germplasm, comprised of accessions originating from diﬀerent regions across North America (Elsadr et al., 2019). A high and signiﬁcant correlation between FW and FDIA (0.92) was previously observed in peach (da Silva Linge et al., 2015; Abdelghafar et al., 2020), as well as the positive correlation between RD and DAB (Elsadr et al., 2019) and the negative correlation between TA and pH (Abidi et al., 2011). In addition, the high estimated narrow sense heritability coeﬃcients observed in this study ranging from 0.68 to 0.87, suggesting that the phenotypic variations of all traits are mainly aﬀected by genetic factors, and therefore this dataset can be used for further genetic analyses. The mean observed heterozygosity (Ho = 0.36) in the U.S. peach germplasm was greater than that observed in the germplasm from four European, one Chinese and one Brazilian peach collections reported in previous studies (Micheletti et al., 2015; Thurow et al., 2019). In addition, the mean inbreeding coeﬃcient of 0.05 indicated a low level of inbreeding. The low mean of the inbreeding coeﬃcient observed in this study could be attributed to the diverse material, including F1 and F2 populations with diﬀerent genetic backgrounds. One of the main goals of breeding programs is the development of commercial varieties with predictable bloom time to adapt to various target environments. Therefore, understanding the genetic architecture of phenology-related traits represents a key prerequisite to enable the development of varieties adapted to diﬀerent climates (Gogorcena et al., 2020). Reliable QTNs associated with bloom date (BD), the qtnBD_1.1, qtnBD_1.2, qtnBD_1.3, qtnBD_1.4, qtnBD_4.1, qtnBD_4.3, qtnBD_7.1 collocate near or in the same regions previously reported using QTL mapping and pedigree-based analysis (PBA) (Fan et al., 2010; Romeu et al., 2014; Bielenberg et al., 2015; Hernández Mora et al., 2017; Rawandoozi et al., 2020a). DISCUSSION RD QTLs and associated SNPs on chromosome 1 were previously reported in peach at approximately 12.0 Mbp (Fresnedo-Ramírez et al., 2015), 35 Mbp (Romeu et al., 2014; Font I Forcada et al., 2019), and 40.0–47.0 Mbp (Romeu et al., 2014; Hernández Mora et al., 2017; Font I Forcada et al., 2019; Nuñez-Lillo et al., 2019). The reliable qtnRD_1.1 detected in this study using mrMLM and FarmCPU was located at 30.9 Mbp. DISCUSSION The fact that these regions were identiﬁed following diﬀerent approaches (linkage analysis, PBA analysis and GWAS) in diverse genetic material, makes them an interesting source of allelic variation for BD in peach. p p g g The multidimensional scaling (MDS) clustered material into two groups, a group of individuals from TX fresh market breeding program related to ‘Tropic Beauty’, ‘TX2293_3’, ‘TX2B136’, ‘TXW1293_1’ and ‘TXW1490_1’, and the second group, comprised of individuals from the AR and SC fresh market breeding programs. Breeding material from AR and SC clustered in one main group, due to the common founders or pedigree- linkages of some breeding populations. Nevertheless, the second group could further be separated in two clusters in which the ﬁrst cluster grouped the individuals linked to ‘A_663’, ‘A_760’, and ‘Bolinha’, and the second cluster contained individuals linked to ‘Clayton’ and/or ‘O’Henry’. The population structure indicated by fastSTRUCTURE, between K = 2 and 19, supported MDS clustering, as the K = 3 reﬂected grouping based on the pedigree background, and number of the breeding programs in the panel. The population structure inﬂuences LD patterns within the genome (Thurow et al., 2019). The LD detected in this study decayed much slower in comparison with the observed by Thurow et al. (2019) and faster than the observed by Micheletti et al. (2015). The diﬀerence could be explained by the genetic material analyzed and the methods used for analyzing the LD decay. In addition, a slower decay of LD is expected in selﬁng materials. The LD decay over distance also determines the number of markers required to cover the genome. Considering the LD decay of our dataset (540 Kb), approximately 421 SNPs covering the total peach genome (227.4 Mb) should be suﬃcient to perform the GWAS. However, domestication regions containing key genes require more SNPs due to the faster LD decay (Cao et al., 2016). The multidimensional scaling (MDS) clustered material into two groups, a group of individuals from TX fresh market breeding program related to ‘Tropic Beauty’, ‘TX2293_3’, ‘TX2B136’, ‘TXW1293_1’ and ‘TXW1490_1’, and the second group, comprised of individuals from the AR and SC fresh market breeding programs. Breeding material from AR and SC clustered in one main group, due to the common founders or pedigree- linkages of some breeding populations. Multi-Locus GWAS In order to control the false positive rate in GWAS analysis, conservative correction methods such as false discovery rate (FDR) and Bonferroni correction are frequently adopted in association studies. However, these corrections are often too February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 12 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. (Hernández Mora et al., 2017); 41 Mbp (Zeballos et al., 2016) and 43 Mbp (da Silva Linge et al., 2015; Abdelghafar et al., 2020). Concerning position of qtnFW_1.2 and qtnFW_6.2 matched the QTL interval identiﬁed in European peach germplasm (Hernández Mora et al., 2017). The FW QTN qtnFW_2.1 and qtnFW_3.1 could be the FW QTLs mapped in an interspeciﬁc cross between peach and a wild relative Prunus davidiana (Quilot et al., 2004; Desnoues et al., 2016). On chromosome 4 qtnFW_4.1 was close to the FW QTL reported by da Silva Linge et al. (2015) and qtnFW_4.2 was in the same genetic interval of the QTL identiﬁed by Shi et al. (2020). Moreover, the qtnRD_6.1 (chr6: 8.8 Mb) and qtnRD_8.1 (chr8: 2.5 Mbp) were in close proximity of the SNPs associated with RD detected on chromosomes 6 (SNP_IGA_630302; 8.3 Mbp) and 8 (SNP_IGA_806528; 2.9 Mbp) in the Spanish germplasm (Font I Forcada et al., 2019). QTL clusters for RD and DAB were commonly detected in peach (Fresnedo-Ramírez et al., 2015; Hernández Mora et al., 2017; Elsadr et al., 2019; Rawandoozi et al., 2020a). In this study, two reliable QTNs (qtnDAB_4.1 and qtnDAB_4.2) associated with DAB overlapped with the QTNs associated with RD. The position of the qtnDAB_4.1 (chr 4: 10.7 Mbp) matched the associated SNPs identiﬁed in a panel of 132 peach accessions genotypically characterized via genotyping by sequencing (GBS) approach (Elsadr et al., 2019). In addition, the SNP_IGA_410398 was emphasized as a predictive SNP for RD and DAB in haplotype analysis in a DAB QTL detected using PBA approach (Rawandoozi et al., 2020a). The qtnDAB_4.2 (10.9 Mbp) was close to QTL for DAB detected using PBA in the European germplasm (Hernández Mora et al., 2017). Thus, our results conﬁrmed the location of RD and DAB associated regions in the peach genome, and due to their importance for breeding, could be useful in selection of various phenology patterns in future studies. Fruit ﬁrmness (FF) represents an essential indicator of fruit quality for peach consumers. Multi-Locus GWAS For this reason, several authors have investigated the genetic mechanisms controlling this trait in peach (Peace et al., 2005; Eduardo et al., 2011, 2015; Martínez- García et al., 2013; Nuñez-Lillo et al., 2015; Zeballos et al., 2016; Serra et al., 2017; Carrasco-Valenzuela et al., 2019). The qtnFF_4.1 and qtnFF_4.2, reported in this study, were located in chromosome 4 with the position matching the QTL interval associated with ﬁrmness loss mapped by Serra et al. (2017). Moreover, the qtnFF_4.3 was in the same genetic region in which Carrasco-Valenzuela et al. (2019) detected a QTL signiﬁcantly associated with softening rate and Zeballos et al. (2016) detected a QTL for fruit ﬁrmness. Previous QTL analyses in peach have identiﬁed a major QTL for blush on chromosome 3 accounting, in average, for 63.7% of observed phenotypic variation (Frett et al., 2014). The qtnBlush_3.1, qtnBlush_3.2 and qtnBlush_3.3 were located within the genetic interval of the major QTL for blush on chromosome 3. QTL regions for blush on chromosome 4 were mapped approximately at 10.5 – 11.5 Mbp (Rawandoozi et al., 2020b); 11.2 – 14.1 Mbp (Hernández Mora et al., 2017); and at 19.8 and 28.6 Mb (Shi et al., 2020). The qtnBlush_4.1 was detected at 6.6 Mbp and accounted for the highest phenotypic variation observed. Moreover, we identiﬁed two QTNs on chromosome 1 (qtnBlush_1.1: 26.9 Mbp; qtnBlush_1.2: 2.5 Mbp). Analyzing an F1 peach population derived from the cross between “Shahong” and “Hongfurong,” Shi et al. (2020) also observed a QTL associated with blush on chromosome 1. However, the genetic interval was approximately at 21.5 Mbp. Although the percentage of the phenotypic variation explained was low, a QTN associated with blush (qtnBlush_5.1) on chromosome 5 was also detected. Flesh adherence to the pit (ADH) is another factor determining overall peach fruit quality, with consumers preferring freestone or semi freestone characteristics (Olmstead et al., 2015). Previous studies have shown that ADH is inherited and controlled by the Freestone-Melting (F-M) locus on chromosome 4, with genes encoding endopolygalacturonase (endoPG) associated with this trait (Peace et al., 2005; Gu et al., 2016). We detected qtnADH_4.1 and qtnADH_4.2 explaining the majority of the phenotypic variation close to the genetic region where the endoPG gene is located. Similar to Shi et al. (2020), we detected signiﬁcant genetic regions associated to ADH in diﬀerent regions of chromosome 4 and also in other chromosomes. Candidate Genes y Soluble solid concentration is one of the most important quality traits in peach, with consumers expecting an enhanced sugar content or sweetness perception for the low acid types (Cirilli et al., 2016). Therefore, SSC has been a target trait in several studies involving intra- and interspeciﬁc progenies and germplasm to access the genetic potential and consequently improve the sugar content in new cultivars. We detected QTNs associated with SSC on chromosome 1, 4, 5, and 6. The qtnSSC_5.1 on chromosome 5 that explained the majority of the phenotypic variation was in agreement with the QTL interval reported by Hernández Mora et al. (2017) using a PBA analysis in European peach germplasm. In the same chromosome, we also identiﬁed qtnSSC_5.2 (0.7 Mb) and qtnSSC_5.3 (13.0 Mb) whose positions matched QTLs mapped in previous studies using diﬀerent germplasm and approaches (Nuñez-Lillo et al., 2019; Abdelghafar et al., 2020; Rawandoozi et al., 2020b). Furthermore, qtnSSC_4.1 on chromosome 4 (6.6 Mb) was close to the QTL (qSSC.V-Ch4-2007a) detected by Zeballos et al. (2016), while qtnSSC_4.3 was near MD locus reported by Eduardo et al. (2011). On the other hand, the qtnSSC_1.1 (chromosome 1: 17.5 Mb) and qtnSSC_6.1 (chromosome 6: 21.2 Mbp) were located in a diﬀerent region in comparison to the QTLs or associated markers previously detected on those chromosomes (Fresnedo-Ramírez et al., 2015; Cao et al., 2016; Hernández Mora et al., 2017; Li et al., 2019; Shi et al., 2020). Concerning the traits pH and TA, the qtnTA_5.1 and qtnpH_5.1 collocated with the major locus for low-acid fruit (D-locus) previously reported in peach (Boudehri et al., 2009). Candidate genes (566) were identiﬁed within the haploblock regions encompassing the QTNs detected using the mrMLM 4.0 and FarmCPU, and the GO enrichment approach narrowed down the initial CG list (222) and revealed over-representation of certain GO terms (68). RNA binding proteins and serine- type endopeptidase inhibitor-related genes were identiﬁed, and previous studies revealed involvement in the regulation of ﬂowering time (Steﬀen et al., 2019; Zhang F. et al., 2019). In addition, genes functionally annotated as 2-oxoglutarate- dependent dioxygenase, drug transmembrane transport, antiporter activity, pyridine nucleotide biosynthetic process, and chromatin assembly or disassembly were associated with fruit ripening in tomato, apricot, grape, peach, apple, and strawberry (Hanana et al., 2007; Farinati et al., 2017; Decros et al., 2019; Ding et al., 2020; García-Gómez et al., 2020). Multi-Locus GWAS (2016) also detected a region signiﬁcantly associated with PW on chromosome 6; however, the location was approximately at 26.9 Mbp. also demonstrated the necessity to promote further studies for the hotspot regions in chromosome 1, 3, 5, 6, and 8. Fruit Quality Hotspots in Peach Genome Fruit Quality Hotspots in Peach Genome Hotspot regions detected on chromosomes 1, 3, 4, 5, 6, and 8 controlled several fruit quality traits. The detection of hotspots in the genome indicates that genes related to certain traits are more densely concentrated in certain genomic regions (Zhang X. et al., 2019). The main hotspot on chromosome 4 (9.0–12.5 Mbp) included reliable QTNs for DAB, FW, RP, ADH, RD, FF, and SSC detected in diﬀerent seasons and/or approaches and represents a target region for future breeding studies in peach. A QTL hotspot associated with quality traits was previously reported in peach on chromosome 4 (Cantín et al., 2010; Eduardo et al., 2011). However, the study was performed using SSR markers and the QTLs were detected in low-density linkage maps. Using high- density SNP maps, Rawandoozi et al. (2020a) reported QTLs for DAB and RD within the genetic region detected. Likewise, Hernández Mora et al. (2017) detected a hotspot for blush, SSC, RD and DAB in a wider genetic interval located at 11.2– 14.1 Mbp in European germplasm. Moreover, Desnoues et al. (2016) identiﬁed a QTL hotspot in the same chromosome related to individual sugars and FW, although in a diﬀerent location. In addition, the hotspot on chromosome 5 (0.3 to 3.7 Mbp) matched with the QTL hotspot for SSC and TA reported by Hernández Mora et al. (2017). Therefore, this study reinforces the importance of breeding programs targeting the improvement of fruit quality traits in peach focusing on the chromosome 4 and Multi-Locus GWAS The QTNs associated with redness around the pit (RP) (qtnRP_3.1 and qtnRP_3.2) located approximately at 18.2 and 18.7 Mb, on chromosome 3, matched the position of the associated signals to ﬂesh color around the stone detected in the recent GWAS using genome structural variations (SVs) (Guo et al., 2020). In addition, the Cs locus associated with red color around the pit was previously mapped in the middle of chromosome 3 (Yamamoto et al., 2001). Interestingly, the SNP_IGA_341962 (qtnRP_3.1) was also associated with blush (qtnBlush_3.2). Therefore, the QTNs associated with RP identiﬁed on chromosome 4 (qtnRP_4.1, qtnRP_4.2 and qtnRP_4.3) were close to the associated signals detected by Guo et al. (2020) and in a diﬀerent region of the associated SNPs reported by Cao et al. (2016), while the QTNs detected on chromosome 6 (qtnRP_6.1) and 8 (qtnRP_8.1) were located in diﬀerent regions when compared with previous studies (Cao et al., 2016; Guo et al., 2020). with blush (qtnBlush_5.1) on chromosome 5 was also detected. Understanding the genetic control of fruit diameter and weight is an important goal of breeding programs due to the importance of these traits for the fresh market (Yue et al., 2014). QTL regions associated with fruit diameter and weight have been detected in all chromosomes (da Silva Linge et al., 2015; Fresnedo-Ramírez et al., 2016; Zeballos et al., 2016; Hernández Mora et al., 2017; Cao et al., 2019; Abdelghafar et al., 2020; Shi et al., 2020). In this study, we identiﬁed two reliable QTNs associated with FDIA (qtnFDIA_7.1 and qtnFDIA_1.1). The qtnFDIA_7.1, on chromosome 7, is in the vicinity to fruit width and fruit depth QTLs (qP-Fwd7.2 and qP-Fd7.2) reported by da Silva Linge et al. (2015) using F2 progeny resulting from a cross between an ornamental peach PI91459 (“NJ Weeping”) × “Bounty.” The qtnFDIA_1.1 (36.9 Mbp) was identiﬁed in a diﬀerent region of chromosome 1 when compared with previous linkage analyses, where QTLs were located approximately at 11 Mbp (da Silva Linge et al., 2015); 27–28 Mbp Concerning pit weight (PW), the qtnPW_6.1, identiﬁed on chromosome 6, is close to the QTL (qSW6; 24.6 Mb) mapped in the interspeciﬁc cross between almond × peach population February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 13 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. (Donoso et al., 2016). Cao et al. Candidate Genes Furthermore, previous studies have shown that molybdopterin cofactor plays an important role in the metabolic control of avocado fruit growth and ﬁnal fruit size (Cowan et al., 2001) and ion/H+ exchanger genes (GO: ion transmembrane transport) were critical for providing pH regulation (Pittman, 2012). Lastly, among the CG, Prupe.4G262200 and Prupe.4G261900 coding for endopolygalacturonases (GO: polygalacturonase activity) were previously involved in the inheritance of fruit texture and ﬂesh adherence to the stone in peach (Peace et al., 2005; Gu et al., 2016). Several CGs detected in our study have already been reported for productivity and fruit-related traits in peach. Prupe.1G531600 (DAM5), Prupe.1G531700 (DAM6), Prupe.1G531500, Prupe.1G549600, Prupe.1G548000, Prupe.1G554100 were considered potential CG for bloom date in peach (Rawandoozi et al., 2020a). These genes are located within the hotspot region detected on chromosome 1 associated with BD and FF. Similarly, Prupe.3G163100 (18.2 Mbp) located in the genetic interval of the hotspot on chromosome 3 was previously associated with blush and RP (redness around the stone) in peach (Frett et al., 2014; Zhang et al., 2018b; Guo et al., 2020). The main hotspot on chromosome 4 (9.0–12.5 Mbp) collocates with: Prupe.4G186800, the major locus controlling fruit ripening (Pirona et al., 2013) and CG for fruit ﬂesh softening rate (Carrasco-Valenzuela et al., 2019); Prupe.4G179900, CG for RD and DAB (Elsadr et al., 2019); and Prupe.4G185800 and Prupe.4G187100 involved in anthocyanin biosynthesis and CGs for blush (Rawandoozi et al., 2020b). In addition, Prupe_5G008400, a CG controlling fruit acidity, is located within the hotspot (0.3–3.7 Mbp) on chromosome 5 (Wang et al., 2020). Frontiers in Plant Science \| www.frontiersin.org REFERENCES genes associated with fruit domestication traits in peach. Plant Biotechnol. J. 17, 1954–1970. doi: 10.1111/pbi.13112 genes associated with fruit domestication traits in peach. Plant Biotechnol. J. 17, 1954–1970. doi: 10.1111/pbi.13112 Abdelghafar, A., da Silva Linge, C., Okie, W. R., and Gasic, K. (2020). Mapping Qtls for phytochemical compounds and fruit quality in peach. Mol. Breed. 40: 32. doi: 10.1007/s11032-020-01114-y Cao, K., Zhou, Z., Wang, Q., Guo, J., Zhao, P., Zhu, G., et al. (2016). Genome- wide association study of 12 agronomic traits in peach. Nat. Commun. 7:13246. doi: 10.1038/ncomms13246 Abidi, W., Jiménez, S., Moreno, M. Á, and Gogorcena, Y. (2011). Evaluation of antioxidant compounds and total sugar content in a nectarine [Prunus persica (L.) Batsch] progeny. Int. J. Mol. Sci. 12, 6919–6935. doi: 10.3390/ijms12106919 Carrasco-Valenzuela, T., Muñoz-Espinoza, C., Riveros, A., Pedreschi, R., Arús, P., Campos-Vargas, R., et al. (2019). Expression QTL (eQTLs) analyses aeveal candidate genes associated with fruit ﬂesh softening rate in peach [Prunus persica (L.) Batsch]. Front. Plant Sci. 10:1581. doi: 10.3389/fpls.2019. 01581 Bielenberg, D. G., Rauh, B., Fan, S., Gasic, K., Abbott, A. G., Reighard, G. L., et al. (2015). Genotyping by sequencing for SNP-based linkage map construction and QTL analysis of chilling requirement and bloom date in peach [Prunus persica (L.) Batsch]. PLoS One 10:e0139406. doi: 10.1371/journal.pone.0139406 Chang, C. C., Chow, C. C., Tellier, L. C., Vattikuti, S., Purcell, S. M., and Lee, J. J. (2015). Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience 4:7. doi: 10.1186/s13742-015-0047-8 Boudehri, K., Belka, M. A., Cardinet, G., Capdeville, G., Renaud, C., Tauzin, Y., et al. (2009). Toward the isolation of the d gene controlling the acidity of peach fruit by positional cloning. Acta Hortic. 814, 507–510. doi: 10.17660/ ActaHortic.2009.814.85 Ciacciulli, A., Cirilli, M., Chiozzotto, R., Attanasio, G., da Silva Linge, C., Pacheco, I., et al. (2018). Linkage and association mapping for the slow softening (SwS) trait in peach (P. persica L. Batsch) fruit. Tree Genet. Genom. 14, 1–10. doi: 10.1007/s11295-018-1305-6 Brachi, B., Morris, G. P., and Borevitz, J. O. (2011). Genome-wide association studies in plants: the missing heritability is in the ﬁeld. Genome Biol. 12:232. doi: 10.1186/gb-2011-12-10-232 Cirilli, M., Bassi, D., and Ciacciulli, A. (2016). Sugars in peach fruit: a breeding perspective. Hortic. Res. 3:15067. doi: 10.1038/hortres.2015.67 Covarrubias-Pazaran, G. (2016). Genome-assisted prediction of quantitative traits Covarrubias-Pazaran, G. (2016). FUNDING This work was funded by USDA’s National Institute of Food and Agriculture-Specialty Crop Research Initiative Projects, “RosBREED: Enabling marker-assisted breeding in Rosaceae” (2009-51181-05858) and “RosBREED: Combining disease resistance and horticultural quality in new rosaceous cultivars” (2014-51181-22378). This work was funded by USDA’s National Institute of Food and Agriculture-Specialty Crop Research Initiative Projects, “RosBREED: Enabling marker-assisted breeding in Rosaceae” (2009-51181-05858) and “RosBREED: Combining disease resistance and horticultural quality in new rosaceous cultivars” (2014-51181-22378). DATA AVAILABILITY STATEMENT The authors would like to thank Terrence Frett, Ralph Burrell, and Musser Fruit Research Farm staﬀat Clemson University; Natalie Anderson, Pamela Hornby, Tim Hartman, and Silvia Carpendo at Texas A&M University; and Paul Sandefur at University of Arkansas and Arkansas System Division of Agriculture Fruit Research Station staﬀfor their help with orchard maintenance and phenotypic data acquisition. The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: www.rosaceae.org, tfGDR1048/b. SUPPLEMENTARY MATERIAL CS: formal analysis and writing – original draft. LC: SNP data curation and review. WF: candidate genes analysis. MW, JC, and DB: resources and writing – review & editing. ZR: phenotypic analysis and review. KG: conceptualization, funding acquisition, The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2021. 644799/full#supplementary-material CONCLUSION We successfully performed a multi-locus GWAS using mrMLM 4.0 and FarmCPU in 620 individuals from three public fresh market peach breeding programs. A total of 88 reliable QTNs were consistently detected in at least two seasons and/or in February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org Frontiers in Plant Science \| www.frontiersin.org 14 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. resources, supervision, and writing – review & editing. All authors have read and approved the ﬁnal manuscript. diﬀerent methods. Hotspots for quality traits were identiﬁed on chromosomes 1, 3, 4, 5, 6, and 8. Candidate genes for quality traits were identiﬁed in the vicinity of the reliable QTNs detected using mrMLM 4.0 and FarmCPU. Furthermore, we observed that the position of the previously reported candidate genes for fruit-related traits (BD, Blush, DAB, ADH, RP, pH, and TA) matched with the position of the hotspots detected on chromosomes 1, 3, 4, and 5. Therefore, the information reported in this study supports the development of DNA tools for MAS in peach. Moreover, the importance of chromosome 4 hotspot in breeding for improvement of fruit quality is reinforced, and also emphasized the necessity to further study the hotspot regions on chromosomes 1, 3, 5, 6, and 8. REFERENCES doi: 10.1007/ s10681-015-1445-9 Jung, S., Lee, T., Cheng, C.-H., Buble, K., Zheng, P., Yu, J., et al. (2018). 15 years of GDR: Nnew data and functionality in the genome database for Rosaceae. Nucleic Acids Res. 47, D1137–D1145. doi: 10.1093/nar/gky1000 Lambert, P., Campoy, J. A., Pacheco, I., Mauroux, J.-B., da Silva Linge, C., Micheletti, D., et al. (2016). Identifying SNP markers tightly associated with six major genes in peach [Prunus persica (L.) Batsch] using a high-density SNP array with an objective of marker-assisted selection (MAS). Tree Genet. Genom. 12:121. doi: 10.1007/s11295-016-1080-1 Elsadr, H., Sherif, S., Banks, T., Somers, D., and Jayasankar, S. (2019). Reﬁning the genomic region containing a major locus controlling fruit maturity in peach. Sci. Rep. 9:7522. Fan, S., Bielenberg, D. G., Zhebentyayeva, T. N., Reighard, G. L., Okie, W. R., Holland, D., et al. (2010). Mapping quantitative trait loci associated with chilling requirement, heat requirement and bloom date in peach (Prunus persica). N. Phytol. 185, 917–930. doi: 10.1111/j.1469-8137.2009.03119.x Laurens, F., Aranzana, M. J., Arus, P., Bassi, D., Bink, M., Bonany, J., et al. (2018). An integrated approach for increasing breeding eﬃciency in apple and peach in Europe. Horticult. Res. 5:11. doi: 10.1038/s41438-018-0016-3 Farinati, S., Rasori, A., Varotto, S., and Bonghi, C. (2017). Rosaceae fruit development, ripening and post-harvest: an epigenetic perspective. Front. Plant Sci. 8:1247. doi: 10.3389/fpls.2017.01247 Li, Y., Cao, K., Zhu, G., Fang, W., Chen, C., Wang, X., et al. (2019). Genomic analyses of an extensive collection of wild and cultivated accessions provide new insights into peach breeding history. Genome Biol. 20:36. doi: 10.1186/s13059- 019-1648-9 Font I Forcada, C., Guajardo, V., Chin-Wo, S. R., and Moreno, M. Á (2019). Association mapping analysis for fruit quality traits in Prunus persica using SNP markers. Front. Plant Sci. 9:2005. doi: 10.3389/fpls.2018.02005 Ligges, U., and Mächler, M. (2002). Scatterplot 3d – an R Package for Visualizing Multivariate Data. Available online at: http://hdl.handle.net/10419/77160 Food and Agricultural Organization of the United Nations (FAOSTAT) (2018). Food Balance Sheets [Online]Food and Agricultural Organization of the United Nations (FAOSTAT), 2018. Available online at: http://www.fao.org/faostat/en/ #data/FBS (accessed November, 2020). Lipka, A. E., Tian, F., Wang, Q., Peiﬀer, J., Li, M., Bradbury, P. J., et al. (2012). GAPIT: genome association and prediction integrated tool. Bioinformatics 28, 2397–2399. doi: 10.1093/bioinformatics/bts444 Liu, X., Huang, M., Fan, B., Buckler, E. S., and Zhang, Z. (2016). Iterative usage of ﬁxed and random eﬀect models for powerful and eﬃcient genome- wide association studies. REFERENCES Genome-assisted prediction of quantitative traits using the r package sommer PLoS One 11:e0156744 doi: 10 1371/journal pone Bradbury, P. J., Zhang, Z., Kroon, D. E., Casstevens, T. M., Ramdoss, Y., and Buckler, E. S. (2007). TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23, 2633–2635. doi: 10.1093/ bioinformatics/btm308 Covarrubias-Pazaran, G. (2016). Genome-assisted prediction of quantitative traits using the r package sommer. PLoS One 11:e0156744. doi: 10.1371/journal.pone. 0156744 using the r package sommer. PLoS One 11:e0156744. doi: 10.1371/journal.pone. 0156744 Cowan, A. K., Cripps, R. F., Richings, E. W., and Taylor, N. J. (2001). Fruit size: towards an understanding of the metabolic control of fruit growth using avocado as a model system. Physiol. Plant. 111, 127–136. doi: 10.1034/j.1399- 3054.2001.1110201.x Cantín, C. M., Crisosto, C. H., Ogundiwin, E. A., Gradziel, T., Torrents, J., Moreno, M. A., et al. (2010). Chilling injury susceptibility in an intra-speciﬁc peach [Prunus persica (L.) Batsch] progeny. Postharvest Biol. Technol. 58, 79–87. doi: 10.1016/j.postharvbio.2010.06.002 da Silva Linge, C., Antanaviciute, L., Abdelghafar, A., Arús, P., Bassi, D., Rossini, L., et al. (2018). High-density multi-population consensus genetic linkage map for peach. PLoS One 13:e0207724. doi: 10.1371/journal.pone.0207724 Cao, K., Li, Y., Deng, C. H., Gardiner, S. E., Zhu, G., Fang, W., et al. (2019). Comparative population genomics identiﬁed genomic regions and candidate February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 15 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. ﬂesh texture and stone adhesion in peach. J. Exp. Bot. 67, 1993–2005. doi: 10.1093/jxb/erw021 da Silva Linge, C., Bassi, D., Bianco, L., Pacheco, I., Pirona, R., and Rossini, L. (2015). Genetic dissection of fruit weight and size in an F2 peach (Prunus persica (L.) Batsch) progeny. Mol. Breed. 35:271. doi: 10.1007/s11032-015-0271-z ﬂesh texture and stone adhesion in peach. J. Exp. Bot. 67, 1993–2005. doi: 10.1093/jxb/erw021 Guo, J., Cao, K., Deng, C., Li, Y., Zhu, G., Fang, W., et al. (2020). An integrated peach genome structural variation map uncovers genes associated with fruit traits. Genome Biol. 21:258. doi: 10.1186/s13059-020-02169-y Decros, G., Beauvoit, B., Colombié, S., Cabasson, C., Bernillon, S., Arrivault, S., et al. (2019). Regulation of pyridine nucleotide metabolism during tomato fruit development through transcript and protein proﬁling. Front. Plant Sci. 10:1201. doi: 10.3389/fpls.2019.01201 Hammer, O., Harper, D., and Ryan, P. (2001). PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4, 1–9. REFERENCES Desnoues, E., Baldazzi, V., Génard, M., Mauroux, J.-B., Lambert, P., Confolent, C., et al. (2016). Dynamic QTLs for sugars and enzyme activities provide an overview of genetic control of sugar metabolism during peach fruit development. J. Exp. Bot. 67, 3419–3431. doi: 10.1093/jxb/erw169 Hanana, M., Cagnac, O., Yamaguchi, T., Hamdi, S., Ghorbel, A., and Blumwald, E. (2007). A grape berry (Vitis vinifera L.) cation/proton antiporter is associated with berry ripening. Plant Cell Physiol. 48, 804–811. doi: 10.1093/pcp/pcm048 Hernández Mora, J. R., Micheletti, D., Bink, M., Van de Weg, E., Cantín, C., Nazzicari, N., et al. (2017). Integrated QTL detection for key breeding traits in multiple peach progenies. BMC Genom. 18:404. doi: 10.1186/s12864-017- 3783-6 Ding, Q., Wang, F., Xue, J., Yang, X., Fan, J., Chen, H., et al. (2020). Identiﬁcation and expression analysis of hormone biosynthetic and metabolism genes in the 2OGD family for identifying genes that may be involved in tomato fruit ripening. Int. J. Mol. Sci. 21:5344. doi: 10.3390/ijms2115 5344 Huang, M., Liu, X., Zhou, Y., Summers, R. M., and Zhang, Z. (2018). BLINK: a package for the next level of genome-wide association studies with both individuals and markers in the millions. GigaScience 8:giy154. doi: 10.1093/ gigascience/giy154 Donoso, J. M., Picañol, R., Serra, O., Howad, W., Alegre, S., Arús, P., et al. (2016). Exploring almond genetic variability useful for peach improvement: mapping major genes and QTLs in two interspeciﬁc almond × peach populations. Mol. Breed. 36:16. doi: 10.1007/s11032-016-0441-7 g g g y Iezzoni, A., Weebadde, C., Luby, J., Chengyan, Y., van de Weg, E., Fazio, G., et al. (2010). RosBREED: enabling marker-assisted breeding in rosaceae. Acta Hortic. 859, 389–394. Eduardo, I., Pacheco, I., Chietera, G., Bassi, D., Pozzi, C., Vecchietti, A., et al. (2011). QTL analysis of fruit quality traits in two peach intraspeciﬁc populations and importance of maturity date pleiotropic eﬀect. Tree Genet. Genom. 7, 323–335. doi: 10.1007/s11295-010-0334-6 Iezzoni, A. F., McFerson, J., Luby, J., Gasic, K., Whitaker, V., Bassil, N., et al. (2020). RosBREED: bridging the chasm between discovery and application to enable DNA-informed breeding in rosaceous crops. Horticult. Res. 7:177. doi: 10.1038/s41438-020-00398-7 Eduardo, I., Picañol, R., Rojas, E., Batlle, I., Howad, W., Aranzana, M. J., et al. (2015). Mapping of a major gene for the slow ripening character in peach: co-location with the maturity date gene and development of a candidate gene- based diagnostic marker for its selection. Euphytica 205, 627–636. REFERENCES PLoS Genet. 12:e1005767. doi: 10.1371/journal.pgen. 1005767 Fresnedo-Ramírez, J., Bink, M. C. A. M., van de Weg, E., Famula, T. R., Crisosto, C. H., Frett, T. J., et al. (2015). QTL mapping of pomological traits in peach and related species breeding germplasm. Mol. Breed. 35:166. doi: 10.1007/s11032- 015-0357-7 Martínez-García, P. J., Parﬁtt, D. E., Ogundiwin, E. A., Fass, J., Chan, H. M., Ahmad, R., et al. (2013). High density SNP mapping and QTL analysis for fruit quality characteristics in peach (Prunus persica L.). Tree Genet. Genom. 9, 19–36. doi: 10.1007/s11295-012-0522-7 Fresnedo-Ramírez, J., Frett, T. J., Sandefur, P. J., Salgado-Rojas, A., Clark, J. R., Gasic, K., et al. (2016). QTL mapping and breeding value estimation through pedigree-based analysis of fruit size and weight in four diverse peach breeding programs. Tree Genet. Genom. 12, 25. doi: 10.1007/s11295-016-0985-z Meneses, C., and Orellana, A. (2013). Using genomics to improve fruit quality. Biol. Res. 46, 347–352. Frett, T. J., Reighard, G. L., Okie, W. R., and Gasic, K. (2014). Mapping quantitative trait loci associated with blush in peach [Prunus persica (L.) Batsch]. Tree Genet. Genom. 10, 367–381. doi: 10.1007/s11295-013-0692-y Micheletti, D., Dettori, M. T., Micali, S., Aramini, V., Pacheco, I., da Silva Linge, C., et al. (2015). Whole-genome analysis of diversity and SNP-major gene association in peach germplasm. PLoS One 10:e0136803. doi: 10.1371/journal. pone.0136803 García-Gómez, B. E., Ruiz, D., Salazar, J. A., Rubio, M., Martínez-García, P. J., and Martínez-Gómez, P. (2020). Analysis of metabolites and gene expression changes relative to apricot (Prunus armeniaca L.) fruit quality during development and ripening. Front. Plant Sci. 11:1269. doi: 10.3389/fpls.2020. 01269 Minas, I. S., Tanou, G., and Molassiotis, A. (2018). Environmental and orchard bases of peach fruit quality. Sci. Horticult. 235, 307–322. doi: 10.1016/j.scienta. 2018.01.028 Gogorcena, Y., Sánchez, G., Moreno-Vázquez, S., Pérez, S., and Ksouri, N. (2020). “Genomic-based breeding for climate-smart peach varieties,” in Genomic Designing of Climate-Smart Fruit Crops, ed. C. Kole (Cham: Springer International Publishing), 271–331. Nuñez-Lillo, G., Balladares, C., Pavez, C., Urra, C., Sanhueza, D., Vendramin, E., et al. (2019). High-density genetic map and QTL analysis of soluble solid content, maturity date, and mealiness in peach using genotyping by sequencing. Sci. Horticult. 257:108734. doi: 10.1016/j.scienta.2019.108734 g Gu, C., Wang, L., Wang, W., Zhou, H., Ma, B., Zheng, H., et al. (2016). Copy number variation of a gene cluster encoding endopolygalacturonase mediates Nuñez-Lillo, G., Cifuentes-Esquivel, A., Troggio, M., Micheletti, D., Infante, R., Campos-Vargas, R., et al. (2015). REFERENCES BMC Plant Biol. 13:166. doi: 10.1186/1471- 2229-13-166 Verde, I., Jenkins, J., Dondini, L., Micali, S., Pagliarani, G., Vendramin, E., et al. (2017). The peach v2.0 release: high-resolution linkage mapping and deep resequencing improve chromosome-scale assembly and contiguity. BMC Genom. 18:225. doi: 10.1186/s12864-017-3606-9 Pittman, J. (2012). Multiple transport pathways for mediating intracellular pH homeostasis: the contribution of H+/ion exchangers. Front. Plant Sci. 3:11. doi: 10.3389/fpls.2012.00011 Quilot, B., Wu, B. H., Kervella, J., Génard, M., Foulongne, M., and Moreau, K. (2004). QTL analysis of quality traits in an advanced backcross between Prunus persica cultivars and the wild relative species P. davidiana. Theor. Appl. Genet. 109, 884–897. doi: 10.1007/s00122-004-1703-z Wang, L., Jiang, X., Zhao, L., Wang, F., Liu, Y., Zhou, H., et al. (2020). A candidate PpRPH gene of the D locus controlling fruit acidity in peach. Plant Mol. Biol. 105, 1–12. doi: 10.1007/s11103-020-01089-6 Wang, S.-B., Feng, J.-Y., Ren, W.-L., Huang, B., Zhou, L., Wen, Y.-J., et al. (2016). Improving power and accuracy of genome-wide association studies via a multi-locus mixed linear model methodology. Sci. Rep. 6:19444. doi: 10.1038/ srep19444 Raj, A., Stephens, M., and Pritchard, J. K. (2014). fastSTRUCTURE: variational inference of population structure in large SNP data sets. Genetics 197, 573–589. doi: 10.1534/genetics.114.164350 Wen, Y.-J., Zhang, H., Ni, Y.-L., Huang, B., Zhang, J., Feng, J.-Y., et al. (2017). Methodological implementation of mixed linear models in multi-locus genome-wide association studies. Brief. Bioinform. 19, 700–712. doi: 10.1093/ bib/bbw145 Rawandoozi, Z. J., Hartmann, T. P., Carpenedo, S., Gasic, K., da Silva Linge, C., Cai, L., et al. (2020a). Mapping and characterization qtls for phenological traits in seven pedigree-connected peach families. Res. Square [preprint]. doi: 10.21203/rs.3.rs-75550/v1 Xu, Y., Yang, T., Zhou, Y., Yin, S., Li, P., Liu, J., et al. (2018). Genome- wide association mapping of starch pasting properties in maize using single- locus and multi-locus models. Front. Plant Sci. 9:1311. doi: 10.3389/fpls.2018. 01311 Rawandoozi, Z. J., Hartmann, T. P., Carpenedo, S., Gasic, K., da Silva Linge, C., Cai, L., et al. (2020b). Identiﬁcation and characterization of QTLs for fruit quality traits in peach through a multi-family approach. BMC Genom. 21:522. doi: 10.1186/s12864-020-06927-x Yamamoto, T., Shimada, T., Imai, T., Yaegaki, H., Haji, T., Matsuta, N., et al. (2001). Characterization of morphological traits based on a genetic linkage map in peach. Breed. Sci. 51, 271–278. doi: 10.1270/jsbbs.51.271 Ren, W.-L., Wen, Y.-J., Dunwell, J. M., and Zhang, Y.-M. (2018). REFERENCES Identiﬁcation of candidate genes associated February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 16 Fruit Quality Hotspots in Peach Genome da Silva Linge et al. with mealiness and maturity date in peach [Prunus persica (L.) Batsch] using QTL analysis and deep sequencing. Tree Genet. Genomes 11:86. doi: 10.1007/ s11295-015-0911-9 population structure, and linkage disequilibrium in Brazilian peach breeding germplasm. Tree Genet. Genom. 16:10. doi: 10.1007/s11295-019-1406-x Vanderzande, S., Howard, N. P., Cai, L., da Silva Linge, C., Antanaviciute, L., Bink, M. C. A. M., et al. (2019). High-quality, genome-wide SNP genotypic data for pedigreed germplasm of the diploid outbreeding species apple, peach, and sweet cherry through a common workﬂow. PLoS One 14:e0210928. doi: 10.1371/journal.pone.0210928 Olmstead, M. A., Gilbert, J., Colquhoun, T., Clark, D. G., and Sims, C. (2015). In pursuit of the perfect peach: Consumer-assisted selection of peach fruit traits. HortScience 50, 1202–1212. Peace, C. P., Crisosto, C. H., and Gradziel, T. M. (2005). Endopolygalacturonase: a candidate gene for freestone and melting ﬂeshin peach. Mol. Breed. 16, 21–31. doi: 10.1007/s11032-005-0828-3 Vendramin, E., Pea, G., Dondini, L., Pacheco, I., Dettori, M. T., Gazza, L., et al. (2014). A unique mutation in a MYB gene cosegregates with the nectarine phenotype in peach. PLoS One 9:e90574. doi: 10.1371/journal.pone.00 90574 Peace, C. P., Luby, J. J., van de Weg, W. E., Bink, M. C. A. M., and Iezzoni, A. F. (2014). A strategy for developing representative germplasm sets for systematic QTL validation, demonstrated for apple, peach, and sweet cherry. Tree Genet. Genom. 10, 1679–1694. doi: 10.1007/s11295-014-0788-z Verde, I., Abbott, A. G., Scalabrin, S., Jung, S., Shu, S., Marroni, F., et al. (2013). The high-quality draft genome of peach (prunus persica) identiﬁes unique patterns of genetic diversity, domestication and genome evolution. Nat. Genet. 45, 487–494. doi: 10.1038/ng.2586 Peakall, R., and Smouse, P. E. (2012). GenAlEx 6.5: genetic analysis in Excel. Population genetic software for teaching and research–an update. Bioinformatics 28, 2537–2539. doi: 10.1093/bioinformatics/bts460 Verde, I., Bassil, N., Scalabrin, S., Gilmore, B., Lawley, C. T., Gasic, K., et al. (2012). Development and evaluation of a 9k SNP array for peach by internationally coordinated snp detection and validation in breeding germplasm. PLoS One 7:e35668. doi: 10.1371/journal.pone.0035668 Pirona, R., Eduardo, I., Pacheco, I., da Silva Linge, C., Miculan, M., Verde, I., et al. (2013). Fine mapping and identiﬁcation of a candidate gene for a major locus controlling maturity date in peach. REFERENCES pKWmEB: integration of Kruskal-Wallis test with empirical Bayes under polygenic background control for multi-locus genome-wide association study. Heredity 120, 208–218. doi: 10.1038/s41437-017-0007-4 Young, M. D., Wakeﬁeld, M. J., Smyth, G. K., and Oshlack, A. (2010). Gene ontology analysis for RNA-seq: accounting for selection bias. Genome Biol. 11:R14. doi: 10.1186/gb-2010-11-2-r14 Romeu, J. F., Monforte, A. J., Sánchez, G., Granell, A., García-Brunton, J., Badenes, M. L., et al. (2014). Quantitative trait loci aﬀecting reproductive phenology in peach. BMC Plant Biol. 14:52. doi: 10.1186/1471-22 29-14-52 Yu, J., Pressoir, G., Briggs, W. H., Vroh Bi, I., Yamasaki, M., Doebley, J. F., et al. (2006). A uniﬁed mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat. Genet. 38, 203–208. doi: 10.1038/ng1702 Rosenberg, N. A. (2004). distruct: a program for the graphical display of population structure. Mol. Ecol. Notes 4, 137–138. doi: 10.1046/j.1471-8286.2003.00566.x Yue, C., Gallardo, K. R., Luby, J. J., Rihn, A. L., McFerson, J. R., McCracken, V., et al. (2014). An evaluation of U.S. peach producers’ trait prioritization: evidence from audience surveys. HortSci. Horts 49, 1309–1314. doi: 10.21273/HORTSCI. 49.10.1309 Serra, O., Giné-Bordonaba, J., Eduardo, I., Bonany, J., Echeverria, G., Larrigaudière, C., et al. (2017). Genetic analysis of the slow-melting ﬂesh character in peach. Tree Genet. Genom. 13:77. doi: 10.1007/s11295-017-1160-x Zeballos, J. L., Abidi, W., Giménez, R., Monforte, A. J., Moreno, M. Á, and Gogorcena, Y. (2016). Mapping QTLs associated with fruit quality traits in peach [Prunus persica (L.) Batsch] using SNP maps. Tree Genet. Genom. 12:37. doi: 10.1007/s11295-016-0996-9 Shi, P., Xu, Z., Zhang, S., Wang, X., Ma, X., Zheng, J., et al. (2020). Construction of a high-density SNP-based genetic map and identiﬁcation of fruit-related QTLs and candidate genes in peach [Prunus persica (L.) Batsch]. BMC Plant Biol. 20:438. doi: 10.1186/s12870-020-02557-3 Steﬀen, A., Elgner, M., and Staiger, D. (2019). Regulation of ﬂowering time by the RNA-binding proteins AtGRP7 and AtGRP8. Plant Cell Physiol. 60, 2040–2050. doi: 10.1093/pcp/pcz124 Zhang, C., Dong, S.-S., Xu, J.-Y., He, W.-M., and Yang, T.-L. (2018a). PopLDdecay: a fast and eﬀective tool for linkage disequilibrium decay analysis based on variant call format ﬁles. Bioinformatics 35, 1786–1788. doi: 10.1093/ bioinformatics/bty875 Supek, F., Bošnjak, M., Škunca, N., and Šmuc, T. (2011). REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS One 6:e21800. doi: 10.1371/ journal.pone.0021800 Zhang, C., Ma, R., Xu, J., Yan, J., Guo, L., Song, J., et al. (2018b). Zhu, C., Gore, M., Buckler, E. S., and Yu, J. (2008). Status and prospects of association mapping in plants. Plant Genome 1, 5–20. doi: 10.3835/ plantgenome2008.02.0089 REFERENCES Genome-wide identiﬁcation and classiﬁcation of MYB superfamily genes in peach. PLoS One 13:e0199192. doi: 10.1371/journal.pone.0199192 Tamba, C. L., Ni, Y.-L., and Zhang, Y.-M. (2017). Iterative sure independence screening EM-Bayesian LASSO algorithm for multi-locus genome-wide association studies. PLoS Comput. Biol. 13:e1005357. doi: 10.1371/journal.pcbi. 1005357 Zhang, F., Liu, X., Zhang, A., Jiang, Z., Chen, L., and Zhang, X. (2019). Genome- wide dynamic network analysis reveals a critical transition state of ﬂower development in Arabidopsis. BMC Plant Biol. 19:11. doi: 10.1186/s12870-018- 1589-6 Tamba, C. L., and Zhang, Y.-M. (2018). A fast mrMLM algorithm for multi- locus genome-wide association studies. bioRxiv [Preprint] doi: 10.1101/341784 bioRxiv:341784 Zhang, J., Feng, J. Y., Ni, Y. L., Wen, Y. J., Niu, Y., Tamba, C. L., et al. (2017). pLARmEB: integration of least angle regression with empirical Bayes for multilocus genome-wide association studies. Heredity 118, 517–524. doi: 10. 1038/hdy.2017.8 Thurow, L. B., Gasic, K., Bassols Raseira, M. D. C., Bonow, S., and Marques Castro, C. (2019). Genome-wide SNP discovery through genotyping by sequencing, Frontiers in Plant Science \| www.frontiersin.org Frontiers in Plant Science \| www.frontiersin.org February 2021 \| Volume 12 \| Article 644799 17 da Silva Linge et al. Fruit Quality Hotspots in Peach Genome Zhang, X., Hina, A., Song, S., Kong, J., Bhat, J. A., and Zhao, T. (2019). Whole- genome mapping identiﬁed novel “QTL hotspots regions” for seed storability in soybean (Glycine max L.). BMC Genom. 20:499. doi: 10.1186/s12864-019- 5897-5 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Zhang, Y.-W., Tamba, C. L., Wen, Y.-J., Li, P., Ren, W.-L., Ni, Y.-L., et al. (2020). mrMLM v4.0: an R platform for multi-locus genome-wide association studies. Genom. Proteom. Bioinform. 18. doi: 10.1016/j.gpb.2020. 06.006 Copyright © 2021 da Silva Linge, Cai, Fu, Clark, Worthington, Rawandoozi, Byrne and Gasic. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Zhu, C., Gore, M., Buckler, E. S., and Yu, J. (2008). Status and prospects of association mapping in plants. Plant Genome 1, 5–20. Zhang, X., Hina, A., Song, S., Kong, J., Bhat, J. A., and Zhao, T. (2019). Whole- genome mapping identiﬁed novel “QTL hotspots regions” for seed storability in soybean (Glycine max L.). BMC Genom. 20:499. doi: 10.1186/s12864-019- 5897-5 Frontiers in Plant Science \| www.frontiersin.org REFERENCES doi: 10.3835/ plantgenome2008.02.0089 February 2021 \| Volume 12 \| Article 644799 Frontiers in Plant Science \| www.frontiersin.org 18
https://openalex.org/W4249857451	https://europepmc.org/articles/pmc3358195?pdf=render	English	null	Methicillin-Resistant<i>Staphylococcus aureus</i>, Samoa, 2007–2008	Emerging infectious diseases	2,011	cc-by	5,846	Methicillin-Resistant Staphylococcus aureus, Samoa, 2007–2008 James Alesana-Slater, Stephen R. Ritchie, Helen Heffernan, Tracy Camp, Alice Richardson, Peter Herbison, and Pauline Norris Little is known about the epidemiology of methicillin- resistant Staphylococcus aureus (MRSA) in most Paciﬁ c Island nations. Relatively high rates of MRSA have been reported in Polynesian people living outside the Paciﬁ c Islands. To determine the prevalence and characteristics of MRSA, we assessed wound swabs from 399 persons with skin and soft tissue infection living in Samoa. MRSA was isolated from 9% of study participants; 34 of the 196 S. aureus isolates were MRSA. Five MRSA genotypes were identiﬁ ed; the 3 most common were USA300, the Queensland clone, and a sequence type 1 MRSA strain that shares <85% homology with the sequence type 1 MRSA strain common in the region (WA MRSA-1). The Southwest Paciﬁ c MRSA clone was identiﬁ ed but accounted for only 12% of MRSA isolates. The high prevalence of MRSA in Samoa provides impetus for initiatives to improve antimicrobial drug resistance surveillance, infection control, and antimicrobial drug use in Paciﬁ c Island nations. The prevalence and genetic variation of MRSA infection in most Paciﬁ c Island nations remain unknown. The only study of MRSA in the Paciﬁ c was conducted after a report in 2004 of an increased incidence of MRSA infection in Polynesian people in Hawaii (5). This study showed that most MRSA infections in Hawaii were caused by the USA300 MRSA strain (6). ( ) Relatively high rates of MRSA infection have been reported in Polynesian people living outside the Paciﬁ c Islands region, including in Alaska, Australia, and New Zealand, and have been attributed to infection with the Southwest Paciﬁ c clone of MRSA (sequence type [ST] 30 SCCmec type IV, also referred to as Western Samoan phage pattern [WSPP] MRSA or the Oceania strain) (7– 10). The Southwest Paciﬁ c clone of community-associated MRSA was identiﬁ ed in New Zealand in 1992 (11). The ﬁ rst isolate and many subsequent isolates of this strain were from persons in New Zealand who had some association with Samoa. The Southwest Paciﬁ c MRSA clone has now spread as far as Europe and South America (4,12). This MRSA strain has been postulated to have arisen from a pandemic penicillin-resistant S. aureus strain, known as phage type 80/81, that caused serious hospital- and community-acquired infections during the 1950s (13). Author afﬁ liations: University of Otago, Dunedin, New Zealand (J. Alesana-Slater, P. Herbison, P. Norris); University of Auckland, Auckland, New Zealand (S.R. Ritchie); Institute of Environmental Science and Research Limited, Wellington, New Zealand (H. Heffernan, A. Richardson); and Auckland City Hospital, Auckland (T. Camp) Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 Methods During the summer of 2007–2008, a study investigator (J.A.S.) collected S. aureus isolates from persons with skin and soft tissue infection. Review of patient notes was used to identify hospital patients who had skin lesions. All ambulatory patients who attended a clinic during the visit by J.A.S. and any family members who accompanied them were asked if they had a skin lesion. All persons with a skin lesion were invited to participate in the study; <5 persons declined to participate. At hospitals, participants included inpatients and outpatients, but this information was not recorded. No attempt was made to categorize infections as hospital acquired or community acquired, although the source of the wound was recorded. Figure. Map of Samoa, showing the 2 main islands, Upolu and Savai’i, and the capital Apia. Reproduced with permission from Oxford Cartographers (www.oxfordcartographers.com). All participants had skin and soft tissue infection, but this infection was not always the primary reason for contact with health services. Skin and soft tissue infection was deﬁ ned as increasing pain, warmth, induration, erythema, or purulent discharge. Swabs were collected twice from patients who had separate lesions on the upper and lower body. If a patient had >1 infected site in the same region (e.g., upper body), only the larger lesion was swabbed. referral hospital in Apia; in addition, Samoa has 7 district hospitals (14). Several outreach and integrated community health services provide primary health care services, such as clinics and vaccinations. Patient data collected included demographic information (age, sex, occupation, and village of residence) and recent exposure to health services (use of antimicrobial drugs in the last month and contact with health services in last 3 months). We attempted to include persons of different ages, both sexes, and various occupations. Samoa gained independence from New Zealand in 1962; a large Samoan population resides in New Zealand. In New Zealand, infections caused by S. aureus, whether methicillin susceptible or resistant, are more common in Maori and Polynesian persons than in other ethnic groups (15–17). Antimicrobial drug resistance is routinely monitored in New Zealand through testing of bacterial isolates from specimens taken for diagnostic purposes. In 2008, 9% of S. aureus isolates were methicillin resistant, but this rate varied considerably among different parts of the country, with rates as high as 13%–16% in hospitals in the Auckland area (18). Methicillin-Resistant Staphylococcus aureus, Samoa, 2007–2008 I nfections caused by methicillin-resistant Staphylococcus aureus (MRSA) have become a global health concern during the past 2–3 decades. The epidemiology of MRSA has demonstrated marked geographic variation in the prevalence and genotypes of MRSA (1,2), and recent reports from many parts of the world indicate that the prevalence and diversity of MRSA continue to increase (3,4). Studies of the global epidemiology frequently have not included MRSA obtained from persons living in developing nations. I Samoa is an independent nation in the Southwest Paciﬁ c, with a population of ≈180,000 persons (Samoan Statistics Department, www.spc.int/prism/wstest/index. htm; Figure). Samoa has a small and developing economy predominantly comprising remittances from Samoan persons living overseas, agriculture, and tourism. The gross domestic product per capita was US $2,987.90 in 2008 (http://data.un.org). Life expectancy in Samoa is 74.9 years for women and 68.5 years for men, and the infant mortality rate is ≈22.3 per 1,000 live-born infants. The publicly funded National Health Service is based at the national DOI: 10.3201/eid1706.101083 1023 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 RESEARCH Figure. Map of Samoa, showing the 2 main islands, Upolu and Savai’i, and the capital Apia. Reproduced with permission from Oxford Cartographers (www.oxfordcartographers.com). Methods Even though we did not categorize infections as hospital acquired or community acquired, we were able to infer likely place of acquisition from the description of the wound: 20 (5%) of 399 represented infections of surgical sites, which were hospital acquired; 335 (84%) represented community- acquired infections, such as impetigo; po’u; boils; or infections of wounds resulting from dog bites, machete cuts, or assaults. We were not able to further classify the remaining 44 (11%) infections. vancomycin in accordance with Clinical and Laboratory Standards Institute methods and interpretive standards (19–21). MRSA isolates were tested by disk diffusion for susceptibility to ciproﬂ oxacin and clindamycin, including inducible clindamycin resistance. Additional fusidic acid and mupirocin disk diffusion susceptibility testing was performed when required to assist with identiﬁ cation of MRSA strains. MRSA isolates were tested for the genes encoding for Panton-Valentine leukocidin (PVL) by PCR (22). Multilocus sequence typing (23); spa typing (24); and if necessary, pulsed-ﬁ eld gel electrophoresis (PFGE) of SmaI-digested genomic DNA were used to identify and characterize MRSA strains. StaphType software version 1.5 (Ridom GmbH, Würzburg, Germany) was used to assign spa types. We compared spa types by using the BURP (based upon repeat pattern) algorithm and excluding spa types with <5 repeats and setting a maximum cost of 4 between members of a spa group cluster. PFGE banding patterns were analyzed with BioNumerics software version 5.1 (Applied Maths, Saint-Martens-Latern, Belgium) by using the Dice coefﬁ cient and unweighted pair group method with arithmetic averages, at settings of 0.5% optimization and 1.5% position tolerance. S. aureus was isolated from 196 (46%, 95% CI 42%– 51%) of the 423 wound swabs and from 187 (47%, 95% CI 42%–52%) of the 399 study participants. S. aureus was more likely to be obtained from children <5 years of age than from persons in other age groups (37 [70%] of 53 compared with 150 [43%] of 346; p<0.001) and less likely in persons currently or recently treated with antimicrobial drugs (84 [38%] of 224, compared with 103 [59%] of 175 persons who were not; p<0.001) (Table 1). p p ) ( ) Thirty-four (17%, 95% CI 13%–23%) of the 196 S. aureus isolates were methicillin resistant. MRSA was isolated from 8% (95% CI 6%–11%) of the 423 wound swabs and from 9% (95% CI 6%–12%) of the 399 study participants. Almost 1 in 5 of the 187 participants with S. Methods The Lower South Ethics Committee in New Zealand, and the Health Research Council in Samoa gave ethical approval for the study. Participants were given an information sheet about the research and a small card explaining, in Samoan, how to take care of wounds. The study investigator explained the purpose of the study in Samoan, and the consent form was printed in Samoan and in English. In Samoa, as in many other developing countries, antimicrobial drug resistance is not systematically monitored. The microbiology service at the main public hospital in Apia performs antimicrobial drug susceptibility testing only when speciﬁ cally requested by the treating doctor. This testing is infrequent, and the results are not regularly collated (V. Kini, pers. comm.). Testing for antimicrobial drug susceptibility requires equipment, resources, and a high level of expertise and quality assurance—requirements that are beyond the means of many laboratories in the developing world. If a high rate of antimicrobial drug resistance exists in Samoa, it would substantially affect this relatively poor country with relatively high rates of infectious diseases. We performed a systematic study to assess the prevalence and characteristics of MRSA isolated from wound swabs from persons with skin and soft tissue infection living in Samoa. Swabs were taken after the wound was cleaned with sterile water. Swabs were placed into Amies transport media (Fort Richard Laboratories, Auckland, New Zealand), stored on ice, and transported to the laboratory at 1 of the 2 main hospitals. All swabs were processed within 12 hours after collection. Swabs were spread onto sheep blood agar containing aztreonam (6 mg/L) and incubated for 16–24 hours. A catalase test was performed on colonies suspected to be S. aureus, and catalase-positive colonies were identiﬁ ed by using a latex agglutination test (BBL Staphyloslide Latex Test, Becton Dickinson, Sparks, MD, USA). S. aureus isolates were transported on a nutrient agar slope to LabPLUS (Auckland, New Zealand). In New Zealand, antimicrobial drug susceptibility testing was performed by using disk diffusion for penicillin, and agar dilution breakpoint testing for cotrimoxazole, doxycycline, erythromycin, gentamicin, oxacillin, and Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1024 MRSA, Samoa boils and carbuncles. Many were described as po’u, a Samoan term for idiopathic tropical ulcer. Twenty-four persons each had swabs from 2 separate lesions. Methods aureus had MRSA (34/187, 18%, 95% CI 13%–24%). The basic demographic features of persons with MRSA did not differ substantially from those with methicillin-susceptible S. aureus (MSSA). Recent antimicrobial drug use and recent health care were not associated with an increase in MRSA infections. For example, the proportion of persons with recent health care exposure from whom MRSA was isolated (23/262, 9%) did not differ signiﬁ cantly from the proportion of persons without recent health care exposure (11/137, 8%). Categorical data are expressed as proportions and 95% conﬁ dence intervals (CIs), calculated by the modiﬁ ed Wald method. Statistical signiﬁ cance of comparative categorical data was examined by using the χ2 test. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1025 eus positive vs. study population >5 y of age. eus positive vs. no antimicrobial drug treatment in the previous month. n-susceptible Staphylococcus aureus, MRSA, methicillin-resistant S. aureus; CI, confidence interval. range) of study participants: All, 32 (0–89); S. aureus positive, 24 (0–84); MSSA positive, 26 (0–84); MRSA positive, 36 (0–76). eus positive vs. study population >5 y of age. Discussion Except for the MRSA isolates identiﬁ ed as type ST1 by multilocus sequence typing, all other MRSA belonged to a known MRSA strain (Table 2). Seven of the 9 ST1 isolates had indistinguishable PFGE proﬁ les. The ST1 MRSA isolates had 82% homology by PFGE typing with another ST1 MRSA strain common in the region (the Australian WA MRSA-1). In Samoa, children <5 years of age were more likely than persons in other age groups to have skin and soft tissue infection caused by S. aureus; however, the prevalence of MRSA infections were similar in children, adults, and elderly persons. Participants who had reported recent antimicrobial drug use were less likely to have S. aureus, but their prevalence of MRSA did not differ signiﬁ cantly from those who did not report recent antimicrobial drug use. The lack of association between MRSA prevalence and health care exposure or antimicrobial drug use was surprising; however, the data on antimicrobial drug use Antimicrobial drug resistance proﬁ le was associated with strain (Table 2). All isolates of the USA300 strain were ciproﬂ oxacin resistant, and most also were erythromycin resistant. All of the Queensland clone and Southwest Paciﬁ c clone MRSA, and 8 of the 9 ST1 MRSA isolates, were resistant only to β-lactams. Table 2. Strains identified among the MRSA isolates, Samoa, summer 2007–2008* No. (%) isolates Strain† MLST type No. (%) PVL-positive isolates spa type (% of strain) Antimicrobial drug susceptibility (% of strain) 10 (29) USA300 ST8 9 (90) t008 (100) Resistant to ciprofloxacin and erythromycin (80); resistant to ciprofloxacin (20) 9 (26) Queensland clone ST93 9 (100) t3949 (56), t202 (44)‡ Resistant only to ȕ-lactams (100) 9§ (26) – ST1 0 t1853 (78), t6080 (11)¶ Resistant only to ȕ-lactams (89); resistant to erythromycin (11) 4 (12) Southwest Pacific/ WSPP/Oceania clone ST30 4 (100) t019 (100) Resistant only to ȕ-lactams (100) 2 (6) AK3 ST5 0 t002 (50), t1265 (50)# Resistant to erythromycin** (50); resistant only to ȕ-lactams (50) MRSA, methicillin-resistant S. aureus; MLST, multilocus sequence typing; PVL, Panton-Valentine leukocidin; WSPP, Western Samoa Phage Pattern. †International MRSA strain designations except for AK3 which is a New Zealand designation for a community-associated MRSA strain common in New Table 2. Strains identified among the MRSA isolates, Samoa, summer 2007–2008 No. (%) isolates Strain† MLST type No. Discussion throughout Samoa and was isolated from persons from 27 villages. There did not appear to be any geographic clustering associated with residence close to the main public hospital in Apia, although MRSA was more likely to be isolated from residents of Upolu, the main island on which Apia is located, than from residents of Savai’i, the less-populated and less-developed island. MRSA was isolated from 11% of participants from Upolu, but from only 3% of participants from Savai’i. throughout Samoa and was isolated from persons from 27 villages. There did not appear to be any geographic clustering associated with residence close to the main public hospital in Apia, although MRSA was more likely to be isolated from residents of Upolu, the main island on which Apia is located, than from residents of Savai’i, the less-populated and less-developed island. MRSA was isolated from 11% of participants from Upolu, but from only 3% of participants from Savai’i. Considerable concern exists internationally about increasing levels of antimicrobial drug resistance (25). Consequences include treatment failure, the need to use newer antimicrobial drugs to achieve treatment goals, the adverse effects frequently associated with these newer drugs, increased expenditure, and longer hospital stays. In developing countries, such as Samoa, an additional concern exists: when resistance to less expensive antimicrobial drugs becomes widespread, the more expensive antimicrobial drugs are simply not available (26). As in many other developing countries (27), use of antimicrobial drugs is extremely high in Samoa. More than 60% of all prescriptions dispensed in hospitals and private pharmacies include an antimicrobial drug (28). No current data are available on the prevalence of antimicrobial drug resistance in Samoa. In other countries, high levels of antimicrobial drug use correlate with high levels of resistance (29). In addition, inadequate dosing and poor adherence by patients may also increase the development of resistance (25). Anecdotal evidence indicates these practices also are common in Samoa. Of the 153 MSSA isolates from 187 participants, 124 (81%) isolates were resistant to penicillin, but resistance to any other antimicrobial drug was uncommon. Of the 34 MRSA isolates, 22 were resistant only to β-lactams. Of the remaining 12 MRSA, 8 were ciproﬂ oxacin and erythromycin resistant, 2 were erythromycin resistant with inducible clindamycin resistance, and 2 were ciproﬂ oxacin resistant. All MRSA isolates were susceptible to cotrimoxazole, doxycycline, gentamicin, and vancomycin, and none had constitutive clindamycin resistance. Results A total of 423 swabs were obtained from skin and soft tissue infections of 399 persons at 8 locations on the 2 main islands of Samoa (Upolu and Savai’i) and a smaller island, Manono. The 8 locations were the main public hospital in Apia (the capital on Upolu), the main public hospital on Savai’i, 5 district hospitals around the main islands, and community-based clinics on Manono. Participants lived in 165 villages around Samoa. We detected S. aureus in participants from 102 villages. In terms of place of domicile, MRSA was widespread Infections in skin and soft tissue were in wounds from falls, cuts, dog bites, car accidents, and burns; as well as Table 1. Demographic characteristics of study participants and prevalence of MSSA and MRSA, Samoa, summer 2007–2008* Characteristic Total study population No. (%, 95% CI) S. aureus positive MSSA MRSA No. participants 399 187 (47, 42–52) 153 (38, 34–43) 34 (9, 6–12) Male sex 263 121 (46, 40–52) 95 (36, 31–42) 26 (10, 7–14) Age, y† <5 53 37 (70, 56–81)‡ 33 (62, 49–74) 4 (8, 3–18) 5–15 93 50 (54, 44–64) 44 (47, 38–57) 6 (7, 3–14) 16–59 195 79 (34–48) 60 (31, 25–38) 19 (10, 6–15) >60 58 21 (36, 25–49) 16 (28, 18–40) 5 (9, 3–19) Antimicrobial drug treatment in previous month 224 84 (38, 31–44)§ 67 (30, 24–36) 17 (8, 5–12) Health care contact in previous 3 months 262 108 (41, 35–47) 85 (32, 27–38) 23 (9, 6–13) MSSA, methicillin-susceptible Staphylococcus aureus, MRSA, methicillin-resistant S. aureus; CI, confidence interval. †Median age, y (range) of study participants: All, 32 (0–89); S. aureus positive, 24 (0–84); MSSA positive, 26 (0–84); MRSA positive, 36 (0–76). ‡p<0.001, S. aureus positive vs. study population >5 y of age. §p<0.001, S. aureus positive vs. no antimicrobial drug treatment in the previous month. Table 1. Demographic characteristics of study participants and prevalence of MSSA and MRSA, Samoa, su Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1025 RESEARCH Discussion (%) PVL-positive isolates spa type (% of strain) Antimicrobial drug susceptibility (% of strain) 10 (29) USA300 ST8 9 (90) t008 (100) Resistant to ciprofloxacin and erythromycin (80); resistant to ciprofloxacin (20) 9 (26) Queensland clone ST93 9 (100) t3949 (56), t202 (44)‡ Resistant only to ȕ-lactams (100) 9§ (26) – ST1 0 t1853 (78), t6080 (11)¶ Resistant only to ȕ-lactams (89); resistant to erythromycin (11) 4 (12) Southwest Pacific/ WSPP/Oceania clone ST30 4 (100) t019 (100) Resistant only to ȕ-lactams (100) 2 (6) AK3 ST5 0 t002 (50), t1265 (50)# Resistant to erythromycin* (50); resistant only to ȕ-lactams (50) MRSA, methicillin-resistant S. aureus; MLST, multilocus sequence typing; PVL, Panton-Valentine leukocidin; WSPP, Western Samoa Phage Pattern. †International MRSA strain designations, except for AK3, which is a New Zealand designation for a community-associated MRSA strain common in New Zealand. ‡spa type t3949 (11–17–23–17–17–17–16–16–25) is a single-repeat variant of t202 (11–17–23–17–17–16–16–25). The extra repeat is shown in boldface. §Only 8/9 ST1 MRSA isolates were available for PVL PCR, spa typing, and pulsed-field gel electrophoresis (PFGE) typing. Seven of the 8 isolates had indistinguishable PFGE profiles, and the eighth shared 87% homology. ¶spa type t6080 (07–23–21–17–13–13–34–16–13–33–13) is a single-repeat variant of t1853 (07–23–21–17–13–34–16–13–33–13). The extra repeat is shown in boldface. #spa type t1265 (26–23–17–34–17–20–17–12–12–12–16) is a variant of t002 (26–23–17–34–17–20–17–12–17–16). The differing repeats are shown in boldface. This isolate also had inducible clindamycin resistance. *This isolate also had inducible clindamycin resistance. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1026 MRSA, Samoa MRSA, Samoa USA400 ST1 MRSA strain with which they shared only 80% homology by PFGE. The Samoa ST1 MRSA strain might have originated in the Paciﬁ c. However, MRSA with spa type t1853 have also been isolated in New Zealand, mainly from patients in the Auckland area, since at least 2008 (34). may not be entirely reliable. A previous study found that many Samoans are unclear about which medicines were antimicrobial drugs (30). Most of the health care exposure reported in our study related to primary care exposure, and most of the MRSA strains isolated are typically associated with community acquisition (31). Even though most participants had community-onset skin and soft tissue infection, more detailed information about diagnosis and place of acquisition would have been valuable. Discussion Three of the 5 strains that we identiﬁ ed among MRSA isolates in Samoa (the USA300, Queensland, and Southwest Paciﬁ c strains) typically are associated with community acquisition (31). The Samoa ST1 MRSA strain is also likely to be predominantly associated with community- acquired infections. ST1 is recognized as a prominent genetic background of community-associated MRSA (35). The Samoa ST1 MRSA shares >80% homology by PFGE typing with 2 other ST1 community-associated MRSA strains: WA MRSA-1 and USA400. WA MRSA-1 caused the ﬁ rst cases of community-associated MRSA infection in previously healthy persons in Western Australia in the early 1990s, and USA400 was the strain isolated from the ﬁ rst cases of community-acquired MRSA in the United States (31). The Samoa ST1 MRSA strains are not multiresistant, another feature typical of community-associated MRSA strains. On the other hand, none of the isolates of this strain carried the PVL genes, which are commonly found in community-associated MRSA. However, WA MRSA-1 is also PVL negative. Clearly, further characterization of the Samoa ST1 MRSA is warranted. Few data are available on the prevalence of MRSA in the community in nonindustrialized countries. The World Health Organization has identiﬁ ed a need to strengthen monitoring of antimicrobial drug prescriptions and resistance and has funded pilot programs to achieve this (25,32). In this study in Samoa, MRSA was isolated from 9% of all participants, and 18% of participants from whom S. aureus was isolated had MRSA. Although we attempted to include a wide range of persons in the study, we cannot be sure that some groups in our sample were not overrepresented or underrepresented. Carrying out such research is difﬁ cult in a country such as Samoa, where research and laboratory infrastructure are not ideal. Ability to speak Samoan, knowledge of Samoan culture, and previous experience of living in Samoa were essential for the success of the project. We found that the diversity of MRSA isolates in Samoa that caused skin and soft tissue infection was similar to that in Denmark, a country of >5 million persons (3). We expected that isolates of the Southwest Paciﬁ c clone MRSA would be the predominant MRSA strain, but these were in the minority. The high prevalence of USA300 and Queensland clone MRSA might reﬂ ect the amount of travel between Samoa and the United States and Australia, respectively. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 References 1. Vandenesch F, Naimi T, Enright M, Lina G, Nimmo G, Heffernan H, et al. Community-acquired methicillin-resistant Staphylococ- cus aureus carrying Panton-Valentine leukocidin genes: worldwide emergence. Emerg Infect Dis. 2003;9:978–84. 2. Diekema DJ, Pfaller MA, Turnidge J, Verhoef J, Bell J, Fluit AC, et al. Genetic relatedness of multidrug-resistant, methicillin (oxa- cillin)–resistant Staphylococcus aureus bloodstream isolates from SENTRY Antimicrobial Resistance Surveillance Centers world- wide, 1998. Microb Drug Resist. 2000;6:213–21. doi:10.1089/ mdr.2000.6.213 The results of our study, together with future surveillance efforts, can be used to provide information for local prescribing; the prevalence of MRSA in Samoa is high, and empiric prescription of antimicrobial drugs needs to account for this high prevalence. For example, we advocate that any patient in Samoa suspected to have serious, invasive S. aureus infection have adequate cultures and antimicrobial drug susceptibility testing performed. S. aureus infection in such a patient should be treated with vancomycin and a β-lactamase–stable penicillin drug until laboratory results are available. Patients with uncomplicated skin and soft tissue infections requiring antimicrobial drug treatment should receive cotrimoxazole or, if the patient cannot tolerate sulfonamides, clindamycin. Boils or furuncles should be treated by drainage, infection control, and wound care, with antimicrobial drugs reserved for complications. 3. Larsen AR, Stegger M, Bocher S, Sorum M, Monnet DL, Skov RL. Emergence and characterization of community-associated methicil- lin-resistant Staphylococcus aureus infections in Denmark, 1999 to 2006. J Clin Microbiol. 2009;47:73–8. doi:10.1128/JCM.01557-08 4. Scribel LV, Silva-Carvalho MC, Souza RR, Superti SV, Kvitko CH, Figueiredo AM, et al. Clinical and molecular epidemiology of methicillin-resistant Staphylococcus aureus carrying SCCmecIV in a university hospital in Porto Alegre, Brazil. Diagn Microbiol Infect Dis. 2009;65:457–61. doi:10.1016/j.diagmicrobio.2009.08.012 5. Centers for Disease Control and Prevention. Community-associated methicillin-resistant Staphylococcus aureus infections in Paciﬁ c Islanders—Hawaii, 2001–2003. MMWR Morb Mortal Wkly Rep. 2004;53:767–70. 6. Estivariz CF, Park SY, Hageman JC, Dvorin J, Melish MM, Ar- pon R, et al. Emergence of community-associated methicillin re- sistant Staphylococcus aureus in Hawaii, 2001–2003. J Infect. 2007;54:349–57. doi:10.1016/j.jinf.2006.08.002 7. Munckhof WJ, Schooneveldt J, Coombs GW, Hoare J, Nimmo GR. Emergence of community-acquired methicillin-resistant Staphy- lococcus aureus (MRSA) infection in Queensland, Australia. Int J Infect Dis. 2003;7:259–64. doi:10.1016/S1201-9712(03)90104-4 We identiﬁ ed a wide range of genotypes of MRSA that were causing wound infections in a small Paciﬁ c Island nation. Discussion The Queensland clone is now common in the Australian states of New South Wales and the Northern Territory in addition to Queensland (33). However, the large number of isolates of the Samoa ST1 MRSA clone and the isolation of almost equal numbers of 3 different MRSA clones suggest that the situation may not be that simple. Evidence suggests that community-associated MRSA strains, particularly USA300, are more easily transmitted and might be more virulent than other S. aureus strains (31). In North America, approximately one third of persons with community-acquired USA300 MRSA infection require hospital admission (36). Moreover, community- associated MRSA is no longer just a problem in the community; it also has become a common cause of health care–associated infections (37). Hospital-acquired USA300 infections are more likely than community-acquired USA300 infections to be invasive and be associated with treatment failures (36). Thus, increases in the prevalence, severity, and complexity of diseases caused by globally successful community-associated MRSA strains are likely to be associated with increased illness, death, and cost. Economic analyses have consistently demonstrated that MRSA infections are associated with higher cost than are MSSA infections; although these studies have focused primarily on the costs associated with hospital care (38). None of these studies have investigated the consequences of MRSA in the developing world, yet the effects of disease are considerable (39). The Samoa ST1 MRSA isolates were distinct by both PFGE and spa typing from the community-associated ST1 MRSA strain commonly found in Australia and New Zealand, WA MRSA-1 (33). The Samoa ST1 MRSA isolates shared only 82% homology by PFGE with WA MRSA-1. WA MRSA-1 isolates are typically spa type t127, which does not cluster by BURP analysis with the spa types (t1853 and t6080) of the Samoa ST1 MRSA. In addition, the WA MRSA-1 strain is characterized by fusidic acid resistance and often also mupirocin or erythromycin resistance, whereas the Samoa ST1 MRSA isolates generally were resistant only to β-lactams. The Samoa ST1 MRSA isolates also were distinct from the Although reducing MRSA infection in the Samoan community is desirable, no controlled trials have demonstrated effective means of reducing community- associated MRSA infections in a community setting. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1027 1027 RESEARCH Mr Alesana-Slater is a ﬁ fth-year medical student at the University of Otago, New Zealand. His research focuses on MRSA. Discussion Recommendations to reduce transmission of community- associated MRSA include washing hands, caring for and covering wounds, not sharing contaminated personal items, appropriately disposing of contaminated waste, and appropriately prescribing antimicrobial drugs (40). Any interventions have substantial resource implications for a developing nation but must start with reliable surveillance of antimicrobial drug susceptibility, which is essential to monitor, control, and manage antimicrobial drug resistance. Thus, there is a clear need to assist developing countries with performing quality antimicrobial drug susceptibility testing and surveillance. References Our hope is that this study will provide a starting point for future research into antimicrobial drug resistance in the Paciﬁ c and provide impetus for initiatives to improve antimicrobial drug use in Paciﬁ c Island nations. Antimicrobial drug resistance is a global concern that does not respect national boundaries; consequently, countries need to assist each other in addressing the problem. 8. Rings T, Findlay R, Lang S. Ethnicity and methicillin-resistant S. aureus in South Auckland. N Z Med J. 1998;111:151. 9. Riley D, MacCulloch D, Morris A. Methicillin-resistant S. aureus in the suburbs. N Z Med J. 1998;111:59. 10. Castrodale LJ, Beller M, Gessner BD. Over-representation of Samo- an/Paciﬁ c Islanders among patients with methicillin-resistant Staph- ylococcus aureus (MRSA) infections at a large family practice clinic in Anchorage, Alaska, 1996–2000. Alaska Med. 2004;46:88–91. 11. Heffernan H, Davies H, Brett M. MRSA increasing in New Zealand. N Z Public Health Rep. 1995;2:97–9. Acknowledgments Annual sur- vey of methicillin-resistant Staphylococcus aureus (MRSA), 2009 [cited 2010 Nov 20]. http://www.surv.esr.cri.nz/PDF_surveillance/ Antimicrobial/MRSA/aMRSA_2009.pdf. 20. Clinical and Laboratory Standards Institute. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically: approved standard. 7th ed. Wayne (PA): The Institute; 2006. 35. David MZ, Daum RS. Community-associated methicillin-resistant Staphylococcus aureus: epidemiology and clinical consequences of an emerging epidemic. Clin Microbiol Rev. 2010;23:616–67. doi:10.1128/CMR.00081-09 21. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; 18th informational supple- ment. Wayne (PA): The Institute; 2008. 36. Moore CL, Hingwe A, Donabedian SM, Perri MB, Davis SL, Haque NZ, et al. Comparative evaluation of epidemiology and outcomes of methicillin-resistant Staphylococcus aureus (MRSA) USA300 infec- tions causing community- and healthcare-associated infections. Int J Antimicrob Agents. 2009;34:148–55. doi:10.1016/j.ijantimicag. 2009.03.004 22. Lina G, Piémont Y, Godail-Gamot F, Bes M, Peter M, Gauduchon V, et al. Involvement of Panton-Valentine leukocidin-producing Staphylococcus aureus in primary skin infections and pneumonia. Clin Infect Dis. 1999;29:1128–32. doi:10.1086/313461 23. Enright MC, Day NP, Davies CE, Peacock SJ, Spratt BG. Multilo- cus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J Clin Mi- crobiol. 2000;38:1008–15. 37. Tenover FC. Community-associated methicillin-resistant Staphylo- coccus aureus: it’s not just in communities anymore. Clinical Mi- crobiology Newsletter. 2006;28:33–6. doi:10.1016/j.clinmicnews. 2006.02.001 24. Strommenger B, Braulke C, Heuck D, Schmidt C, Pasemann B, Nubel U, et al. spa typing of Staphylococcus aureus as a frontline tool in epidemiological typing. J Clin Microbiol. 2008;46:574–81. doi:10.1128/JCM.01599-07 38. Gould IM, Reilly J, Bunyan D, Walker A. Costs of healthcare asso- ciated methicillin-resistant Staphylococcus aureus (MRSA) and its control. Clin Microbiol Infect. 2010;16:1721–8. doi:10.1111/j.1469- 0691.2010.03365.x. 25. World Health Organization. WHO global strategy for containment of antimicrobial resistance. Geneva: The Organization; 2001. 39. Nickerson EK, Hongsuwan M, Limmathurotsakul D, Wuthiekanun V, Shah KR, Srisomang P, et al. Staphylococcus aureus bacter- aemia in a tropical setting: patient outcome and impact of antibi- otic resistance. PLoS ONE. 2009;4:e4308. doi:10.1371/journal. pone.0004308 26. Okeke IN, Laxminarayan R, Bhutta ZA, Duse AG, Jenkins P, O’Brien TF, et al. Antimicrobial resistance in developing coun- tries. Part I: recent trends and current status. Lancet Infect Dis. 2005;5:481–93. doi:10.1016/S1473-3099(05)70189-4 27. Hart CA, Kariuki S. Antimicrobial resistance in developing coun- tries. BMJ. 1998;317:647–50. p 40. Barton M, Hawkes M, Moore D, Conly J, Nicolle L, Upton A, et al. Guidelines for the prevention and management of community-asso- ciated methicillin-resistant Staphylococcus aureus: a perspective for Canadian health care practitioners. Acknowledgments 12. Łuczak-Kadłubowska A, Sulikowska A, Empel J, Piasecka A, Orc- zykowska M, Kozinska A, et al. Countrywide molecular survey of methicillin-resistant Staphylococcus aureus strains in Poland. J Clin Microbiol. 2008;46:2930–7. doi:10.1128/JCM.00869-08 We thank Southern Community Laboratories Ltd. for its help in training J.A.S.; the management and staff of the laboratory of the Samoan National Health Service, particularly Vaimalo Kini; Sally Roberts for guidance and supervision of susceptibility testing; Tai Sopoaga, Mikaele Filipo Lauﬁ so, and the late Sandy Smith for their advice on the design of the study; and Mark Thomas for helpful advice concerning the manuscript. 13. Robinson DA, Kearns AM, Holmes A, Morrison D, Grundmann H, Edwards G, et al. Re-emergence of early pandemic Staphylococcus aureus as a community-acquired methicillin-resistant clone. Lancet. 2005;365:1256–8. doi:10.1016/S0140-6736(05)74814-5 14. Samoa Ministry of Health, Samoa Bureau of Statistics. Samoa De- mographic and Health Survey 2009. Apia (Samoa): The Ministry; 2009. This study received funding from the University of Otago Research Grants. The New Zealand Ministry of Health funded the work undertaken at the Institute of Environmental Health and Research. S.R.R. is a Clinical Research Training Fellow of the Health Research Council of New Zealand. 15. Hill PC, Birch M, Chambers S, Drinkovic D, Ellis-Pegler RB, Everts R, et al. Prospective study of 424 cases of Staphylococcus aureus bacteraemia: determination of factors affecting incidence and mortality. Intern Med J. 2001;31:97–103. doi:10.1111/j.1444- 0903.2001.00029.x Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1028 MRSA, Samoa MRSA, Samoa 31. Deleo FR, Otto M, Kreiswirth BN, Chambers HF. Community- associated meticillin-resistant Staphylococcus aureus. Lancet. 2010;375:1557–68. doi:10.1016/S0140-6736(09)61999-1 16. Finger F, Rossaak M, Umstaetter R, Reulbach U, Pitto R. Skin infec- tions of the limbs of Polynesian children. N Z Med J. 2004;117:U847. 17. Rossaak M, Pitto R. Osteomyelitis in Polynesian children. Int Or- thop. 2005;29:55–8. doi:10.1007/s00264-004-0597-3 ( ) 32. Holloway K. Community-based surveillance of antimicrobial use and resistance in resource-constrained settings. Geneva: World Health Organization; 2009. 18. Institute of Environmental Science and Research Limited. Antimicro- bial resistance data from hospital and community laboratories, 2008 [cited 2010 Nov 20]. http://www.surv.esr.cri.nz/PDF_surveillance/ Antimicrobial/AR/National_AR_2008.pdf 33. Nimmo GR, Coombs GW. Community-associated methicillin-resis- tant Staphylococcus aureus (MRSA) in Australia. Int J Antimicrob Agents. 2008;31:401–10. doi:10.1016/j.ijantimicag.2007.08.011 19. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial disk susceptibility tests: approved standard. 9th ed. Wayne (PA): The Institute; 2006. 34. Richardson A, Pope C, Couper J, Desai U, Heffernan H. Auckland 1142, New Zealand; email: s.ritchie@auckland.ac.nz Auckland 1142, New Zealand; email: s.ritchie@auckland.ac.nz Acknowledgments Can J Infect Dis Med Microbiol. 2006;17(Suppl C):4–24C. 28. Norris P, Nguyen H. Consumption of antibiotics in a small Paciﬁ c Island nation: Samoa. Pharmacy Practice. 2007;5:36–41. 29. Goossens H, Ferech M, Vander Stichele R, Elseviers M. Outpatient antibiotic use in Europe and association with resistance: a cross- national study. Lancet. 2005;365:579–87. 30. Norris P, Vaai C, Faalau F, Churchward M, Arroll B. Pain, infection and colds and ﬂ u: Samoan people’s views about antibiotics. Res So- cial Adm Pharm. 2011;7:81–92. Address for correspondence: Stephen R. Ritchie, Department of Molecular Medicine and Pathology, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand; email: s.ritchie@auckland.ac.nz Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 17, No. 6, June 2011 1029
https://openalex.org/W2948494052	https://infoscience.epfl.ch/record/268250/files/Final_Version.pdf	English	null	Special Issue on Smart Sustainable Manufacturing Systems	Applied sciences	2,019	cc-by	1,379	Special Issue on Smart Sustainable Manufacturing Systems Gökan May * and Dimitris Kiritsis * École Polytechnique Fédérale de Lausanne, ICT for Sustainable Manufacturing, EPFL SCI-STI-DK, Station 9, 1015 Lausanne, Switzerland * Correspondence: gokan.may@epﬂ.ch (G.M.); dimitris.kiritsis@epﬂ.ch (D.K.) Gökan May * and Dimitris Kiritsis * École Polytechnique Fédérale de Lausanne, ICT for Sustainable Manufacturing, EPFL SCI-STI-DK, Station 9, 1015 Lausanne, Switzerland * Correspondence: gokan.may@epﬂ.ch (G.M.); dimitris.kiritsis@epﬂ.ch (D.K.) Received: 24 May 2019; Accepted: 27 May 2019; Published: 31 May 2019 1. Introduction With the advent of disruptive digital technologies, companies are facing unprecedented challenges and opportunities. Advanced manufacturing systems are of paramount importance in making key enabling technologies and new products more competitive, aﬀordable and accessible as well as fostering their economic and social impact. The manufacturing industry also serves as an innovator for sustainability since automation coupled with advanced manufacturing technologies have helped to transition manufacturing practices to the circular economy [1]. In this context, shifting paradigms comprehend the 360-degree makeover of factories, from shop-ﬂoor to supply chain, from blue collar staﬀto top management, from employee to stakeholder. In that regard, the objective of smart and sustainable manufacturing systems of the future is to enable clean and competitive manufacturing systems irrespective of factories’ location or size, and to ﬁnd opportunities based on sustainability issues to grow beyond their borders [2]. To that end, this special issue of the journal Applied Sciences devoted to the broad ﬁeld of Smart Sustainable Manufacturing Systems was introduced to explore recent research into the concepts, methods, tools and applications for smart sustainable manufacturing in order to advance and promote the development of modern and intelligent manufacturing systems. applied sciences applied sciences Appl. Sci. 2019, 9, 2264; doi:10.3390/app9112264 2. Smart Sustainable Manufacturing Systems In light of the above, this special issue collects the latest research on relevant topics, and addresses present challenging issues with the introduction of smart sustainable manufacturing systems. There were 24 papers submitted to this special issue, and 9 papers were accepted (i.e., a 37.5% acceptance rate). Various topics have been addressed in this special issue, mainly on design of sustainable production systems and factories; industrial big data analytics and cyber physical systems; intelligent maintenance approaches and technologies for increased operating life of production systems; zero-defect manufacturing strategies, tools and methods towards on-line production management; and connected smart factories. The first paper, Production and Maintenance Planning for a Deteriorating System with Operation-Dependent Defectives, authored by H. Rivera-Gómez, O. Montaño-Arango, J. Corona-Armenta, J. Garnica-González, E. Hernández-Gress, and I. Barragán-Vite provides new insights into the area of sustainable manufacturing systems by analysing the novel paradigm of integrated production logistics, quality and maintenance design, and investigates the optimal production and maintenance switching strategy of an unreliable deteriorating manufacturing system. The paper presents a model that deﬁnes the joint production and maintenance switching strategies minimizing the total cost over an inﬁnite planning horizon [3]. The second paper, Kernel-Density-Based Particle Defect Management for Semiconductor Manufacturing Facilities, proposes a particle defect management method for the reduction of the Appl. Sci. 2019, 9, 2264; doi:10.3390/app9112264 www.mdpi.com/journal/applsci www.mdpi.com/journal/applsci 2 of 3 Appl. Sci. 2019, 9, 2264 defect ratio in semiconductor manufacturing facilities, and presents a kernel-density-based particle map that can overcome the limitations of the conventional method [4], authored by S. Park, S. Kim, and J.-G. Baek. The third paper of the Special Issue, A Multi-Usable Cloud Service Platform: A Case Study on Improved Development Pace and Eﬃciency, authored by J. Lindström, A. Hermanson, F. Blomstedt, and P. Kyösti, addresses a micro small and medium-sized enterprise (SME) in Sweden and its journey of developing and operating a multi-usable cloud service platform for big data collection and analytics [5]. The article An Integrated Open Approach to Capturing Systematic Knowledge for Manufacturing Process Innovation Based on Collective Intelligence by G. Wang, Y. Hu, X. Tian, J. Geng, G. Hu, and M. Zhang [6] builds a novel holistic paradigm of process innovation knowledge capture based on collective intelligence as a foundation for the future knowledge-inspired computer-aided process innovation and smart process planning. The next two articles focus on approaches and methods on supply chain level. 3. Future Research Although the special issue has been closed, more in-depth research in smart sustainable manufacturing systems is expected. In particular, demonstrative scenarios that pertain to smart design, smart machining, smart control, smart monitoring, and smart scheduling to highlight key enabling technologies and their possible applications to Industry 4.0 smart manufacturing systems could complement the research aspects covered within this Special Issue. Acknowledgments: We would like to take this opportunity to thank all the authors, reviewers, and dedicated editorial team of Applied Sciences. The special issue would not have been possible without the contributions and generous support of them. Conﬂicts of Interest: The authors declare no conﬂict of interest. 2. Smart Sustainable Manufacturing Systems The ﬁrst one, A Modiﬁed Method for Evaluating Sustainable Transport Solutions Based on AHP and Dempster–Shafer Evidence Theory by L. Chen and X. Deng, presents a transport sustainability index (TSI) as a primary measure to determine whether transport solutions have a positive impact on city sustainability [7]. The subsequent paper, Dynamic Supply Chain Design and Operations Plan for Connected Smart Factories with Additive Manufacturing authored by B. Chung, S.I. Kim and J.S. Lee suggests a general planning framework and various optimization models for dynamic supply chain design and operations plan [8]. The seventh paper in this Special Issue, The Role of Managerial Commitment and TPM Implementation Strategies in Productivity Beneﬁts written by J. Díaz-Reza, J. García-Alcaraz, L. Avelar-Sosa, J. Mendoza-Fong, J. Sáenz Diez-Muro, and J. Blanco-Fernández [9], proposes a structural equation model to integrate four latent variables: Managerial commitment, preventive maintenance, total productive maintenance, and productivity beneﬁts. Subsequently, Opportunities for Industry 4.0 to Support Remanufacturing by S. Yang, A.M.R., J. Kaminski, and H. Pepin reviews the challenges encountered by the remanufacturing sector and discusses how the Industry 4.0 revolution could help to eﬀectively address these issues and unlock the potential of remanufacturing [10]. Last but not least, the ﬁnal article Hybrid Laminate for Haptic Input Device with Integrated Signal Processing of R. Schmidt, A. Graf, R. Decker, V. Kräusel, W. Hardt, D. Landgrebe, and L. Kroll presents a new tool concept for joining and forming hybrid laminates in a manufacturing process [11]. References 1. 2018 World Manufacturing Forum Report, Recommendations for the Future of Manufacturing. Available online: https://www.worldmanufacturingforum.org/report (accessed on 24 May 2019). 2. May, G.; Stahl, B.; Taisch, M. Energy management in manufacturing: Toward eco-factories of the future—A focus group study. Appl. Energy 2016, 164, 628–638. [CrossRef] Appl. Sci. 2019, 9, 2264 3 of 3 3. Rivera-Gómez, H.; Montaño-Arango, O.; Corona-Armenta, J.; Garnica-González, J.; Hernández-Gress, E.; Barragán-Vite, I. Production and Maintenance Planning for a Deteriorating System with Operation-Dependent Defectives. Appl. Sci. 2018, 8, 165. [CrossRef] 4. Park, S.; Kim, S.; Baek, J.-G. Kernel-Density-Based Particle Defect Management for Semicon Manufacturing Facilities. Appl. Sci. 2018, 8, 224. [CrossRef] 5. Lindström, J.; Hermanson, A.; Blomstedt, F.; Kyösti, P. A Multi-Usable Cloud Service Platform: A Case Study on Improved Development Pace and Eﬃciency. Appl. Sci. 2018, 8, 316. [CrossRef] 6. Wang, G.; Hu, Y.; Tian, X.; Geng, J.; Hu, G.; Zhang, M. An Integrated Open Approach to Capturing Systematic Knowledge for Manufacturing Process Innovation Based on Collective Intelligence. Appl. Sci. 2018, 8, 340. [CrossRef] 7. Chen, L.; Deng, X. A Modiﬁed Method for Evaluating Sustainable Transport Solutions Based on AHP and Dempster–Shafer Evidence Theory. Appl. Sci. 2018, 8, 563. [CrossRef] 8. Chung, B.; Kim, S.I.; Lee, J.S. Dynamic Supply Chain Design and Operations Plan for Connected Smart Factories with Additive Manufacturing. Appl. Sci. 2018, 8, 583. [CrossRef] 9. Díaz-Reza, J.; García-Alcaraz, J.; Avelar-Sosa, L.; Mendoza-Fong, J.; Sáenz Diez-Muro, J.; Blanco-Fernández, J. The Role of Managerial Commitment and TPM Implementation Strategies in Productivity Beneﬁts. Appl. Sci. 2018, 8, 1153. [CrossRef] 10. Yang, S.; MR, A.; Kaminski, J.; Pepin, H. Opportunities for Industry 4.0 to Support Remanufacturing. Appl. Sci. 2018, 8, 1177. [CrossRef] 11. Schmidt, R.; Graf, A.; Decker, R.; Kräusel, V.; Hardt, W.; Landgrebe, D.; Kroll, L. Hybrid Laminate for Haptic Input Device with Integrated Signal Processing. Appl. Sci. 2018, 8, 1261. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4366813050	https://zenodo.org/records/7794827/files/%D0%97%D1%80%D0%B8%D0%B1%D0%BD%D1%94%D0%B2%D0%B0%20%D0%86%D1%80%D0%B8%D0%BD%D0%B0%20%D0%9F%D0%B0%D0%B2%D0%BB%D1%96%D0%B2%D0%BD%D0%B0.pdf	Ukrainian	null	Економічні та правові аспекти трансферу технологій в Україні: результати, проблеми, перспективи	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	4,245	1 доктор економічних наук, доцент кафедри маркетингу, інновацій та регіонального розвитку, Чернівецький національний університет імені Юрія Федьковича, Україна, м. Чернівці, вул. Коцюбинського, 2, https://orcid.org/0000-0003-4156-4702 2 кандидат економічних наук, доцент кафедри соціальної філософії та управління, Державний податковий університет, 08205, Київська область, м. Ірпінь, вул. Університетська, 31, https://orcid.org/0000-0002-4585-3440 3 старший науковий співробітник, Український науково-дослідний інститут спеціальної техніки та судових експертиз Служби безпеки України, 03113, Україна, м. Київ, вул. Миколи Василенка, 3, https://orcid.org/0000-0002-1447-6275 4 кандидат економічних наук, доцент, доцент кафедри економіки та економічної кібернетики, фінансово- економічний факультет, Національний технічний університет «Дніпровська політехніка», 49005 м. Дніпро, пр. Дмитра Яворницького, 19, https://orcid.org/0000-0002-8982-9725 Ліцензовано за умовами Creative Commons BY 4.0 International license Анотація. Метою цієї статті є аналіз проблем та перспектив економічних та правових аспектів трансферу технологій в Україні, зокрема зважаючи на виклики сьогодення, пов’язані з війною та енергетичною кризою. На його основі зроблено наступні висновки. Українські інноваційні технологічні стартапи демонструють стійкість і високі результати. ІТ-послуги становлять майже половину загального обсягу експорту послуг країни. Очікується, що стартапи стануть фундаментом для розвитку компаній з новим рівнем технологій, що дозволить створити значну додану вартість економіці та привабить закордонні інвестиції. Залучення іноземного капіталу саме таким чином має залишитися пріоритетом повоєнної відбудови України, оскільки трансфер сучасних технологій та розвиток інновацій однозначно прискорить відновлення української економіки та сприятиме її зростанню. Водночас існує треба вдосконалити та зробити більш простими та прозорими правові аспекти трансферу. Ключові слова: трансфер технологій, інновації, економічний розвиток, стартапи. Опубліковано Секція УДК 03.04.2023 Економіка 330.341.11 DOI: http://dx.doi.org/10.5281/zenodo.7794827 Ліцензовано за умовами Creative Commons BY 4.0 International license Економічні та правові аспекти трансферу технологій в Україні: результати, проблеми, перспективи Опубліковано Секція УДК 03.04.2023 Економіка 330.341.11 DOI: http://dx.doi.org/10.5281/zenodo.7794827 Опубліковано Секція УДК 03.04.2023 Економіка 330.341.11 DOI: http://dx.doi.org/10.5281/zenodo.7794827 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 Економічні та правові аспекти трансферу технологій в Україні: результати, проблеми, перспективи Економічні та правові аспекти трансферу технологій в Україні: результати, проблеми, перспективи АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 country's total export of services. Startups are expected to become the foundation for the development of companies of a new level of technology, which will create significant added value to the economy and attract foreign capital. Startups are expected to become the foundation for the development of companies of a new level of technology, which will create significant added value to the economy and attract foreign capital. Development of industrial parks is important for the implementation of modern technological solutions and the promotion of innovation. To this end, the government of Ukraine introduced significant state incentives for the development of industrial parks, including exemption from paying certain taxes, full or partial compensation of interest rates on loans, compensation for joining engineering networks even in wartime. Attracting foreign capital in this way should remain a priority for the post-war reconstruction of Ukraine, since the transfer of modern technologies and the development of innovations will definitely accelerate the recovery of the Ukrainian economy and contribute to its post-war growth. Legal aspects are an important problem of technology transfer in Ukraine. It is necessary to improve and make the legal aspects of the transfer simpler and more transparent. country's total export of services. Startups are expected to become the foundation for the development of companies of a new level of technology, which will create significant added value to the economy and attract foreign capital. Startups are expected to become the foundation for the development of companies of a new level of technology, which will create significant added value to the economy and attract foreign capital. Development of industrial parks is important for the implementation of modern technological solutions and the promotion of innovation. To this end, the government of Ukraine introduced significant state incentives for the development of industrial parks, including exemption from paying certain taxes, full or partial compensation of interest rates on loans, compensation for joining engineering networks even in wartime. Attracting foreign capital in this way should remain a priority for the post-war reconstruction of Ukraine, since the transfer of modern technologies and the development of innovations will definitely accelerate the recovery of the Ukrainian economy and contribute to its post-war growth. Legal aspects are an important problem of technology transfer in Ukraine. It is necessary to improve and make the legal aspects of the transfer simpler and more transparent. Keywords: technology transfer, innovation, economic development, startups. Economic and legal aspects of technology transfer in Ukraine: results, problems, prospects Annotation. The main aim of this article is to analyze the problems and prospects of the economic and legal aspects of technology transfer in Ukraine, taking into account, in particular, the current challenges associated with the war and the energy crisis. Due to this analysis, the following conclusions were made. The trend of widespread use of technology transfer is characteristic of the entire world economy today. Ukrainian innovative technology startups demonstrate sustainability and high results. IT services make up almost half of the АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 федеральних відомств США, що призведе до збільшення винагороди платникам податків у вигляді ліцензійних гонорарів. У своїй роботі Peña-Vinces J. та Audretsch D. [12] пишуть, що майже в усіх країнах Латинської Америки, що розвиваються, підприємницька діяльність за кордоном пов’язана в основному з експортом товарів та продуктів видобувної промисловості та сільського господарства. Таким чином, зрушення такої підприємницької моделі в бік високих технологій є серйозним викликом як для бізнесменів, так і для політиків. В роботі, з точки зору екології та промисловості, досліджено, чи може вища освіта й наука стати рушійною силою зростання компаній-експортерів високих технологій. Більшість поточних досліджень цього питання зосереджено на рівні фірм. Розвідка Peña-Vinces J. та Audretsch D. [12] охоплює цілу країну. Їхня економетрична модель емпірично підтримує гіпотезу, що вища освіта й наука разом є рушійними силами економічного зростання. В Україні важливою проблемою трансферу технологій являються правові аспекти. Як зазначено в роботі Svitlak I. та Huts N. [1], на сьогодні проблематика установчих документів підприємства привертає увагу як науковців, так і практиків, але деякі питання залишаються невирішеними ще на етапі його створення. Більшість установчих документів юридичних осіб не передбачають конкретного порядку їх легітимації. Та для деяких із них, наприклад, тих, що надають банківські чи юридичні послуги, передбачено, що їхній статус, повноваження, цілі діяльності, порядок заснування можуть бути визначені законом, а їхня реорганізація та ліквідація регулюються відповідними спеціальними законодавчими актами конкретної країни. Саме ці нормативно-правові акти за своїм змістом є статутами конкретних юридичних осіб. Не можуть бути визнані законними статути юридичних осіб, створених в інших організаційно-правових формах. Регулювання діяльності компаній усіх форм приєднання в інших країнах здійснюється за допомогою окремих установчих документів. Закон України «Про державне регулювання діяльності у сфері трансферу технологій» [6] визначає правові, економічні, організаційні та фінансові засади державного регулювання діяльності у сфері трансферу технологій і спрямований на забезпечення ефективного використання науково-технічного та інтелектуального потенціалу України, технологічності виробництва продукції, охорони майнових прав на вітчизняні технології та/або їхні складові на території держав, де планується або здійснюється їх використання, розширення міжнародної науково-технічної співпраці в цій сфері. Метою цієї статті є аналіз проблем та перспектив економічних та правових аспектів трансферу технологій в Україні, зокрема зважаючи на виклики сьогодення, пов’язані з війною та енергетичною кризою. Вступ у На сьогодні жоден процес не обходиться без застосування технологій, що стрімко розвиваються. Причому більшість сучасних технологій володіють важливою особливістю, яку треба брати до уваги, – вони швидко змінюються, вдосконалюються та модифікуються. Як зазначено в роботі Sanetra-Półgrabi S. та Tetłaka Z. [2], цифровий ринок XXI століття диктує певні вимоги до інновацій, адресованих споживачам, учасникам економічних процесів, постачальникам матеріально-технічних ресурсів, комунікаційним агентам ринку, сторонам-виробникам (або продавцям). Їх дотримання необхідне для побудови відносин між ними, які залучають учасників взаємодії до уявних ринкових обмінів як способів збереження досягнутої конкурентної позиції на ринку та виживання в стратегічній перспективі при виникненні негативних економічних явищ. Технології розвиваються настільки стрімко, що в межах одного підприємства, чи навіть однієї країни або ж низки країн, неможливо самотужки втілити в життя всі необхідні кроки, щоб кінцева продукція відповідала світовим стандартам. Отже, трансфер технологій з часом стає все актуальнішим для всіх видів діяльності. р ф р у Тенденція широкого застосування трансферу технологій притаманна на сьогодні всій світовій економіці. Наприклад, Leech D.P. та Scott J.T. [11] у своїй роботі відзначають важливість іноземних патентів для передачі технологічних винаходів, створених у лабораторіях федеральних агентств США. Описано патентні портфоліо 11 таких установ, причому 98 відсотків досліджень виконувались в лабораторіях усіх федеральних агентств США. Також оцінено функцію розподіленого лагу, яка показує вплив на ліцензійний дохід історії заявок агентства на патенти на винаходи поза територією США. Виявилось, що ці ефекти залежать від того, чи агентство отримало також іноземний патентний захист для своїх винаходів. Оцінено модель динамічних панельних даних доходів від ліцензій в якості функції історії заявок і виданих патентів. Докази підтверджують ідею, що агентство, яке отримує патенти США на свої винаходи, але не має патентного захисту за кордоном, ставить у невигідне становище корпорації, які ліцензують технології агентства. Згодом ці корпорації змушені конкурувати з фірмами, які копіюють їхні технології та виводять їх на міжнародний ринок з меншими витратами, оскільки вони не вкладають кошти в розробку з нуля та не платять роялті за ліцензування. Зростання кількості іноземних патентів підвищить готовність компаній брати на себе усі витрати, необхідні для отримання успішних комерційних продуктів, а також покращить передачу технологічних винаходів із лабораторій АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 жодним чином не суперечить ідеї трансферу відповідних технологій. Сьогодні прогресивною вважається тенденція не просто зростання експортного потенціалу, а передовсім збільшення частки наукоємних високотехнічних товарів у загальній структурі експорту. Такі тенденції діють у різних галузях, зокрема в деревопереробній промисловості, як зазначено в роботі Olianych V., Olianych L. та Kryński A. [3]. Проте Svitlak I. та Huts N. [1] відмічають, що типові статути, які існують у більшості країн, у тому числі в Україні, не є гнучкими за своєю природою, оскільки чинне законодавство унеможливлює внесення до них змін або передбачає складну процедуру таких перетворень, зокрема їх державної реєстрації, що ускладнює правовий аспект трансферу технологій. У законі України «Про державне регулювання діяльності у сфері трансферу технологій» [6] визначено, що для забезпечення виконання функцій, пов'язаних зі створенням, використанням, патентуванням та ліцензуванням прав на технології та їхні складові, уповноважений орган з питань формування та забезпечення реалізації державної політики у сфері трансферу технологій, інші центральні органи виконавчої влади, Національна академія наук України та галузеві академії наук створюють у межах чисельності та фонду оплати праці структурні підрозділи, які б займалися питаннями трансферу технологій, інноваційної діяльності та інтелектуальної власності. Також, відповідно до своїх повноважень, створюють такі підрозділи на підприємствах, у наукових установах, організаціях та вищих навчальних закладах, що належать до сфери управління відповідного центрального органу виконавчої влади або перебувають у віданні відповідної академії наук. В інформаційно-аналітичній записці щодо впливу діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій у 2021 році [7] зазначається, що, для виконання пункту 3 постанови Кабінету Міністрів України від 04 грудня 2019 р. № 1030, Міністерство освіти й науки України (МОН) узагальнило відповідну інформацію, надану головними розпорядниками бюджетних коштів – шістьма національними академіями наук і 70 органами державної влади. Форми для надання цих відомостей було укладено з урахуванням Методологічних положень з організації державного статистичного спостереження щодо інноваційної діяльності підприємств (у редакції наказу Державної служби статистики України від 25 жовтня 2019 р. № 350), згідно з якими до показників про кількість придбаних/переданих нових ій ф б / технологій за формами придбання/передання належать:  придбання/передання права на патенти, ліцензії на використання винаходів, промислових зразків, корисних моделей;  придбання/передання права на патенти, ліцензії на використання винаходів, промислових зразків, корисних моделей;  придбання/передання результатів досліджень та розробок;  ноу-хау, угоди на придбання/передачу технологій;  придбання/продаж устаткування;  цілеспрямований прийом/перехід на роботу кваліфікованих фахівців;  цілеспрямований прийом/перехід на роботу кваліфікованих фахівців;  інші. Результати Важливість трансферу технологій та його значимість для розвитку економічних та соціальних процесів складно переоцінити. Зрибнєва І. П. [4; 5] зазначає, що концепція побудови економічної політики забезпечення конкурентоспроможності суб’єктів інноваційного підприємництва формується на макро- й мікрорівні. Водночас на кожному рівні управління необхідний особливий організаційно-економічний підхід, який забезпечив би перетворення наявних ресурсів у конкурентні переваги шляхом використання тих чи інших компетенцій (на мікрорівні – це ключові компетенції менеджменту, на макрорівні – це інституційні компетенції, що ґрунтуються, наприклад, на використанні кластеризації). Також відзначається, що на макрорівні економічна політика забезпечення конкурентоспроможності суб’єктів інноваційного підприємництва формується для створення міжнародної конкурентоспроможності, що АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023  цілеспрямований прийом/перехід на роботу кваліфікованих фахівців; Чистий дохід від реалізації продукції підприємствами, установами та організаціями, які здійснювали трансфер технологій, у 2021 році склав 431,58 млн грн, що майже в 5 разів більше, ніж у 2020 році, в тому числі заклади вищої освіти (ЗВО) – 203,05 млн грн, що в 17,6 разів більше за аналогічний показник 2020 року. В таблицях 1 та 2 наочно відображена відповідна статистика. За результатами, наведеними в табл. 1, видно, що в 2020 році найбільша кількість установ, що здійснювала трансфер технологій, підпорядкована Національній академії аграрних наук України. В цьому бізнесі одним із основних способів ефективної конкуренції є підвищення економічного виробництва сільськогосподарських культур, ціна на які на світових ринках все більше знижується впродовж останніх років. Саме тому передові представники агробізнесу не шкодують інвестицій у новітні технології, АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 які дозволяють мінімізувати використання ЗЗР та добрив, отримувати чіткіші прогнози на основі обробки накопичених даних, а також слідувати новому тренду – ставати екологічнішими. які дозволяють мінімізувати використання ЗЗР та добрив, отримувати чіткіші прогнози на основі обробки накопичених даних, а також слідувати новому тренду – ставати екологічнішими. Таблиця 1 ЗВО, установи, підприємства та організації, які здійснювали передачу технологій та інших об’єктів права інтелектуальної власності у 2021 році Порядковий номер організації, якій підпорядковано установи Назва організації, якій підпорядковано установи Кількість установ Дохід, млн грн 1 Національна академія наук України 5 131,18 2 Національна академія медичних наук України 2 0,42 3 Національна академія аграрних наук України 26 2,37 4 Державна служба спеціального зв'язку та захисту інформації України 2 2,10 5 Міністерство освіти і науки України 26 61,81 6 Міністерство соціальної політики України 2 0,13 Джерело: складено авторами на основі аналізу [7]. ЗВО, установи, підприємства та організації, які здійснювали передачу ологій та інших об’єктів права інтелектуальної власності у 2021 році Таблиця 2 ЗВО, установи, підприємства та організації, які здійснювали передачу технологій та інших об’єктів права інтелектуальної власності у 2020 році Порядковий номер організації, якій підпорядковано установи Назва організації, якій підпорядковано установи Кількість установ Дохід, млн грн 1 Національна академія наук України 4 77,64 2 Національна академія медичних наук України 2 0,23 3 Національна академія аграрних наук України 101 2,99 4 Державна служба спеціального зв'язку та захисту інформації України 0 0 5 Міністерство освіти і науки України 17 19,82 6 Міністерство соціальної політики України 2 0,14 Джерело: складено авторами на основі аналізу [13]. АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 позитивних моментів відзначено зростання витрат на проведення оцінки науково- технічного рівня технологій та/або їхніх складових як об'єкта комерціалізації, закупівлю інструментів, обладнання та устаткування, необхідних для проведення тестування, прототипування, проведення експериментальних досліджень, створення та випробування дослідних зразків, виготовлення моделей. Тобто наукові установи та ЗВО використали можливість оновити свою інфраструктуру та обладнання, а також збільшилися темпи патентування. у Зазначимо, що до 2020 року на базі ДП "Центр науково-технічної інформації та сприяння інноваційному розвитку України" працювала Українська інтегрована система трансферу технологій (УІСТТ) – платформа, яка надавала клієнтам послуги, спрямовані на пошук українських і зарубіжних партнерів у сфері трансферу технологій, а також допомагала налагодити технологічне співробітництво між зацікавленими сторонами (продавцем і покупцем технології), а саме: у Зазначимо, що до 2020 року на базі ДП "Центр науково-технічної інформації та сприяння інноваційному розвитку України" працювала Українська інтегрована система трансферу технологій (УІСТТ) – платформа, яка надавала клієнтам послуги, спрямовані на пошук українських і зарубіжних партнерів у сфері трансферу технологій, а також допомагала налагодити технологічне співробітництво між зацікавленими сторонами (продавцем і покупцем технології), а саме: o допомагала формулювати технологічні пропозиції та запити (ТП/ТЗ); o проводила технологічний аудит – перевір спроможність і готовність до трансферу; o проводила технологічний аудит – перевірку пропозицій та запитів на техніч спроможність і готовність до трансферу; o розміщувала профілі технологічних запитів та пропозицій в Системі; o розміщувала профілі технологічних запитів та пропозицій в Системі; o інформувала про нові ТП/ТЗ, що відповідають потребам клієнта; o інформувала про нові ТП/ТЗ, що відповідають потребам клієнта; o здійснювала пошук потенційних партнерів та сприяла встановленню прямих контактів і проведенню переговорів щодо технологічного співробітництва; б o надавала послуги технологічного брокера; o здійснювала інформаційну та консультаційну підтримку клієнтів щодо участі в Рамкових програмах ЄС (насамперед – FP7) та допомагала знайти партнера; o інформувала про конкурси й програми, що надають фінансову підтримку проектам комерціалізації технологій, та консультувала щодо питань участі в таких конкурсах і програмах. o займалась просуванням ТП/ТЗ клієнтів на різних заходах, спрямованих на пошук партнерів (виставках, ярмарках, брокерських заходах тощо) o займалась просуванням ТП/ТЗ клієнтів на різних заходах, спрямованих на пошук партнерів (виставках, ярмарках, брокерських заходах тощо) o надавала інші види послуг (технологічний аудит, маркетинг, консалтингові послуги у сфері захисту інтелектуальної власності, проведення науково-технічної експертизи та інші). o надавала інші види послуг (технологічний аудит, маркетинг, консалтингові послуги у сфері захисту інтелектуальної власності, проведення науково-технічної експертизи та інші). Проте з початком війни ситуація істотно змінилася. Sanetra-Półgrabi S. та Tetłaka Z. АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 ЗВО, установи, підприємства та організації, які здійснювали передачу ологій та інших об’єктів права інтелектуальної власності у 2020 році АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 Використання технологій міжмашинної взаємодії (Machine‐to‐Machine, M2M) та Інтернету речей (Internet of Things, IoT) допомагає сучасним підприємцям в агробізнесі використовувати недоступні раніше можливості – автоматизовувати й оптимізовувати складні процеси, підвищувати продуктивність роботи та раціонально розподіляти ресурси. Підприємці отримали змогу збирати великі масиви даних і зберігати їх на хмарних платформах за допомогою сенсорів, дронів, «розумних» датчиків, польових метеостанцій та інших пристроїв Інтернету речей. Далі ця інформація аналізується з використанням методів data science, що дає основу для прийняття виважених рішень і набагато знижує ризики для розвитку бізнесу. Завдяки правильному використанню IoT-рішень агропідприємці можуть реалізувати RTK-технологію точного землеробства, що дозволяє зменшити витрати дорогих добрив, покращити якість оброблення ґрунту та збирання врожаю, збільшити врожайність майже на третину, а також мінімізувати витрати на насіння й полив. Інтернет речей надає великі переваги не лише у сфері рослинництва. IoT-рішення довели високу ефективність також й у тваринництві, тепличному бізнесі тощо. Проте бачимо, що в аграрній галузі відбувся істотний спад у період 2020-2021 років, та в інших помітно зростання. На рис. 1 наведено відповідну порівняльну діаграму. Рис. 1. Порівняння кількості установ, підприємств та організацій, які здійснювали передачу технологій та інших об’єктів права інтелектуальної власності у 2020 та у 2021 роках Джерело: власна розробка авторів. 4 2 101 0 17 2 5 2 26 2 26 2 0 20 40 60 80 100 120 Національна академія наук України Національна академія медичних наук України Національна академія аграрних наук України Державна служба спеціального зв'язку та захисту інформації України Міністерство освіти і науки України Міністерство соціальної політики України Кількість установ 2020 2021 4 2 101 0 17 2 5 2 26 2 26 2 0 20 40 60 80 100 120 Кількість установ Рис. 1. Порівняння кількості установ, підприємств та організацій, які здійснювали передачу технологій та інших об’єктів права інтелектуальної власності у 2020 та у 2021 роках Джерело: власна розробка авторів. В інформаційно-аналітичній записці щодо напрямів використання коштів, одержаних у результаті трансферу технологій, створення яких профінансовано з державного бюджету у 2021 році та в період 2014—2021 років [8], зазначено, що адаптивні карантинні обмеження, пов’язані із пандемією COVID-19 в Україні, та запровадження повних локдаунів у м. Києві та низці областей України у 2021 році, уповільнили діяльність у цій сфері. Так, наукові установи й ЗВО у 2021 році переважно передавали не готові технології, а результати досліджень та розробок. Серед АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 які покликані створити значну додану вартість в економіці та залучити іноземний капітал. З цією метою Український фонд стартапів (УФС) протягом останніх 3 років виділив $6,5 млн на їхнє фінансування. Насамкінець, розвиток індустріальних парків є важливим для впровадження сучасних технологічних рішень та сприяння інноваціям. Для цього уряд України запровадив значні державні стимули, зокрема звільнення від сплати певних податків, повну або часткову компенсацію процентних ставок за кредитами, компенсацію за приєднання до інженерних мереж навіть у воєнний час. Відповідно до Закону України № 5018-VI державна підтримка надається індустріальним паркам у сфері переробної промисловості, перероблення промислових та/або побутових відходів (крім їх розміщення), науково-дослідних робіт (R&D) та ІТ. Висновки З огляду на нагальну необхідність повоєнного відновлення України такий потужний інструмент як трансфер технологій в жодному випадку не можна залишати без уваги. Водночас найбільший прибуток приносить саме експорт технологій. В цьому напрямку варто сконцентрувати увагу на найсильніших галузях української економіки. Це насамперед ІТ-сектор. У підсумку, залучення іноземного капіталу саме таким чином має залишитися пріоритетом повоєнної відбудови України, оскільки трансфер сучасних технологій та розвиток інновацій однозначно прискорить відновлення української економіки та сприятиме її зростанню. Також потрібно вдосконалити та зробити більш простими та прозорими правові аспекти трансферу. Перспективою подальших наукових досліджень цього питання може стати вивчення досвіду інших країн, що пов’язаний зі сферою трансферу технологій, та ретельний аналіз економічних та правових аспектів трансферу технологій в Україні в майбутньому. Список використаних джерел 1. Svitlak, I., & Huts, N. (2022). The legal charter of the company of the future: the order of creation. Futurity Economics & Law, 2(1), 22–27. https://doi.org/10.57125/FEL.2022.03.25.03 2. Sanetra-Półgrabi, S., & Tetłaka, Z. (2022). Protection of consumer rights in the advertising of the future in the context of economic instability. Futurity Economics & Law, 2(3), 12–18. https://doi.org/10.57125/FEL.2022.09.25.02 g 3. Olianych, V., Olianych, L., & Kryński, A. (2022). Woodworking industries in Kharkiv region during the NEP years: a historical aspect. Scientific Journal of Polonia University, 53(4), 83-89. https://doi.org/10.23856/5310 3. Olianych, V., Olianych, L., & Kryński, A. (2022). Woodworking industries in Kharkiv region during the NEP years: a historical aspect. Scientific Journal of Polonia University, 53(4), 83-89. https://doi.org/10.23856/5310 ( ) p // g/ / 4. Зрибнєва, І., П. (2021). Стратегія економічного розвитку технологічного підприємництва в Україні: теорія, методологія, практика (монографія). Запоріжжя: Гельветика. 4. Зрибнєва, І., П. (2021). Стратегія економічного розвитку технологічного підприємництва в Україні: теорія, методологія, практика (монографія). Запоріжжя: Гельветика. 5. Зрибнєва, І., П. (2021) Концепція побудови економічної політики забезпечення конкурентоспроможності суб’єктів інноваційного підприємництва на макро- і макрорівні. Науковий погляд: економіка та управління, 1 (71), 83–88. 5. Зрибнєва, І., П. (2021) Концепція побудови економічної політики забезпечення конкурентоспроможності суб’єктів інноваційного підприємництва на макро- і макрорівні. Науковий погляд: економіка та управління, 1 (71), 83–88. 6. Про державне регулювання діяльності у сфері трансферу технологій. № 143-V. (2006). https://zakon.rada.gov.ua/laws/show/143-16#Text 6. Про державне регулювання діяльності у сфері трансферу технологій. № 143-V. (2006). https://zakon.rada.gov.ua/laws/show/143-16#Text 7. Інформаційно-аналітична записка щодо впливу діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій у 2021 році. (2022). https://mon.gov.ua/storage/app/media/nauka/2022/Transefer.tekhnolohiy.2022/16. 08.2022/Zvit.vplyv.na.finstan.u.2021.rotsi.sayt.MON-16.08.22.pdf 7. Інформаційно-аналітична записка щодо впливу діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій у 2021 році. (2022). https://mon.gov.ua/storage/app/media/nauka/2022/Transefer.tekhnolohiy.2022/16. АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 [2] констатують, що вплив кризових явищ на економіки країн світу призводить до негативної реакції на їхню економічну діяльність. Споживач завжди орієнтований на реальну пропозицію з боку виробників, а виробник – на попит, який є головним рушієм економіки. Споживчий попит повинен помірно випереджати виробничі потужності, і якщо цей процес не відбувається, відповідно інвестиції не спрямовуються на розвиток виробничої галузі. Крім того, за умов зниження доходів населення, виробник намагається знайти маркетингові рішення, які б дозволили йому заощадити на різних складових його виробничо-збутової діяльності. Передовсім йдеться про економію на технологічних процесах, на якості виробів, на складових сировини й продукції, що випускається, на просуванні товарів, на стимулюванні збуту, на інформаційному супроводі пропонованих товарів (послуг), про використання різноманітних замінників окремих інгредієнтів. Попри виклики, спричинені війною, українські інноваційні технологічні стартапи демонструють стійкість і високі результати. За перші вісім місяців 2022 року ІТ- експорт України зріс на 16% проти 2021 року. ІТ-послуги становлять майже половину загального обсягу експорту послуг країни. Державна підтримка та потужне зростання свідчать про те, що настав час інвестувати в український ІТ-сектор. Очікується, що стартапи мають стати фундаментом для розвитку компаній нового рівня технологій, АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 Висновки 08.2022/Zvit.vplyv.na.finstan.u.2021.rotsi.sayt.MON-16.08.22.pdf 8. Інформаційно-аналітична записка щодо напрямів використання коштів, одержаних у результаті трансферу технологій, створених за рахунок коштів 8. Інформаційно-аналітична записка щодо напрямів використання коштів, одержаних у результаті трансферу технологій, створених за рахунок коштів АКАДЕМІЧНІ ВІЗІЇ Випуск 18/2023 державного бюджету у 2021 році та у період 2014-2021 років. (2022). https://mon.gov.ua/storage/app/media/nauka/2022/Transefer.tekhnolohiy.2022/16. 08 2022/Inf-analit zapyska napr vykor koshtiv 2014-2021-16 08 2022 pdf державного бюджету у 2021 році та у період 2014-2021 років. (2022). https://mon.gov.ua/storage/app/media/nauka/2022/Transefer.tekhnolohiy.2022/16. 08.2022/Inf-analit.zapyska.napr.vykor.koshtiv.2014-2021-16.08.2022.pdf / g / g / pp/ / / / 2/Inf-analit.zapyska.napr.vykor.koshtiv.2014-2021-16.08.2022.pdf 9. Про затвердження порядку подання та узагальнення відомостей про вплив діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій. № 1558. (2022). https://mon.gov.ua/ua/nauka/innovacijna-diyalnist- ta-transfer-tehnologij/transfer-tehnologij 9. Про затвердження порядку подання та узагальнення відомостей про вплив діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій. № 1558. (2022). https://mon.gov.ua/ua/nauka/innovacijna-diyalnist- ta-transfer-tehnologij/transfer-tehnologij 10. Dimitropoulou, A., Giotopoulos, I., Protogerou, A., & Tsakanikas, A. (2023). Does the innovativeness of creative firms help their business clients to innovate?. The Journal of Technology Transfer, (48), 1–32. https://doi.org/10.1007/s10961-021-09901-1 10. Dimitropoulou, A., Giotopoulos, I., Protogerou, A., & Tsakanikas, A. (2023). Does the innovativeness of creative firms help their business clients to innovate?. The Journal of Technology Transfer, (48), 1–32. https://doi.org/10.1007/s10961-021-09901-1 11. Leech, D.P., & Scott, J.T. (2022). Foreign patents for the technology transfer from laboratories of U.S. federal agencies. The Journal of Technology Transfer, (47), 937–978. https://doi.org/10.1007/s10961-021-09857-2 11. Leech, D.P., & Scott, J.T. (2022). Foreign patents for the technology transfer from laboratories of U.S. federal agencies. The Journal of Technology Transfer, (47), 937–978. https://doi.org/10.1007/s10961-021-09857-2 p // g/ / 12. Peña-Vinces, J., Audretsch, D. (2021). Tertiary education and science as drivers of high- technology exporting firms growth in developing countries. The Journal of Technology Transfer, (46), 1734–1757. https://doi.org/10.1007/s10961-020-09807-4 p // g/ / 12. Peña-Vinces, J., Audretsch, D. (2021). Tertiary education and science as drivers of high- technology exporting firms growth in developing countries. The Journal of Technology Transfer, (46), 1734–1757. https://doi.org/10.1007/s10961-020-09807-4 13. Інформаційно-аналітична записка щодо впливу діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій у 2020 році. (2021). https://mon.gov.ua/storage/app/media/nauka/2021/Transefer.tekhnolohiy/06/23/Zv it.vplyv.na.finstan.u.2020.rotsi.sayt.MON-23.06.21.pdf 13. Інформаційно-аналітична записка щодо впливу діяльності у сфері трансферу технологій на фінансовий стан підприємств, установ та організацій у 2020 році. (2021). https://mon.gov.ua/storage/app/media/nauka/2021/Transefer.tekhnolohiy/06/23/Zv it.vplyv.na.finstan.u.2020.rotsi.sayt.MON-23.06.21.pdf
https://openalex.org/W2793337761	https://www.nature.com/articles/s41419-018-0389-0.pdf	English	null	Indoleamine 2,3-dioxygenase expression regulates the survival and proliferation of Fusobacterium nucleatum in THP-1-derived macrophages	Cell death and disease	2,018	cc-by	11,416	The Author(s) 2018 OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. A R T I C L E A R T I C L E O p e n A c c e s s Abstract b Abstract Fusobacterium nucleatum (Fn) is a tumor-associated obligate anaerobic bacterium, which has a role in the progression of colorectal cancer (CRC). Fn can invade and promote colon epithelial cells proliferation. However, how Fn survives and proliferates in its host cells remains largely unknown. In this study, we aimed to determine the molecular mechanisms underlying the morphology, survival, and proliferation of Fn in THP-1-derived macrophages (dTHP1). For the ﬁrst time, we found that Fn is a facultative intracellular bacterium that can survive and limited proliferate in dTHP1 cells up to 72 h, and a live Fn infection can inhibit apoptosis of dTHP1 cells by activating the PI3K and ERK pathways. Both Fn bacteria and dTHP1 cells exhibit obvious morphological changes during infection. In addition, Infection of Fn-induced indoleamine 2,3-dioxygenase (IDO) expression by TNF-α-dependent and LPS-dependent pathway in a time-dependent and dose-dependent manner, and the IDO-induced low tryptophan and high kynurenine environment inhibited the intracellular multiplication of Fn in dTHP1 cells. IDO expression further impaired the function of peripheral blood lymphocytes, permitting the escape of Fn-infected macrophages from cell death. IDO inhibition abrogated this effect caused by Fn and relieved immune suppression. In conclusion, we identiﬁed IDO as an important player mediating intracellular Fn proliferation in macrophages, and inhibition of IDO may aggravate infection in Fn-associated tumor immunotherapy. Indoleamine 2,3-dioxygenase expression regulates the survival and proliferation of Fusobacterium nucleatum in THP-1-derived macrophages Ying Xue1,2, Han Xiao2, Songhe Guo2, Banglao Xu3, Yuehua Liao2, Yixian Wu2 and Ge Zhang1,2 Ying Xue1,2, Han Xiao2, Songhe Guo2, Banglao Xu3, Yuehua Liao2, Yixian Wu2 and Ge Zhan Introduction interaction of some aerobic or facultative anaerobic intracellular bacteria, including Listeria monocytogenes, Legionella pneumophila, Salmonella typhi, Mycobacter- ium tuberculosis, and Chlamydia trachomatis2. However, little is known about obligate anaerobic intracellular bacteria and their interaction with host cells. Some aggressive intracellular bacteria can survive and multiply in the cytoplasm of infected macrophages1. These facultative intracellular bacteria are shielded from humoral antibodies and can only be eliminated by a cel- lular immune response2. The treatment of intracellular bacteria is an ongoing clinical problem. Currently, numerous studies have conﬁrmed the host–pathogen Fusobacterium nucleatum (F. nucleatum, Fn) is an opportunistic commensal obligate anaerobic Gram-negative bacterium that is indigenous to the human oral cavity and has a role in periodontal disease. Fn has previously been reported to be involved in different infectious processes3. Recently, accumulated evidence has demonstrated that Fn is associated with the development and carcinogenesis, and promote metastasis in colorectal cancer (CRC)4–6. Fn can Correspondence: Ge Zhang (zhangge@mail.sysu.edu.cn) 1School of Public Health (Shenzhen), Sun Yat-sen University, Guangdong, China 2Department of Microbial and Biochemical Pharmacy, School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China Full list of author information is available at the end of the article These authors contributed equally: Ying Xue, Han Xiao, Songhe Guo. Edited by H-U. Simon Xue et al. Cell Death and Disease (2018) 9:355 DOI 10.1038/s41419-018-0389-0 Xue et al. Cell Death and Disease (2018) 9:355 DOI 10.1038/s41419-018-0389-0 Cell Death & Disease Cell Death & Disease Ofﬁcial journal of the Cell Death Differentiation Association © The Author(s) 2018 ( ) OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Ofﬁcial journal of the Cell Death Differentiation Association Page 2 of 14 Xue et al. Cell Death and Disease (2018) 9:355 Cy3 Merge c a b Merge DAPI Cy3 FITC DAPI FITC Cy3 Merge Merge d * Fig. 1 F. nucleatum invades THP-1-derived macrophages. THP-1-derived macrophages (dTHP1) were infected with F. nucleatum (Fn) at a MOI of 10:1 (bacteria:cells) for 48 h. Immunoﬂuorescence staining of live Fn infection (a) and heat-killed Fn infection (b) were observed by confocal microscope (×60). c After 72 h co-culture, the recovery colonies numbers of average cell lysis and supernatant liquid. d Gram staining of Fn bacteria (×100) and Fn-infected dTHP1 cells (×20) were observed by light microscope. Bacteria external to the host cell were labeled with both Cy3 (red) and FITC (green), bacteria inside the cells were labeled with Cy3 (appearing red when channels were merged). Scale bar = 10 μm. P < 0.001 b Merge DAPI Cy3 FITC b a Cy3 Cy3 FITC Cy3 c Merge d c d Fig. 1 F. nucleatum invades THP-1-derived macrophages. THP-1-derived macrophages (dTHP1) were infected with F. nucleatum (Fn) at a MOI of 10:1 (bacteria:cells) for 48 h. Immunoﬂuorescence staining of live Fn infection (a) and heat-killed Fn infection (b) were observed by confocal microscope (×60). c After 72 h co-culture, the recovery colonies numbers of average cell lysis and supernatant liquid. d Gram staining of Fn bacteria (×100) and Fn-infected dTHP1 cells (×20) were observed by light microscope. Bacteria external to the host cell were labeled with both Cy3 (red) and FITC (green), bacteria inside the cells were labeled with Cy3 (appearing red when channels were merged). Scale bar = 10 μm. P < 0.001 intracellular bacteria, such as L. monocytogenes or M. Ofﬁcial journal of the Cell Death Differentiation Association © The Author(s) 2018 tuberculosis, is associated with IDO induction in various tissues and cell types15,16. Interestingly, some obligate intracellular bacteria, such as C. trachomatis and Tox- oplasma gondii, are tryptophan auxotroph. Tryptophan deprivation causes Chlamydia to enter a persistent growth17. Previous studies have reported that tryptophan is required to stimulate the growth of Fn, and Fn tryp- tophanase degrades tryptophan to indole, which can inhibit the growth of Fn in vitro18. Furthermore, IDO inhibitors, such as 1-MT (Indoximod), are promising drugs for cancer immunotherapy. Given that a tryptophan-deﬁcient environment caused by IDO in infected macrophages may inhibit the growth of intra- cellular Fn, IDO inhibitors may aggravate infection during Fn-associated tumor therapy. To date, live Fn infection of macrophages is poorly understood, and whether Fn infection can induce the expression of IDO in macro- phages and the effects of Fn-induced IDO on macrophage immune functions have not been investigated. adhere to and invade epithelial cells7, and the interaction of Fn with CRC cells has been found to promote host cell proliferation8. Interestingly, our recent study showed that the overload of Fn elicits high levels of Fn-speciﬁc anti- bodies in patients with CRC, implying that Fn may escape host humoral immune responses by developing inside host cells9. Macrophages provide the ﬁrst line of defense against invading pathogens. Thus, whether Fn can survive and multiply in macrophages and its effects on immune func- tions in host cells need to be explored. An immunomodulatory role for the enzyme indolea- mine 2,3-dioxygenase (IDO), which catalyzes the con- version of tryptophan into kynurenine, has been suggested to have a role in macrophage functions10. Increased IDO activity is often associated with tumors and infectious diseases11. Several studies have described IDO-dependent T-cell suppression by antigen-presenting cells under many infectious and inﬂammatory conditions, indicating that biochemical changes due to tryptophan catabolism have a profound effect on T-cell proliferation and effector functions in tissue microenvironments12,13. To elucidate the interactions between Fn and macro- phages, we investigated the survival of both Fn and macrophages during Fn infection and identiﬁed a possible IDO expression can be induced in macrophages by some bacterial infections14. Infection with facultative Xue et al. © The Author(s) 2018 Cell Death and Disease (2018) 9:355 Page 3 of 14 * * * * * * Relative protein expression f Time post-infection proliferation rate dTHP-1 THP-1 dTHP-1+Fn dTHP-1+Heated-killed Fn p-AKT AKT — 1.0h 1.5h 2.0h 4.0h p-ERK ERK GAPDH 130 120 Kda 60 50 60 50 50 40 40 30 a b c d e dTHP-1 in Annexin V-FITC/PI PI 0.69% 4.32% 0.86% 2.78% 2.43% 0.71% Annexin V-FITC Fn (24 h) Heat-killed Fn 1.65% 4.00% Fn (72 h) Mock Apoptotic cell (%) Fig. 2 F. nucleatum infection exhibits little or no effect on the cell viability of THP-1-derived macrophages. dTHP1 cells were infected with dead F. nucleatum (Fn) (heat-killed-Fn, grey) or live Fn (Fn, dark grey) at a MOI of 10:1 (bacteria:cells) for the indicated time-points. a Morphology was observed at 72 h; b cell viability was measured by an MTT assay; c, d the apoptotic cells were analyzed by ﬂow cytometry at 72 h; e the PI3K/AKT and ERK signaling pathway were analyzed by western blot from 1 h to 4 h and f quantitation was performed using pixel density analysis. Data indicate the mean ± standard deviation (SD) of triplicate-infected cultures. Bars represent the mean ± SD of the results from replicate measurements. P < 0.05. Scale bar = 20 μm Time post-infection proliferation rate b dTHP-1 THP-1 dTHP-1+Fn dTHP-1+Heated-killed Fn a d Apoptotic cell (%) c PI 0.86% 2.78% 2.43% 0.71% Heat-killed Fn Mock d d c PI P 0.69% 4.32% C Fn (24 h) 1.65% 4.00% Fn (72 h) Mock dTHP-1 in Annexin V-FITC/PI * * * * * * Relative protein expression f dTHP-1 in Annexin V-FITC/PI Annexin V-FITC p-AKT AKT — 1.0h 1.5h 2.0h 4.0h p-ERK ERK GAPDH 130 120 Kda 60 50 60 50 50 40 40 30 e e Fig. 2 F. nucleatum infection exhibits little or no effect on the cell viability of THP-1-derived macrophages. dTHP1 cells were infected with dead F. nucleatum (Fn) (heat-killed-Fn, grey) or live Fn (Fn, dark grey) at a MOI of 10:1 (bacteria:cells) for the indicated time-points. a Morphology was observed at 72 h; b cell viability was measured by an MTT assay; c, d the apoptotic cells were analyzed by ﬂow cytometry at 72 h; e the PI3K/AKT and ERK signaling pathway were analyzed by western blot from 1 h to 4 h and f quantitation was performed using pixel density analysis. F. nucleatum infection induces classical activation of THP- 1-derived macrophages, but has no effects on INF-γ expression To evaluate whether live Fn infection leads to the induction of activated or polarized dTHP1 cells, an array of ﬁve different cytokines, two chemokine receptors, and two MHC class II cell surface receptors, which are typi- cally produced during bacterial infections, was estab- lished, and mRNA expression was analyzed by qRT-PCR. The relative fold change of mRNA was calculated relative to the uninfected dTHP1 cells. The pro-inﬂammatory cytokines IL-6, IL-12p40, and TNFα (Fig. 3a); the che- mokine receptors CCR7 and CXCR4 (Fig. 3b); and the MHC class II cell surface receptors HLA-DR and CD80 (Fig. 3b) were highly induced by live/heat-killed Fn, whereas the expression of INF-γ and anti-inﬂammatory cytokine IL-10 remained unaffected (Fig. 3a). Moreover, macrophage mannose receptor CD206 and scavenger receptor CD163, both of which are M2-polarized phe- notype markers, exhibited reduced expression in Fn- dTHP1 cells (Fig. 3b). In addition, western blot and ﬂow cytometry assays revealed that the expression of the macrophage M1 phenotype markers CCR7, CXCR4, and HLA-DR were increased with live Fn or heated-kill Fn treated (Fig. 3c, d). These results indicated that infection with Fn induces classically activated (M1-polarized) macrophages with the exception of INF-γ expression. To further assay the survival of intracellular Fn, the Fn- infected dTHP1 were collected and lysed after 48 h post- infection, the culture supernatants and cell lysates were cultured in CDC blood agar anaerobically. Interestingly, large numbers of Fn bacterial colonies were observed from the cell lysates, whereas the culture supernatants of Fn-infected dTHP1 showed no bacterial growth (Fig. 1c). The Gram- negative bacteria which were isolated from the infected dTHP1 cells showed a fusiform morphology (Fig. 1d). Those results indicated that Fn can invade and survival in the dTHP1 cell with the changed morphology. More importantly, those ﬁnding provided a convenient method for the co-culture of anaerobic intracellular bacteria and host cells under aerobic culture condition. F. nucleatum can invade and survive in THP-1-derived macrophages To investigate whether Fn can adhere to and invade macrophages, human THP-1-derived macrophages (dTHP1) were treated with live Fn bacteria at an multi- plicity of infection (MOI) of 10:1 (bacteria:cells) and were incubated with the conventional cell culture method at 37 °C with 5% CO2. Bacteria invasion assays were carried out using an antibody-based differential staining method, all invasion experiments were performed under the aerobic condition. The speciﬁc immunoﬂuorescence staining of Fn bacteria was conﬁrmed by using mouse and human Fn polyclonal primary antibody respectively (Fig. S1). As shown in Fig. 1a, bacteria inside the cells were labeled with Cy3 (red), whereas Fn bacteria external to the host cell were labeled with both Cy3 and FITC (green, appearing yellow when channels were merged). Intracel- lular Fn were distributed mainly around the cell nucleus, and exhibited obvious morphological changes into short rod or spheres shapes in the cytoplasm of dTHP1 cells, whereas extracellular Fn showed normal fusiform rod shapes (Fig. 1a). In contrast, heat-killed Fn were not observed to enter host cells (Fig. 1b). © The Author(s) 2018 Data indicate the mean ± standard deviation (SD) of triplicate-infected cultures. Bars represent the mean ± SD of the results from replicate measurements. P < 0.05. Scale bar = 20 μm Ofﬁcial journal of the Cell Death Differentiation Association Page 4 of 14 Xue et al. Cell Death and Disease (2018) 9:355 signiﬁcant difference in dTHP1 cell viability whether they were infected with either live or heat-killed Fn (Fig. 2b). In addition, in the presence of live or heat-killed Fn treated, dTHP1 cells exhibited no signiﬁcant differences in the frequency of early apoptotic (FITC+PI−) or late apop- totic/necrotic (FITC+PI+) cells compared with unin- fected cells according to ﬂow cytometry (Fig. 2c, d). Furthermore, western blot assays revealed that live Fn infection induced a signiﬁcant increase in Akt phos- phorylation (p-AKT, ser473) and ERK phosphorylation (p-ERK, ERK1/2) after 1.5 h and 2 h of co-incubation, respectively. Maximal increases in the p-AKT and p-ERK levels occurred after 2–4 h (Fig. 2e, f). These results indicated that treated with live/heat-killed Fn exerted differentiation-like morphological changes, but not sig- niﬁcant cytotoxicity on dTHP1 cells, and that macrophage apoptosis was inhibited during infection by activated PI3K/Akt and ERK signaling pathways. role for Fn-mediated IDO induction in limiting Fn mul- tiplication inside macrophages and creating a micro- environment with suppressed lymphocyte immune responses to kill the infected host cells. Ofﬁcial journal of the Cell Death Differentiation Association F. nucleatum infection induces IDO expression in THP-1- derived macrophages To investigate whether Fn infection inﬂuences the survival of macrophages, dTHP1 cells were treated with Fn bacteria (MOI 10:1) and were incubated at 37 °C with 5% CO2. The dTHP1 cells exhibited obvious morpholo- gical changes into spindle shapes when infected with live or heat-killed Fn compared with the uninfected cells (Fig. 2a). However, MTT assays revealed that there was no IDO expression in dTHP1 cells treated with live or heat- killed Fn was assessed by qRT-PCR and western blot analysis, respectively. The IDO mRNA and protein assays revealed that IDO can be induced in dTHP1 cells by treatment with live/heat-killed Fn. Moreover, live Fn Ofﬁcial journal of the Cell Death Differentiation Association Page 5 of 14 Xue et al. Cell Death and Disease (2018) 9:355 HLA-DR CD80 CD206 CD163 CCR-7 CXCR-4 -10 -5 0 5 10 40 80 Mock Heat-killed Fn Fn * * * * * * * * * * * * * d l o f A N R m e v it a l e R a b IFN-γ TNF-α IL-6 IL-12p40 IL-10 0 10 20 30 40 60 80 Mock Heat-killed Fn Fn * * * * *** * * * d l o f A N R m e v it a l e R Mock Fn Heat-killed Fn Mock Fn Heat-killed Fn CCR-7 HLA-DR Mock Fn Heat-killed Fn CD206 CCR-7 HLA-DR CD206 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Mock Heat-killed Fn Fn ** * * * * * * * r e k r a m e vitale r e p yt o n e h p 1 M n ois s e r p x e e c n e c s e r o ulf d CCR-7 CXCR-4 HLA-DR 0.0 0.2 0.4 0.6 0.8 1.0 Mock Heat-killed Fn Fn ** * * * * * * * * n ois s e r p x e nie t o r p e vit ale R CCR-7 HLA-DR GAPDH CXCR-4 30 20 Kda 50 40 40 30 40 30 c Fig. 3 F. nucleatum infection induces classically activated THP-1-derived macrophages. F. nucleatum infection induces IDO expression in THP-1- derived macrophages The expression of the M1- polarized phenotype marker CCR7, CXCR4, and HLA-DR in dTHP1 cells with live/heated-killed Fn at an MOI of 10:1 for 48 h was analyzed by ﬂow cytometry (c) and western blot (d). Bars represent the mean ± SD of the results from replicate measurements. P < 0.05, P < 0.01, *P < 0.001 infection exhibited higher expression of IDO than treat- ment with heat-killed Fn (Fig. 4a–h). Infection with low- dose live or heat-killed Fn at an MOI of 1:1 for 24 h was able to induce IDO expression. IDO was further increased in a time-dependent manner, peaking at 48 h (Fig. 4g, h), and a dose-dependent manner, peaking at an MOI of 100:1 (Fig. 4c, d). The enzymatic activity of IDO was also investigated by HPLC. The enzymatic activity of IDO was almost undetectable in the supernatant of dTHP1 cells, but was observed in cells infected with live or heat-killed Fn (Fig. 4b, f). These enzymatic activity results are consistent with the qPCR and western blotting results. Figure 4i–j demonstrates that some Fn-dTHP1 cells exhibited positive cytoplasmic IDO staining by immunohistochemical analy- sis. Those results conﬁrmed that live or heat-killed Fn treated can induce IDO expression in human macrophages and that live Fn infection induced a higher level of IDO. dTHP1), cytokines were measured using a cytokine bead array. As shown in Table S1, the INF-γ, IL-2, IL-4, and IL- 10 concentrations were below the limits of detection until 72 h in the supernatant of dTHP1 cells infected with live or heat-killed Fn. High levels of IL-6 and TNF-α were detected in the supernatants of both live Fn-dTHP1 and heat-killed Fn-dTHP1 cells. Moreover, IL-6 and TNF-α were signiﬁcantly increased in the supernatant of live Fn- dTHP1 cells relative to the supernatant of heat-killed Fn- dTHP1 cells (P < 0.001) (Fig. 5a, b). These ﬁndings are consistent with the IDO expression levels measured in treated live or heated-killed Fn by western blot assay at 48 h (Fig. 5c). In addition, TNF-α (1043.78 ± 53.23 pg/ml) levels reached their peak at 24 h post-live Fn infection and substantially decreased. However, IL-6 was substantially elevated and reached its peak at 72 h post-live Fn infection (4256.35 ± 135.63 pg/ml) (Fig. 5a, b). F. nucleatum infection induces IDO expression in THP-1- derived macrophages The mRNA levels of a cytokines (INF-γ, TNF-α, IL-6, IL- 12p40, IL-10); b MHC class II cell surface receptors (HLA-DR, CD80); M2-polarized phenotype markers (CD206, CD163); and chemokine receptors (CCR7, CXCR4) were assessed by qRT-PCR in dTHP1 cells infected with heat-killed-Fn or live Fn at an MOI of 10:1 for 24 h. The expression of the M1- polarized phenotype marker CCR7, CXCR4, and HLA-DR in dTHP1 cells with live/heated-killed Fn at an MOI of 10:1 for 48 h was analyzed by ﬂow cytometry (c) and western blot (d). Bars represent the mean ± SD of the results from replicate measurements. P < 0.05, P < 0.01, P < 0.001 a IFN-γ TNF-α IL-6 IL-12p40 IL-10 0 10 20 30 40 60 80 Mock Heat-killed Fn Fn * * * *** * * * d l o f A N R m e v it a l e R Mock Fn Heat-killed Fn CCR-7 c c CCR 7 HLA-DR Mock Fn Heat-killed Fn CD206 CCR-7 HLA-DR CD206 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Mock Heat-killed Fn Fn ** * * * * * * * r e k r a m e vitale r e p yt o n e h p 1 M n ois s e r p x e e c n e c s e r o ulf HLA-DR CD80 CD206 CD163 CCR-7 CXCR-4 -10 -5 0 5 10 40 80 Mock Heat-killed Fn Fn * * * * * * * * * * * * * * d l o f A N R m e v it a l e R b b d CCR-7 CXCR-4 HLA-DR 0.0 0.2 0.4 0.6 0.8 1.0 Mock Heat-killed Fn Fn ** * * * * * * * * n ois s e r p x e nie t o r p e vit ale R CCR-7 HLA-DR GAPDH CXCR-4 30 20 Kda 50 40 40 30 40 30 Fig. 3 F. nucleatum infection induces classically activated THP-1-derived macrophages. The mRNA levels of a cytokines (INF-γ, TNF-α, IL-6, IL- 12p40, IL-10); b MHC class II cell surface receptors (HLA-DR, CD80); M2-polarized phenotype markers (CD206, CD163); and chemokine receptors (CCR7, CXCR4) were assessed by qRT-PCR in dTHP1 cells infected with heat-killed-Fn or live Fn at an MOI of 10:1 for 24 h. F. nucleatum infection induces IDO expression in THP-1- derived macrophages Given that lipopo- lysaccharide (LPS) induces the production of TNF-α and IL-6, which are potent inducers of IDO expression, we examined the possible involvement of LPS or these two cytokines in IDO expression induced by live Fn infection Ofﬁcial journal of the Cell Death Differentiation Association F. nucleatum undergoes limited proliferation in the low- tryptophan environment of IDO-induced THP-1-derived macrophages To assess the ability of Fn to multiply inside macro- phages, a gentamycin protection assay was performed. Fn was able to survive up to 72 h in dTHP1 cells, although the mean surviving intracellular bacteria were depleted to a small fraction of the inoculums (0.073 living intracel- lular bacteria per macrophage after 72 h post-infection of 10 originally inoculated bacterial cells per macrophage, Fig. 6a). Whereas, strictly extracellular anaerobic bacteria would die if they were no invasion. Most interestingly, the IDO-speciﬁc inhibitor 1-methyl-L-tryptophan (1-MT) relieved the inhibition of Fn proliferation inside macro- phages. With the treatment of 1-MT (100 µM), Fn grew quickly from 24 to 48 h post-infection, resulting in up to >100 living intracellular bacteria per macrophage after 72 h post-infection (Fig. 6a, b). Immunoﬂuorescence staining showed the marked increase of intracellular Fn (red) after 72 h post-infection with the treatment of 1-MT (Fig. 6c). In addition, Fn-dTHP1 cells treated with 1-MT exhibited only a slight elevation in the rates of apoptosis and necrosis after 72 h post-infection according to ﬂow cyto- metry (Fig. 6d). Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum infection of THP-1-derived macrophages These data indicate that Fn infection is able to induce several pro- inﬂammatory cytokines by LPS in dTHP1 cells with the exception of IFN-γ, and LPS, and that the expression TNF-α stimulated by Fn infection was the most involved cytokine in the induction of IDO. Those results indicate that Fn is able to survive and undergo limited intracellular proliferation in macro- phages. The growth of Fn inside macrophages indicated that the intracellular proliferation of Fn was limited due to the concentration of tryptophan and kynurenine induced by the expression of IDO. Moreover, blocking IDO enzyme activity markedly promoted intracellular pro- liferation of Fn. Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum infection of THP-1-derived macrophages a IDO mRNA expression assessed by qRT-PCR, b induction of IDO enzymatic activity by HPLC, and c, d representative western blots for IDO protein expression in dTHP1 cells infected with live Fn or heat-killed-Fn for 48 h at different dosages; e IDO mRNA expression assessed by qRT-PCR, f induction of IDO enzymatic activity by HPLC and g, h representative western blots for IDO protein expression in dTHP1 cells infected with live Fn or heat-killed-Fn at a MOI of 10:1 for the indicated time-points; i Fn-infected dTHP1 cells exhibit positive cytoplasmic staining for IDO at a MOI of 10:1 for 48 h and j negative staining for uninfected-dTHP1 cells as detected by immunohistochemistry. Bars represent the mean ± SD of the results from replicate measurements. ^Below the detection limit. P < 0.05, P < 0.01, P < 0.001. Scale bar = 20 μm Given that IDO generates a tryptophan-deﬁcient environment to inhibit Fn proliferation, we cultured Fn in tryptophan-enriched or kynurenine-enriched BHI medium in vitro. Tryptophan promotes the growth of Fn in a dose-dependent manner; the minimum effective tryptophan concentration is 0.02 mg/ml, and the effect peaks at 0.32 mg/ml (Fig. 7a). The growing Fn population doubles with enriched tryptophan from 0.02 to 0.18 mg/ ml, and an about 1.5-fold increase is observed from 0.18 to 0.36 mg/ml. Furthermore, kynurenine has the opposite effect and inhibits the growth of Fn in a dose-dependent manner (Fig. 7b). Fn growth decreased by about 1.5-fold at the low Kyn concentration from 0.01 to 0.04 mg/ml, decreased about twofold at the high concentration from 0.04 to 0.16 mg/ml and was almost completely inhibited at the dose of 0.36 mg/ml. using neutralizing assays. LPS neutralized by polymyxin B (PMB, >25 μg/ml) decreased the IL-6 mRNA level, but had no effect on the TNF-α level, even at the highest dose of PMB (100 μg/ml) (Fig. 5d). Higher levels of PMB (>100 μg/ml) were cytotoxic to dTHP1 cells (date not show). Furthermore, LPS neutralized by the highest dose of PMB (100 μg/ml) markedly reduced the expression of IDO and enzyme activity, but was unable to completely inhibit IDO expression (Fig. 5e, f). Moreover, neutralizing antibodies against TNF-α signiﬁcantly blocked the increased IDO expression and enzyme activity induced by live Fn infection, and the combination of PMB and TNF-α antibodies almost completely blocked the induction of IDO protein and its enzyme activity (Fig. 5e, f). Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum infection of THP-1-derived macrophages Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum infection of THP-1-derived macrophages To investigate the cytokines that are involved in the induction of IDO in Fn-infected dTHP1 cells (Fn- Ofﬁcial journal of the Cell Death Differentiation Association Xue et al. Cell Death and Disease (2018) 9:355 Page 6 of 14 * Heat-killed Fn IDO GAPDH Fn IDO GAPDH 0 24h 48h 72h Kda 50 40 40 30 50 40 40 30 Relative IDO protein g h * Time post-infection * Time post-infection A N R m O D I e v it a l e R * * * e f i j Fn-infected dTHP-1 cells dTHP-1 cells a b Kyn/Trp(uM/uM) * * Initial infective dose * A N R m O D I e v it a l e R Initial infective dose * * Kyn/Trp(uM/uM) Relative IDO protein Heat-killed Fn IDO GAPDH Fn IDO GAPDH 0 1:1 10:1 100:1 Kda 50 40 40 30 50 40 40 30 c d a A N R m O D I e v it a l e R Initial infective dose * * b Kyn/Trp(uM/uM) * * Initial infective dose * b Heat-killed Fn IDO GAPDH Fn IDO GAPDH 0 1:1 10:1 100:1 Kda 50 40 40 30 50 40 40 30 c Relative IDO protein d d c Fn A N R m O D I e v it a l e R * * * e * * f Kyn/Trp(uM/uM) f e h h * Relative IDO protein h Time post-infection Heat-killed Fn IDO GAPDH Fn IDO GAPDH 0 24h 48h 72h Kda 50 40 40 30 50 40 40 30 g Time post-infection g 0 Fn i i Fn-infected dTHP-1 cells j dTHP-1 cells j Fn-infected dTHP-1 cells Fn-infected dTHP-1 cells Ofﬁcial journal of the Cell Death Differentiation Association Page 7 of 14 Xue et al. Cell Death and Disease (2018) 9:355 (see ﬁgure on previous page) Fig. 4 F. nucleatum infection induces IDO expression in THP-1-derived macrophages in a dose-dependent and time-dependent manner. Ofﬁcial journal of the Cell Death Differentiation Association F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes To investigate whether IDO produced by Fn-dTHP1 has an effect on the proliferation activity of T lymphocyte cells, an MTT assay was conducted using the supernatant of dTHP1 cells as conditioned medium (CM). CM from live Fn-dTHP1 (Fn-CM) exhibited IDO enzyme activity, whereas CM from uninfected dTHP1 exhibited almost undetectable IDO enzyme activity (Fig. 8a). In addition, the functional IDO enzyme activity in 1-MT-treated CM from Fn-dTHP1 cells was markedly inhibited by 1-MT. Moreover, CM from Fn-dTHP1 cells markedly inhibited the proliferation of human Jurkat T lymphocyte cells (P < 0.001), and the proliferation was almost completely restored (82.8%) by CM from Fn-dTHP1 cells with the addition of 1-MT (Fig. 8b). To investigate the anti-Fn effect of cytotoxicity against infected macrophages by peripheral blood leukocyte Ofﬁcial journal of the Cell Death Differentiation Association Page 8 of 14 Xue et al. Cell Death and Disease (2018) 9:355 (PBL) an LDH release assay was conducted using remarkably reduced when the PBLs cells were exposed to a c TNF-α )l m / g p ( IL- )l m / g p ( 6 b Time post-infection Time post-infection Kyn/Trp(uM/uM) Relative mRNA fold PMB dose (μg/ml) e f 0 1500 3000 4500 0 h 24 h 48 h 72 h Heat-killed Fn Fn 0 300 600 900 1200 0 h 24 h 48 h 72 h Heat-killed Fn Fn GAPDH IDO IDO GAPDH d -3 -2 -1 0 1 2 TNF-α IL-6 0 10 25 50 75 100 ^ ^ ^ ^ * * Kda 50 40 40 30 Kda 50 40 40 30 Mock Fn Fn+PMB Fn+ Fn+PMB aTNFa +aTNFa 0 0.2 0.4 0.6 0.8 1 1.2 1.4 IDO 0 0.5 1 1.5 2 2.5 IDO Mock Fn Heate-killed TNFa INF-γ Fn n i e t o r p e v it a l e R expression Relative protein expression Mock Fn Fn+PMB Fn+ Fn+PMB aTNFa +aTNFa ^ ^ ^ ^ ^ ^ * ** Fig. 5 Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum-infected THP-1 cells. TNF-α (a) and IL-6 (b) levels in the supernatants of Fn- or heat-killed Fn-infected dTHP1 cells at an MOI of 10:1 for the indicated times. c IDO expression was assessed by Western blot at 48 h. d mRNA expression of IL-6 and TNF-α level in the presence of PMB with the indicated dosages. F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes IDO enzymatic activity in supernatants (e) and representative western blots for IDO in cells lysates (f) of live Fn-infected dTHP1 cells (MOI: 10:1) in the presence of neutralizing antibodies to TNF-α and/or PBM for 48 h. Bars represent the mean ± SD of the results from triplicate determinations. ^Below the detection limit. P < 0.05, P < 0.01, P < 0.001 a TNF-α )l m / g p ( Time post-infection Relative mRNA fold PMB dose (μg/ml) 0 300 600 900 1200 0 h 24 h 48 h 72 h Heat-killed Fn Fn d -3 -2 -1 0 1 2 TNF-α IL-6 0 10 25 50 75 100 ^ ^ Kyn/Trp(uM/uM) PMB dose (μg/ml) e -3 Mock Fn Fn+PMB Fn+ Fn+PMB aTNFa +aTNFa IL- )l m / g p ( 6 b Time post infection 0 1500 3000 4500 0 h 24 h 48 h 72 h Heat-killed Fn Fn ^ ^ * * b f IDO GAPDH Kda 50 40 40 30 c GAPDH IDO Kda 50 40 40 30 f c 0 0.2 0.4 0.6 0.8 1 1.2 1.4 IDO Relative protein expression Mock Fn Fn+PMB Fn+ Fn+PMB aTNFa +aTNFa ^ ^ ^ ^ * 0 0.5 1 1.5 2 2.5 IDO Mock Fn Heate-killed TNFa INF-γ Fn n i e t o r p e v it a l e R expression ^ ^ ** Fig. 5 Involvement of IL-6 and TNF-α in the induction of IDO by F. nucleatum-infected THP-1 cells. TNF-α (a) and IL-6 (b) levels in the supernatants of Fn- or heat-killed Fn-infected dTHP1 cells at an MOI of 10:1 for the indicated times. c IDO expression was assessed by Western blot at 48 h. d mRNA expression of IL-6 and TNF-α level in the presence of PMB with the indicated dosages. IDO enzymatic activity in supernatants (e) and representative western blots for IDO in cells lysates (f) of live Fn-infected dTHP1 cells (MOI: 10:1) in the presence of neutralizing antibodies to TNF-α and/or PBM for 48 h. Bars represent the mean ± SD of the results from triplicate determinations. ^Below the detection limit. P < 0.05, P < 0.01, P < 0.001 (PBL), an LDH release assay was conducted using dTHP1 cells infected by Fn for 48 h as targets. Ofﬁcial journal of the Cell Death Differentiation Association F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes Columns indicate the mean of six replicate measurements, and bars indicate the SD (P < 0.01, **P < 0.001) ﬁcial journal of the Cell Death Differentiation Association Fig. 6 F. nucleatum survives and undergoes limited intracellular proliferation in THP-1-derived macrophages. Intracellular bacteria proliferation was assessed by gentamycin protection assay. Fn-dTHP1 cells with or without treated by 1-MT were lysed at the indicated time-points after infection, and the numbers of total viable bacteria (a) and viable bacteria per macrophage (b) were determined by the serial dilution method. c Immunoﬂuorescence staining of intracellular Fn (red) was observed by confocal microscope at 24 h and 72 h (×60). d The apoptotic cells were analyzed by ﬂow cytometry at 72 h. Bars represent the mean ± SD of the results from triplicate determinations. P < 0.05, P < 0.01, P < 0.001. Scale bar = 20 μm survives and undergoes limited intracellular proliferation in THP-1-derived macrophages. Intracellular bacteria Fig. 6 F. nucleatum survives and undergoes limited intracellular proliferation in THP-1-derived macrophages. Intracellular bacteria proliferation was assessed by gentamycin protection assay. Fn-dTHP1 cells with or without treated by 1-MT were lysed at the indicated time-points after infection, and the numbers of total viable bacteria (a) and viable bacteria per macrophage (b) were determined by the serial dilution method. c Immunoﬂuorescence staining of intracellular Fn (red) was observed by confocal microscope at 24 h and 72 h (×60). d The apoptotic cells were analyzed by ﬂow cytometry at 72 h. Bars represent the mean ± SD of the results from triplicate determinations. P < 0.05, P < 0.01, P < 0.001. Scale bar = 20 μm a b 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 Concentration of tryptophan (mg/ml) OD595 OD595 Concentration of kynurenine (mg/ml) * * * * * * * * * *** Fig. 7 Growth of F. nucleatum is regulated by tryptophan and kynurenine in a dose-dependent manner. a Fn grew in a tryptophan dose- dependent manner. b The growth of Fn was inhibited with kynurenine in a dose-dependent manner. Bacterial growth was assessed in enriched BHI broth by spectrophotometry. F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes PBLs from three healthy volunteers whose feces exhibited high levels of Fn-DNA by PCR detection were stimulated by IL-2 for 96 h as effectors. PBLs lysed the target cells at 50:1 E/T ratios when exposed to CM, whereas the lysis rate was remarkably reduced when the PBLs cells were exposed to CM from Fn-dTHP1 cells (P < 0.001). The cytolytic activity of PBLs cells was effectively restored when exposed to CM from Fn-dTHP1 cells in the presence of 1- MT (Fig. 8c). Next, the mixture of targets and effectors cells were collected and lysed to release intracellular Ofﬁcial journal of the Cell Death Differentiation Association Page 9 of 14 Xue et al. Cell Death and Disease (2018) 9:355 acteria. The CFU counts of Fn (0.072 per macrophage) ere increased when PBLs cells were exposed to CM from n-dTHP1 cells, whereas the CFU counts of Fn (0.031 per macrophage) were effectively restored when exposed to CM from Fn-dTHP1 cells in the presence of 1-MT (Fig. 8d). These results indicated that exposure to the micro- environment created by Fn-infected macrophages severely Fig. 6 F. nucleatum survives and undergoes limited intracellular proliferation in THP-1-derived macrophages. Intracellular bacteria proliferation was assessed by gentamycin protection assay. Fn-dTHP1 cells with or without treated by 1-MT were lysed at the indicated time-points after infection, and the numbers of total viable bacteria (a) and viable bacteria per macrophage (b) were determined by the serial dilution method. c mmunoﬂuorescence staining of intracellular Fn (red) was observed by confocal microscope at 24 h and 72 h (×60). d The apoptotic cells were analyzed by ﬂow cytometry at 72 h. Bars represent the mean ± SD of the results from triplicate determinations. P < 0.05, P < 0.01, P < 0.001. Scale bar = 20 μm a b 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 Concentration of tryptophan (mg/ml) OD595 OD595 Concentration of kynurenine (mg/ml) * * * * * * * * * * * Fig. 7 Growth of F. nucleatum is regulated by tryptophan and kynurenine in a dose-dependent manner. a Fn grew in a tryptophan dose- dependent manner. b The growth of Fn was inhibited with kynurenine in a dose-dependent manner. Bacterial growth was assessed in enriched BHI broth by spectrophotometry. Ofﬁcial journal of the Cell Death Differentiation Association F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes Columns indicate the mean of six replicate measurements, and bars indicate the SD (P < 0.01, P < 0.001) a 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 Concentration of tryptophan (mg/ml) OD595 * * * * b 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.01 0.02 0.04 0.08 0.16 0.32 0.64 OD595 Concentration of kynurenine (mg/ml) * * * * * * b Concentration of tryptophan (mg/ml) Concentration of kynurenine (mg/ml) Fig. 7 Growth of F. nucleatum is regulated by tryptophan and kynurenine in a dose-dependent manner. a Fn grew in a tryptophan dose- dependent manner. b The growth of Fn was inhibited with kynurenine in a dose-dependent manner. Bacterial growth was assessed in enriched BHI broth by spectrophotometry. Columns indicate the mean of six replicate measurements, and bars indicate the SD (P < 0.01, *P < 0.001) bacteria. The CFU counts of Fn (0.072 per macrophage) were increased when PBLs cells were exposed to CM from Fn-dTHP1 cells, whereas the CFU counts of Fn (0.031 per macrophage) were effectively restored when exposed to CM from Fn-dTHP1 cells in the presence of 1-MT (Fig. 8d). These results indicated that exposure to the micro- environment created by Fn-infected macrophages severely Ofﬁcial journal of the Cell Death Differentiation Association Xue et al. Cell Death and Disease (2018) 9:355 Page 10 of 14 0 0.02 0.04 0.06 0.08 0.1 CM Fn-CM Fn-CM+1-MT a b 0 2 4 6 8 10 CM Fn-CM Fn-CM+1-MT 0% 20% 40% 60% 80% 100% 120% CM Fn-CM Fn-CM+1-MT c Proliferation rate ) M u / M u ( p r T / n y K s i s y l e v it a l e R Numbers of viable Fn Per macrophage d * *** * 0% 20% 40% 60% 80% ET 50:1 CM Fn-CM Fn-CM+1-MT * *** Fig. 8 Effects of IDO in the conditional media (CM) of Fn-dTHP1 cells on the proliferation and cytolytic activity of lymphocytes cells. a Concentrations of tryptophan (Kyn) and kynurenine (Try)were measured by HPLC in the supernatants of dTHP1 or Fn-dTHP1 cells (MOI 10:1), supernatants of dTHP1 cells (CM), supernatants of Fn-infected cells (CM-Fn), or infected cell supernatants plus 1-MT (100 μM) (Fn-CM + 1-MT) for 48 h. b MTT proliferation assay of Jurkat T cells with indicated CM at 48 h. F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes c PBLs derived from PBMCs were activated by IL-2 and cultured in the CM, CM-Fn, or the CM with 1-MT (100 μM) (Fn-CM + 1-MT) for 48 h. Cytolytic activity against the target cell:Fn-dTHP1 was evaluated using a standard LDH release assay. d Live Fn numbers were determined by a gentamycin protection assay in a mixture of Fn-dTHP1 and PBLs. The E:T ratios are 50:1. Bars represent the mean ± SD of the results from three or six independent experiments. P < 0.05, P < 0.01, P < 0.001 a 0 2 4 6 8 10 CM Fn-CM Fn-CM+1-MT ) M u / M u ( p r T / n y K b 0% 20% 40% 60% 80% 100% 120% CM Fn-CM Fn-CM+1-MT Proliferation rate * * c s i s y l e v it a l e R 0% 20% 40% 60% 80% ET 50:1 CM Fn-CM Fn-CM+1-MT * *** 0 0.02 0.04 0.06 0.08 0.1 CM Fn-CM Fn-CM+1-MT Numbers of viable Fn Per macrophage d * ** d c Fig. 8 Effects of IDO in the conditional media (CM) of Fn-dTHP1 cells on the proliferation and cytolytic activity C i f h (K ) d k i (T ) d b HPLC i h f dTHP F dT in the conditional media (CM) of Fn-dTHP1 cells on the proliferation and cytolytic activity of lymphocytes cells. a h (K ) d k i (T ) d b HPLC i h f dTHP1 F dTHP1 ll (MOI 10 1) Fig. 8 Effects of IDO in the conditional media (CM) of Fn-dTHP1 cells on the proliferation and cytolytic activity of lymphocytes cells. a Concentrations of tryptophan (Kyn) and kynurenine (Try)were measured by HPLC in the supernatants of dTHP1 or Fn-dTHP1 cells (MOI 10:1), supernatants of dTHP1 cells (CM), supernatants of Fn-infected cells (CM-Fn), or infected cell supernatants plus 1-MT (100 μM) (Fn-CM + 1-MT) for 48 h. b MTT proliferation assay of Jurkat T cells with indicated CM at 48 h. c PBLs derived from PBMCs were activated by IL-2 and cultured in the CM, CM-Fn, or the CM with 1-MT (100 μM) (Fn-CM + 1-MT) for 48 h. Cytolytic activity against the target cell:Fn-dTHP1 was evaluated using a standard LDH release assay. d Live Fn numbers were determined by a gentamycin protection assay in a mixture of Fn-dTHP1 and PBLs. The E:T ratios are 50:1. F. nucleatum-infected dTHP1 cells escape killing by impairing the cytolytic function of peripheral blood lymphocytes Bars represent the mean ± SD of the results from three or six independent experiments. P < 0.05, P < 0.01, **P < 0.001 reduced lymphocyte cell survival and impaired the cyto- toxic function, providing a potential mechanism for the immune evasion and spread of Fn bacteria via infected macrophages. more, Fn exhibits limited proliferation when it is co- cultured with host cells under aerobic conditions, which demonstrates that the intracellular environment is more suitable than the extracellular environment for the sur- vival of Fn in an un-strict anaerobic environment. This ﬁnding provides a convenient research methodology for the interaction of anaerobic pathogens with host cells. Ofﬁcial journal of the Cell Death Differentiation Association Materials and methods Bacterial culture F. nucleatum (Fn) strain ATCC 25586 was purchased from the China General Microbiological Culture Collec- tion Center (CGMCC, Beijing, China). The organisms were grown anaerobically (AnaeroPack, Bio-Merieux, France) at 37 °C for 72 h on CDC anaerobic blood agar plates (Guangzhou detgerm Microbiology Technology Co. Ltd, Guangzhou, China) or 48 h in brain heart infu- sion (BHI, Oxoid, Hampshire, UK) broth medium before harvesting. Heat-killed (dead) Fn was made by heating at 100 °C for 10 min. Then, live/heat-killed Fn were cen- trifuged and suspended to 1 × 108 colony-forming units (CFUs)/ml with RPMI 1640 (Hyclone Labs, Logan, UT) for infection experiments IDO activation in APCs can potently inhibit the immune response by which tryptophan is depleted in cells in response to the infection, which may reﬂect an anti- parasitic mechanism in humans26. As a tumor-associated bacterium, Fn exhibits similar nutritional regulation with respect to tryptophan metabolism, and both Fn and tumors are sensitive to tryptophan-depleted micro- environments. It is interesting that kynurenine exhibits an inhibitory effect on the growth of Fn, which is also observed in L. monocytogenes33. The IDO-induced high kynurenine and low-tryptophan microenvironment lim- ited the proliferation of Fn and impaired the function of T-cells, which might downregulate anti-Fn or anti-CTL cell responses, leading to Fn persistence in vivo. The inﬂuence of tryptophan (Trp, Sigma-Aldrich, St. Louis, MO) and kynurenine (Kyn, Sigma-Aldrich) on the growth of Fn was assayed using a method described as follows: Fn were harvested in the exponential growth phase and subsequently resuspended and diluted with BHI medium to the appropriate concentration. The twofold serially diluted samples with Try or Kyn (0.01–0.64 mg/ml) were placed into a ﬂat-bottom 96-well microtiter plate and incubated anaerobically for 48 h at 37 °C. After incubation, the bacterial cell culture was measured by spectrophotometry with an iMark micro- plate absorbance reader (Bio-Rad, Philadelphia, USA) at 595 nm to assess bacterial growth. Immune checkpoint blockade therapy has opened a new therapeutic era for cancer therapy by activating anti- tumor T-cell immunity. IDO inhibitors such as D-1-MT, which can suppress the growth of tumor cells by relieving T-cell suppression, have been applied in clinical trials in some tumors34. Given that IDO blockade relieved intra- cellular bacterial and T-cell suppression, we asked whe- ther IDO inhibitors affected Fn multiplication. Discussion In this study, we identiﬁed Fn as an obligate anaerobe that can invade and multiply inside host cells under aerobic co-culture conditions. Fn has been reported pre- viously to survive during the invasion of oxygenated tis- sues in host oral cavities by adapting to oxidative stress with enhanced pathogenicity19. Generally, intracellular oxygen restriction can impair the mitochondrial respira- tory chain to inhibit aerobic bactericidal activity inside macrophages, such as in the case of L. monocytogenes or M. tuberculosis infection20, but this response may provide a suitable hypoxic microenvironment for anaerobic intracellular bacteria. In our study, intracellular Fn exhibited morphological changes into the short rod or spheres shapes that may be responsive to intracellular environmental stress. Similarly, intracellular Helicobacter pylori have been observed the morphology transition from normal helical bacillary to a coccoid shape21. What’s Microbial infections often elicit programmed cell death as part of the host defense system or as a part of the survival strategy of the pathogen, but some intracellular bacteria manipulate the host death and survival pathways to enhance their replication and survival by blocking or inhibiting the macrophage apoptotic responses of the mitochondrial pro-death, NF-κB–dependent pro-survival, and inﬂammasome-dependent host cell death pathways during infection22. In our study, Fn infection kept host cells alive up to 72 h, wheraes they reproduced by sup- pressing host cells apoptosis via the PI3K pathway, sug- gesting macrophage-mediated dissemination of Fn infection in the body. Canonical Th1 cytokine IFN-γ is critical for innate and adaptive immunity, especially in response to intracellular bacterial infections23. Macrophages activated by IFN-γ Ofﬁcial journal of the Cell Death Differentiation Association Page 11 of 14 Xue et al. Cell Death and Disease (2018) 9:355 growth of bacteria, largely inside macrophages. Similarly, some anti-inﬂammatory drug, such as TNFα blockers, which have been used for the treatment of certain patients with rheumatoid arthritis and Crohn’s disease, contain warnings for serious infections from some associated disease-causing bacterial pathogens, including Salmo- nella, Legionella, and Listeria35,36. These data suggest that therapy with IDO inhibitors might result in the exacer- bation of infection and bacteria spread throughout the body via infected macrophages in some Fn-associated cancers, such as CRC. exhibit increased pinocytosis and receptor-mediated phagocytosis as well as an enhanced microbial killing ability to restrain intracellular bacterial replication. Discussion Fn- infected macrophages were activated in a classically M1- polarized manner in our study, but with the absence IFN- γ secretion, which alleviated the effect of host cells against intracellular bacteria. Similar results were observed in Fn- infected natural killer (NK) cells with inhibition of IFN-γ secretion; infection inhibited NK cell cytotoxicity via the interaction of the Fap2 protein of Fn with TIGIT 24. Similarly to other intracellular bacteria, Fn infection induced high TNF-α and IL-6 secretion. A continuous increase of IL-6 secretion during infection mediates host defense and cell survival in different bacterial infections25. Furthermore, TNF-α is produced early in bacterial infection as a Trojan horse to ensure their intracellular replication26. A wide spectrum of microbes has acquired elegant mechanisms to overcome or deﬂect the host responses mediated by TNF-α27. TNF-α-mediated acti- vation of IDO has been reported in microbial infections28. Our data indicate that Fn induces IDO expression in response to TNF-α and LPS stimuli in dTHP1 cells. These results are consistent with previous studies that infection with viruses, including HIV and EBV, as well as with some bacteria, such as Haemophilus ducreyi and L. mono- cytogenes, induced IDO expression in macrophages29–32. In addition, TNF-α altered type 1 immune activation in part by suppressing T-cell proliferation during myco- bacterial infection 26. Taken together, these ﬁndings indicate that Fn is able to survival and undergo limited intracellular proliferation in macrophages, and the induction of IDO expression cre- ates a tryptophan-deﬁcient and kynurenine-rich toxic microenvironment inside macrophages to inhibit the proliferation of both intracellular Fn bacteria and T lymphocytes in the microenvironment, thereby allowing Fn-infected macrophages to escape attack by CTL cells. Moreover, Fn inside infected macrophages evade cell- intrinsic death by activating the PI3K and ERK pathways to inhibit host cell apoptosis. In conclusion, Fn escape is regulated by nutritional requirements that are similar to those of T cells. Our ﬁndings suggest that IDO inhibitors may aggravate infection in Fn-associated tumor therapy. Ofﬁcial journal of the Cell Death Differentiation Association Materials and methods Bacterial culture Indeed, we found that Fn-infected dTHP1 cells exhibited a sig- niﬁcantly greater bacterial load after treatment with 1- MT, demonstrating that repression of IDO relieved the Page 12 of 14 Xue et al. Cell Death and Disease (2018) 9:355 CM treatment dTHP1 cells were cultured in 6-well plates (1.5 × 106 cells per well) in the absence or presence of live/heat- killed Fn for 24 h, and the medium was then replaced by fresh medium, with or without 100 µM 1-methyl-D-tryp- tophan (1-MT, Sigma-Aldrich). Twenty-four hours after medium replacement, the culture medium was harvested as CM and used for the incubation of Jurkat T cells or PBLs. Quantitative RT-PCR Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque plus gradient centrifugation of leukopacks derived from three healthy volunteers with positive Fn-DNA in stool. Informed consent was obtained from the volunteers prior to participation in accordance with the human experimentation guidelines of the Insti- tute Research Ethics Committee of the Cancer Centre, Sun Yat-Sen University (No: GZR2012-123). Monocytes and phagocytes were removed by adherence to plastic by culturing PBMCs in RPMI 1640 containing 50 U/ml IL-2 (Peprotech, Rocky Hill, USA) for 4–5 h. The non- adherent peripheral blood lymphocyte (PBL) fraction was harvested and cultured with 100 U/ml IL-2 in com- plete RPMI 1640 medium (containing 10% FCS, 100 U/ml penicillin and 100 μg/ml streptomycin). The total mRNA of the cells was extracted after treat- ment for the indicated time. First strand cDNA synthesis was carried out from 800 ng of total RNA. The quantiﬁ- cation of the target and reference (18s RNA) genes was performed in triplicate using a LightCycler® 480 II (Roche Diagnostics, Mannheim, Germany). The primers used in the real-time PCR reaction are presented in Table S2. Immunoblotting Total protein extracts were extracted by using a lysis buffer and protease inhibitor (Beyotime Biotechnology) from cultured cells after treatment for the indicated time intervals. Equivalent protein amounts were separated by SDS-PAGE and transferred onto polyvinylidene diﬂuoride membranes. After being blocked with 5% non-fat dry milk in PBS containing 0.05% Tween-20, the blotted mem- branes were incubated with polyclonal antibodies against CCR7, CXCR4 (1:800, Bioworld, St. Louis, MN), HLA-DR (1:800, Abcam, Cambridge, UK) or monoclonal antibodies against IDO (1:1000, Cell Signaling Technology Inc., Beverly, MA), AKT (1:1000, Abcam), p-AKT (1:1000, Cell Signaling Technology Inc.), ERK1/2 (1:1000, Cell Signal- ing Technology Inc.), or p-ERK1/2 (1:1000, Cell Signaling Technology Inc.), as well as a horseradish peroxidase- conjugated secondary antibody. The GAPDH protein levels were also determined using a speciﬁc antibody (1:5000, Bioworld) as a loading control. Western blot signal was quantiﬁed by ImageJ software. The human monocyte cell line THP-1 and T lympho- cyte cell line Jurkat (Cell Bank of Chinese Academy of Sciences, Shanghai, China) were grown in RPMI 1640 supplemented with 10% fetal bovine serum (FBS). THP-1 monocytes were differentiated into macrophages (THP-1 derived macrophages, dTHP1) by treatment with 10 nM PMA (phorbol 12-myristate 13-acetate, Sigma- Aldrich) for 48 h. For bacterial and host cells co-culture experiments, dTHP1 cells were added at the indicated concentrations to live/heat-killed Fn and cultured for the indicated times under a humidiﬁed 5 % CO2 atmosphere at 37 °C in a CO2 incubator. For blocking experiments, dTHP1 cells were pre-incubated with 10 µg/ml of neutralizing antibodies against TNF-α mouse IgG1 [mIgG1] (eBioscience, San Diego, CA) or 100 µg/ml of polymyxin B (PMB, Sigma- Aldrich) for 2 h at 37 °C before treated. Measurement of IDO activity The cell culture medium was mixed with trichloroacetic acid and then centrifuged. Subsequently, the supernatant was injected onto a C-18 column and eluted with KH2PO4. The concentrations of Trp and Kyn were ana- lyzed by HPLC (Waters). Trp was measured by the detection of its native ﬂuorescence at 285-nm excitation and 365-nm emission wavelength. Kyn was detected by UV absorption at the 360-nm wavelength in the same chromatographic run, and the results were processed using Breeze version 3.30 SPA software. IDO activity was determined by calculating the Kyn to Try ratio (Kyn/Trp, μM/μM). Ofﬁcial journal of the Cell Death Differentiation Association Flow cytometry dTHP1 cells that were 80% conﬂuent were treated with Fn for 24 h. Cells were observed under an inverted microscope (Nikon TE 300), harvested in 5 mM EDTA in PBS and washed. For analysis of CCR7, CD206, and HLA- DR, cells were resuspended with 300 µl of PBS and 5 µl of PE-conjugated anti-human CCR7 and CD206 (eBioscience) or FITC-conjugated anti-human leukocyte antigen (HLA)-DR (BD Biosciences, Erembodegem, Bel- gium) for 15 min and then analyzed by a FACS Calibur (Beckman-Coulter, Miami, USA). For analysis of cell apoptosis, cells were resuspended with annexin-binding buffer and then stained with annexin V and propidium iodide (PI) according to the manufacturer’s instructions (BD Biosciences). Apoptotic cells were analyzed by a FACS Calibur (Beckman-Coulter). Immunohistochemistry The sections were immersed in a 3% hydrogen peroxide solution for 10 min to block endogenous peroxidase activity and were incubated with the primary antibody rabbit anti-human IDO (1:700, Cell Signaling Technology, Inc.) at 4 °C overnight. A negative control was performed by replacing the primary antibody with PBS. The sections were then incubated with a horseradish peroxidase- labeled secondary antibody (1:100, Boster, Wuhan, China) at room temperature for 120 min. Finally, the signal was developed for visualization with 3,3′-diamino- benzidine tetrahydrochloride, and all of the slides were counterstained with hematoxylin. Intracellular survival assays were permeabilized by the addition of 0.5% Triton X-100, and incubated with prepared mouse anti-Fn polyclonal antibodies (home-made, mouse immunized by heat-killed Fn, 1:100). Following this, cells were labeled with Cy3- labeled goat anti-mouse IgG (1:400, Boster, China) as well as FITC-labeled goat anti-human (1:50, Boster, China) for 30 min at 37 °C. DAPI staining for nuclear standing <10 min. After every step, cells were washed three times with PBS. Finally, cells were imaged at ×60 magniﬁcation using the Zeiss LSM710 confocal micro- scope (Zeiss, Oberkochen, Germany). Using this protocol, bacteria external to the host cell were labeled with both Cy3 and FITC, whereas bacteria inside the cells were labeled with Cy3 only (appearing only red when channels were merged). y Bacterial infection for intracellular entry and prolifera- tion was assessed according to the gentamycin protection assay. Brieﬂy, dTHP1 cells were seeded in a 24-well plate at a density of 1 × 106 cells per well in RPMI 1640 with 10% FBS. After 24 h, cells were infected for 2 h with live Fn bacteria at a MOI of 10:1. For blocking experiments, dTHP1 cells were pre-incubated with 1-MT (100 μM) for 2 h prior to infection. After infection, dTHP1 cells were washed three times with RPMI 1640 and then incubated for 2 h in RPMI 1640 containing 10% FBS and gentamicin (50 μg/ml) to remove extracellular Fn. Subsequently, the culture wells were transferred to RPMI medium con- taining a lower dose of gentamycin (10 μg/ml) and 0 or 10 µM 1-MT and cultured for the indicated times. Ulti- mately, dTHP1 cells were lysed with 0.1% Triton X-100 and 0.01% SDS in PBS, and serial dilutions were spread onto anaerobic blood agar plates to quantify the number of internalized (intracellular) bacteria with a mean of six replicate measurements. Cytotoxicity assay The cytotoxic activity of the PBLs was determined by a standard lactate dehydrogenase (LDH) release assay using a CytoTox 96® kit (Beyotime Biotechnology, China) fol- lowing the manufacturer’s instructions. Brieﬂy, the Fn- infected dTHP1 cells (5 × 104 cells/well) were cultured as target cells. IL-2-stimulated PBLs from three healthy volunteers were incubated in different CM as the treated effector cells. The target cell and effector cell suspensions were co-cultured at effector-to-target (E:T) ratios of 50:1. After 4 h of incubation, the release of LDH into the supernatant was quantiﬁed by recording the absorbance at 490 nm. The percentage of cytotoxicity was calculated as described by the manufacturer. Cytokine analysis 1School of Public Health (Shenzhen), Sun Yat-sen University, Guangdong, China. 2Department of Microbial and Biochemical Pharmacy, School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China. 3Department of Clinical Laboratory Medicine, Guangzhou First Municipal People’s Hospital, Guangzhou Medical University, Guangzhou, China The qualiﬁcation of cytokines in the CM was performed using a BD™Cytometric Bead Array panel kit (BD Bios- ciences). The analytes included in the 6-plex kit were as follows: IL-2, IL-4, IL-6, IL-10, gamma interferon (IFN-γ), and TNF-α. The 6 cytokines were measured by ﬂow cytometry according to the manufacturer’s instructions. Statistical analysis All values are presented as the mean ± SD. Paired t tests were used to analyse data unless otherwise indicated. We report the nominal P value for each comparison without adjusting for multiple testing. A P value < 0.05 was con- sidered statistically signiﬁcant. Acknowledgements g This work was supported by the National Natural Science Foundation of China (No. 81372573) Cell viability assay Bacteria invasion assays were carried out using a dif- ferential staining immunoﬂuorescence procedure as pre- viously similar described34. Brieﬂy, Fn-infected cells were washed with PBS at least three times to remove non- adherent bacteria and then ﬁxed with 4% paraformalde- hyde, and blocked in 10% (v/v) normal goat serum. Human anti-Fn polyclonal antibody (home-made, puriﬁed IgG from serum of a CRC patient with high anti-Fn level35, 1:100) was incubated overnight at 4 °C. Then, cells Cell proliferation or viability was measured using an MTT cell proliferation kit (Beyotime Biotechnology, Shanghai, China) following the manufacturer’s instructions. Brieﬂy, cells were seeded into 96-well plates at a density of 5 × 104 cells per well and were cultured for 12 h. Then, cells were infected with live/heat-killed bacteria or treated with CM for the indicated times. The absorbance was measured at 570 nm by an iMark Microplate Absorbance Reader. Ofﬁcial journal of the Cell Death Differentiation Association Page 13 of 14 Xue et al. Cell Death and Disease (2018) 9:355 Ofﬁcial journal of the Cell Death Differentiation Association References 6. Bullman, S. et al. Analysis of Fusobacterium persistence and antibiotic response in colorectal cancer. Science 358, 1443–1448 (2017). 6. Bullman, S. et al. Analysis of Fusobacterium persistence and antibiotic response in colorectal cancer. Science 358, 1443–1448 (2017). 28. Schmidt, S. V. & Schultze, J. L. New insights into IDO biology in bacterial and viral infections. Front. Immunol. 5, 384 (2014). 7. Rubinstein, M. R. et al. Fusobacterium nucleatum promotes colorectal carci- nogenesis by modulating E-cadherin/β-catenin signaling via its FadA adhesin. Cell Host. Microbe 14, 195–206 (2013). 7. Rubinstein, M. R. et al. Fusobacterium nucleatum promotes colorectal carci- nogenesis by modulating E-cadherin/β-catenin signaling via its FadA adhesin. Cell Host. Microbe 14, 195–206 (2013). 29. Fu, X., Lawson, M. A., Kelley, K. W. & Dantzer, R. HIV-1 Tat activates indoleamine 2,3-dioxygenase in murine organotypic hippocampal slice cultures in a p38mitogen-activated protein kinase-dependent manner. J. Neuroinﬂamm. 8, 88 (2011). 8. Yang, Y. et al. Fusobacterium nucleatum increases proliferation of colorectal cancer cells and tumor development in mice by activating toll-like receptor 4 signaling to nuclear factor-κB, and up-regulating expression of microRNA-21. Gastroenterology 152, 851–866 (2017). 30. Liu, W. L. et al. Epstein-Barr virus infection induces indoleamine 2,3-dioxy- genase expression in human monocyte-derived macrophages through p38/ mitogen-activated protein kinase and NF-κB pathways: impairment in T cell functions. J. Virol. 88, 6660–6671 (2014). gy 9. Wang, H. F. et al. Evaluation of antibody level against Fusobacterium nucleatum in the serological diagnosis of colorectal cancer. Sci. Rep. 6, 33440 (2016). 31. Yeung, A. W. et al. Flavivirus infection induces indoleamine 2,3-dioxygenase in human monocyte-derived macrophages via tumor necrosis factor and NF-κB. J. Leukoc. Biol. 91, 657–666 (2012). g g p 10. Johnson, T. S. & Munn, D. H. Host indoleamine 2,3-dioxygenase: contribution to systemic acquired tumor tolerance. Immunol. Invest. 41, 765–797 (2012). 11. Uyttenhove, C. et al. Promoter methylation modulates indoleamine 2,3-diox- ygenase 1 induction by activated T cells in human breast cancers. Cancer Immunol. Res. 5, 330–344 (2017). 32. Li, W., Katz, B. P. & Spinola, S. M. Haemophilus ducreyi lipooligosaccharides induce expression of the immunosuppressive enzyme indoleamine 2,3-diox- ygenase via type I interferons and tumor necrosis factor alpha in human dendritic cells. Infect. Immun. 79, 3338–3347 (2011). 12. Fallarino, F. et al. T cell apoptosis by kynurenines. Adv. Exp. Med. Biol. 527, 183–190 (2003). 33. Niño-Castro, A. et al. Publisher's note 17. Ibana, J. A. et al. Inhibition of indoleamine 2,3-dioxygenase activity by levo-1- methyl tryptophan blocks gamma interferon-induced Chlamydia trachomatis persistence in human epithelial cells. Infect. Immun. 79, 4425–4437 (2011). Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. 18. Sasaki-Imamura, T., Yano, A. & Yoshida, Y. Production of indole from l- tryptophan and effects of these compounds on bioﬁlm formation by Fuso- bacterium nucleatum ATCC 25586. Appl. Environ. Microbiol. 76, 2460–2468 (2010). Conﬂict of interest Th h d l Conﬂict of interest The authors declare that they have no conﬂict of interest. The authors declare that they have no conﬂict of interest. Ofﬁcial journal of the Cell Death Differentiation Association Page 14 of 14 Page 14 of 14 Xue et al. Cell Death and Disease (2018) 9:355 Xue et al. Cell Death and Disease (2018) 9:355 Xue et al. Cell Death and Disease (2018) 9:355 Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41419-018-0389-0). Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41419-018-0389-0). 19. Silva, V. L. et al. Enhanced pathogenicity of Fusobacterium nucleatum adapted to oxidative stress. Microb. Pathog. 39, 131–138 (2005). Received: 8 October 2017 Revised: 4 February 2018 Accepted: 5 February 2018 20. Wiese, M. et al. Hypoxia-mediated impairment of the mitochondrial respiratory chain inhibits the bactericidal activity of macrophages. Infect. Immun. 80, 1455–1466 (2012). 21. Amieva, M. R. & El-Omar, E. M. Host-bacterial interactions in Helicobacter pylori infection. Gastroenterology 134, 306–323 (2008). References 22. Ashida, H. et al. Cell death and infection: a double-edged sword for host and pathogen survival. J. Cell Biol. 195, 931–942 (2011). 1. Halstead, S. B., Mahalingam, S., Marovich, M. A., Ubol, S. & Mosser, D. M. Intrinsic antibody-dependent enhancement of microbial infection in macrophages: disease regulation by immune complexes. Lancet Infect. Dis. 10, 712–722 (2010). 1. Halstead, S. B., Mahalingam, S., Marovich, M. A., Ubol, S. & Mosser, D. M. Intrinsic antibody-dependent enhancement of microbial infection in macrophages: disease regulation by immune complexes. Lancet Infect. Dis. 10, 712–722 (2010). 23. Schoenborn, J. R. & Wilson, C. B. Regulation of interferon-gamma during innate and adaptive immune responses. Adv. Immunol. 96, 41–101 (2007). 23. Schoenborn, J. R. & Wilson, C. B. Regulation of interferon and adaptive immune responses. Adv. Immunol. 96, 4 24. Gur, C. et al. Binding of the Fap2 protein of Fusobacterium nucleatum to human inhibitory receptor TIGIT protects tumors from immune cell attack. Immunity 42, 344–355 (2015). 2. Hornef, M. W., Wick, M. J., Rhen, M. & Normark, S. Bacterial strategies for overcoming host innate and adaptive immune responses. Nat. Immunol. 3, 1033–1040 (2002). 25. Quinton, L. J. et al. Alveolar epithelial STAT3, IL-6 family cytokines, and host defense during Escherichia coli pneumonia. Am. J. Respir. Cell Mol. Biol. 38, 699–706 (2008). 3. Allen-Vercoe, E., Strauss, J. & Chadee, K. Fusobacterium nucleatum: an emerging gut pathogen? Gut Microbes 2, 294–298 (2011). 3. Allen-Vercoe, E., Strauss, J. & Chadee, K. Fusobacterium nucleatum: an emerging gut pathogen? Gut Microbes 2, 294–298 (2011). 4. Yu, T. et al. Fusobacterium nucleatum promotes chemoresistance to colorectal cancer by modulating autophagy. Cell 170, 548–563 (2017). 4. Yu, T. et al. Fusobacterium nucleatum promotes chemoresistance to colorectal cancer by modulating autophagy. Cell 170, 548–563 (2017). 4. Yu, T. et al. Fusobacterium nucleatum promotes chemoresistan 26. Zganiacz, A. et al. TNF-alpha is a critical negative regulator of type 1 immune activation during intracellular bacterial infection. J. Clin. Invest. 113, 401–413 (2004). cancer by modulating autophagy. Cell 170, 548–563 (2017). 5. Kostic, A. D. et al. Fusobacterium nucleatum potentiates intestinal tumorigen- esis and modulates the tumor-immune microenvironment. Cell Host. Microbe 14, 207–215 (2013). 5. Kostic, A. D. et al. Fusobacterium nucleatum potentiates intestinal tumorigen- esis and modulates the tumor-immune microenvironment. Cell Host. Microbe 14, 207–215 (2013). 27. Rahman, M. M. & McFadden, G. Modulation of tumor necrosis factor by microbial pathogens. PLoS Pathog. 2, e4 (2006). References The IDO1-induced kynurenines play a major role in the antimicrobial effect of human myeloid cells against Listeria monocytogenes. Innate. Immun. 20, 401–411 (2014). 13. de Souza Sales, J. et al. The role of indoleamine 2,3-dioxygenase in lepromatous leprosy immunosuppression. Clin. Exp. Immunol. 165, 251–263 (2011). 34. Hou, D. Y. et al. Inhibition of indoleamine 2,3-dioxygenase in dendritic cells by stereoisomers of 1-methyl-tryptophan correlates with antitumor responses. Cancer Res. 67, 792–801 (2007). 14. Zelante, T., Fallarino, F., Bistoni, F., Puccetti, P. & Romani, L. Indoleamine 2,3- dioxygenase in infection: the paradox of an evasive strategy that beneﬁts the host. Microbes Infect. 11, 133–141 (2009). 35. Lanternier, F. et al. Incidence and risk factors of Legionella pneumophila pneumonia during anti-tumor necrosis factor therapy: a prospective French study. Chest 144, 990–998 (2013). 15. Popov, A. et al. Indoleamine 2,3-dioxygenase-expressing dendritic cells form suppurative granulomas following Listeria monocytogenes infection. J. Clin. Invest. 116, 3160–3170 (2006). 36. Davies, R. et al. Inﬂuence of anti-TNF patient warning regarding avoidance of high risk foods on rates of Listeria and salmonella infections in the UK. Ann. Rheum. Dis. 72, 461–462 (2013). 16. Blumenthal, A. et al. M. tuberculosis induces potent activation of IDO-1, but this is not essential for the immunological control of infection. PLoS One 7, e37314 (2012). Ofﬁcial journal of the Cell Death Differentiation Association
https://openalex.org/W2756960894	https://discovery.dundee.ac.uk/ws/files/32811074/W17_5105.pdf	English	null	Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models	null	2,017	cc-by	6,875	University of Dundee Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models Lawrence, John; Reed, Chris Published in: Proceedings of the 4th Workshop on Argument Mining DOI: 10.18653/v1/W17-5105 Publication date: 2017 Licence: CC BY Document Version Publisher's PDF, also known as Version of record Link to publication in Discovery Research Portal Citation for published version (APA): Lawrence, J., & Reed, C. (2017). Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models. In Proceedings of the 4th Workshop on Argument Mining (pp. 39-48). Article W17-5105 Association for Computational Linguistics. https://doi.org/10.18653/v1/W17-5105 University of Dundee Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models L J h R d Ch i Document Version Publisher's PDF, also known as Version of record Citation for published version (APA): Lawrence, J., & Reed, C. (2017). Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models. In Proceedings of the 4th Workshop on Argument Mining (pp. 39-48). Article W17-5105 Association for Computational Linguistics. https://doi.org/10.18653/v1/W17-5105 General rights Copyright and moral rights for the publications made accessible in Discovery Research Portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Download date: 24. Oct. 2024 Download date: 24. Oct. 2024 Mining Argumentative Structure from Natural Language text using Automatically Generated Premise-Conclusion Topic Models John Lawrence and Chris Reed Centre for Argument Technology, University of Dundee, UK Abstract within a piece of natural language text. By auto- matically identifying this structure and its associ- ated premises and conclusions, we are able to tell not just what views are being expressed, but also why those particular views are held. Argument mining has recently been enjoying rapid growth, propelled by three drivers: ﬁrst, the academic and commercial success of opinion mining and sen- timent analysis techniques upon which argument mining builds; second, a strong commercial ap- petite for such technologies from companies such as IBM; and third, the development of infrastruc- ture and tools for (Bex et al., 2013), and theoretical understanding of (Budzynska et al., 2014), argu- ment structure in both monologue and dialogue. This paper presents a method of extract- ing argumentative structure from natural language text. The approach presented is based on the way in which we under- stand an argument being made, not just from the words said, but from existing contextual knowledge and understanding of the broader issues. We leverage high- precision, low-recall techniques in order to automatically build a large corpus of infer- ential statements related to the text’s topic. These statements are then used to produce a matrix representing the inferential rela- tionship between different aspects of the topic. From this matrix, we are able to determine connectedness and directional- ity of inference between statements in the original text. By following this approach, we obtain results that compare favourably to those of other similar techniques to classify premise-conclusion pairs (with re- sults 22 points above baseline), but with- out the requirement of large volumes of annotated, domain speciﬁc data. The intuition underlying the work presented here is that there are rich and predictable thematic and lexical commonalities present in the expres- sion of human reasoning, and that these common- alities can be identiﬁed in helping to extract the structure of reasoning. For example, in debates concerning abortion, arguments are carefully mar- shalled on both sides, with religious themes more typically appearing on one side, and feminist phi- losophy themes more typically on the other. For a debate on the construction of a new road, we may ﬁnd environmental issues on one side and eco- nomic concerns on the other. If such generalisa- tions are possible at a coarse scale, perhaps they are similarly possible at a more ﬁne-grained scale. Proceedings of the 4th Workshop on Argument Mining, pages 39–48 Copenhagen, Denmark, September 8, 2017. c⃝2017 Association for Computational Linguistics Proceedings of the 4th Workshop on Argument Mining, pages 39–48 Copenhagen, Denmark, September 8, 2017. c⃝2017 Association for Computational Linguistics 39 1 Introduction The continuing growth in the volume of data which we produce has driven efforts to unlock the wealth of information this data contains. Auto- matic techniques such as Opinion Mining and Sen- timent Analysis (Liu, 2010) allow us to determine the views expressed in a piece of textual data, for example, whether a product review is positive or negative. Existing techniques struggle, however, to identify more complex structural relationships between concepts. These themes are represented in terms of both the topics discussed and the language used to ex- press them: an anti-abortion stance is likely to cover, not just feminist philosophy themes in gen- eral, but to use speciﬁc terminology more fre- quently, perhaps mentioning ‘choice’ or ‘free- dom’ more than views expressed on the other side. When humans hear such a debate, they understand the structure of the arguments being made not only based on the content of the argument itself, but Argument Mining is the automatic identiﬁ- cation of the argumentative structure contained 39 on a broad general knowledge of the topic and the way in which such arguments are commonly pre- sented. (Houngbo and Mercer, 2014), a straightforward feature of co-referring text using the word “this” is used to build a self-annotating corpus extracted from a large biomedical research paper dataset. This is achieved by collecting pairs of sequential sentences where the second sentence begins with “This method...”, “This result...”, or “This conclu- sion...”, and then categorising the ﬁrst sentence in each pair respectively as Method, Result or Con- clusion sentences. The argument mining technique which we present in this paper takes the commonly occur- ring terms in the original text and then uses these terms to gather data from the web on the same topic. This large volume of additional data can be considered as contextual knowledge, and is pro- cessed to ﬁnd pairs of text spans which have an inferential relationship. We then use these pairs to create premise-conclusion topic models, reﬂecting the ways in which one topic or phraseology is used to support another. Similarly, in (Habernal and Gurevych, 2015), unsupervised features are developed for argument component identiﬁcation which exploit clustering of unlabelled argumentative data from online de- bate portals. Al-Khatib et al. (2016) likewise leverage online debate portals, applying distant su- pervision to automatically create a large annotated corpus with argumentative and non-argumentative text segments from several domains. 1 Introduction Previous work (Lawrence and Reed, 2015) has shown that discourse indicators such as because and therefore are very reliable predictors of argu- ment structure. Unfortunately they are also rather rare, occurring with fewer than 10% of argumen- tative inference steps. With a high-precision/low- recall technique such as is provided by these in- dicators, it becomes possible to process large amounts of text to extract a dataset in which we can have high conﬁdence. This dataset can be used to capture topical regularities in the argument structure which can then be exploited in analysing text which does not beneﬁt from the presence of indicators. Our approach to expanding the data available on the topic under discussion relies on the high precision identiﬁcation of inferential relationships shown by the presence of discourse indicators. Discourse indicators are explicitly stated linguis- tic expressions of the relationship between state- ments (Webber et al., 2011), and, when present, can provide a clear indication of argumentative structure. For example, if we take the sentence “Britain should disarm because it would set a good example for other countries”, then this can be split into two separate propositions “Britain should dis- arm” and “it [disarming] would set a good exam- ple for other countries”. The presence of the word “because” between these two propositions clearly tells us that the second is a reason for the ﬁrst. 2 Related Work The results presented show that discourse markers are again important features for the discrimination of claims and premises in German as well as English language texts. There are many different ways in which indi- cators can appear, and a wide range of relations which they can suggest (Knott, 1996). For auto- matic corpus construction, the ability to identify all of these connections is not relevant and we are able to concentrate solely on those indicators of- fering a very high chance of describing an infer- ential relationship. Using discourse indicators to build such a cor- pus is supported by the work done in identify- ing implicit discourse relations, for example (Lin et al., 2009; Park and Cardie, 2012), where a range of relations labelled in the Penn Discourse Tree- Bank (Prasad et al., 2008), but not explicitly in- dicated, were identiﬁed using features from those relations where an explicit indicator did occur. These implicit relations were identiﬁed with accu- racies of between 70-80% in one-vs-others tests, clearly suggesting that studying cases where indi- cators are present can give a strong indication of a relationship in those cases where they are omitted. Identifying the argumentative structure con- tained within a piece of text can be viewed as a two-step process: Firstly, identifying the in- dividual units of discourse which the text con- tains (commonly referred to as ‘Argumentative Discourse Units’ or ADUs (Peldszus and Stede, 2013)); and then, determining the ways in which these propositions are connected. Figure 1 shows the AIF compliant representa- tion of a fragment of the Moral Maze dialogue. In this ﬁgure, the blue boxes represent individual ADUs, while the arrows show connections, and the diamonds detail the nature of these connec- tions. In this case, the conclusion “I know bankers who behave absolutely splendidly” is supported by the individual premises “who are major benefac- tors”, “who spend their Christmases manning soup kitchens”, and “Think about Bill Gates and all the wonderful things that his money is doing”. The relationship between the topics being ex- pressed in a piece of text and the argumentative structure which it contains have been previously explored in (Lawrence et al., 2014), where a La- tent Dirichlet Allocation (LDA) topic model is used to determine the topical similarity of consec- utive propositions in a piece of text. 2 Related Work The majority of the work carried out to date in the ﬁeld of argument mining, has used either a super- vised learning approach (e.g. (Palau and Moens, 2009; Feng and Hirst, 2011; Stab and Gurevych, 2014)), or a linguistic rule-based approach ((Vil- lalba and Saint-Dizier, 2012; Pallotta and Del- monte, 2011; Wyner et al., 2012)), to determine argumentative function. In both cases these efforts are limited by a lack of consistently annotated ar- gument data. Whilst resources such as the Internet Argument Corpus (IAC) (Walker et al., 2012) and AIFdb (Lawrence et al., 2012), offer rapidly grow- ing volumes of high quality argument analyses, they do not provide the large volumes of data re- quired to train a robust classiﬁer, particularly when considered in the context of a speciﬁc topic or do- main. Discourse indicators have been previously used as a component of argument mining techniques, for example in (Stab and Gurevych, 2014) indi- cators are used as a feature in multiclass classiﬁ- cation of argument components, with each clause classiﬁed as a major claim, claim, premise or non-argumentative. Similar indicators are used in (Wyner et al., 2012), along with domain terminol- ogy (e.g. camera names and properties) to high- light potential argumentative sections of online product reviews. In (Eckle-Kohler et al., 2015) a German language corpus is annotated with ar- guments according to the common claim-premise model of argumentation and the connection be- tween these annotated connections and the pres- ence of discourse indicators (or discourse markers Attempts have been made to mitigate this con- straint by the automatic creation of argument cor- pora, however, the datasets produced are limited to very speciﬁc types of data. For example, in 40 Maze1. Speciﬁcally, we look at the episode from July 4th 20122 on the morality of the banking system. Manual argumentative analy- sis was performed on the transcript, using the OVA+ (Online Visualisation of Argument) anal- ysis tool (Janier et al., 2014) to create a se- ries of argument maps capturing the structure using the Argument Interchange Format (AIF) (Ches˜nevar et al., 2006). A corpus containing the full manual analysis of the transcript can be found online at http://corpora.aifdb. org/bankingsystem. The corpus comprises 5,768 words, split across 327 propositions, with 128 inferential connections (premise/conclusion relations) between them. as they are referred to here) is investigated. 1http://www.bbc.co.uk/programmes/ b006qk11 2http://www.bbc.co.uk/programmes/ b01kbj37 2 Related Work The intuition is that if a proposition is similar to its predecessor then there exists some argumentative link between them, whereas if there is low similarity between a proposition and its predecessor, the author is go- ing back to address a previously made point and, in this case, the proposition is compared to all those preceding it to determine whether they should be connected. Using this method a precision of 0.72, and recall of 0.77 are recorded when comparing the resulting structure to a manual analysis, how- ever it should be noted that what is being identiﬁed here is merely that an inference relationship ex- ists between two propositions, and no indication is given as to the direction of this inference. We can see from this example that the broad concept of charitable works is being used to sup- port the idea that bankers are good people. The knowledge that these premises are both themati- cally related and support the character of a group of people, whilst clear to a human analyst, is not explicitly indicated in the original text. For our purposes, we are aiming to identify inferential connections between pairs of ADUs. Whilst a complete argument mining pipeline would require the automation of this segmenta- tion, this is outside the scope of this paper, and the focus of much additional research within the ar- gument mining ﬁeld (Lawrence et al., 2014; Mad- 1http://www.bbc.co.uk/programmes/ 4 Implementation An overview of the methodology used can be seen in Figure 2. Starting with raw, natural language text, manual segmentation is performed to split the text into ADUs. From here these segments are examined in order to ﬁnd those unigrams and bi- grams which occur most frequently throughout the text, giving an indication of the overall theme of the text which we are working with. 3 Experimental Data The data used in this paper is taken from a transcript of the BBC Radio 4 program Moral 41 Figure 1: Fragment of Manually Analysed Argument Structure from the BBC Radio 4 program Moral Maze Figure 1: Fragment of Manually Analysed Argument Structure from the BBC Radio 4 program Moral Maze nani et al., 2012; Saint-Dizier, 2012). As such, we use the same segmentation carried out for the manual analysis, and split the possible ADU pairs into those which are connected by an inferential relationship, and those which are not. An alternative approach would be to use the premise/conclusion dataset as training data for a supervised machine learning approach. This is limited by the fact that we only obtain positive examples, and, whilst techniques such as PU- learning (Learning from Positive and Unlabelled examples) (Liu et al., 2003) provide a way of deal- ing with only positively labelled data, we do not have sufﬁcient quantities of unlabelled examples for these techniques to be applied. In future work, the ability to identify arbitrary ADUs in text could be used to extract large volumes of unlabelled ex- amples, and such approaches may then become more suitable. 4.1 Obtaining Premise/Conclusion Pairs The next step is then to build a corpus of re- lated documents by searching the web for those unigram and bigram terms identiﬁed as being in- dicative of the theme. From this extended corpus, we then extract sentences which contain an infer- ential relationship by searching for those discourse indicators which we have found to have the highest precision. This search results in a large collection of pairs of text fragments where one of the pair is a premise supporting the other. The ﬁrst step in the pipeline described above is to determine the overall theme of the text being anal- ysed. This was performed by looking for those un- igrams and bigrams which occur most frequently throughout the text. With the text previously seg- mented into ADUs, we calculated the number of unique ADUs in which each unigram or bigram appeared. This list is then sorted and ﬁltered to re- move common stop words. The resulting lists of terms can be seen in Table 1 and Table 2. Using these fragments as documents, we then generate a Latent Dirichlet Allocation (LDA) (Blei et al., 2003) topic model, and from this create a matrix showing the probability of support between each of the identiﬁed topics. By matching pairs of ADUs from the original text against the probabil- ities in this matrix, we are then able to determine the probability that there is an inferential relation- ship between them, and by thresholding these val- ues, we can then categorise ADU pairs as being ‘inferential’ or ‘non-inferential’. Having identiﬁed keywords describing the topic, a corpus of related documents was created by searching the web for combinations of these terms. The top ten terms of each kind were com- bined into search queries by taking all possible combinations of two and three unigrams as well as each bigram both on its own and paired with each unigram. Using these queries, the ﬁrst 200 Google search results for each were compiled. Af- ter ﬁltering the list of related documents to remove 42 Raw Text ADUs Unigrams & Bigrams Manual Segmentation ADU Frequency Corpus of Related Documents Web Search Extract Inferential Statements x1 because y1 x2 because y2 ... 3http://www.crummy.com/software/ BeautifulSoup/ 4http://www.nltk.org/ 4.1 Obtaining Premise/Conclusion Pairs Create Topic Model Using x's and y's LDA Topic Model Topical Inference Matrix ADU Topic Probabilities Classified Support Relations Figure 2: Overview of the Implementation Methodology for Creating Extended Corpus, Creating a Top- ical Inference Matrix and Classifying Support Relations Corpus of Related Documents Unigrams & Bigrams Manual Segmentation ADUs Topical Inference Matrix ADU Topic Probabilities Classified Support Relations Figure 2: Overview of the Implementation Methodology for Creating Extended Corpus, Creating a Top- ical Inference Matrix and Classifying Support Relations Figure 2: Overview of the Implementation Methodology for Creating Extended Corpus, Creating a Top- ical Inference Matrix and Classifying Support Relations Unigram Count investment 39 banking 35 banks 28 money 27 problem 16 capitalism 13 culture 12 behaviour 12 rules 12 ethical 10 Table 1: Top ten unigrams by number of ADUs in which they appear duplicates, a total of 6,981 pages remained. Although the pages identiﬁed in the previous step are high ranking search results for the terms identiﬁed, such pages commonly contain material unrelated to the topic, for example, advertisements and summaries of other articles. In order to ex- tract those sections of the documents most likely to contain the body of an article, the Python Beau- tiful Soup library3 was used to parse the HTML and extract consecutive paragraphs of text. These paragraphs were then split into sentences, using the NLTK4 tokeniser, and each of the re- sulting sentences searched for the presence of a discourse indicator. Previous work using dis- course indicators to identify argumentative struc- ture (Lawrence and Reed, 2015) has shown that, although not common enough to give a full repre- sentation of the structure, when present, discourse indicators give a very clear indication of the argu- mentative connection between two spans of text. As our aim is to extract only those sentences most likely to contain an inferential relationship, we ﬁrst looked more closely at the relative perfor- mance of different indicators. 4.1 Obtaining Premise/Conclusion Pairs Based on analysis of a separate Moral Maze episode, we identiﬁed those indicators showing the highest precision (the Table 1: Top ten unigrams by number of ADUs in which they appear Bigram Count investment banks 18 investment banking 12 common good 5 immoral behaviour 3 free market 3 banking industry 3 wealth creation 3 redeemed capitalism 2 moral code 2 dutch bankers 2 Table 2: Top ten bigrams by number of ADUs in which they appear Bigram Count investment banks 18 investment banking 12 common good 5 immoral behaviour 3 free market 3 banking industry 3 wealth creation 3 redeemed capitalism 2 moral code 2 dutch bankers 2 Table 2: Top ten bigrams by number of ADUs in which they appear 43 Indicator Precision Recall therefore 0.95 0.0004 because 0.91 0.0031 consequently 0.82 0.0001 hence 0.76 0.0001 accordingly 0.74 0.0002 so 0.73 0.0005 after 0.69 0.0011 since 0.65 0.0008 then 0.58 0.0013 for 0.57 0.0006 Table 3: Top ten discourse indicators sorted by precision Indicator Precision Recall therefore 0.95 0.0004 because 0.91 0.0031 consequently 0.82 0.0001 hence 0.76 0.0001 accordingly 0.74 0.0002 so 0.73 0.0005 after 0.69 0.0011 since 0.65 0.0008 then 0.58 0.0013 for 0.57 0.0006 Table 3: Top ten discourse indicators sorted by precision topic model, the sentences were ﬁrst split where the indicator occurred, giving two documents for each sentence (one representing a premise, and the other, the conclusion). For our experiments, the model was created with forty topics using 20 passes over the supplied corpus. From the probability distributions for each pair of conclusion (C) and premise (P), a topical infer- ence matrix (T) was created, where the i,jth entry in the matrix corresponds to the product of prob- abilities that the premise has topic i and the con- clusion topic j. For example, in the simplest case, if there is a probability of 1.0 that the premise has topic m and the conclusion topic n, then the ma- trix will contain 1.0 at m,n and zero for all other possible pairings. So, given topic distributions θC for the conclusion, and θP for the premise, T is deﬁned thus: Table 3: Top ten discourse indicators sorted by precision precision and recall for the top ten indicators can be seen in Table 3). 4.1 Obtaining Premise/Conclusion Pairs These results show that, when present, “therefore” and “because” give the high- est indication of inference with a signiﬁcant drop in accuracy for the remaining indicators. As such, we limited our generated corpus to only those sen- tences containing one of these two words. ti,j = θP i ∗θC j (1) (1) To investigate the validity of our assumption that there would be a noticeable pattern in the re- lationships between topic and inference, we ﬁrst created a combined topical inference matrix for each of the because relations identiﬁed, by sum- ming all of the matrices resulting from these rela- tions. We then looked at the entropy of this matrix calculated as the sum of the differences between each value in the matrix and the mean of all val- ues. For the because matrix, the mean score was 3.67 and the total difference was 2275.58, giving an average difference of 1.42 for each item in the matrix from the mean value (with no relationship between topic and inference, this difference would be ∼0). Where the number of words either before or after the matching indicator was less than 5, the sentence was discarded. After carrying out this process, a total of 7,162 inferential sentences were identiﬁed (6,288 containing “because” and 874 containing “therefore”), giving a dataset of premise conclusion pairs, either premise therefore conclusion or conclusion because premise. Whilst we do not have 100% precision for ei- ther of the discourse indicators used, the impact of this is mitigated by the way in which the result- ing pairs are subsequently used. The use of the topic models described in the next section means that we neither need all of the inferential rela- tions contained within our search results, or for every premise conclusion pair to be correctly la- belled as such. The models which we produce may have a small amount of noise generated by false- positives, but these either comprise topics which are not then matched to elements from the original text, or add a small number of lower importance terms to a valid topic. A corresponding matrix was then produced for the therefore relations, and the distance between the because and therefore matrices calculated. 4.1 Obtaining Premise/Conclusion Pairs This calculation was performed by ﬁrst scaling the values in each matrix to a value between zero and one, and then calculating the distance between the resulting matrices: d(A, B) = v u u t n X i=1 n X j=1 (ai,j −bi,j)2 (2) (2) 5https://radimrehurek.com/gensim/ 5.1 Using the Topical Inference Matrix to determine directionality The manual analysis of our original text contained 128 premise conclusion pairs. As an initial ex- periment, we investigated how well the produced topical inference matrix could determine the direc- tion of the inference between these pairs. This was achieved by creating a test set containing each pair (a,b) and its reverse (b,a). Two alternative methods were tested to clas- sify these pairs as being ‘inferential’ or ‘non- inferential’. In each case, the topic probabili- ties for the ADUs were ﬁrst inferred from the LDA model and a score determined as to whether there was an inferential relationship. For the ﬁrst method, (MaxTopic), the score was calculated by taking the highest probability topic for each ADU and using these to look up the corresponding value in the overall topical inference matrix: 5 Experiments Table 4: Results for the MaxTopic and TopicDist methods to determine directionality of inferential connections compared to the random baseline Table 4: Results for the MaxTopic and TopicDist methods to determine directionality of inferential connections compared to the random baseline Table 4: Results for the MaxTopic and TopicDist methods to determine directionality of inferential connections compared to the random baseline In order to test our original hypotheses that the thematic commonalities present in the expression of human reasoning can be identiﬁed and used to help determine the structure of that reasoning, a number of experiments were carried out to explore the effectiveness of using this data to determine both the direction of inference between two ADUs that are known to have an inferential relationship, and the connectedness of pairs of arbitrary ADUs. The results for directionality can be seen in Ta- ble 4. The results show an improvement over the random baseline for both methods, however the improvement in precision is low when just looking at the highest scoring topic. One reason for this is that a reasonable percentage of pairs (twenty- ﬁve out of one hundred and twenty-six) have the same highest scoring topic for both items (i.e. a conclusion is being supported by a premise that is closely related). When these same topic pairs are removed, the precision increases to 0.56, compa- rable to the results for the weighted topic distribu- tion. The results for using the weighted topic dis- tribution are better, and suggest that even in cases where the main topic is similar, there is enough of a difference in the secondary topics to determine the directionality of the pair. 4.2 Creating the Topical Inference Matrix To extract the topical nature of the premise conclu- sion pairs previously identiﬁed, a Latent Dirich- let allocation (LDA) topic model was created us- ing the Python gensim library5. To produce this For identical matrices, this distance would be zero, for a pair of 40 × 40 matrices where all en- tries have maximal difference, the distance would be 40, and for a pair of 40 × 40 matrices where all entries have an average difference of 0.5 (in- 44 dicating no correlation between the two), the dis- tance would be 28.29. The distance between the because and therefore matrices was calculated as 18.32, suggesting a positive correlation between the two. We are not aware of any other technique that can be used to quantify the signiﬁcance be- tween such datasets: our analysis indicates merely that there is indeed some pattern beyond random chance linking the two concepts. STopicDist = n X i=1 n X j=1 ti,j ∗θP i ∗θC j (4) (4) For each of these two methods, the resulting scores were then compared against the mean of all values in the matrix (mean = 3.15), over which a pair would be classiﬁed as being ‘inferential’, and below which, ‘non-inferential’. Finally, the because and therefore matrices were summed to give an overall topical inference ma- trix. Method Precision Recall F1-score Random Baseline 0.5 0.5 0.5 MaxTopic 0.51 0.82 0.63 TopicDist 0.57 0.83 0.67 Table 4: Results for the MaxTopic and TopicDist methods to determine directionality of inferential connections compared to the random baseline Method Precision Recall F1-score Random Baseline 0.5 0.5 0.5 MaxTopic 0.51 0.82 0.63 TopicDist 0.57 0.83 0.67 5.3 Thresholding Topical Values The experiments presented so far have looked at the likelihood that one topic supports another in terms of its score relative to all other scores in the matrix. However, it is possible that for some top- ics the scores will generally be higher. For exam- ple, if a large number of propositions have a high probability of corresponding to topic n, then all the values in column n of the matrix will be dispropor- tionately high. To overcome any problems caused by this kind of topical skew, we took each column of the matrix and divided each value by the sum of values in that column. This resulting scaled matrix was then used to perform the same experiments as previously. The results for both experiments com- bined are shown in Table 6. Although we focus on identifying patterns of in- ference within a single debate, there is nothing in- trinsic to the approach that makes it a better ﬁt for this domain than any other. The automatic deter- mination of the domain being discussed requires only the original text, and from this we are able to build a dataset speciﬁc to that domain which, due to the reliability of discourse indicators, contains domain speciﬁc pairs that we can say with high conﬁdence have an inferential relationship. Method Precision Recall F1-score Directionality Random Baseline 0.5 0.5 0.5 MaxTopic 0.61 0.77 0.68 TopicDist 0.65 0.78 0.71 Connectedness Random Baseline 0.5 0.5 0.5 MaxTopic 0.59 0.75 0.66 TopicDist 0.64 0.83 0.72 Table 6: Results for the MaxTopic and TopicDist methods to determine connectedness and direc- tionality using a thresholded inference matrix Method Precision Recall F1-score Directionality Random Baseline 0.5 0.5 0.5 MaxTopic 0.61 0.77 0.68 TopicDist 0.65 0.78 0.71 Connectedness Random Baseline 0.5 0.5 0.5 MaxTopic 0.59 0.75 0.66 TopicDist 0.64 0.83 0.72 7 Conclusion Table 6: Results for the MaxTopic and TopicDist methods to determine connectedness and direc- tionality using a thresholded inference matrix Table 6: Results for the MaxTopic and TopicDist methods to determine connectedness and direc- tionality using a thresholded inference matrix This work has demonstrated how by automatically creating large, high-conﬁdence datasets of infer- ential pairs related to a speciﬁc topic, we can closely mirror one of the ways in which humans understand the complex interactions between the individual propositions expressed in a debate. In all cases, we can see that the precision is slightly improved, though (with the exception of the TopicDist results for connectedness) this is at the expense of recall. The approach presented is effective in tackling the challenging high-level pragmatic task of iden- tifying both connectedness and directionality be- tween argumentative discourse units, with results 22 points above baseline. 5.2 Using the Topical Inference Matrix to determine connectedness The second experiment performed looked at whether the produced topical inference matrix could determine inferential connections between arbitrary pairs of ADUs. For this task, a dataset was created containing the known 126 premise conclusion pairs and an equal number of random, unconnected ADUs. The same two methods of classifying these pairs as being ‘inferential’ or ‘non-inferential’ were used as in the ﬁrst experi- ment, and the results can be seen in Table 5. SMaxTopic = tmax(θP ),max(θC) (3) (3) SMaxTopic = tmax(θP ),max(θC) For the second method, (TopicDist), the values in the matrix were multiplied by the correspond- ing probabilities for each item in the pair and then summed to give an overall score. The results show that the precision is increased for classifying pairs as being connected over the previous results for directionality. 45 Method Precision Recall F1-score Random Baseline 0.5 0.5 0.5 MaxTopic 0.58 0.79 0.67 TopicDist 0.60 0.82 0.69 Table 5: Results for the MaxTopic and Top- icDist methods to determine connectedness of ADU pairs Method Precision Recall F1-score Random Baseline 0.5 0.5 0.5 MaxTopic 0.58 0.79 0.67 TopicDist 0.60 0.82 0.69 Table 5: Results for the MaxTopic and Top- icDist methods to determine connectedness of ADU pairs similarity in topics for which the ADUs are con- nected (in a signiﬁcant percentage of cases the maximum probability topic was the same). Currently, the identiﬁcation of relationships is limited to inferential relationships, and one area of development would be to extend this by examining those discourse indicators which show a conﬂict relationship. Additionally, no account is taken of the polarity or sentiment of the ADUs. Where we have a conclusion, ‘C’, and a premise, ‘P’, then there would be a high topical similarity between P and ‘not P’, and as such, an inference relationship would be assigned between them. This problem could be overcome by applying sentiment classi- ﬁcation to the ADUs as a preliminary step, and where there is negation of one item in the pair, replacing an inference relationship with conﬂict. Expanding the scope of this technique to give a fuller indication of relations will be carried out in future work. Table 5: Results for the MaxTopic and Top- icDist methods to determine connectedness of ADU pairs 6 Discussion The results we have presented show in all cases that there is some correlation identiﬁed between the topics that a pair of ADUs have, and the na- ture of their potential inferential relationship. By looking at the topics of each item in the pair, we have been able to determine both connectivity and directionality of inference. Overall, the results are better for identifying connectedness than di- rectionality, predominantly resulting from higher This outcome represents strong performance for this level of task (cf., for example, (Feng and Hirst, 2011; Peldszus, 2014)), giving results compara- ble to those of (Palau and Moens, 2009), where each Argument sentence was classiﬁed as either premise or conclusion with F1-scores of 0.68 for classiﬁcation as premise and 0.74 for conclusion. Furthermore, where existing approaches are often 46 constrained in their generality by a lack of ap- propriately annotated, domain-speciﬁc, data, the same requirement does not apply in this case. Vanessa Wei Feng and Graeme Hirst. 2011. Clas- sifying arguments by scheme. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies-Volume 1. Association for Computa- tional Linguistics, pages 987–996. The results show a clear link between the words used to express an argument and its underlying structure, and strongly support the intuition that understanding the structure of an argument re- quires not only consideration of the text itself, but contextual knowledge and understanding of the broader issues. We see this work as a key compo- nent in a larger ensemble approach (Lawrence and Reed, 2015), mirroring the complex process fol- lowed by a human annotator whereby general do- main knowledge, understanding of linguistic cues and familiarity with common patterns of reasoning are combined to understand the arguments being made. Ivan Habernal and Iryna Gurevych. 2015. Exploiting debate portals for semi-supervised argumentation mining in user-generated web discourse. In Pro- ceedings of the 2015 conference on Empirical Meth- ods in Natural Language Processing (EMNLP). pages 2127–2137. Hospice Houngbo and Robert Mercer. 2014. An au- tomated method to build a corpus of rhetorically- classiﬁed sentences in biomedical texts. In Proceed- ings of the First Workshop on Argumentation Min- ing. Association for Computational Linguistics, Bal- timore, Maryland, pages 19–23. Mathilde Janier, John Lawrence, and Chris Reed. 2014. OVA+: An argument analysis interface. In S. Par- sons, N. Oren, C. Reed, and F. 6 Discussion Cerutti, editors, Proceedings of the Fifth International Conference on Computational Models of Argument (COMMA 2014). IOS Press, Pitlochry, pages 463–464. Acknowledgements This work was funded in part by EPSRC in the UK under grant EP/N014871/1. Alistair Knott. 1996. A data-driven methodology for motivating a set of coherence relations. Ph.D. the- sis, Department of Artiﬁcial Intelligence, University of Edinburgh. References Khalid Al-Khatib, Henning Wachsmuth, Hagen Matthias Stein, Jonas K¨ohler, and Benno Stein. 2016. Cross-domain mining of argumentative text through distant supervision. In Proceedings of NAACL-HLT. pages 1395–1404. John Lawrence, Floris Bex, Chris Reed, and Mark Snaith. 2012. AIFdb: Infrastructure for the argu- ment web. In Proceedings of the Fourth Interna- tional Conference on Computational Models of Ar- gument (COMMA 2012). pages 515–516. Floris Bex, John Lawrence, Mark Snaith, and Chris Reed. 2013. Implementing the argument web. Com- munications of the ACM 56(10):66–73. John Lawrence and Chris Reed. 2015. Combining ar- gument mining techniques. In Proceedings of the 2nd Workshop on Argumentation Mining. Associ- ation for Computational Linguistics, Denver, CO, pages 127–136. David M. Blei, Aandrew Y. Ng, and Michael I. Jordan. 2003. Latent dirichlet allocation. The Journal of Machine Learning Research 3:993–1022. Katarzyna Budzynska, Mathilde Janier, Chris Reed, Patrick Saint-Dizier, Manfred Stede, and Olena Yaskorska. 2014. A model for processing illocu- tionary structures and argumentation in debates. In Proceedings of the 9th edition of the Language Re- sources and Evaluation Conference (LREC). pages 917–924. John Lawrence, Chris Reed, Colin Allen, Simon McAl- ister, and Andrew Ravenscroft. 2014. Mining ar- guments from 19th century philosophical texts us- ing topic based modelling. In Proceedings of the First Workshop on Argumentation Mining. Asso- ciation for Computational Linguistics, Baltimore, Maryland, pages 79–87. Carlos Ches˜nevar, Sanjay Modgil, Iyad Rahwan, Chris Reed, Guillermo Simari, Matthew South, Gerard Vreeswijk, Steven Willmott, et al. 2006. Towards an argument interchange format. The Knowledge Engi- neering Review 21(04):293–316. Ziheng Lin, Min-Yen Kan, and Hwee Tou Ng. 2009. Recognizing implicit discourse relations in the penn discourse treebank. In Proceedings of the 2009 Conference on Empirical Methods in Natural Lan- guage Processing: Volume 1-Volume 1. Association for Computational Linguistics, pages 343–351. Judith Eckle-Kohler, Roland Kluge, and Iryna Gurevych. 2015. On the role of discourse markers for discriminating claims and premises in argumen- tative discourse. In Proceedings of the 2015 Con- ference on Empirical Methods in Natural Language Processing (EMNLP). pages 2236–2242. Bing Liu. 2010. Sentiment analysis and subjectivity. Handbook of natural language processing 2:627– 666. 47 Marilyn A Walker, Jean E Fox Tree, Pranav Anand, Rob Abbott, and Joseph King. 2012. A corpus for research on deliberation and debate. In Proceedings of the 8th edition of the Language Resources and Evaluation Conference (LREC). pages 812–817. Bing Liu, Yang Dai, Xiaoli Li, Wee Sun Lee, and Philip S Yu. 2003. References Building text classiﬁers using positive and unlabeled examples. In Proceedings of the Third IEEE International Conference on Data Mining (ICDM-03). IEEE, pages 179–186. Bonnie Webber, Markus Egg, and Valia Kordoni. 2011. Discourse structure and language technology. Natu- ral Language Engineering 18(4):437–490. Nitin Madnani, Michael Heilman, Joel Tetreault, and Martin Chodorow. 2012. Identifying high-level or- ganizational elements in argumentative discourse. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies. Association for Computational Linguistics, pages 20–28. Adam. Wyner, Jodi. Schneider, Katie. Atkinson, and Trevor. Bench-Capon. 2012. Semi-automated argu- mentative analysis of online product reviews. In Proceedings of the Fourth International Conference on Computational Models of Argument (COMMA 2012). pages 43–50. Raquel M. Palau and Marie-Francine Moens. 2009. Argumentation mining: the detection, classiﬁcation and structure of arguments in text. In Proceedings of the 12th international conference on artiﬁcial in- telligence and law. ACM, pages 98–107. Vincenzo Pallotta and Rodolfo Delmonte. 2011. Auto- matic argumentative analysis for interaction mining. Argument & Computation 2(2-3):77–106. Joonsuk Park and Claire Cardie. 2012. Improving im- plicit discourse relation recognition through feature set optimization. In Proceedings of the 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue. Association for Computational Lin- guistics, pages 108–112. Andreas Peldszus. 2014. Towards segment-based recognition of argumentation structure in short texts. In Proceedings of the First Workshop on Argumen- tation Mining. Association for Computational Lin- guistics, Baltimore, Maryland, pages 88–97. Andreas Peldszus and Manfred Stede. 2013. From ar- gument diagrams to argumentation mining in texts: a survey. International Journal of Cognitive Infor- matics and Natural Intelligence (IJCINI) 7(1):1–31. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Milt- sakaki, Livio Robaldo, Aravind K Joshi, and Bon- nie L Webber. 2008. The penn discourse treebank 2.0. In Proceedings of the 6th Language Resources and Evaluation Conference (LREC-2008). Patrick Saint-Dizier. 2012. Processing natural lan- guage arguments with the <TextCoop> platform. Argument & Computation 3(1):49–82. Christian Stab and Iryna Gurevych. 2014. Identify- ing argumentative discourse structures in persuasive essays. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguis- tics, Doha, Qatar, pages 46–56. Maria Paz G. Villalba and Patrick Saint-Dizier. 2012. Some facets of argument mining for opinion anal- ysis. In Proceedings of the Fourth International Conference on Computational Models of Argument (COMMA 2012). pages 23–34. 48
https://openalex.org/W2805691916	https://www.nature.com/articles/s41598-018-20302-7.pdf	English	null	Extracellular DNA as a genetic recorder of microbial diversity in benthic deep-sea ecosystems	Scientific reports	2,018	cc-by	7,557	Extracellular DNA as a genetic recorder of microbial diversity in benthic deep-sea ecosystems C Corinaldesi 1 M Tangherlini2,3 E Manea2 &A Dell’Anno2 Received: 26 June 2017 Accepted: 17 January 2018 Published: xx xx xxxx Received: 26 June 2017 Accepted: 17 January 2018 Published: xx xx xxxx C. Corinaldesi 1, M. Tangherlini2,3, E. Manea2 & A. Dell’Anno2 Extracellular DNA in deep-sea sediments represents a major repository of genes, which previously belonged to living organisms. However, the extent to which these extracellular genes influence current estimates of prokaryotic biodiversity is unknown. We investigated the abundance and diversity of 16S rDNA sequences contained within extracellular DNA from continental margins of different biogeographic regions. We also compared the taxonomic composition of microbial assemblages through the analysis of extracellular DNA and DNA associated with living cells. 16S rDNA contained in the extracellular DNA pool contributed up to 50% of the total 16S rDNA copy number determined in the sediments. Ca. 4% of extracellular Operational Taxonomic Units (OTUs) were shared among the different biogeographic regions revealing the presence of a core of preserved OTUs. A higher fraction of OTUs was exclusive of each region potentially due to its geographic and thermohaline characteristics. Ca. one third of the OTUs identified in the extracellular DNA were absent from living prokaryotic assemblages, possibly representing the signatures of past assemblages. Our findings expand the knowledge of the contribution of extracellular microbial sequences to current estimates of prokaryotic diversity obtained through the analyses of “environmental DNA”, and open new perspectives for understanding microbial successions in benthic ecosystems. Extracellular DNA in surface deep-sea sediments is by far the largest reservoir of DNA of the world oceans1. The main sources of extracellular DNA in such ecosystems are represented by in situ DNA release from dead benthic organisms, and/or other processes including cell lysis due to viral infection, cellular exudation and excretion from viable cells, virus decomposition, and allochtonous inputs from the water column1–4. Previous studies provided evidence that an important fraction of extracellular DNA can escape degradation processes, remaining preserved in the sediments5,6. This DNA represents, potentially, a genetic repository that records biological processes occur- ring over time7,8. g Recent investigations revealed that DNA preserved in marine sediments is characterized by a large number of highly diverse gene sequences7–10. In particular, extracellular DNA has been used to reconstruct past prokaryotic and eukaryotic diversity in benthic ecosystems characterized by low temperatures and/or permanently anoxic conditions10–14. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports/ Figure 1. Total extracellular DNA concentrations in the sediments of continental margins from NE Atlantic and Arctic Oceans and Mediterranean Sea (A); copy number of 16S rRNA genes in the extracellular and intracellular DNA pools from the sediments of continental margins (B). Figure 1. Total extracellular DNA concentrations in the sediments of continental margins from NE Atlantic and Arctic Oceans and Mediterranean Sea (A); copy number of 16S rRNA genes in the extracellular and intracellular DNA pools from the sediments of continental margins (B). and with the gene sequences belonging to living microbial assemblages. Findings reported here provide new insights on ubiquitous and exclusive prokaryotic signatures preserved in different biogeographic regions, and their contribution to the estimates of the current diversity. Results E i Environmental variables. Temperature ranged from −0.84 °C to 13.1 °C (in the Arctic and Mediterranean sites, respectively) and salinity from 34.84 to 38.49 (in the NE Atlantic 1 and Mediterranean sites, respectively). The lowest concentrations of biopolymeric carbon were found in the site NE Atlantic 2 whereas the highest one in the Arctic margin (1.33 ± 0.17 and 4.14 ± 0.53 and mgC g−1 of sediments, respectively, Table S1). Total extracellular DNA concentrations and 16S rDNA copy number. The concentrations of total extracellular DNA in surface sediments of the continental margins investigated ranged from 9.4 ± 3.0 μg DNA g−1 to 22.5 ± 4.8 μg DNA g−1 (in the sites NE Atlantic 1 and 2, respectively; Fig. 1A). The concentrations of total extracellular DNA determined in the deepest site of the NE Atlantic 2 were significantly higher than those of NE Atlantic 1 and Arctic sites (p < 0.01), but not significantly different from those of the Mediterranean Sea. p gi yf The copy number of 16S rDNA ranged from 3.2 ± 0.3 × 107 to 32.5 ± 4.1 × 107 g−1 within the extracel- lular DNA pools (in the NE Atlantic 1 and Mediterranean sites, respectively) and from 5.4 ± 0.2 × 107 to 78.5 ± 1.2 × 107 g−1 within the intracellular DNA pools (in the NE Atlantic 1 and Mediterranean Sea, respectively; Fig. 1B). The copy number of 16S rDNA contained in the extracellular and intracellular DNA pools varied signifi- cantly among all the samples investigated (p < 0.01). In all benthic sites the abundances of prokaryotic 16S rDNA in the extracellular DNA pools were significantly lower than in the intracellular DNA pools (p < 0.01), except for the Arctic site where no significant differences were found.h gif The contribution of extracellular 16S rRNA gene copies to the total pool of 16S rDNA copies (as the sum of 16S rRNA gene copies contained in extracellular and intracellular DNA pools) ranged from 11% to 50%, in NE Atlantic site 2 and Arctic site, respectively. Prokaryotic diversity associated with extracellular DNA pools. The number of sequences obtained for each extracellular DNA sample at each step of quality check and during the QIIME analysis are reported in Table S2. The number of OTUs after normalisation to the same number of sequences ranged from 1003 to 1158 in the NE Atlantic 2 and Mediterranean samples, respectively (Table S3, Figure S1). Extracellular DNA as a genetic recorder of microbial diversity in benthic deep-sea ecosystems C Corinaldesi 1 M Tangherlini2,3 E Manea2 &A Dell’Anno2 Previous studies suggested that the preservation of DNA might be also favoured in benthic systems character- ised by high organic matter inputs and sedimentation rates, such as continental margins15,16. These systems, which represent ca. 15% of the global seafloor, are also hotspots of benthic prokaryotic diversity17–19, and therefore they could represent optimal sites to investigate the prokaryotic diversity preserved within extracellular DNA. Spatial distribution of prokaryotic diversity has been intensively studied in benthic deep-sea ecosystems20–23 through the analysis of “environmental DNA” (i.e., the genetic material obtained directly from environmental samples without any obvious signs of biological source material24). However, the extent to which gene sequences contained within extracellular DNA can alter the estimates of the diversity of the present-day prokaryotic assem- blages is unknown25. g In the present study, we utilised the extracellular DNA pool as a recorder of the prokaryotic diversity in the sediments of different sites of continental margins (Atlantic and Arctic Ocean, and Mediterranean Sea). The prokaryotic genetic signatures contained in the different extracellular DNA pools were compared among them, 1Dipartimento di Scienze e Ingegneria della Materia, dell’Ambiente ed Urbanistica, Polytechnic University of Marche, Via Brecce Bianche, 60131, Ancona, Italy. 2Dipartimento di Scienze della Vita e dell’Ambiente, Polytechnic University of Marche, Via Brecce Bianche, 60131, Ancona, Italy. 3Stazione Zoologica A. Dohrn, Villa Comunale, Naples, Italy. Correspondence and requests for materials should be addressed to C.C. (email: c.corinaldesi@univpm.it) ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 1 www.nature.com/scientificreports/ Results E i ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 2 www.nature.com/scientificreports/ Figure 2. Taxonomic composition of prokaryotic assemblages in the extracellular DNA pools from deep- sea benthic ecosystems of continental margins. The taxa, whose OTUs contribute for at least 1% to the whole prokaryotic assemblages, are shown. Figure 2. Taxonomic composition of prokaryotic assemblages in the extracellular DNA pools from deep- sea benthic ecosystems of continental margins. The taxa, whose OTUs contribute for at least 1% to the whole prokaryotic assemblages, are shown. Extracellular DNA pools from all sites investigated were characterized by 16S rDNA gene sequences affiliating to a number of prokaryotic families ranging from 201 (in the Mediterranean Sea) to 236 (NE Atlantic 1; Table S3). Overall, 316 families were identified in the extracellular DNA pools, 39% of which were shared among all pools. From ca. 4% to 8% of the families were exclusive of each site (in Arctic and NE Atlantic 2 sites, respectively). k bl l d b b h b Prokaryotic assemblages were mainly represented by JTB255 marine benthic group (Gammaproteobacteria), which contributed from 6% to 14% to the total number of sequences, unclassified Sh765B-TzT-29 (Deltaproteobacteria, from 8% to 12%) and uncultured bacterium of Subgroup 21 (Acidobacteria, from 4% to 9%), followed by other bacterial and archaeal families including Sva0725 (Subgroup 10, Acidobacteria), Flammeovirgaceae (Cytophagales, Bacteroidetes) and unclassified/ambiguous taxa belonging to Marine Group I (Thaumarchaeota, Fig. 2).hf h The network-based analysis showed the presence of core OTUs (i.e. shared among all the different extracellu- lar DNA pools) belonging to different taxa mainly affiliated with Acidobacteria Subgroup 21, Xantomonadaceae and JTB255 marine benthic group (Fig. 3). Such core OTUs accounted for only ca. 4% to the total OTU number. This analysis also showed the presence of a large fraction of partially shared OTUs (i.e., shared among two or three DNA pools) contributing for ca. 61% to the total number of OTUs. Finally, all extracellular DNA pools were characterized by several OTUs exclusive of each site, accounting, on average, for 35% to the total number of OTUs. SIMPER analyses revealed that prokaryotic assemblages contained within the extracellular DNA pool from the Mediterranean sample were largely dissimilar from those from NE Atlantic 1 or Arctic samples (ca. 58–66%). Results E i OTU network of the extracellular DNA pools from the sediments of the different continental margins. Dot size within the network is proportional to the abundance of sequences for each OTU. Dots circled in red represent extracellular core OTUs, dot circled in yellow are partially shared (among two or more pools) OTUs, dots circled in black are OTUs exclusive of each pool. The core OTUs contributing at least for 20 sequences are shown. The numbers in parentheses represent the number of connections among OTUs and samples: 1 for exclusive OTUs, 2–3 for partially shared OTUs and 4 for core OTUs. Figure 3. OTU network of the extracellular DNA pools from the sediments of the different continental margins. Dot size within the network is proportional to the abundance of sequences for each OTU. Dots circled in red represent extracellular core OTUs, dot circled in yellow are partially shared (among two or more pools) OTUs, dots circled in black are OTUs exclusive of each pool. The core OTUs contributing at least for 20 sequences are shown. The numbers in parentheses represent the number of connections among OTUs and samples: 1 for exclusive OTUs, 2–3 for partially shared OTUs and 4 for core OTUs. In the NE Atlantic site 1, we found some exclusive prokaryotic taxa of extracellular DNA pool belonging to Burkholderia and Thermoplasmatales, whereas no exclusive family was found in the intracellular DNA pools (Fig. 4A). g In the NE Atlantic site 2, exclusive prokaryotic taxa of the intracellular DNA pool mainly belonged to TM6 (although their sequence contribution to the whole assemblage was <0.1%), whereas taxa affiliated to Burkholderiaceae and OD1 mainly belonged to the extracellular DNA pool (Fig. 4B). y g p g Cluster analyses indicated that the assemblages of prokaryotic OTUs of the extracellular and intracellular DNA pools in both the NE Atlantic sites clustered together although with a high level of dissimilarity (ca. 52% on average, as calculated by the SIMPER analysis and 64.5% with binary Jaccard dissimilarity).h g y y y y The analysis of intracellular DNA pools showed that a set of OTUs were shared from the two pools from the NE Atlantic sites (6.5%). This fraction was very similar to the percentage of core OTUs of the extracellu- lar DNA pools from the two NE Atlantic sites (7.4%). Results E i The results of the cluster analysis based on the Jaccard dissimilarity indicated that the different extracellular SIMPER analyses revealed that prokaryotic assemblages contained within the extracellular DNA pool from the Mediterranean sample were largely dissimilar from those from NE Atlantic 1 or Arctic samples (ca. 58–66%). The results of the cluster analysis based on the Jaccard dissimilarity indicated that the different extracellular DNA pools (in terms of OTU presence/absence) from NE Atlantic 1 and NE Atlantic 2 sites clustered together SIMPER analyses revealed that prokaryotic assemblages contained within the extracellular DNA pool from the Mediterranean sample were largely dissimilar from those from NE Atlantic 1 or Arctic samples (ca. 58–66%). The results of the cluster analysis based on the Jaccard dissimilarity indicated that the different extracellular DNA pools (in terms of OTU presence/absence) from NE Atlantic 1 and NE Atlantic 2 sites clustered together whereas Arctic and, particularly, the Mediterranean samples were separated (Figure S2). The results of the cluster analysis based on the Jaccard dissimilarity indicated that the different extracellular DNA pools (in terms of OTU presence/absence) from NE Atlantic 1 and NE Atlantic 2 sites clustered together whereas Arctic and, particularly, the Mediterranean samples were separated (Figure S2). Comparison between the prokaryotic diversity associated with extracellular and intracellular DNA pools. The number of sequences obtained for each extracellular and intracellular DNA sample at each step of quality check and during the QIIME analysis are reported in Table S2. The OTU richness and number of families (obtained using the same number of sequences) identified within each DNA pool are reported in Table S4 while rarefaction curves are shown in Figure S1. g Intracellular DNA pools from the NE Atlantic sites were characterized by 16S rDNA gene sequences belonging to a total of 262 prokaryotic families in both sites (Figure S3) while prokaryotic signatures contained in the extra- cellular DNA pools were affiliated with 268 families. Xanthomonadales, Deltaproteobacteria and Acidobacteria were the dominant prokaryotic taxa either in intracellular or extracellular DNA pools.fi In the NE Atlantic sites 1 and 2, 43% and 46% of the prokaryotic OTUs (mainly affiliated with Subgroup 21, Sh765B-TzT-29 and JTB255) were respectively shared between extracellular and intracellular DNA pools (Fig. 4A,B). Ca. one third of the prokaryotic OTUs were exclusively present in extracellular DNA from each of the two NE Atlantic sites. ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 3 www.nature.com/scientificreports/ Figure 3. Results E i Core OTUs of the intracellular DNA pools belonged to the same prokaryotic taxa as the core OTUs of extracellular DNA pools (JTB255, Acidobacteria Subgroup 21, Xantomonadaceae). Discussion The OTUs, which contribute most to the prokaryotic assemblage in terms of sequences, are colored according to their taxonomic affiliation (at the lowest identifiable taxonomic level). contained in the extracellular DNA pool were not correlated to the total extracellular DNA concentrations. This could be explained by the heterogeneous composition of the extracellular DNA pool, which includes genes differ- ent from the 16S rDNA, and/or its level of fragmentation/damage, which is influenced by a variety of biotic and abiotic factors25.fi We also found that the extracellular 16S rDNA sequences were highly diverse and affiliated to both archaea nd bacterial families, most of which have not been classified yet.h i y More than one third of the prokaryotic families were common to all of the extracellular DNA pools. These families, including those affiliated with the unclassified Subgroup 21 Acidobacteria, the unclassified Sh765BTz-29 Deltaproteobacteria and JTB255 marine benthic group (Xanthomonadales), are typical of benthic deep-sea eco- systems26–29. However, the analysis conducted at the OTU level revealed that the similarity among different extra- cellular DNA pools was very low, suggesting the presence of a low but highly represented (in terms of sequence abundance) number of core OTUs28–30. This was also highlighted by the network analysis, which showed a core of extracellular OTUs (mainly affiliated with Acidobacteria Subgroup 21 and JTB255 marine benthic group), char- acterized by the highest connectivity among samples. Such OTUs were thus present in all biogeographic regions despite their different temperature and salinity values, trophic conditions, sedimentation rates and consequently age of the sediment layer considered (from decades to several hundreds of years). These findings allow us to hypothesize that the prokaryotic core signatures might belong to resident taxa in all sites of continental margins for several decades. On the other hand, we also observed the presence of groups of extracellular OTUs exclusive of each biogeographic region suggesting that different environmental conditions could select specific taxa31 and/or favour the preservation of their genetic signatures12,25. We found that extracellular OTU pools from the two sites of the NE Atlantic margin grouped together, whereas the extracellular OTU pool from the Mediterranean site clustered apart, potentially due to the wide differences in temperature and salinity regimes. Discussion Previous studies suggested that the preservation of DNA in deep-sea sediments may be favoured by an array of factors including anoxic conditions, low temperatures, and high sedimentation rates, as observed in continental margins6,15. In the present study, we found that the total extracellular DNA concentrations, in the sediments of different continental margins, were among the highest values reported so far for deep-sea ecosystems1,2. In par- ticular, we provide evidence that the deepest sediments among those investigated (4902 m depth in the Atlantic Ocean) were characterized by the highest amount of extracellular DNA, suggesting that benthic ecosystems along continental margins, even at abyssal depth, can be a major repository of extracellular DNA. Investigations conducted in subsurface anoxic sediments and in surface sediments of coastal and deep-sea ecosystems revealed that extracellular DNA can preserve sequences of dead prokaryotes and eukaryotes7,8,25. Our findings obtained from surface sediments of continental margins show that extracellular 16S rRNA gene copies accounted for a variable fraction of the total number of 16S rRNA gene copies (as the sum of extracellular and intracellular 16S rDNA gene copies) depending on the investigated site. The number of 16S rRNA gene copies ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 4 www.nature.com/scientificreports/ Figure 4. OTUs shared between extracellular and intracellular DNA pools and exclusive OTUs of each pool from the sediments of NE Atlantic 1 (A) and NE Atlantic 2 (B) sites. Dot size is proportional to the contribution of sequences to each OTU belonging to the different prokaryotic families. The OTUs, which contribute most to the prokaryotic assemblage in terms of sequences, are colored according to their taxonomic affiliation (at the lowest identifiable taxonomic level). Figure 4. OTUs shared between extracellular and intracellular DNA pools and exclusive OTUs of each pool from the sediments of NE Atlantic 1 (A) and NE Atlantic 2 (B) sites. Dot size is proportional to the contribution of sequences to each OTU belonging to the different prokaryotic families. The OTUs, which contribute most to the prokaryotic assemblage in terms of sequences, are colored according to their taxonomic affiliation (at the lowest identifiable taxonomic level). Figure 4. OTUs shared between extracellular and intracellular DNA pools and exclusive OTUs of each pool from the sediments of NE Atlantic 1 (A) and NE Atlantic 2 (B) sites. Dot size is proportional to the contribution of sequences to each OTU belonging to the different prokaryotic families. ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 Materials and Methods d d li Study areas and sampling strategy. Surface sediment samples were collected using a multiple corer in different continental margins of the Mediterranean Sea, NE Atlantic and Arctic Oceans. In particular, sediment samples were collected in the NW Mediterranean Sea at 2342 m depth (42.080 N, 4.682E), in the NE Atlantic Ocean, at two sites at 3475 m (40.167 N, 9.983E, hereafter defined NE Atlantic 1) and 4902 m depth (40.167 N, 10.984E; hereafter defined NE Atlantic 2) and in the Arctic Ocean at 2482 m depth (79.067 N, 4.170E; Table S1). At each site, three sediment cores were collected by independent deployments (n = 3) of multiple corers, which allow the collection of hermetically-closed samples. Immediately after retrieval, the sediment cores were sliced vertically and the top layer (0–2 cm) was stored at −80 °C until laboratory analyses. Moreover, at each site temperature and salinity of bottom waters were measured by CTD.hf The high variability in the sedimentation rates across the investigated sites (Table S1) resulted in different ages (i.e., 12 years for the Mediterranean site, 67 and 667 years for the sites NE Atlantic 1 and 2, respectively, and 105 years for the Arctic site) of the sediments considered25. Biochemical composition of organic matter. The concentrations of proteins, carbohydrates and lipids in the sediment were determined spectrophotometrically36 and expressed as bovine serum albumin, glucose and tripalmitine equivalents, respectively. The sum of the carbohydrate, protein and lipid concentrations converted into carbon equivalents (using the conversion factors of 0.40, 0.49 and 0.75 mg C mg−1, respectively) was defined as the biopolymeric organic carbon37. Total extracellular DNA concentrations. Working conditions and precautions during extracellular DNA analyses are described in Supplementary Information. y pp y To determine the concentrations of total extracellular DNA in the sediment we used a procedure based on the hydrolysis of the extracellular DNA (using commercial nucleases) which does not allow the recovery of the DNA for subsequent molecular studies25,38,39; therefore, contextual aliquots of the same samples were also processed to provide extracellular DNA that was suitable for molecular analyses. p y For the total extracellular DNA analyses, 2.5 mL 0.1 M Tris, 0.1 M NaCl, 1 mM CaCl2, 10 mM MgCl2, pH 7.5, was added to 1 g sediment (wet weight; i.e., buffer: sediment ratio of 2.5:1 [v/w]). www.nature.com/scientificreports/ www.nature.com/scientificreports/ a low level of similarity of the composition of prokaryotic assemblages contained therein (as also revealed by the Jaccard index). Indeed, we observed a high fraction of OTUs exclusive either of the intracellular DNA or of the extracellular DNA, which even contained OTUs affiliated to families not encountered in any of the two intracel- lular DNA pools of the Atlantic margin (such as those belonging to TM6, Burkholderiales and others).h p g ( g g ) The low level of similarity between OTUs of the extracellular and intracellular DNA pools cannot be attributed to methodological artefacts/biases related with cell lysis induced by the procedure here used, and/or sequencing effort applied to our samples. Indeed, the analyses performed indicate that the amount of intracellular DNA/ prokaryotic genes released due to sample manipulation and/or active growth of the cells is negligible33. At the same time, despite the sequencing effort in our samples was not completely exhaustive, the rarefaction curves obtained were not far from reaching the saturation, suggesting that prokaryotic diversity was largely described. Therefore, the presence of OTUs exclusive of the intracellular DNA pool could be explained by different degra- dation rates of the prokaryotic sequences once outside the cells or the sporadic inputs of microbial taxa from the water column8,32. On the other hand, the presence of prokaryotic taxa exclusively found in the extracellular DNA pool suggests that they are no longer part of the living microbial fraction possibly representing the signatures of past assemblages8,12.h p g The exclusive extracellular prokaryotic taxa found in our study have been already reported in other benthic ecosystems through extraction procedures of total sedimentary DNA (without discriminating between DNA associated with extracellular and microbial fractions34,35). However, in these cases, it cannot be excluded that, due to the potential co-extraction of past prokaryotic signatures, the current diversity has been over-estimated. Therefore, we claim for the need of considering the contribution of extracellular DNA in the current estimates of benthic prokaryotic biodiversity obtained through the analyses conducted on the so called “environmental DNA“24. Moreover, the contextual analysis of prokaryotic signatures associated with intracellular DNA and extra- cellular DNA might represent a useful and effective tool for understanding microbial successions from the past to present-day. Discussion Since the sedimenta- tion rates determined in the two NE Atlantic sites were very different, we exclude that the age of the sediment, in the order of hundreds of years, can represent a key factor influencing the preservation of extracellular prokaryotic signatures as previously reported in other benthic ecosystems25.h g p y p y The comparison between extracellular and intracellular DNA pools revealed that they contained a rather simi- lar number of OTUs, which fall within the range of values reported so far for benthic deep-sea ecosystems27,32. Ca. one third of the total OTUs were shared between extracellular and intracellular DNA pools in each site indicating ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 5 Materials and Methods d d li The pellets containing prokaryotic cells were treated with DNase I to exclude any possible contamination from residual extracellular DNA, and then processed for DNA extraction by using the UltraClean soil DNA isolation kit (MoBio Laboratories Inc., CA, USA) according to the manufacturer’s instructions.hl ) g The extracellular DNA contained in the supernatant fluid was recovered by adding 1 volume of cetyltrimeth- ylammonium bromide solution (1% CTAB in 50 mM, Tris 10 mM EDTA, pH 8.0) and further precipitation after incubation at 65 °C for 30 min, cooling on ice and centrifugation at 10000 × g for 15 min at 4 °C. The pellet was suspended in high-salt TE buffer (pH 8.0) added to 0.6 volumes of cold isopropanol, incubated for 2 h at −20 °C and centrifuged at 10000 × g for 15 min at 4 °C. The pellets containing the extracellular DNA were suspended in MilliQ water and purified using the Wizard PCR clean-up system (Promega).h Q pi g p y ( g ) The details of the analyses conducted to test the absence of cell lysis potentially induced by the procedure described above are provided in the Supplementary Information. Quantitative PCR analyses. Quantitative PCR (qPCR) analyses of the prokaryotic 16S rDNA gene copies were carried out both in the extracellular and intracellular DNA pools that were extracted con- textually from the same sediment samples39. qPCR analyses were performed using the TaqMan tech- nology40. The prokaryotic 16S rDNA sequences were amplified using the universal primers Uni340F (5′-CCTACGGGRBGCASCAG-3′) and Uni806R (5′-GGACTACNNGGGTATCTAAT-3′). The TaqMan probe was Uni516F (5′-TGYCAGCMGCCGCGGTAAHACVNRS-3′), which contained a fluorescent reporter dye (6-carboxyfluorescein) covalently attached to the 5′-end, and a fluorescent quencher dye (6-carboxytetramethylrhodamine) attached six or more bases downstream of the reporter dye41. All of the real-time PCR was performed in a volume of 25 μL with an iQ5-icycler (Bio-Rad) using iQ Supermix (2×; Bio-Rad) contain- ing 40 mM Tris-HCl, pH 8.4, 100 mM KCl, 0.4 mM each dNTP (dATP, dCTP, dGTP, dTTP), 50 U/mL hot-start iTaq DNA polymerase, and 6 mM MgCl2. To amplify prokaryotic genes, 40 PCR cycles were used: as 95 °C for 15S, and 57 °C for 5 min, which were preceded by 3 min of Taq activation at 95 °C. Negative controls were performed by using the same reaction mixture without DNA template. To quantify the 16S rDNA, calibration curves were obtained from a standard solution of E. Materials and Methods d d li Aliquots of the sediment samples (n = 3) had additions of DNase I (1.9 U mL−1), nucleases P1 and S1 (4.0 and 2.3 U mL−1, respectively), and exonuclease-3 (1.9 U mL−1); another set of replicates was added to an equal volume of buffer (without the enzymes) and used as a control. The samples were incubated at room temperature for 2 h under gentle agitation, and then centrifuged at 2,000 × g for 5 min, with the resultant supernatants used to determine the DNA released from the sediments. The supernatants were dried under vacuum and analysed fluorometrically using diami- nobenzoic acid. The fluorescence of the hydrolysed DNA was converted into concentrations using calibration curves obtained from standard solutions of Escherichia coli DNA (from 0.05 to 5.0 µg DNA mL−1). The amounts of extracellular DNA hydrolysed by the nucleases were obtained from the differences between the DNA concen- trations of the enzyme-treated samples and the control samples.h The extracellular DNA concentrations were expressed as micrograms of total extracellular DNA per gram dry ediment. Extraction of extracellular and intracellular DNA pools suitable for molecular analyses. The extra- cellular and intracellular DNA used for molecular analyses were recovered contextually from the same sediment sample following the protocol developed by Corinaldesi et al.39, with some modifications to increase the extraction efficiency. The robustness and reliability of this protocol for extracellular DNA extraction has been shown to exclude ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 6 www.nature.com/scientificreports/ any possible contamination due to cell lysis for co-extraction of DNA contained in the microbial cells10,39. Briefly, the sediment samples were treated with acid-washed polyvinylpolypyrrolidone (0.05% final concentration) and SDS (final concentration, 0.1%). Then the samples were chilled on ice, centrifuged, and the supernatants were transferred to sterile tubes. The sediment pellets were washed again with sodium phosphate buffer (pH 8.0) and centrifuged. These steps were repeated 6 times. Supernatants were combined and centrifuged for 20 min at 10000 × g (4 °C). After centrifugation, the supernatants containing extracellular DNA were filtered through 0.02-µm-pore-size filters to eliminate any contaminating viruses or prokaryotic cells. Aliquots of the supernatant fluid after filtration were fur- ther checked using epifluorescence microscopy to exclude viral or prokaryotic contamination. Materials and Methods d d li coli (from 0.2 to 200 pg µL−1). Standard concentrations were plotted against the number of cycles at which the fluorescence signal increased above background, or the cycle threshold (the Ct value). The iCycler software analysis programme was used to calculate the Ct values and to determine the sample concentrations based on the standard curves. The copy number of ribosomal gene sequences determined by qPCR was normalised to sediment dry weight for a comparison with available literature information42. Sequencing and bioinformatics. Genetic diversity of the prokaryotic 16S rDNA sequences associated with the extracellular DNA pools was analysed in the sediment of all areas. In addition, we analysed 16S rDNA diversity within the intracellular DNA pool in the two sites of NE Atlantic Ocean. Analyses were conducted by tag-encoded amplicon pyrosequencing of hypervariable regions (from V5 to V9). Bacterial and archaeal 16S rDNA amplicons were generated using the universal primers 789 F (5′ -TAGATACCCSSGTAGTCC-3′) and 1492 R (5′ –GGTTACCTTGTTACGACTT-3′)43 and sequenced on a Genome Sequencer 454 FLX Titanium plat- form (Roche). Three independent PCR analyses were carried out from each replicated extraction (n = 3) of extra- cellular and intracellular DNA. The amplicons obtained from PCR analysis of extracellular DNA were pooled together as well as those for intracellular DNA. Additional details on pyrosequencing analyses are reported in Supplementary Information.i Raw sequences were first subjected to homopolymer removal by means of the Acacia tool44 with standard values, and subsequently quality-trimmed by means of the PRINSEQ tool45 by removing sequences with a mean quality score <20 and further filtered to remove sequences shorter than 100 bp. The number of sequences obtained through each step is reported in Table S2. High-quality amplicon reads were subsequently analysed by the QIIME pipeline46, aligning them against the SILVA database (v119)47 by means of the PyNAST aligner using the open-reference strategy48 and clustering them at 97% of identity. The ChimeraSlayer tool was used to assess the presence of chimeric sequences, which were not found in any sample. All bioinformatic analyses were performed using the same number of sequences to properly compare the different samples. In particular, we used 3400 sequences for comparing extracellular DNA pools from different geographical areas (Mediterranean, Arctic and NE Atlantic margins) and 5000 sequences for comparing intra- cellular and extracellular DNA pools obtained contextually from the two sites of the NE Atlantic margin (the NE Atlantic 1 and 2 sites). ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 Materials and Methods d d li OTU networks were created on the data provided by the make_otu_network script provided by the QIIME pipeline with standard values on the extracellular DNA samples with the gephi tool49 by means of a combination of the Yifan-Hu and Force Atlas 2 algorithms provided by the program, with the Dissuade Hubs and Prevent Overlap flags toggled for ease of visualization. l Comparative analyses between assemblage composition of extracellular and intracellular DNA pools from the sediments of NE Atlantic 1 and 2 sites were carried out to evaluate the contribution of preserved OTUs to the total prokaryotic diversity. To do so we determined exclusive taxa contained within extracellular and intracellular DNA pools of the same site by counting OTUs exclusively found in either extracellular or intracellular DNA pools. ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 7 www.nature.com/scientificreports/ Statistical analyses. Analyses of variance (ANOVA) were carried out to test for differences in the extracel- lular DNA concentrations and prokaryotic 16S rDNA copy number among sampling sites. SIMPER analyses were carried out to assess the similarity among the prokaryotic OTUs contained within the different extracellular DNA pools and between extracellular DNA and their respective intracellular DNA pools50. All analyses were carried out with the PRIMER-E 6 suite50. In addition, to assess dissimilarity among extracellular DNA pools and between extracellular DNA and their respective intracellular pools (in terms of presence/absence of OTU), the binary Jaccard distance was determined by means of the beta_diversity script within QIIME and the results visualized with an UPGMA tree46. Data accessibility. The sequences have been submitted to the MG-RAST server under the project “Continental Margins” accessible for reviewers at the following token: http://metagenomics.anl.gov/mgmain. html?mgpage=token&token=67EuldY6_FfX8tRaYFO8Qxlo6LPnMq8OO2phSra8KwIvMUjy0m. Data accessibility. The sequences have been submitted to the MG-RAST server under the project “Continental Margins” accessible for reviewers at the following token: http://metagenomics.anl.gov/mgmain. html?mgpage=token&token=67EuldY6_FfX8tRaYFO8Qxlo6LPnMq8OO2phSra8KwIvMUjy0m. References 1. Dell’Anno, A. & Danovaro, R. Extracellular DNA plays a key role in deep-sea ecosystem functioning. Science. 309, 2179 (2005). 1. Dell’Anno, A. & Danovaro, R. Extracellular DNA plays a key role in deep-sea ecosystem functioning. Science. 309, 2179 (2005). 1. DellAnno, A. & Danovaro, R. Extracellular DNA plays a key role in deep-sea ecosystem functioning. Science. 309, 2179 (2005). 2. Corinaldesi, C., Dell’Anno, A. & Danovaro, R. Viral infection plays a key role in extracellular DNA dynamics in marine anoxic systems. Limnol Oceanogr. 52, 508–516 (2007). y g 3. Dell’Anno, A., Corinaldesi, C. & Danovaro, R. Virus decomposition provides an important contribution to benthic deep-sea ecosystem functioning. Proc Natl Acad Sci USA 112, 2014–2019 (2015).f 3. Dell’Anno, A., Corinaldesi, C. & Danovaro, R. Virus decomposition provides an important contribution to benthic deep-sea ecosystem functioning. Proc Natl Acad Sci USA 112, 2014–2019 (2015).f y g 4. Nielsen, K. M., Johnsen, P. J., Bensasson, D. & Daffonchio, D. Release and persistence of extracellular DNA in the environment. Environ Biosafety Res. 6, 37–53 (2007). f y 5. Coolen, M. J. L. & Overmann, J. Analysis of subfossil molecular remains of purple sulfur bacteria in a lake sediment. Appl Environ Microbiol. 64, 4513–4521 (1998). 6. Coolen, M. J. L. et al. Combined DNA and lipid analyses of sediments reveal changes in Holocene haptophyte and diatom populations in an Antarctic lake. Earth Planet Sci Lett. 223, 225–239 (2004). p p 7. Coolen, M. J. L. et al. Evolution of the plankton paleome in the Black Sea from the Deglacial to Anthropocene. Proc Natl Acad Sci USA 110, 8609–8614 (2013). 8. Corinaldesi, C., Tangherlini, M., Luna, G. M. & Dell’Anno, A. Extracellular DNA can preserve the genetic signatures of present and past viral infection events in deep hypersaline anoxic basins. Proc R Soc Lond B Biol Sci. 281, 20133299 (2014).l p p yp 9. Boere, A. C. et al. Exploring preserved fossil dinoflagellate and haptophyte DNA signatures to infer ecological and environm changes during deposition of sapropel S1 in the eastern Mediterranean. Paleoceanography. 26, PA2204 (2011). 0. Corinaldesi, C., Barucca, M., Luna, G. M. & Dell’Anno, A. Preservation, origin and genetic imprint of extracellular DNA in permanently anoxic deep-sea sediments. Mol Ecol. 20, 642–654 (2011). y 1. Coolen, M. J. L. et al. Ongoing modification of Mediterranean Pleistocene sapropels mediated by prokaryotes. Science. 296 2407–2410 (2002). 12. Coolen, M. J. L. & Overmann, J. References 217 000-year-old DNA sequences of green sulphur bacteria in Mediterranean sapropels and their implications for the reconstruction of the paleoenvironment. Environ Microbiol. 9, 238–249 (2007).iij 3. Coolen, M. J. L. et al. Identification of organic matter sources in sulfidic Late Holocene Antarctic fjord sediments from fossil rDNA sequence analysis. Paleoceanography. 22(2), PA2211 (2007). 4. Coolen, M. J. L. et al. DNA and lipid molecular stratigraphic records of haptophyte succession in the Black Sea during the Holocene Earth Planet Sci Lett. 284, 610–621 (2009). 15. Dell’Anno, A., Corinaldesi, C., Stavrakakis, V. L. & Danovaro, R. Pelagic-benthic coupling and diagenesis of nucleic acids in a deep- sea continental margin and an open-slope system of the Eastern Mediterranean. Appl Environ Microbiol. 71(10), 6070–6076 (2005). 16 Levin L A et al Comparative biogeochemistry ecosystem human interactions on dynamic continental margins J Marine Syst 15. Dell’Anno, A., Corinaldesi, C., Stavrakakis, V. L. & Danovaro, R. Pelagic-benthic coupling and diagenesis of nucleic acids in a deep- sea continental margin and an open-slope system of the Eastern Mediterranean. Appl Environ Microbiol. 71(10), 6070–6076 (2005). 16. Levin, L. A. et al. Comparative biogeochemistry–ecosystem–human interactions on dynamic continental margins. J Marine Syst. 141, 3–17 (2015). 16. Levin, L. A. et al. Comparative biogeochemistry–ecosystem–human interactions on dynamic continen 141, 3–17 (2015). ( ) 17. Jørgensen, B. B. & Boetius, A. Feast and famine—microbial life in the deep-sea bed. Nat Rev Microbiol. 5, 770–781 (2007). 18. Orcutt, B. N., Sylvan, J. B., Knab, N. J. & Edwards, K. J. Microbial ecology of the dark ocean above, at, and below the s Microbiol Mol Biol Rev. 75, 361–422 (2011). 19. Corinaldesi, C. New perspectives in benthic deep-sea microbial ecology. Front Mar Sci. 2, 1–12 (2015).l al. Global patterns of bacterial beta-diversity in seafloor and seawa . Zinger, L. et al. Global patterns of bacterial beta-diversity in seaflo 20. Zinger, L. et al. Global patterns of bacterial beta-diversity in seafloor and seawater ecosystems. PLoS One. 6, e24570 (2011). 21. Bienhold, C., Boetius, A. & Ramette, A. The energy-diversity relationship of complex bacterial communities in Arctic deep-sea sediments. ISME J. 6, 724–732 (2012). g p yl y 21. Bienhold, C., Boetius, A. & Ramette, A. The energy-diversity relationship of complex bacterial communities in A sediments. ISME J. 6, 724–732 (2012). 2. Zinger, L., Boetius, A. & Ramette, A. Bacterial taxa-area and distance-decay relationships in marine environments. Mol Ecol. 23 954–964 (2014). 23. Drummond, A. ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 www.nature.com/scientificreports/ Bastian, M., Heymann, S. & Jacomy, M. Gephi: an open source software for exploring and manipulating networks. ICW 361–362. (2009. 50 Cl k K R N i l i i l f h i i A t J E l 18 117 143 (1993) 50. Clarke, K. R. Non-parametric multivariate analyses of changes in community structure. Aust J Ecol. 18, 117–143 (1993) 50. Clarke, K. R. Non-parametric multivariate analyses of changes in community structure. Aust J Ecol. 18, 117–143 (1993). Acknowledgementsh c o edge e ts This study was co-funded by the National project EXPLODIVE (FIRB 2008, contract no. I31J10000060001) and the Flagship Project RITMARE – The Italian Research for the Sea – coordinated by the Italian National Research Council and funded by the Italian Ministry of Education, University and Research within the National Research Program 2011–2013. This study was conducted in the framework of the fieldwork activities of the European funded project HERMIONE (Hotspot Ecosystem Research and Man’s Impact On European Seas; contract n. 226354). We thank Flavia Tarquinio for her support in the laboratory analyses during the Master Thesis Degree. This study was co-funded by the National project EXPLODIVE (FIRB 2008, contract no. I31J10000060001) and the Flagship Project RITMARE – The Italian Research for the Sea – coordinated by the Italian National Research Council and funded by the Italian Ministry of Education, University and Research within the National Research Program 2011–2013. This study was conducted in the framework of the fieldwork activities of the European funded project HERMIONE (Hotspot Ecosystem Research and Man’s Impact On European Seas; contract n. 226354). We thank Flavia Tarquinio for her support in the laboratory analyses during the Master Thesis Degree. Author Contributions C.C. and A.D. conceive the study. M.T. and E.M. performed laboratory analyses. M.T. carried out bioinformatics analyses. C.C., A.D., M.T. and E.M. wrote the manuscript. www.nature.com/scientificreports/ www.nature.com/scientificreports/ 37. Pusceddu, A., Dell’Anno, A., Fabiano, M. & Danovaro, R. Quantity and bioavailability of sediment organic matter as signatures of benthic trophic status. Mar Ecol Prog Ser. 375, 41–52 (2009). ll’ d fi b d f f ll l d 8. Dell’Anno, A., Bompadre, S. & Danovaro, R. Quantification, base composition, and fate of extracellular DNA in marine sediments Limnol Oceanogr. 47, 899–905 (2002). g ( ) 39. Corinaldesi, C., Danovaro, R. & Dell’Anno, A. Simultaneous recovery of extracellular and intracellular DNA suitable for molecular studies from marine sediments. Appl Environ Microbiol. 71, 46–50 (2005).i pp ( ) 40. Takai, K. & Horikoshi, K. Rapid detection and quantification of members of the Archaeal community by quantitative PCR using fluorogenic probes. Appl Environ Microbiol. 66, 5066–5072 (2000). i fluorogenic probes. Appl Environ Microbiol. 66, 5066–5072 (2000). l g p pp 1. Wang, Y. & Qian, P. Y. Conservative fragments in bacterial 16S rRNA genes and primer design for 16S ribosomal DNA amplicons in metagenomic studies. PLoS One. 4, e7401 (2009). g 42. Sun, D.-L., Jiang, X., Wu, Q. L. & Zhou, N.-Y. Intragenomic heterogeneity of 16S r RNA genes causes overestimation of prokaryotic diversity. Appl Environ Microbiol. 79, 5962–5969 (2013).i y pp 43. Brown, M. V. et al. Microbial community structure in the North Pacific Ocean. ISME J. 3, 1374–1386 (2009). i 44. Bragg, L. et al. Fast, accurate error-correction of amplicon pyrosequences using Acacia. Nat Methods. 9, 425–426 (2 45. Schmieder, R. & Edwards, R. Quality control and preprocessing of metagenomic datasets. Bioinformatics. 27(6), 863 864 (201 46. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 7(5), 335–336 (20h y g g f 46. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 7(5), 335–336 (2010). 47. Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41, 590–596 (2013).i 46. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 7(5), 335– y g g y g 7. Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41, 590–596 (2013).i 48. Rideout, J. R. et al. Subsampled open-reference clustering creates consistent, comprehensive OTU definitions and scales to billions of sequences. PeerJ. 2, e545 (2014). q 49. References J. et al. Evaluating a multigene environmental DNA approach for biodiversity assessment. GigaScience. 4(1), 46 (2015). 4. Thomsen, P. F. & Willerslev, E. Environmental DNA—An emerging tool in conservation for monitoring past and presen biodiversity. Biol Cons. 183, 4–18 (2015). y 5. Corinaldesi, C., Beolchini, F. & Dell’Anno, A. Damage and degradation rates of extracellular DNA in marine sediments: implications for the preservation of gene sequences. Mol Ecol. 17, 3939–3951 (2008).ii p g q 26. Jorgensen, S. L. et al. Correlating microbial community profiles with geochemical data in highly stratified sediments from the Arctic Mid-Ocean Ridge. Proc Natl Acad Sci USA 109(42), 2846–2855 (2012).l 26. Jorgensen, S. L. et al. Correlating microbial community profiles with geochemical Mid-Ocean Ridge. Proc Natl Acad Sci USA 109(42), 2846–2855 (2012). 27. Bienhold, C., Zinger, L., Boetius, A. & Ramette, A. Diversity and Biogeography of Bathyal and Abyssal Seafloor Bacteria. PLoS One. 11, e0148016 (2016). 8. Mußmann, M., Pjevac, P., Krüger, K. & Dyksma, S. Genomic repertoire of the Woeseiaceae/JTB255, cosmopolitan and abundan core members of microbial communities in marine sediments. ISME J 11, 1276–1281 (2017).i 29. Rivière, D. et al. Towards the definition of a core of microorganisms involved in anaerobic digestion of sludge. ISME J. 3, 700–714 (2009). 30. Huse, S. M., Ye, Y., Zhou, Y. & Fodor, A. A. A core human microbiome as viewed through 16S rRNA sequence cluster. PLoS One 7(6), e34242 (2012). 31. Martiny, J. B. H. et al. Microbial biogeography: putting microorganisms on the map. Nat Rev Microbiol. 4(2), 102–112 (2006).l 32. Walsh, E. A. et al. Bacterial diversity and community composition from seasurface to subseafloor. ISME J. 10, 979–989 (2015) l 33. Corinaldesi, C., Dell’Anno, A., Magagnini, M. & Danovaro, R. Viral decay and viral production rates in continental-shelf and deep- sea sediments of the Mediterranean Sea. FEMS Microbiol Ecol. 72, 208–218 (2010).l 34. Walsh, E. A. et al. Relationship of bacterial richness to organic degradation rate and sediment age in subseafloor sediment. Appl Environ Microbiol. 82(16), 4994–4999 (2016). 5. Schauer, R. et al. Bacterial diversity and biogeography in deep-sea surface sediments of the South Atlantic Ocean. ISME J. 4, 159–170 (2010). ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 8 Additional Information Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-018-20302-7. Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2018 ScIentIfIc REporTs \| (2018) 8:1839 \| DOI:10.1038/s41598-018-20302-7 9
https://openalex.org/W4323073880	https://digital.csic.es/bitstream/10261/334797/1/maternalpathways.pdf	English	null	Maternal obesity and gestational diabetes reprogram the methylome of offspring beyond birth by inducing epigenetic signatures in metabolic and developmental pathways	Cardiovascular diabetology	2,023	cc-by	13,913	Cardiovascular Diabetology Cardiovascular Diabetology Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 https://doi.org/10.1186/s12933-023-01774-y Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 https://doi.org/10.1186/s12933-023-01774-y Open Access © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Maternal obesity and gestational diabetes reprogram the methylome of offspring beyond birth by inducing epigenetic signatures in metabolic and developmental pathways Juan José Alba‑Linares1,2,3,4,5†, Raúl F. Pérez1,2,3,4,5†, Juan Ramón Tejedor1,2,3,4,5, David Bastante‑Rodríguez1,2,3,4,5, Francisco Ponce6,7, Nuria García Carbonell6,8, Rafael Gómez Zafra6,8, Agustín F. Fernández1,2,3,4,5, Mario F. Fraga1,2,3,4,5* and Empar Lurbe6,7,8* Abstract Background Obesity is a negative chronic metabolic health condition that represents an additional risk for the devel‑ opment of multiple pathologies. Epidemiological studies have shown how maternal obesity or gestational diabetes mellitus during pregnancy constitute serious risk factors in relation to the appearance of cardiometabolic diseases in the offspring. Furthermore, epigenetic remodelling may help explain the molecular mechanisms that underlie these epidemiological findings. Thus, in this study we explored the DNA methylation landscape of children born to mothers with obesity and gestational diabetes during their first year of life. Methods We used Illumina Infinium MethylationEPIC BeadChip arrays to profile more than 770,000 genome-wide CpG sites in blood samples from a paediatric longitudinal cohort consisting of 26 children born to mothers who suffered from obesity or obesity with gestational diabetes mellitus during pregnancy and 13 healthy controls (meas‑ urements taken at 0, 6 and 12 month; total N = 90). We carried out cross-sectional and longitudinal analyses to derive DNA methylation alterations associated with developmental and pathology-related epigenomics. Results We identified abundant DNA methylation changes during child development from birth to 6 months and, to a lesser extent, up to 12 months of age. Using cross-sectional analyses, we discovered DNA methylation biomarkers maintained across the first year of life that could discriminate children born to mothers who suffered from obesity or obesity with gestational diabetes. Importantly, enrichment analyses suggested that these alterations constitute epi‑ genetic signatures that affect genes and pathways involved in the metabolism of fatty acids, postnatal developmental processes and mitochondrial bioenergetics, such as CPT1B, SLC38A4, SLC35F3 and FN3K. Finally, we observed evidence of an interaction between developmental DNA methylation changes and maternal metabolic condition alterations. †Juan José Alba-Linares and Raúl F. Pérez are co-first authors and contributed equally to this work *Correspondence: Mario F. Fraga mffraga@cinn.es Empar Lurbe empar.lurbe@uv.es Full list of author information is available at the end of the article Correspondence: Mario F. Fraga mffraga@cinn.es Empar Lurbe empar.lurbe@uv.es Full list of author information is available at the end of the article Background underlie these strong epidemiological findings. Further- more, it remains unclear whether childhood obesity is simply the result of the unhealthy eating behaviour instilled by parents during postnatal growth or whether the intrauterine environment may be capable of affect- ing children of obese mothers, predisposing them to the development of cardiometabolic diseases. In line with the theory of the developmental origins of health and dis- ease, the intrauterine period is crucial to understand the adult risk to experience cardiovascular events [18], but at the same time, early postnatal development has gained relevance in recent decades. Longitudinal studies which trace a life course perspective are needed to address these questions, exploring the interrelations between the influ- ence of the intrauterine environment and developmental processes [19].l Obesity is a negative health condition characterized by an excessive accumulation of fat that is diagnosed at a body mass index (BMI) equal to or greater than 30 kg/m2 [1]. In recent decades, obesity has become a central pub- lic health concern due to its increasing prevalence, which has reached worldwide pandemic levels, and because it represents a leading risk factor for the development of cardiovascular, metabolic diseases (e.g., hypertension, myocardial infarction, stroke, diabetes mellitus) and cer- tain types of cancer, amongst other illnesses [1, 2]. Interestingly, epidemiological studies have shown how maternal obesity during pregnancy constitutes an important risk factor related to the appearance of cardiometabolic diseases in the offspring [3], includ- ing obesity, elevated blood pressure, impaired insulin/ glucose homoeostasis, increased inflammatory markers and altered lipid profiles [4–8]. This is a serious concern since prevalence rates of obesity in pregnant women can exceed 30% [9]. Furthermore, a recent longitudinal girl study with a follow-up from birth to 10 years has found that birth weight and maternal obesity are the main risk factors responsible for the appearance of obesity at 5 years, while at 10 years the only significant related con- dition is maternal obesity [10].h To date, the mechanisms governing the influence of the intrauterine environment on the offspring’s health are still under discussion. For instance, extracellular vesi- cles are posited to play roles in the systemic regulation of physiological processes and pathologies [20]. In this line, we propose that epigenetic remodelling, which is especially sensitive to extrinsic and intrinsic influences during early life [21, 22], could constitute the molecular mechanism through which intrauterine stimuli can affect the biology of the cell. © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 2 of 18 Conclusions Our observations highlight the first six months of development as being the most crucial for epige‑ netic remodelling. Furthermore, our results support the existence of systemic intrauterine foetal programming linked to obesity and gestational diabetes that affects the childhood methylome beyond birth, which involves alterations related to metabolic pathways, and which may interact with ordinary postnatal development programmes. Keywords Metabolism, DNA methylation, Epigenetics, Obesity, Gestational diabetes, Newborn, Development, Longitudinal Background In fact, it is well known that DNA methylation patterns are key regulators of genes involved in pancreatic β-cell homeostasis, including insulin sig- nalling and secretion [23]. At the same time, metabolic imbalance can disrupt epigenetic mechanisms through alterations in the levels of tricarboxylic acid cycle inter- mediates and the redox balance, which subsequently can have an impact on gene regulation and DNA dam- age and repair [24]. Ten-eleven translocation methylcy- tosine dioxygenases, responsible for DNA demethylation processes, are sensitive to these metabolic dysfunctions, such as obesity and diabetes mellitus, establishing a crosstalk between metabolism, epigenetics and genome stability [24]. In agreement with this, gestational diabe- tes [25, 26], maternal obesity [27] and hypertension [19] have been consistently associated with DNA methylation alterations in placenta, blood and tissue samples from offspring. To shed light on this issue, longitudinal studies are of great value for several reasons, among them: (a) in The increased risk for these conditions is maintained not only in childhood, but also in terms of adulthood morbidity and mortality [11–13], which evinces pro- found implications for the design of public health policies and interventions, especially in the case of cardiovascular diseases. Moreover, the most common medical complica- tion of pregnancy, gestational diabetes mellitus (GDM, hyperglycaemia that develops during pregnancy and resolves after birth), represents another risk factor for the development of obesity and cardiovascular disease in both the mother and child [14]. A recent cohort study has evinced that previous GDM leads to a higher incidence of dyslipidemia in women [15]. What is more, maternal obesity is also correlated with metabolic complications in pregnancy such as GDM, gestational hypertension and pre-eclampsia, highlighting the fact that obesity and GDM are very intertwined [16]. However, due to obesity’s multifactorial nature [17], little is known about the molecular mechanisms that Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 3 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 The exclusion criteria employed were: multiple gesta- tions, overweight (rather than obese) women (BMI 25.0– 29.9 kg/m2), underweight women (BMI < 18.5 kg/m2) and any complication during gestation apart from GDM. The general characteristics of gestation and delivery were obtained from routine obstetric records. The subjects were followed-up at the General Hospital of Valencia Outpatient Clinic for the first year of the child’s life. Anthropometric parameters At birth and during the follow-up, weight and length were measured by trained nurses. Length was meas- ured in the supine position using a paediatric measuring device. Weight was measured in the Maternity Unit using an ADE scale model M112600 (GmbH & Co.) and in the Outpatient Clinic using a Seca 354 scale (GmbH & Co.). Body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters for mothers and using WHO AnthroPlus software for children. Background the field of developmental biology, they allow the defini- tion of the key time frame when epigenetic alterations are most dynamic, and therefore, when external stimuli can have the greatest influence at the molecular level; (b) they can define epigenetic patterns that are acquired through uterine exposure but are maintained over time, acting as reliable childhood biomarkers.h This study presents, to the best of our knowledge, the first longitudinal genome-wide analysis of the methylome of whole blood samples from a paediatric cohort of chil- dren born to mothers suffering from obesity or obesity with GDM during pregnancy and healthy controls. We performed longitudinal measurements throughout the first year of life (0, 6 and 12 months) on 39 subjects (total N = 90) using Illumina Infinium MethylationEPIC Bead- Chip arrays to profile more than 770,000 CpG sites. The design allowed us to carry out both cross-sectional and longitudinal analyses in order to derive DNA methylation alterations associated with developmental and pathol- ogy-related epigenomics. Selection of study subjectsh Blood samples were taken from 39 subjects at three time points (birth, 6 and 12 months of age; total N = 90). At birth, samples were collected from the umbilical cord, while at 6 and 12 months they were taken from periph- eral venous blood. This cohort is part of a prospective and ongoing study begun in 2018. The present research includes data from April 2018 to February 2020. Newborns born at term (gestational age ≥ 37 weeks) at the General Hospital of Valencia were randomly recruited to participate in the study. At or before birth, all parents gave informed con- sent for their children to participate in the study, which was approved by the Clinical Research Ethics Commit- tee of the Consorcio Hospital General Universitario de Valencia. Genomic DNA was extracted from whole blood cells with the RealPure kit (RealPure, REAL, Durviz) and quantified with the Nanodrop-2000C Spectrophotom- eter. Next, the DNA was bisulphite converted using the EZ-96 DNA Methylation Kit conversion protocol (Zymo Research). Finally, the Illumina Infinium HD Methylation Assay protocol was performed by hybridising processed DNA samples to Infinium MethylationEPIC BeadChips. Three groups were established on the basis of moth- ers’ BMI and the presence or absence of GDM: children of obese mothers with GDM, children of obese mothers without diabetes and children of control mothers (nor- mal weight BMI 18.5–24.9 kg/m2, without pathology). Obesity in pregnant women was defined as BMI ≥ 30 kg/ m2 at the beginning of the pregnancy. The screening of GDM consisted of a 50 g oral glucose load (glucose chal- lenge test or GCT) followed by a plasma sugar level test 1 h later when the women’s pregnancies were at between 24 and 28 weeks of gestation. A level of more than 7.8 mmol/L (140 mg/dL) indicated the need for full diag- nostic testing with an oral glucose tolerance test (OGTT, 100 g oral glucose load, testing during 3 h). The diagnosis of GDM required any two of the four plasma glucose val- ues to be equal to or above the following values: (a) after overnight fast: 105 mg/dL (5.8 mmol/L); (b) 1 h: 190 mg/ dL (10.6 mmol/L); (c) 2 h: 165 mg/dL (9.6 mmol/L); (d) 3 h: 145 mg/dL (8.1 mmol/L). Pathway enrichment analyses Pathway enrichment analyses were performed using the missMethyl package (v.1.22.0) [43] on the gene sets from the Molecular Signatures Database (MSigDB) [44] accessed via the msigdbr package (v.7.2.1). The gsameth function was used to interrogate the functionality of the DMPs identified, while the gsaregion function was used to analyse DMRs. Both take into account the number of probes mapping to each gene as a bias factor for the enrichment analyses. To visualise pathway enrichment results, several networks of gene-set similarity were built using the EnrichmentMap application [45] in Cytoscape (v.3.9.1) [46] using the RCy3 package (v.2.8.1) [47] with the default combined similarity cutoff. To discover distinct methylation clusters, significant DMPs (0 > 6 and 6 > 12 longitudinal comparisons) were clustered by using Spearman correlation distances to group their scaled methylation values. The optimal num- ber of clusters was determined using the within-cluster sum of squared error method. Probe‑level differential methylation analyses For the annotation of regions, the probes belonging to each region were first individually annotated as described above. A single annotation was then assigned to each region according to the following criteria: (1) for CGI status, “Island” > “N_Shore” > “S_Shore” > “N_Shelf” > “S_ Shelf” > “OpenSea”; and (2) for gene locations, “TSS1500 ” > “TSS200” > “5’UTR” > “1stExon” > “Body” > “ExonBnd” > “3’UTR” > “Intergenic”. f y y First, β-values were logit-transformed to M-values with the beta2m function of the lumi package (v.2.40.0) in order to achieve greater homoscedasticity in the differen- tial methylation analyses [40]. Then, linear mixed mod- els were built using the limma package (v.3.44.3) [41] to detect differentially methylated probes (DMPs). Several statistical models were designed by fitting M-values as the dependent variable. All models included fixed covari- ates that accounted for possible experimental batch effects (array plate), sex and cell-type composition from deconvolution analyses, while subject-specific contribu- tions were controlled via random-effects components. Cross-sectional comparisons were performed using Group (Control, Obesity, Obesity + Diab) as the inde- pendent variable, while longitudinal comparisons were carried out using Time (t0, t6, t12) as the independent variable. DMPs were defined by contrasting coefficients using an empirical Bayes-moderated t-test, such that the set of p-values was adjusted for multiple comparisons using the Benjamini–Hochberg method (FDR < 0.05).i Genome‑wide profiling of DNA methylation alterations during the first year of life To characterize DNA methylation alterations that occur during the first year of development, we studied the methylome of whole blood samples from a paediat- ric cohort of 39 subjects consisting of children born to Region‑level differential methylation analyses The “comb-p” method [42] was used to find differentially methylated regions (DMRs) at FDR < 0.05 via the Enmix Array data preprocessing All MethylationEPIC BeadChip data analyses were per- formed using the statistical software R (v.4.0.2). First, IDAT files were imported and processed with the minfi package (v.1.32.0) [28]. Self-reported sex and subject genetic tracking were validated by accessing the array methylation data for sex chromosome probes and SNP probes, using the getSex and getSnpBeta functions from minfi. In addition, probes from sesame package (v.1.4.0) [29] were used to carry out ethnic inference analysis and ensure correct sample tracking. All samples passed the specific quality control of the minfi package for intensity signals both in methylated and unmethylated channels.i Probes were filtered out if: (a) detection p-value was > 0.01 in any sample; (b) they were located in sex chromosomes; (c) they were cross-reactive or Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 4 of 18 multi-mapping [30, 31] and (d) they included SNPs with MAF ≥ 0.01 at their CpG or SBE sites (dbSNP v.147). Moreover, clustered-distribution analysis using the gap- hunter function (threshold = 0.25, outCutoff = 5/90) of the minfi package allowed the detection of experiment- specific conflicting probes (N = 1,065), which were dis- carded for downstream analysis [32]. After this, intensity values were subjected to background correction using the ssNoob method [33] in minfi and extracted β-values were normalised using the BMIQ approach [34] implemented in ChAMP (v.2.16.2) [35]. The final number of probes that passed through all the filters was 772,088. package (v.1.28.2) [37] using default parameters. In order to discover spatially-associated regions of significance, the limma p-values from the DMP analyses were fed into the combp function. These initial regions were first selected under an FDR threshold, and then final signifi- cant DMRs were defined as those with a Sidak-corrected p < 0.05 and containing at least 3 CpG sites. In addition, “mixed” DMRs displaying less than 66% of CpG sites with changes in the same direction were filtered out along with those DMRs whose changes were lower than or equal to 1% of mean methylation value. Cell‑type deconvolution The IlluminaHumanMethylationEPICanno.ilm10b4.hg19 package (v.0.6.0) was used to assign each probe to its CGI (CpG Island) and gene location status. Fisher’s exact tests were used to compare statistically differential propor- tions of annotations and intersections, and odds ratios (ORs) were employed as a measure of the association effect with respect to a particular feature. An appropriate background which included the filtered probes analysed by the EPIC array was used for statistical purposes. Cell-type composition was predicted from DNA methyl- ation data by the Houseman algorithm [36] implemented in the ENmix package (v.1.28.2) [37]. Appropriate and specific reference datasets were used for the cell- type prediction of cord-blood FlowSorted.CordBlood- Combined.450 k [38] and peripheral-blood samples FlowSorted.Blood.EPIC [39]. Region‑level differential methylation analyseshi The “comb-p” method [42] was used to find differentially methylated regions (DMRs) at FDR < 0.05 via the Enmix Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 5 of 18 mothers suffering from obesity or obesity with GDM during pregnancy, and healthy controls (Table 1; Fig. 1A; Additional file 2: Table S1). We performed longitudinal measurements throughout the first year of their infants’ lives (0, cord blood; 6 and 12 months, peripheral blood; total N = 90) using MethylationEPIC arrays to profile 772,088 CpG sites in differential methylation analyses (see “Methods”), in which we considered cell-type com- position biases as we found variation in proportions, especially significant during the first six months of life (Additional file 1: Fig. S1). (Fisher’s test P < 0.001, OR = 25). Moreover, those com- mon changes massively preserved their direction of change at both longitudinal comparisons (hyper: Fish- er’s test P < 0.001, OR = 89; hypo: Fisher’s test P < 0.001, OR = 44) (Fig. 1E), although they were noticeably stronger in the 0 > 6 period (Additional file 1: Fig. S2). i When we studied the genomic distribution of the DMPs identified (Fig. 1F, G), we determined that hypermeth- ylated CpGs tended to be located at CpG island-asso- ciated loci (0 > 6 Fisher’s test P < 0.001, OR = 1.18; 6 > 12 Fisher’s test P < 0.001, OR = 1.93) and promoters (TSS 0 > 6 Fisher’s test P < 0.01, OR = 1.07; TSS 6 > 12 Fisher’s test P < 0.001, OR = 1.20). In fact, this result is particu- larly important since the array background is already enriched in CpG sites that are mainly located at island and promoter regions. Inversely, hypomethylated CpGs tended to be enriched at open sea regions (0 > 6 Fisher’s test P < 0.001, OR = 1.94; 6 > 12 Fisher’s test P < 0.001, OR = 1.81), gene bodies (0 > 6 Fisher’s test P < 0.001, OR = 1.46; 6 > 12 Fisher’s test P < 0.001, OR = 1.18) and 5’UTR regions (0 > 6 Fisher’s test P < 0.001, OR = 1.37; 6 > 12 Fisher’s test P < 0.001, OR = 1.20), revealing that the direction of the methylation change is intrinsically dependent on the genomic context of its target. Region‑level differential methylation analyseshi i First, we sought to describe global aspects of DNA methylation alterations through an exploratory princi- pal component analysis (PCA) (Fig. 1B) and were able to confirm that developmental processes are those with the deepest impact on the paediatric methylome. Nonethe- less, we did not observe noticeable differences in global methylation values during the first year of child develop- ment (Fig. 1C), so we focused on identifying differentially methylated CpG sites (DMPs). To this end, we employed empirical Bayes moderated t tests across the different age groups (FDR < 0.05; see “Methods”) and discovered abundant DNA methylation changes during develop- ment from birth to 6 months (27,475 DMPs) and, to a lesser extent, from 6 to 12 months of age (12,742 DMPs) (Fig. 1D; Additional file 3: Table S2). The direction of DNA methylation changes was well balanced from birth to 6 months in that 14,953 DMPs were found to be hypo- methylated (~ 54%) while 12,522 were hypermethylated (~ 46%). Conversely, this balance disappeared in the fol- lowing 6 months, and at 12 months 10,606 DMPs (~ 83%) were found to be hypomethylated whereas only 2136 DMPs were hypermethylated (~ 17%). Functional dynamics of DNA methylation alterations during the first year of life In order to study the functional context in which DNA methylation changes accumulate during development, we clustered the aforementioned DMPs on the basis of their methylation patterns at 0, 6 and 12 months (see “Meth- ods”). We found six distinct clusters (Fig. 2A) which differed in size and numbers of annotated genes, the number of DMPs and associated genes being correlated (Fig. 2B). The two main clusters (1 and 2) reflected grad- ual longitudinal hypo- and hypermethylation alterations over the first year of life, while clusters 4 and 5 under- went, respectively, marked hyper- and hypomethylation Next, we proceeded to analyse the degree of overlap between those CpG sites that were differentially meth- ylated between 0 and 6 months and between 6 and 12 months (Fig. 1E), finding that a large proportion of DMPs at 0 > 6 months were also altered at 6 > 12 months Table 1 Summary of clinical information Based on maternal BMI/GDM, subjects are classified into controls (C) (maternal BMI < 30), obese (Ob) (maternal BMI > 30) and obese (maternal BMI > 30) with gestational diabetes (ObDia). Children were followed-up at birth (t0), six months (t6) and one year of life (t12) Group Control (C) Obese (Ob) Obese + Diabetes (ObDia) Time point t0 t6 t12 t0 t6 t12 t0 t6 t12 Number of subjects 13 13 4 15 15 4 11 11 4 Sex (M/F) 8/5 8/5 2/2 11/4 11/4 3/1 6/5 6/5 2/2 Maternal pregnancy BMI, mean (SD) 22.58 (1.42) 22.58 (1.42) 22.62 (1.28) 35.46 (5.25) 35.46 (5.25) 37.57 (7.61) 32.87 (2.33) 32.87 (2.33) 32.92 (0.66) Child weight/g, mean (SD) 3202 (404) 7585 (1032) 9078 (248) 3435 (508) 7867 (1098) 10,465 (675) 3404 (409) 8106 (1368) 10,640 (1087) Child height/cm, mean (SD) 49.23 (2.07) 65.31 (3.22) 72.25 (2.87) 49.17 (1.90) 65.47 (1.95) 75.00 (1.29) 49.14 (1.60) 66.09 (5.25) 75.88 (2.66) Table 1 Summary of clinical information Based on maternal BMI/GDM, subjects are classified into controls (C) (maternal BMI < 30), obese (Ob) (maternal BMI > 30) and obese (maternal BMI > 30) with gestational diabetes (ObDia). Children were followed-up at birth (t0), six months (t6) and one year of life (t12) Based on maternal BMI/GDM, subjects are classified into controls (C) (maternal BMI < 30), obese (Ob) (maternal BMI > 30) and obese (maternal BMI > 30) with gestational diabetes (ObDia). Functional dynamics of DNA methylation alterations during the first year of life Children were followed-up at birth (t0), six months (t6) and one year of life (t12) Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 6 of 18 Fig. 1 Development is the main origin of methylation changes during the first year of life. a Schematic of the study design. b Scatter plot showing the PCA of the subjects according to their methylation values at the top 100,000 most variable CpG sites. c Violin plots describing the distribution of methylation values at the 772,088 CpGs analysed by time point. d Barplots depicting the number of hyper- and hypomethylated DMPs (FDR < 0.05) found in 0 > 6 and 6 > 12 longitudinal comparisons. e At the top, a Venn diagram showing the intersections between 0 > 6 and 6 > 12 DMPs. At the bottom, a Venn diagram describing the intersection between hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons. f Barplots showing the relative distribution of hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons according to their CpG island location status. g Barplots showing the relative distribution of hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons according to their gene location status. Rightmost bars represent the background distribution considering all 772,088 probes analysed Fig. 2 Hyper- and hypomethylation changes affect different functional targets during development. a Line plots depicting the scaled methylation patterns at 0, 6 and 12 months of the six clusters identified. b At the top, barplots showing the number of DMPs for each cluster. At the bottom, i (See figure on next page.) Fig. 1 Development is the main origin of methylation changes during the first year of life. a Schematic of the study design. b Scatter plot showing the PCA of the subjects according to their methylation values at the top 100,000 most variable CpG sites. c Violin plots describing the distribution of methylation values at the 772,088 CpGs analysed by time point. d Barplots depicting the number of hyper- and hypomethylated DMPs (FDR < 0.05) found in 0 > 6 and 6 > 12 longitudinal comparisons. e At the top, a Venn diagram showing the intersections between 0 > 6 and 6 > 12 DMPs. At the bottom, a Venn diagram describing the intersection between hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons. Functional dynamics of DNA methylation alterations during the first year of life First, we again performed differential methylation analyses to uncover DMPs (FDR < 0.05) that were able to distinguish our experimental groups throughout the first year of life. We detected moderate DNA methylation changes across all comparisons (Ob.C, ObDia.C, ObDia. Ob) (Fig. 3A; Additional file 4: Table S3). When we stud- ied the degree of overlap between the three biological comparisons (Fig. 3B), we observed large enrichments in shared DMPs (Ob.C U ObDia.C Fisher’s test P < 0.001, OR = 782; ObDia.C U ObDia.Ob Fisher’s test P < 0.001, OR = 2130; Ob.C U ObDia.Ob Fisher’s test P < 0.001, OR = 220), indicating that these metabolic processes target common loci. Overlaps between the comparisons were also observed when separating hyper- and hypo- DMPs (Additional file 1: Fig. S6A). i Next, to increase our power to detect DNA methylation alterations with functional implications for the molecu- lar physiopathology of obesity, we performed differential methylation analyses at the regional level to discover dif- ferentially methylated regions (DMRs; FDR < 0.05; see “Methods”). With this method, we detected even more numerous alterations across all comparisons (Fig. 3C; Additional file 5: Table S4). To analyse the regional over- lap between the three biological comparisons (Fig. 3D), we studied the shared CpGs belonging to the DMRs, find- ing large enrichments for all of them (Ob.C U ObDia.C Fisher’s test P < 0.001, OR = 234; ObDia.C U ObDia. Ob Fisher’s test P < 0.001, OR = 284; Ob.C U ObDia.Ob Fisher’s test P < 0.001, OR = 207), thus indicating a shared epigenetic footprint associated with these metabolic pro- cesses. Interestingly, when we separated the regions into hyper- or hypo-DMRs, we observed that the intersection between ObDia.Ob and Ob.C DMRs disappeared, sug- gesting that GDM also produces directionally-specific DNA methylation alterations which may not be associ- ated with obesity (Additional file 1: Fig. S6B). In contrast, hypomethylated CpG-linked genes from clusters 1 and 5 were mainly involved in immune system activation and maturation pathways (Fig. 2D). In fact, we found a high level of overlap between the significant gene sets belonging to clusters 1 and 5, but nothing between these and cluster 2 (Additional file 1: Fig. S4; CGP C1–C5 Fisher’s test P < 0.001, OR = 42; GO BP C1–C5 Fisher’s test P < 0.001, OR = 296; REACTOME C1–C5 Fisher’s test P < 0.001, OR = 195). Functional dynamics of DNA methylation alterations during the first year of life f Barplots showing the relative distribution of hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons according to their CpG island location status. g Barplots showing the relative distribution of hyper- and hypomethylated DMPs at 0 > 6 and 6 > 12 comparisons according to their gene location status. Rightmost bars represent the background distribution considering all 772,088 probes analysed Fig. 2 Hyper- and hypomethylation changes affect different functional targets during development. a Line plots depicting the scaled methylation patterns at 0, 6 and 12 months of the six clusters identified. b At the top, barplots showing the number of DMPs for each cluster. At the bottom, barplots of the number of associated genes by cluster. c Barplots showing the proportion of significant gene sets (P < 0.05, grey; FDR < 0.05, red) found in the enrichment analyses for each cluster and type of MSigDb database. d Bubble plots showing the top 5 most significant GO BP, CGP and REACTOME gene sets (FDR < 0.05) for clusters 1, 2 and 5 (See figure on next page.) Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 7 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 8 of 18 changes from 0 to 6 months. Finally, clusters 3 and 6 grouped directionally opposite changes for both longitu- dinal comparisons. paediatric cohort: (a) children born to mothers with obe- sity during pregnancy (Ob); (b) children born to moth- ers with obesity and GDM during pregnancy (ObDia); (c) children born to healthy mothers (C).f Next, we performed pathway enrichment analyses against several databases from the MSigDB collection to ascertain which of these clusters were functionally relevant. We concluded that clusters 1, 2 and 5 were the most informative according to the numbers of statisti- cally significant gene sets found (P < 0.05; FDR < 0.05) (Fig. 2C). The main roles of the genes associated with the hypermethylated CpGs from cluster 2 were related to developmental processes, being targets of repression by Polycomb (PRC2, SUZ12) and part of the pluripotency regulation mechanisms that control cell differentiation (Fig. 2D). We then explored the genes related to these pathways that showed a very high number of associated DMPs (Additional File 1: Figure S3), and found several examples with relevant developmental roles, such as transcription factors NFIX and TBX1, growth factors WNT10A and NRG2, hormone CALCA and cadherin CDH23, amongst others. Functional dynamics of DNA methylation alterations during the first year of life When we analysed the genes that showed a high number of associated hypo-DMPs, we distinguished many targets that regulate hematopoi- etic differentiation, especially the development, selec- tion and maturation of T and NK cells (Additional File 1: Figure S5), including transcription factors (FOXP1, ETS1, HIVEP2 and HIVEP3), cell adhesion proteins (CYTH1, ITGAL, ITGB2, HLA-E), apoptotic genes (BCL- 2, TNF family) and epigenetic modifiers (PRDM histone methyltransferases). Therefore, during early postnatal development, the functionality of the affected genes is clearly defined by the direction of the DNA methylation changes. i By exploring the genomic distribution of the DMRs, we concluded that both hyper- and hypo-DMRs were clearly located at CpG islands (Fig. 3E) (Ob.C hyper Fisher’s test P = 0.073, OR = 1.51; Ob.C hypo Fisher’s test P < 0.001, OR = 2.49; ObDia.C hyper Fisher’s test P < 0.001, OR = 2.20; ObDia.C hypo Fisher’s test P < 0.001, OR = 3.38; ObDia.Ob hyper Fisher’s test P < 0.001, OR = 3.08; ObDia.Ob hypo Fisher’s test P < 0.001, OR = 3.30) and promoter regions (Fig. 3F) (Ob.C hyper Fisher’s test P < 0.001, OR = 2.44; Ob.C hypo Fisher’s test P < 0.01, OR = 2.01; ObDia.C hyper Fisher’s test P < 0.01, OR = 1.60; ObDia.C hypo Fisher’s test P < 0.01, OR = 2.48; ObDia.Ob hyper Fisher’s test P = 0.067, OR = 1.40; Maternal obesity and gestational diabetes during pregnancy alter the methylomic landscape of the offspring Once we had described the DNA methylation alterations during the first year of development, we sought to answer several questions: (1) To what extent could maternal influence during pregnancy affect the methylome of the offspring? (2) Could those putative methylome changes be preserved at least during the first year of life? To address these questions, we defined three groups in our Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 9 of 18 ObDia.Ob hypo Fisher’s test P < 0.001, OR = 3.40). As anticipated the regional level analysis of DNA meth concentrated at regulatory sequences and thus could have potential functional implications Fig. 3 Obesity-mediated maternal influence during pregnancy alters the methylome of offspring during the first year of life. a Barplots depicting the number of hyper- and hypomethylated DMPs (FDR < 0.05) found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. b Venn diagram showing the intersection between hyper- and hypomethylated DMPs found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. c Barplots depicting the number of hyper- and hypomethylated DMRs (FDR < 0.05) found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. d Venn Diagram showing the intersection between hyper- and hypomethylated CpGs from DMRs found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. e Barplots depicting the relative distribution of hyper- and hypomethylated DMRs from the Ob.C, ObDia.C and ObDia.Ob comparisons according to their CpG island locations status. F Barplots showing the relative distribution of hyper- and hypomethylated DMRs from the Ob.C, ObDia.C and ObDia.Ob comparisons according to their gene location status. Rightmost bars represent the background distribution considering all 772,088 probes analysed. g On the left, scatter plot showing the PCA of the subjects according to their methylation values at Ob.C, ObDia.C, ObDia.Ob DMPs. On the right, scatter plot representing the PCA of the subjects according to their methylation values at CpGs from Ob.C, ObDia.C, ObDia.Ob DMRs ediated maternal influence during pregnancy alters the methylome of offspring during the first year of life. a Barplots depictin Fig. 3 Obesity-mediated maternal influence during pregnancy alters the methylome of offspring during the first year of life. a Barplots depicting the number of hyper- and hypomethylated DMPs (FDR < 0.05) found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. b Venn diagram showing the intersection between hyper- and hypomethylated DMPs found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. Maternal obesity and gestational diabetes during pregnancy alter the methylomic landscape of the offspring c Barplots depicting the number of hyper- and hypomethylated DMRs (FDR < 0.05) found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. d Venn Diagram showing the intersection between hyper- and hypomethylated CpGs from DMRs found in the Ob.C, ObDia.C and ObDia.Ob cross-sectional comparisons. e Barplots depicting the relative distribution of hyper- and hypomethylated DMRs from the Ob.C, ObDia.C and ObDia.Ob comparisons according to their CpG island locations status. F Barplots showing the relative distribution of hyper- and hypomethylated DMRs from the Ob.C, ObDia.C and ObDia.Ob comparisons according to their gene location status. Rightmost bars represent the background distribution considering all 772,088 probes analysed. g On the left, scatter plot showing the PCA of the subjects according to their methylation values at Ob.C, ObDia.C, ObDia.Ob DMPs. On the right, scatter plot representing the PCA of the subjects according to their methylation values at CpGs from Ob.C, ObDia.C, ObDia.Ob DMRs concentrated at regulatory sequences and thus could have potential functional implications.i ObDia.Ob hypo Fisher’s test P < 0.001, OR = 3.40). As anticipated, the regional-level analysis of DNA meth- ylation alterations allowed us to discover candidates for obesity-mediated maternal influence, which were We also confirmed, via PCA, how these DNA methyla- tion biomarkers were able to distinguish our phenotypes Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 10 of 18 Page 10 of 18 of at the DMP level (Fig. 3G, top left) and, more subtly, at the DMR level (Fig. 3G, top right). Interestingly, we observed that these biomarkers were also affected by developmental-associated changes during the first year of life, as was revealed by grouping the subjects by time point, whereby the main component (PC1) tended to separate t0 subjects from t6 and t12 subjects within each maternal group in the PCA with DMPs (Fig. 3G, bottom left), and across all groups when using DMRs (Fig. 3G, bottom right). Thus, both DMPs and DMRs captured developmental epigenetic alterations, especially during the first six months of life, thereby establishing a relation- ship between maternal metabolic condition and early childhood development. in spite of the limited number of DMRs involved in the analysis. Hypermethylated DMRs were strikingly con- centrated in genes involved in the transport of organic compounds, fundamentally fatty acids, as well as in the metabolism of vitamins and steroids (Fig. 4, left panel). Maternal obesity and gestational diabetes during pregnancy alter the methylomic landscape of the offspring On the other hand, hypomethylated DMRs were located at genes involved in mitochondrial metabolism through aerobic respiration, mitochondrial autophagy and nitric oxide production (Fig. 4, right panel). In addition, DMR alterations within the obesity with gestational diabe- tes group (ObDia.C) were also enriched (unadjusted P < 0.05) in developmental processes for hypermethyla- tion and metabolic pathways for hypomethylation (Addi- tional file 1: Fig. S7).f i Next, we explored several gene candidates affected by the regional DNA methylation alterations linked to obesity and GDM some of which could also present development-associated alterations. (Additional file 1: Fig. S8). Some transporter genes of organic compounds (ATP11A, CPT1B, SLC38A4) were consistently hyper- methylated in the Ob and ObDia groups at all time points and, furthermore, underwent hypermethylation in development between 0 and 6–12 months. In con- trast, the TTYH1 transporter gene was hypermethyl- ated in the Ob and ObDia groups, but did not exhibit hypermethylation during development. We also found diabetes-specific biomarkers as SLC35F3 gene, which was exclusively hypermethylated in the ObDia group. Regarding hypomethylated DMRs (Additional file 1: Fig. S9), they involved genes linked to metabolic control Maternal DNA methylation biomarkers are enriched in functional signatures related to metabolic pathways in functional signatures related to metabolic pathways After establishing the existence of maintained DNA methylation alterations in offspring according to the maternal condition, we sought to explore their functional features and whether they could constitute epigenetic signatures of metabolic interest. Consequently, we per- formed pathway enrichment analyses against the Gene Ontology Biological Process database to clarify whether the Ob.C hyper- and hypo- DMRs were concentrated in genes relevant for the molecular physiopathology of obesity. Using the top 25 most-significant obesity-associ- ated ontologies (unadjusted P < 0.05), we built a network through gene-set similarity in which we observed clus- ters of highly similar pathways of great interest (Fig. 4), Fig. 4 Maternal obesity generates DNA methylation signatures in offspring that are related to the metabolism of organic compounds, fatty acids and mitochondria. Network showing the similarity between the pathways found enriched (unadjusted P < 0.05) in the significant DMR analyses involving the Ob.C comparison. The clusters are coloured to show hyper- (red) and hypomethylation (blue) alterations The interplay between developmental and maternal‑derived DNA methylation alterations in offspringi f After defining the genomic and functional features of the infants’ DNA methylation alterations, we integrated the specific DNA methylation alterations coming from devel- opmental epigenetic remodelling with those produced by maternal metabolic condition during pregnancy to gain insight into the interaction between both processes. We first studied whether development-associated CpGs (defined as those differentially methylated at least in one longitudinal comparison (0 > 6 or 6 > 12) preserv- ing the direction of methylation change during the first year of life; Dev CpGs: 27,491 CpGs; 9593 hyper; 17,898 hypo) were particularly affected by maternal metabolic alterations, and, interestingly, we observed a significant enrichment between Dev CpGs and those from maternal obesity-associated DMRs (Fig. 5A, Fisher’s test P < 0.001, OR = 2.10), particularly in the case of the hypermethyla- tion changes (hyper: Fisher’s test P < 0.001, OR = 9.42; hypo: Fisher’s test P = 0.097, OR = 1.59) (Fig. 5B). We next classified Dev CpGs into three groups based on the mean methylation change that Ob children expe- rienced compared to controls: Hyper (> 1% gain), Hypo (> 1% loss), Equal (≤ 1% change). The majority of Dev CpGs did not change between Ob and C children (18,774 CpGs, ~ 68%), but 5209 (~ 19%) suffered hypometh- ylation while 3508 CpGs underwent hypermethylation (~ 13%) (Fig. 5C). When we excluded those developmen- tal changes that remain Equal for Ob.C comparison, we observed an extremely significant association between the direction of the methylation changes in development and obesity (Fisher’s test P < 0.001, OR = 3.02) (Fig. 5D). In sum, these results support the hypothesis that mater- nal obesity during pregnancy is able to alter postnatal development in a longitudinal fashion.i Next, we characterized the molecular pathways enriched in DNA methylation alterations during post- natal development, taking into account the effect of obesity and obesity with GDM. Enrichment analyses for both the GO BP and CGP MSigDB databases (FDR < 0.05) revealed that maternal-influenced Dev CpGs are located in the same pathways than those that did not suf- fer appreciable changes by maternal condition (Fig. 5E, Additional file 1: Fig. S10E). This result reinforces the idea that the maternal effect on early development tends to be global and untargeted, for both hypermethylation and hypomethylation processes. Finally, we also investigated if maternal-condition CpGs from Ob.C and ObDia.C DMRs were particularly targeted by developmental changes found. (See figure on next page.) Fig. 5 Maternal obesity globally affects postnatal development pathways. a Venn diagram showing the intersection between the DMPs that change at least in one longitudinal comparison preserving the direction of the change (Dev) and the CpGs found in the Ob.C DMRs. b Venn diagrams depicting the overlaps between aforementioned CpGs, separated by the direction of the methylation change. c Sankey diagram describing the distribution of Dev DMPs when considering the changes between Ob and C subjects. d Sankey diagram showing the distribution of those Dev DMPs that experience hyper- and hypo- changes between Ob and C subjects. e Bubble plots showing the top 5 most significant GO BP and CGP gene sets (FDR < 0.05) for the Dev DMPs based on their methylation patterns between Ob and C subjects Fig. 4 Maternal obesity generates DNA methylation signatures in offspring that are related to the metabolism of organic compounds, fatty acids and mitochondria. Network showing the similarity between the pathways found enriched (unadjusted P < 0.05) in the significant DMR analyses involving the Ob.C comparison. The clusters are coloured to show hyper- (red) and hypomethylation (blue) alterations Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 11 of 18 Page 11 of 18 Alba‑Linares et al. Cardiovascular Diabetology of diabetes such as FN3K, RPH3AL and HOX, HDAC4. These candidates were mainly hypomethylated in the Ob and ObDia groups at all time points and did not undergo appreciable changes during development. significant enrichment of common changes (hyper: Fish- er’s test P < 0.01, OR = 2.04; hypo: Fisher’s test P = 0.335; OR = 1.33) (Additional file 1: Fig. S10B). By grouping Dev CpGs in categories based on average DNA meth- ylation change of ObDia.C comparison (± 1% change), we observed 18,503 Equal CpGs (~ 67%), 6014 Hyper (~ 22%) and 2974 Hypo (~ 11%) (Additional file 1: Fig. S10C). Unlike previous analyses for Ob.C, the propor- tion of hypermethylation changes compared to hypo- methylation alterations increased with the appearance of GDM. Considering only those developmental CpGs that change with ObDia condition, we found a strong association between the direction of the methylation changes in development and obesity with GDM (Fisher’s test P < 0.001, OR = 5.55; Additional file 1: Fig. S10D). Therefore, our evidence suggests that maternal metabolic condition during pregnancy influences early child devel- opment though DNA methylation alterations which usu- ally intensify the molecular changes that occur during normal development. Fig. 5 (See legend on previous page.) The interplay between developmental and maternal‑derived DNA methylation alterations in offspringi c Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P < 0.05) for the Ob.C DMR CpGs based on their methylation patterns across development Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing f Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing the distribution of CpGs coming from Ob.C DMRs when considering their methylation differences during development. b Sankey diagram showing the distribution of Ob.C DMR CpGs excluding those that do not change during development. c Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P < 0.05) for the Ob.C DMR CpGs based on their methylation patterns across development Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing the distribution of CpGs coming from Ob.C DMRs when considering their methylation differences during development. b Sankey diagram showing the distribution of Ob.C DMR CpGs excluding those that do not change during development. c Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P < 0.05) for the Ob.C DMR CpGs based on their methylation patterns across development Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing the distribution of CpGs coming from Ob.C DMRs when considering their methylation differences during development. b Sankey diagram showing the distribution of Ob.C DMR CpGs excluding those that do not change during development. c Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P < 0.05) for the Ob.C DMR CpGs based on their methylation patterns across development determined that DMR CpGs that suffered developmen- tal changes did not show any clear differential pattern with those that were not altered, for both Ob.C (Fig. 6C) and ObDia.C (Additional file 1: Fig. S11C) comparisons. Therefore, although DMR CpGs captured developmental alterations, the molecular functions associated to mater- nal effect on offspring prevailed over any consequence the direction of methylation based on the maternal con- dition and the postnatal development was very strong for Ob.C DMR CpGs (Fisher’s test P < 0.001, OR = 5.00; Fig. The interplay between developmental and maternal‑derived DNA methylation alterations in offspringi To this end, we classified those CpGs in three groups based on the average DNA methylation change that occurred between t12 and t0: Hyper (> 1% gain), Hypo (> 1% loss), Equal (≤ 1% change). In the case of Ob.C DMR CpGs, the majority suffered from hypermethylation changes during development (518 CpGs, ~ 61%), while 155 CpGs experi- enced hypomethylation (~ 18%) and 175 remained with- out noticeable changes (~ 21%) (Fig. 6A). We observed the same pattern for ObDia.C DMR CpGs (Hyper: 633, ~ 54%; Hypo: 214, ~ 18%; Equal: 326, ~ 28%) (Addi- tional file 1: Fig. S11A). Again, the association between In the case of gestational diabetes, we did not find evidence for enrichment between ObDia.C altera- tions and Dev CpGs in general (Fisher’s test P = 0.237, OR = 0.81) (Additional File 1: Fig. S10A), though we again found that hypermethylation processes showed a‑Linares et al. Cardiovascular Diabetology (2023) 22:44 g. 5 (See legend on previous page.) Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 12 of 18 ee legend on previous page.) Fig. 5 (See legend on previous page.) Page 13 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 the direction of methylation based on the maternal con- dition and the postnatal development was very strong for Ob C DMR CpGs (Fisher’s test P<0 001 OR=5 00; determined that DMR CpGs that suffered developmen- tal changes did not show any clear differential pattern with those that were not altered for both Ob C (Fig 6C) Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing the distribution of CpGs coming from Ob.C DMRs when considering their methylation differences during development. b Sankey diagram showing the distribution of Ob.C DMR CpGs excluding those that do not change during development. c Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P < 0.05) for the Ob.C DMR CpGs based on their methylation patterns across development Fig. 6 Development modifies the methylation of obesity biomarkers in offspring without altering their functionality. a Sankey diagram describing the distribution of CpGs coming from Ob.C DMRs when considering their methylation differences during development. b Sankey diagram showing the distribution of Ob.C DMR CpGs excluding those that do not change during development. Discussion In this work, we have analysed the paediatric methyl- ome across the first year of life in order to establish the existence of epigenetic signatures that reflect the mater- nal metabolic condition during pregnancy on the off- spring beyond birth. To this end, we employed a cohort formed by longitudinal whole blood samples at 0, 6 and 12 months, whose methylome was profiled using Illu- mina Infinium MethylationEPIC BeadChip arrays. With this strategy, we first characterized those DNA methyla- tion alterations produced during postnatal development, defining the first six months of life as the most dynamic for epigenetic remodelling. In addition, we observed that a significant proportion of the altered loci continued to experience changes until the age of 12 months, preserv- ing the direction of the effect, which highlights their importance in achieving a correct postnatal develop- ment. Likewise, we performed cross-sectional studies to infer systemic DNA methylation alterations that allow to distinguish children born to mothers suffering from obesity or obesity with GDM during pregnancy from infants born to healthy controls. Our results revealed that there are DNA methylation biomarkers at both sin- gle CpG positions and genomic regions that track these maternal differences in pregnancy at least during the first year of infants’ lives. These DNA methylation alterations based on the maternal condition were concentrated at genes that recapitulated metabolic pathways of transport of fatty acids, mitochondrial bioenergetics and several developmental processes. Although an important part of these biomarkers experienced parallel alterations across the first year of life, the functionality of the involved genes remained in the same metabolic signatures, high- lighting the consistency of these biomarkers during development. Interestingly, we found that maternal influence during pregnancy was able to alter postnatal development, especially those changes that implied gain of methylation. In addition, more subtle influences in development confirmed that maternal obesity tended to intensify development processes at the methylation level, in a global way, without apparently targeting any specific developmental process. Therefore, the results presented in this manuscript provide a systematic, comprehensive analysis of the epigenetic dynamics associated with the early postnatal development, the prenatal intrauterine conditioning and the interactions between both pro- cesses in a longitudinal fashion. The interplay between developmental and maternal‑derived DNA methylation alterations in offspringi 6B) and ObDia.C DMR CpGs (Fisher’s test P < 0.001, OR = 1.97; Additional file 1: Fig. S11B). When we evaluated the GO BP pathways that were enriched in DNA methylation alterations (unadjusted P < 0.05), we Page 14 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 14 of 18 related to the direction of the developmental DNA meth- ylation change. related to the direction of the developmental DNA meth- ylation change. especially during the first 3 years [48]. The uncovering of the epigenetic relevance of the first six months of life could lead to new studies that point the importance of earlier lifestyle interventions on adult health, especially at the cardiovascular level. From a functional viewpoint, the genes that are involved in maturation processes of the hematopoietic compartment undergo a regulation programme mediated by hypomethylation changes, thus being a potential source of transcriptional activation, as has been previously reported [49]. On the other hand, loci involved in gene regulatory networks of embryonic development concentrate hypermethylation signatures in their promoter regions, consolidating Polycomb-medi- ated gene repression programmes once developmental processes have concluded [50, 51]. These results agree with previous longitudinal studies carried out from birth up to 10 years [22, 48]. All in all, early postnatal develop- ment is tightly regulated at the epigenetic level according to the direction of the methylation changes. Conclusions Altogether, in this work we have described DNA methyl- ation alterations in whole blood cells of offspring born to mothers suffering from obesity (with or without GDM) during pregnancy, on the basis of a longitudinal follow- up. Importantly, this foetal reprogramming is maintained during the first year of life and displays epigenetic signa- tures enriched in metabolic pathways, thus suggesting a link between the maternal intrauterine environment and the offspring’s genome function. Moreover, some of these alterations are observed to interact with the epigenetic reprogramming occurring during early-life development. The definition of DNA methylation biomarkers in the off- spring which are associated with maternal condition may be of value for estimating the risk for the development of cardiometabolic diseases, but further studies will be needed to assess their feasibility. From a paediatric per- spective, we foresee that our results could pave the way to the design of earlier interventions that limit the appear- ance of cardiometabolic events in adult life. Thus, our longitudinal design could be a baseline for future studies aimed at exploring the potential functional implications of these epigenetic footprints on intergenerational health. In this work, we also seek to draw attention to the inter- play between intrauterine conditioning and postnatal development. Importantly, obesity-mediated epigenetic alterations are maintained regardless of the developmen- tal changes that can occur concurrently at those CpG sites. In fact, those DNA methylation biomarkers that are also modified during development are still related to the same metabolic functional pathways, providing robust evidence for the conservation of these epigenetic signatures with time. From a developmental viewpoint, we observed that maternal obesity with or without GDM tends to magnify the DNA methylation changes which occur with time, although opposite dynamics are also observed. Further longitudinal studies will be needed to ascertain if maternal metabolic condition causes an acceleration of developmental processes, at least from an epigenetic perspective. All in all, these results provide support for the notion that the maternal metabolic influ- ence during pregnancy can affect the epigenetic features of early development, with hitherto unexplored conse- quences for future health. Ideally, public health interven- tions should focus on controlling the maternal metabolic status during pregnancy, but at the same time, our results point that the first six months of infant’s lives are crucial for an adequate postnatal development. Discussion Regarding postnatal development, our results are in line with previous studies that establish the first 5 years y g Using our longitudinal design, we performed a com- prehensive epigenetic profiling at the single-CpG level to uncover reliable DNA methylation biomarkers of maternal obesity (with or without GDM) in infant blood samples. Importantly, these alterations were main- tained beyond birth across the first year of life. Moreo- ver, detecting DNA methylation changes at the regional level constituted a better option for discovering altera- tions enriched at regulatory sequences and thus with potential functional implications. Strikingly, our results revealed a clear enrichment of the alterations in func- tions important for the molecular physiopathology of obesity. For instance, hypermethylated regions affected important deregulated genes in cardiometabolic diseases, the majority related to the transport of organic mole- cules and fatty acids. In obese subjects, CPT1B promoter hypermethylation has been associated with diminished muscular expression in response to lipids leading to a reduced ability to oxidize fatty acids [52]. The SLC38A4 amino acid transporter, which is crucial for the placen- tal nutrition of the embryo, is overexpressed in the pla- centas from human foetuses with macrosomia [53], while its knockout causes foetal weight reduction in mice [54]. Several polymorphisms in this transporter gene are also linked to the appearance of hyperglycaemia [55] even in the placenta of normal-weight newborns [56]. Similarly, Soranzo and colleagues have shown that ATP11A is sig- nificantly associated with the levels of HbA1c, which is used to monitor diabetes [57]. The SLC35F3 gene, a dia- betes-specific biomarker, is a thiamine transporter that exhibits polymorphisms related to increased blood pres- sure and potential higher risk of hypertension [58, 59]. When we considered the hypomethylated regions, we found genes such as FN3K, which has been associated Regarding postnatal development, our results are in line with previous studies that establish the first 5 years of life [22] as the most critical for epigenetic remodelling, Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 15 of 18 Finally, despite the fact that our study does not allow us to infer direct functional consequences, detecting these changes in blood indicates that maternal influence causes a systemic effect in the offspring through epigenetic foot- prints that are preserved beyond birth. with HbA1clevels [57]. Discussion In addition, the RPH3AL and HOX genes are known to suffer DNA methylation altera- tions during adipogenesis in cells from obese patients with and without type II diabetes [60]. We also observed alterations in epigenetic modifiers such as HDAC4, whose mutation impairs β-cell function and insulin secretion, leading to a non-autoimmune paediatric dia- betes [61]. Author contributions MFF and EL conceived, coordinated, and supervised the study. EL, RGZ, NGC and FPZ were involved in the collection of clinical samples. JJAL and RFP produced the methylation data and performed computational analyses. JRT and AFF assisted in the computational analyses. AFF, JRT, DBR, RGZ, NGC and FPZ assisted in data analyses and interpretation. JJAL, RFP, MFF and EL participated in drafting the manuscript. All authors read and approved the final manuscript. The online version contains supplementary material available at https://doi. org/10.1186/s12933-023-01774-y. The online version contains supplementary ma org/10.1186/s12933-023-01774-y. The online version contains supplementary material available at https://doi. org/10.1186/s12933-023-01774-y. Additional file 1: Supplementary Information and Supplementary Figures S1–11. Figure S1. Barplots depicting the average blood cell- type composition for each combination of group (Control, Obese, Obese+Diab) and time point (t0, t6, t12), inferred via the Houseman algorithm from methylation data. Figure S2. Violin plot showing the magnitude of change of those DMPs common to common 0>6 and 6>12 which had the same direction of change at both time points. The magnitude of change is measured as the absolute β-value difference in the average methylation values of each CpG between the longitudinal groups. The P-value from a Wilcoxon rank sum test is shown. Figure S3. Boxplots showing examples of genes that accumulate a high number of hypermethylation alterations (FDR<0.05) during the first year of develop‑ ment. Figure S4. Upset plots depicting intersections of significant results (FDR<0.05) for Chemical and Genetic Perturbations (CGP), Gene Ontology Biological Process (GOBP) and Reactome (REACT) gene set enrichment analyses. Figure S5. Boxplots showing examples of genes that accumu‑ late a high number of hypomethylation alterations (FDR<0.05) during the first year of development. Figure S6. Venn diagrams indicating (A) the overlap of DMPs between all comparisons (Ob.C, ObDia.C, ObDia.Ob) in terms of hypermethylation DMPs and hypomethylation (FDR<0.05) and (B) the overlap of CpGs belonging to DMRs (Sidak P<0.05) between the same comparisons. Figure S7. Network showing the similarities found in the ObDia.C comparison with respect to the pathways found enriched (unadjusted P<0.05) in the significant DMR analyses. Blue clusters relate to hypomethylation alterations and red clusters to hypermethylated DMRs. Figure S8. Boxplots showing examples of DMRs that are hypermethylated in the obesity and/or obesity+diabetes groups (Sidak P-value<0.05). Fig‑ ure S9. Boxplots showing some examples of DMRs that are hypomethyl‑ ated in the obesity and/or obesity+diabetes groups (Sidak P-value<0.05). Figure S10. Conclusions (a) Sankey diagram describing the distribution of CpGs coming from ObDia.C DMRs when considering their methylation differences during development. (b) Sankey diagram showing the distribution of ObDia.C DMR CpGs excluding those that do not change during development. (c) Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P<0.05) for the ObDia.C DMR CpGs based on their methyla‑ tion patterns across development. Additional file 2: Table S1. Expanded clinical information related to the subjects. Additional file 2: Table S1. Expanded clinical information related to the subjects. Additional file 3: Table S2. List and details of the DMPs (FDR<0.05) found in the longitudinal analyses (0>6; 6>12). Additional file 4: Table S3. List and details of the DMPs (FDR<0.05) found in the cross-sectional analyses (Ob.C; ObDia.C; ObDia.Ob). Additional file 5: Table S4. Lists and details of the DMRs (Sidak-corrected P<0.05) found in the cross-sectional analyses (Ob.C; ObDia.C; ObDia.Ob). Conclusions Abbreviations BMI Body mass index GDM Gestational diabetes mellitus DNA Deoxyribonucleic acid CpG Cytosine nucleotide followed by a guanine nucleotide GCT Glucose challenge test OGTT Glucose tolerance test WHO World Health Organisation IDAT Intensity data SNP Single nucleotide polymorphism MAF Minor allele frequency SBE Single base extension DMPs Differentially methylated probes t0/t6/t12 Timepoint (0/6/12 months of life) FDR False discovery rate DMRs Differentially methylated regions CGI CpG Island OR Odds ratio MSigDB Molecular Signature DataBase P P-value PCA Principal component analysis PRC2 Polycomb repressive complex 2 SUZ12 SUZ12 polycomb repressive complex 2 subunit NFIX Nuclear factor IX TBX1 T-Box transcription factor 1 WNT10A Wingless-type MMTV integration site family, member 10A NRG2 Neuregulin 2 CALCA Calcitonin related polypeptide alpha CDH23 Cadherin related 23 CGP Chemical and genetic perturbations (gene sets) q p p That said, our observations present limitations. First, individual genetic traits could explain why our obese- mediated DNA methylation changes are longitudinally maintained during the first year of life [62]. In addi- tion, although our results support the existence of an obesity-mediated intrauterine effect on the childhood methylome, we cannot rule out that the maintenance of the DNA methylation changes across the first year of life is the result of early-life lifestyle factors. Neverthe- less, our study clearly positions epigenetic mechanisms as the molecular link that explains the environmental maternal influence in the offspring. Other works that use methylation arrays also support this conclusion in a non- longitudinal fashion, showing that gestational diabetes is associated with methylation changes of metabolic genes in placenta [25, 26, 63, 64] and blood samples from new- borns, children and adolescents [65–70]. Fewer studies have addressed associations between DNA methylation alterations and maternal obesity, those that exist mainly using only cord blood samples from newborns [71–74]. Alba‑Linares et al. Availability of data and materials All data generated during this study are included in the published article and its supplementary files. The raw IDAT and processed data are also available in the ArrayExpress public repository under accession E-MTAB-12728. Funding This work was supported by: the Spanish Association Against Cancer (Grant number PROYE18061FERN to M.F.F.), the Asturias Government (PCTI) cofound‑ ing 2018–2022/FEDER (Grant number IDI/2018/146 to M.F.F.), the Fundación General CSIC (Grant number 0348_CIE_6_E to M.F.F.), the Institute of Health Carlos III (Plan Nacional de I + D + I) cofounding FEDER (Grant numbers PI18/01527 and PI21/01067 to M.F.F. and A.F.F.; Grant numbers PI17/01517 and PI20/00269 to E.L.), and the Spanish Ministry of Science and Innovation (Grant number SGL2021-03-039/40 to M.F.F.) cofounding NextGenerationEU. J.R.T. is supported by a Juan de la Cierva fellowship from the Spanish Ministry of Science and Innovation (Grant number IJC2018-36825-I). J.J.A.L. is supported by the Spanish Association Against Cancer (Grant number PRDAS21642ALBA). R.F.P. and D.B.R. are supported by the Severo Ochoa program (Grant numbers BP17-114 and BP20-186). We also acknowledge support from the Institute of Oncology of Asturias (IUOPA, supported by Obra Social Cajastur Liberbank, Spain), the Health Research Institute of Asturias (ISPA-FINBA), the Health Research Institute INCLIVA and the Biomedical Research Networking Center on Rare Diseases (CIBERER-ISCIII). Supplementary Information Supplementary Information Author contributions (a) Venn diagram showing the intersection between the DMPs that change at least in one longitudinal comparison preserving the direction of the change (Dev) and the CpGs found in the ObDia.C DMRs. (b) Venn diagrams depicting the overlaps between aforementioned Conclusions Cardiovascular Diabetology (2023) 22:44 Page 16 of 18 GO BP Gene ontology biological process (gene sets) FOXP1 Forkhead box P1 ETS1 ETS proto-oncogene 1, transcription factor HIVEP2 Human immunodeficiency virus type I enhancer-binding prot 2 HIVEP3 Human immunodeficiency virus type I enhancer-binding prot 3 CYTH1 Cytohesin 1 ITGAL Integrin subunit alpha L ITGB2 Integrin subunit beta 2 HLA-E Major histocompatibility complex, class I, E BCL-2 BCL2 apoptosis regulator TNF Tumor necrosis factor PRDM PR/SET domain C Control group Ob Obese group ObDia Obese + Diabetes group Ob.C Obese group vs Control group (comparison) ObDia.C Obese + Diabetes group vs Control group (comparison) ObDia.Ob Obese + Diabetes group vs Obese group (comparison) ATP11A ATPase Phospholipid Transporting 11A CPT1B Carnitine palmitoyltransferase 1B SLC38A4 Solute carrier family 38 member 4 TTYH1 Tweety family member 1 SLC35F3 Solute carrier family 35 member F3 FN3K Fructosamine 3 kinase RPH3AL Rabphilin 3A like (without C2 domains) HOX Homeobox HDAC4 Histone deacetylase 4 HbA1c Glycated hemoglobin Dev Development-associated CpGs CpGs, separated by the direction of the methylation change. (c) Sankey diagram describing the distribution of Dev DMPs when considering the changes between ObDia and C subjects. (d) Sankey diagram showing the distribution of those Dev DMPs that experience hyper- and hypo- changes between ObDia and C subjects. (e) Bubble plots showing the top 5 most significant GO BP and CGP gene sets (FDR<0.05) for the Dev DMPs based on their methylation patterns between ObDia and C subjects. Figure S11. (a) Sankey diagram describing the distribution of CpGs coming from ObDia.C DMRs when considering their methylation differences during development. (b) Sankey diagram showing the distribution of ObDia.C DMR CpGs excluding those that do not change during development. (c) Bubble plots showing the top 5 most significant GO BP and CGP gene sets (unadjusted P<0.05) for the ObDia.C DMR CpGs based on their methyla‑ tion patterns across development. CpGs, separated by the direction of the methylation change. (c) Sankey diagram describing the distribution of Dev DMPs when considering the changes between ObDia and C subjects. (d) Sankey diagram showing the distribution of those Dev DMPs that experience hyper- and hypo- changes between ObDia and C subjects. (e) Bubble plots showing the top 5 most significant GO BP and CGP gene sets (FDR<0.05) for the Dev DMPs based on their methylation patterns between ObDia and C subjects. Figure S11. Acknowledgements We are grateful to Ronnie Lendrum for her manuscript editing and to the members of the Cancer Epigenetics Laboratory for their positive contribu‑ tions. Finally, we thank the patients involved in the study for their generous contribution. References l h 1. Blüher M. Obesity: global epidemiology and pathogenesis. Nat Rev Endocrinol. 2019;15:288–98. 25. Franzago M, Fraticelli F, Stuppia L, Vitacolonna E. Nutrigenetics, epigenet‑ ics and gestational diabetes: consequences in mother and child. Epige‑ netics. 2019;14:215–35. 2. Prospective Studies Collaboration, Whitlock G, Lewington S, Sherliker P, Clarke R, Emberson J, et al. Body-mass index and cause-specific mortal‑ ity in 900 000 adults: collaborative analyses of 57 prospective studies. Lancet. 2009;373:1083–96. 26. Słupecka-Ziemilska M, Wychowański P, Puzianowska-Kuznicka M. Gesta‑ tional diabetes mellitus affects offspring’s epigenome. Is there a way to reduce the negative consequences? Nutrients. 2020;12:2792. 3. Gaillard R. Maternal obesity during pregnancy and cardiovascular devel‑ opment and disease in the offspring. Eur J Epidemiol. 2015;30:1141–52. 3. Gaillard R. Maternal obesity during pregnancy and cardiovascular devel‑ opment and disease in the offspring. Eur J Epidemiol. 2015;30:1141–52. 27. Reichetzeder C. Overweight and obesity in pregnancy: their impact on epigenetics. Eur J Clin Nutr. 2021;75:1710–22. 4. Fraser A, Tilling K, Macdonald-Wallis C, Sattar N, Brion M-J, Benfield L, et al. Association of maternal weight gain in pregnancy with offspring obesity and metabolic and vascular traits in childhood. Circulation. 2010;121:2557–64. 4. Fraser A, Tilling K, Macdonald-Wallis C, Sattar N, Brion M-J, Benfield L, et al. Association of maternal weight gain in pregnancy with offspring obesity and metabolic and vascular traits in childhood. Circulation. 2010;121:2557–64. 28. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics. 2014;30:1363–9. 5. Gaillard R, Steegers EAP, Duijts L, Felix JF, Hofman A, Franco OH, et al. Childhood cardiometabolic outcomes of maternal obesity during preg‑ nancy: the Generation R Study. Hypertension. 2014;63:683–91. 29. Zhou W, Triche TJ, Laird PW, Shen H. SeSAMe: reducing artifactual detec‑ tion of DNA methylation by Infinium BeadChips in genomic deletions. Nucleic Acids Res. 2018;46:e123. y y y 6. Oostvogels AJJM, Stronks K, Roseboom TJ, van der Post JAM, van Eijsden M, Vrijkotte TGM. Maternal prepregnancy BMI, offspring’s early postnatal growth, and metabolic profile at age 5–6 years: the ABCD Study. J Clin Endocrinol Metab. 2014;99:3845–54. 6. Oostvogels AJJM, Stronks K, Roseboom TJ, van der Post JAM, van Eijsden M, Vrijkotte TGM. Maternal prepregnancy BMI, offspring’s early postnatal growth, and metabolic profile at age 5–6 years: the ABCD Study. J Clin Endocrinol Metab. 2014;99:3845–54. 30. References l h Chen Y, Lemire M, Choufani S, Butcher DT, Grafodatskaya D, Zanke BW, et al. Discovery of cross-reactive probes and polymorphic CpGs in the Illumina Infinium HumanMethylation450 microarray. Epigenetics. 2013;8:203–9. 7. Perng W, Gillman MW, Mantzoros CS, Oken E. A prospective study of maternal prenatal weight and offspring cardiometabolic health in mid‑ childhood. Ann Epidemiol. 2014;24:793-800.e1. 31. Pidsley R, Zotenko E, Peters TJ, Lawrence MG, Risbridger GP, Molloy P, et al. Critical evaluation of the Illumina MethylationEPIC BeadChip microarray for whole-genome DNA methylation profiling. Genome Biol. 2016;17:208. 8. Yu Z, Han S, Zhu J, Sun X, Ji C, Guo X. Pre-pregnancy body mass index in relation to infant birth weight and offspring overweight/obesity: a systematic review and meta-analysis. PLoS ONE. 2013;8:e61627. 32. Andrews SV, Ladd-Acosta C, Feinberg AP, Hansen KD, Fallin MD. “Gap hunting” to characterize clustered probe signals in Illumina methylation array data. Epigenet Chromatin. 2016;9:56. 9. Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. JAMA. 2012;307:491–7. 33. Triche TJ, Weisenberger DJ, Van Den Berg D, Laird PW, Siegmund KD. Low-level processing of Illumina Infinium DNA Methylation BeadArrays. Nucleic Acids Res. 2013;41:e90. 10. Lurbe E, Aguilar F, Álvarez J, Redon P, Torró MI, Redon J. Determinants of cardiometabolic risk factors in the first decade of life. Hypertension. 2018;71:437–43. 34. Teschendorff AE, Marabita F, Lechner M, Bartlett T, Tegner J, Gomez- Cabrero D, et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics. 2013;29:189–96. 11. Hochner H, Friedlander Y, Calderon-Margalit R, Meiner V, Sagy Y, Avgil- Tsadok M, et al. Associations of maternal prepregnancy body mass index and gestational weight gain with adult offspring cardiometabolic risk factors: the Jerusalem Perinatal Family Follow-up Study. Circulation. 2012;125:1381–9. 35. Tian Y, Morris TJ, Webster AP, Yang Z, Beck S, Feber A, et al. ChAMP: updated methylation analysis pipeline for Illumina BeadChips. Bioinfor‑ matics. 2017;33:3982–4. 36. Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, et al. DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinform. 2012;13:86. 12. Reynolds RM, Allan KM, Raja EA, Bhattacharya S, McNeill G, Hannaford PC, et al. Maternal obesity during pregnancy and premature mortality from cardiovascular event in adult offspring: follow-up of 1 323 275 person years. BMJ. 2013;347:f4539. 37. Received: 6 December 2022 Accepted: 15 February 2023 23. Kaimala S, Ansari SA, Emerald BS. DNA methylation in the pathogen‑ esis of type 2 diabetes. Vitamins and Hormones. Academic Press; 2022. https://www.sciencedirect.com/science/article/pii/S0083672922000851. Accessed 13 Feb 2023. 24. Davison GW, Irwin RE, Walsh CP. The metabolic-epigenetic nexus in type 2 diabetes mellitus. Free Radical Biol Med. 2021;170:194–206. The authors declare that they have no competing interests. The authors declare that they have no competing interests. 17. Bray GA, Frühbeck G, Ryan DH, Wilding JPH. Management of obesity. Lancet. 2016;387:1947–56. Ethics approval and consent to participate Ethics approval and consent to participate This study was approved by the Clinical Research Ethics Committee of the Consorcio Hospital General Universitario de Valencia in accordance with the Page 17 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Declaration of Helsinki and all parents gave written informed consent for their children to participate in the study. 14. McIntyre HD, Catalano P, Zhang C, Desoye G, Mathiesen ER, Damm P. Gestational diabetes mellitus. Nat Rev Dis Primers. 2019;5:47. McIntyre HD, Catalano P, Zhang C, Desoye G, Mathiesen ER, Damm 15. Christensen MH, Rubin KH, Petersen TG, Nohr EA, Vinter CA, Andersen MS, et al. Cardiovascular and metabolic morbidity in women with previous gestational diabetes mellitus: a nationwide register-based cohort study. Cardiovasc Diabetol. 2022;21:179. Competing interests 16. Huda SS, Brodie LE, Sattar N. Obesity in pregnancy: prevalence and meta‑ bolic consequences. Semin Fetal Neonatal Med. 2010;15:70–6. Author details 18. Barker DJ, Osmond C, Golding J, Kuh D, Wadsworth ME. Growth in utero, blood pressure in childhood and adult life, and mortality from cardiovas‑ cular disease. BMJ. 1989;298:564–7. 1 Cancer Epigenetics and Nanomedicine Laboratory, Nanomaterials and Nano‑ technology Research Center (CINN‑CSIC), University of Oviedo, Oviedo, Spain. 2 Health Research Institute of Asturias (ISPA-FINBA), University of Oviedo, Oviedo, Spain. 3 Institute of Oncology of Asturias (IUOPA), University of Oviedo, Oviedo, Spain. 4 Department of Organisms and Systems Biology (B.O.S.), University of Oviedo, Oviedo, Spain. 5 Biomedical Research Networking Center on Rare Diseases (CIBERER), Institute of Health Carlos III (ISCIII), Madrid, Spain. 6 Health Research Institute INCLIVA, Valencia, Spain. 7 Biomedical Research Networking Center for Physiopathology of Obesity and Nutrition (CIBEROBN), Institute of Health Carlos III (ISCIII), Madrid, Spain. 8 Servicio de Pediatría, Cons‑ orcio Hospital General Universitario de Valencia, Valencia, Spain. 19. Lurbe E, Ingelfinger J. Developmental and early life origins of cardiometa‑ bolic risk factors. Hypertension. 2021;77:308–18. 20. Ormazabal V, Nair S, Carrión F, Mcintyre HD, Salomon C. The link between gestational diabetes and cardiovascular diseases: potential role of extra‑ cellular vesicles. Cardiovasc Diabetol. 2022;21:174. 21. Feil R, Fraga MF. Epigenetics and the environment: emerging patterns and implications. Nat Rev Genet. 2012;13:97–109. 22. Pérez RF, Santamarina P, Tejedor JR, Urdinguio RG, Álvarez-Pitti J, Redon P, et al. Longitudinal genome-wide DNA methylation analysis uncovers persistent early-life DNA methylation changes. J Transl Med. 2019;17:15. Received: 6 December 2022 Accepted: 15 February 2023 References l h Xu Z, Niu L, Li L, Taylor JA. ENmix: a novel background correction method for Illumina HumanMethylation450 BeadChip. Nucleic Acids Res. 2016;44:e20. 13. Schack-Nielsen L, Michaelsen KF, Gamborg M, Mortensen EL, Sørensen TIA. Gestational weight gain in relation to offspring body mass index and obesity from infancy through adulthood. Int J Obes. 2010;34:67–74. Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 Page 18 of 18 Page 18 of 18 Alba‑Linares et al. Cardiovascular Diabetology (2023) 22:44 38. Gervin K, Salas LA, Bakulski KM, van Zelm MC, Koestler DC, Wiencke JK, et al. Systematic evaluation and validation of reference and library selec‑ tion methods for deconvolution of cord blood DNA methylation data. Clin Epigenetics. 2019;11:125. 59. Zhang K, Huentelman MJ, Rao F, Sun EI, Corneveaux JJ, Schork AJ, et al. Genetic implication of a novel thiamine transporter in human hyperten‑ sion. J Am Coll Cardiol. 2014;63:1542–55. 60. Mirzaeicheshmeh E, Zerrweck C, Centeno-Cruz F, Baca-Peynado P, Martinez-Hernandez A, García-Ortiz H, et al. Alterations of DNA methyla‑ tion during adipogenesis differentiation of mesenchymal stem cells isolated from adipose tissue of patients with obesity is associated with type 2 diabetes. Adipocyte. 2021;10:493–504. 39. Salas LA, Koestler DC, Butler RA, Hansen HM, Wiencke JK, Kelsey KT, et al. An optimized library for reference-based deconvolution of whole-blood biospecimens assayed using the Illumina HumanMethylationEPIC Bead‑ Array. Genome Biol. 2018;19:64. 61. Gong M, Yu Y, Liang L, Vuralli D, Froehler S, Kuehnen P, et al. HDAC4 mutations cause diabetes and induce β-cell FoxO1 nuclear exclusion. Mo Genet Genomic Med. 2019;7:e602. 40. Du P, Zhang X, Huang C-C, Jafari N, Kibbe WA, Hou L, et al. Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinform. 2010;11:587. 41. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers dif‑ ferential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43:e47. 62. Villicaña S, Bell JT. Genetic impacts on DNA methylation: research findings and future perspectives. Genome Biol. 2021;22:127. 63. Januar V, Desoye G, Novakovic B, Cvitic S, Saffery R. Epigenetic regulation of human placental function and pregnancy outcome: considerations for causal inference. Am J Obstet Gynecol. 2015;213:S182-196. 42. Pedersen BS, Schwartz DA, Yang IV, Kechris KJ. Comb-p: software for com‑ bining, analyzing, grouping and correcting spatially correlated P-values. Bioinformatics. 2012;28:2986–8. 64. Petropoulos S, Guillemin C, Ergaz Z, Dimov S, Suderman M, Weinstein- Fudim L, et al. References l h Gestational diabetes alters offspring DNA methylation profiles in human and rat: identification of key pathways involved in endocrine system disorders, insulin signaling, diabetes signaling, and ILK signaling. Endocrinology. 2015;156:2222–38. 43. Phipson B, Maksimovic J, Oshlack A. missMethyl: an R package for analyz‑ ing data from Illumina’s HumanMethylation450 platform. Bioinformatics. 2016;32:286–8. 44. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–50. g g gy 65. Canouil M, Khamis A, Keikkala E, Hummel S, Lobbens S, Bonnefond A, et al. Epigenome-wide association study reveals methylation loci associ‑ ated with offspring gestational diabetes mellitus exposure and maternal methylome. Diabetes Care. 2021;44:1992–9. 45. Merico D, Isserlin R, Stueker O, Emili A, Bader GD. Enrichment map: a network-based method for gene-set enrichment visualization and inter‑ pretation. PLoS ONE. 2010;5:e13984. 66. Chen P, Piaggi P, Traurig M, Bogardus C, Knowler WC, Baier LJ, et al. Differ‑ ential methylation of genes in individuals exposed to maternal diabetes in utero. Diabetologia. 2017;60:645–55. 46. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecu‑ lar interaction networks. Genome Res. 2003;13:2498–504. 67. Hjort L, Martino D, Grunnet LG, Naeem H, Maksimovic J, Olsson AH, et al. Gestational diabetes and maternal obesity are associated with epige‑ nome-wide methylation changes in children. JCI Insight. 2018;3:122572. 47. Gustavsen JA, Pai S, Isserlin R, Demchak B, Pico AR. RCy3: network biology using cytoscape from within R. F1000Research. 2019;8:1774. 68. Howe CG, Cox B, Fore R, Jungius J, Kvist T, Lent S, et al. Maternal gesta‑ tional diabetes mellitus and newborn DNA methylation: findings from the pregnancy and childhood epigenetics consortium. Diabetes Care. 2020;43:98–105. 48. Acevedo N, Reinius LE, Vitezic M, Fortino V, Söderhäll C, Honkanen H, et al. Age-associated DNA methylation changes in immune genes, histone modifiers and chromatin remodeling factors within 5 years after birth in human blood leukocytes. Clin Epigenet. 2015;7:34. 69. Kim E, Kwak SH, Chung HR, Ohn JH, Bae JH, Choi SH, et al. DNA methyla‑ tion profiles in sibling pairs discordant for intrauterine exposure to maternal gestational diabetes. Epigenetics. 2017;12:825–32. 49. Calvanese V, Fernández AF, Urdinguio RG, Suárez-Alvarez B, Mangas C, Pérez-García V, et al. A promoter DNA demethylation landscape of human hematopoietic differentiation. Nucleic Acids Res. 2012;40:116–31. 70. References l h Quilter CR, Cooper WN, Cliffe KM, Skinner BM, Prentice PM, Nelson L, et al. Impact on offspring methylation patterns of maternal gestational diabetes mellitus and intrauterine growth restraint suggest common genes and pathways linked to subsequent type 2 diabetes risk. FASEB J. 2014;28:4868–79. f 50. Leeb M, Pasini D, Novatchkova M, Jaritz M, Helin K, Wutz A. Polycomb complexes act redundantly to repress genomic repeats and genes. Genes Dev. 2010;24:265–76. 51. Viré E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al. The Polycomb group protein EZH2 directly controls DNA methylation. Nature. 2006;439:871–4. 71. Martin CL, Jima D, Sharp GC, McCullough LE, Park SS, Gowdy KM, et al. Maternal pre-pregnancy obesity, offspring cord blood DNA methylation, and offspring cardiometabolic health in early childhood: an epigenome- wide association study. Epigenetics. 2019;14:325–40. 52. Maples JM, Brault JJ, Witczak CA, Park S, Hubal MJ, Weber TM, et al. Dif‑ ferential epigenetic and transcriptional response of the skeletal muscle carnitine palmitoyltransferase 1B (CPT1B) gene to lipid exposure with obesity. Am J Physiol Endocrinol Metab. 2015;309:E345–56. 72. Rizzo HE, Escaname EN, Alana NB, Lavender E, Gelfond J, Fernandez R, et al. Maternal diabetes and obesity influence the fetal epigenome in a largely Hispanic population. Clin Epigenet. 2020;12:34. 53. Li Z, Lai G, Deng L, Han Y, Zheng D, Song W. Association of SLC38A4 and system A with abnormal fetal birth weight. Exp Ther Med. 2012;3:309–13. 73. Sharp GC, Salas LA, Monnereau C, Allard C, Yousefi P, Everson TM, et al. Maternal BMI at the start of pregnancy and offspring epigenome-wide DNA methylation: findings from the pregnancy and childhood epigenet‑ ics (PACE) consortium. Hum Mol Genet. 2017;26:4067–85. 54. Matoba S, Nakamuta S, Miura K, Hirose M, Shiura H, Kohda T, et al. Pater‑ nal knockout of Slc38a4/SNAT4 causes placental hypoplasia associated with intrauterine growth restriction in mice. Proc Natl Acad Sci USA. 2019;116:21047–53. 74. Yeung EH, Guan W, Mumford SL, Silver RM, Zhang C, Tsai MY, et al. Measured maternal prepregnancy anthropometry and newborn DNA methylation. Epigenomics. 2019;11:187–98. 55. González-Renteria SM, Loera-Castañeda V, Chairez-Hernández I, Sosa- Macias M, Paniagua-Castro N, Lares-Aseff I, et al. Association of the polymorphisms 292 C>T and 1304 G>A in the SLC38A4 gene with hyperglycaemia. Diabetes Metab Res Rev. 2013;29:39–43. Publisher’s Note 56. Galaviz-Hernández C, Sosa-Macías M, Rodríguez-Morán M, Martínez- Aguilar G, Guerrero-Romero F, González-Rentería SM. Placental SLC38A4 gene polymorphisms 1304 G > A and 292 C > T, and their association with glucose > 95 mg/dl in normal weight full-term healthy newborns. Bol Med Hosp Infant Mex. 2021;78:200–6. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. 57. Soranzo N, Sanna S, Wheeler E, Gieger C, Radke D, Dupuis J, et al. Com‑ mon variants at 10 genomic loci influence hemoglobin A1C levels via glycemic and nonglycemic pathways. Diabetes. 2010;59:3229–39. 58. Zang X-L, Han W-Q, Yang F-P, Ji K-D, Wang J-G, Gao P-J, et al. Association of a SNP in SLC35F3 gene with the risk of hypertension in a Chinese Han Population. Front Genet. 2016;7:108.
https://openalex.org/W1990047055	https://ccforum.biomedcentral.com/counter/pdf/10.1186/cc11910	English	null	Protease-activated receptor-1 impairs host defense in murine pneumococcal pneumonia: a controlled laboratory study	Critical care	2,012	cc-by	6,404	Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 RESEARCH Open Access © 2012 Schouten et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. * Correspondence: m.schouten@amc.uva.nl 1Center for Experimental and Molecular Medicine (CEMM), Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands Full list of author information is available at the end of the article Abstract Introduction: Streptococcus pneumoniae is the most common causative pathogen in community-acquired pneumonia. Protease-activated receptor-1 (PAR-1) is expressed by multiple cell types present in the lungs and can be activated by various proteases generated during acute inflammation. The cellular effect of PAR-1 activation partially depends on the specific protease involved. We here determined the role of PAR-1 in the host response during murine pneumococcal pneumonia. Methods: Wild-type (WT) and PAR-1 knockout (KO) mice were infected intranasally with viable S. pneumoniae and observed in a survival study or euthanized at 6, 24 or 48 hours of infection. Results: PAR-1 KO mice had a better survival early after infection compared to WT mice. Moreover, PAR-1 KO mice had lower bacterial loads in lungs and blood at 24 hours and in spleen and liver at 48 hours after infection. This favorable response was accompanied by lower lung histopathology scores and less neutrophil influx in PAR-1 KO mice. Conclusion: PAR-1 impairs host defense during murine pneumococcal pneumonia. involved in the coagulation system. Intriguingly, activa- tion of PAR-1 can result in opposite cellular effects depending on the protease involved in its proteolytic cleavage: for example high concentrations of thrombin can cause barrier disruptive effects on vascular endothe- lium via activation of PAR-1, whereas the anticoagulant protein activated protein C (APC) exerts barrier protec- tive and anti-inflammatory effects via the same receptor [7-9]. Protease-activated receptor-1 impairs host defense in murine pneumococcal pneumonia: a controlled laboratory study Marcel Schouten1,2, Cornelis van’t Veer1,2, Joris JTH Roelofs3, Marcel Levi4 and Tom van der Poll1,2,5 el Schouten1,2, Cornelis van’t Veer1,2, Joris JTH Roelofs3, Marcel Levi4 and Tom van der Poll1,2,5 Materials and methods Animals Tumor necrosis factor (TNF)-a, interleukin (IL)-6, IL-10, IL-12p70, interferon (IFN)-g and monocyte chemoattrac- tant protein (MCP)-1 were measured by cytometric bead array (CBA) multiplex assay (BD Biosciences, San Jose, CA, USA). Macrophage inflammatory protein (MIP)-2 was measured by ELISA (R&D systems, Minneapolis, MN, USA). Heterozygous PAR-1 KO mice on a C57Bl/6 back- ground were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) [12]. Animals were intercrossed to obtain homozygous PAR-1 KO mice. WT C57BL/6 mice were purchased from Charles River (Maastricht, the Netherlands). All experiments were approved by the Institutional Animal Care and Use Committee of the University of Amsterdam. Survival To determine whether PAR-1 is important for outcome in pneumococcal pneumonia a survival study was performed (Figure 1). PAR-1 KO mice had a significantly delayed Figure 1 Protease-activated receptor-1 accelerates mortality in murine pneumococcal pneumonia. Survival of wild-type (dashed line, open symbols) and protease-activated receptor-1 knockout (solid line, closed symbols) mice in murine pneumococcal pneumonia (14 mice per group). * indicates statistical significance as compared to wild-type (P <0.05, log rank test). Introduction Streptococcus (S.) pneumoniae or the pneumococcus is the number one causative pathogen in community- acquired pneumonia (CAP) [1]. CAP is an important cause of sepsis: in a recent large sepsis trial 35.6% of the patients suffered from severe CAP, with the pneumococ- cus being the most frequent cause [2]. Worldwide S. pneumoniae is responsible for approximately ten mil- lion deaths annually, making pneumococcal pneumonia and sepsis a major health threat [3]. We here considered it of interest to investigate the effect of PAR-1 activation on the course of pneumococcal pneumonia. Thus far, data on the role of PAR-1 in severe bacterial infection are limited to studies using endotoxe- mia or polymicrobial peritonitis induced by cecal ligation and puncture (CLP) as models of severe sepsis. Kaneider et al used a pepducin-based approach to show that acti- vation of PAR-1 is harmful during the early phases of endotoxemia and CLP-induced sepsis, but beneficial at later stages [10]. Somewhat contradicting, Niessen et al showed that PAR-1 is harmful during early as well as late stages of endotoxemia and sepsis induced by CLP, with a pivotal role for dendritic cell signaling [11]. We here for the first time studied the role of PAR-1 in respiratory Protease-activated receptors (PARs) are G protein- coupled receptors that are abundantly expressed in the lungs [4-6]. PARs, of which four family members have been described (PAR-1 to -4), carry their own ligand: proteolytic cleavage leads to exposure of a neo-amino terminus, which serves as a ligand for the same receptor, hereby initiating transmembrane signaling. A variety of proteases can activate PARs, including several proteases Page 2 of 8 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 described [15,16]. Ly-6G stained slides were photographed with a microscope equipped with a digital camera (Leica CTR500, Leica Microsystems, Wetzlar, Germany). Ten random pictures were taken per slide. Stained areas were analyzed with Image Pro Plus (Media Cybernetics, Bethesda, MD, USA) and expressed as percentage of the total surface area. tract infection, using our well-established clinically rele- vant model of pneumococcal pneumonia, comparing sur- vival, antibacterial defense and inflammatory responses in PAR-1 knockout (KO) and normal wild-type (WT) mice. We show that in pneumococcal pneumonia, PAR-1 impairs host defense, as reflected by a reduced lethality and lower bacterial loads, lung histopathology scores and less pulmonary neutrophil influx in PAR-1 KO mice. Statistical analysis Data are expressed as box-and-whisker diagrams depict- ing the smallest observation, lower quartile, median, upper quartile and largest observation, as medians with interquartile ranges or as Kaplan Meier plots. Differences between groups were determined with Mann-Whitney U or log rank test where appropriate. Analyses were per- formed using GraphPad Prism version 4.0 (GraphPad Software, San Diego, CA, USA). P values less than 0.05 were considered statistically significant. Experimental infection and sample harvesting Pneumonia was induced by intranasal inoculation with approximately 5 × 104 colony-forming units (CFU) of S. pneumoniae serotype 3 (American Type Culture Collec- tion, ATCC 6303, Rockville, MD, USA) as described [13,14]. Mice were sacrificed after 6, 24 or 48 hours of infection (n = 8 per group per time point) or observed for 4 days in a survival study (n = 14 per group). On prede- fined time points mice were anesthetized, citrated plasma was prepared from blood drawn from the vena cava infer- ior and left lung homogenates were prepared as described [13,14]. Bacterial loads were determined as described [13,14]. For further measurements, homogenates were diluted 1:2 with lysis buffer (300 mM NaCl, 30 mM Tris, 2 mM MgCl2, 2 mM CaCl2, 1% (v/v) Triton X-100, pH 7.4) with protease inhibitor mix and incubated for 30 minutes on ice, followed by centrifugation at 680 g for 10 minutes. Supernatants were stored at -20ºC until analysis. Bacterial outgrowth To determine whether the difference in survival between PAR-1 KO and WT mice in pneumococcal pneumonia could be attributed to a difference in antibacterial defense, we determined bacterial outgrowth 6, 24 and 48 hours in lungs, blood and distant organs (spleen, liver) (Figure 2). At 6 hours after infection, there were no differences in pulmonary bacterial loads between PAR-1 KO and WT mice (Figure 2A). At this time point, bacteria could not be detected in blood and distant organs (Figure 2B to 2D). At 24 hours, PAR-1 KO mice had markedly lower bacterial burdens in their lungs (Figure 2A) and blood (Figure 2B) with a trend toward lower levels in spleen (P = 0.18) (Figure 2C) as compared to WT mice. Whereas at 48 hours the differences in bacterial outgrowth in lung and blood had subsided (Figure 2A to 2B), PAR-1 KO mice had lower bacterial loads in spleen (Figure 2C) and liver (Figure 2D) as compared to WT mice. To further investigate the role of PAR-1 in the local inflammatory response, we determined levels of various cytokines (TNF-a, IL-6, IL-10, IL-12, IFN-g) and che- mokines (MCP-1, MIP-2) in lung homogenates at 6, 24 and 48 hours after infection (Table 1). During the first 24 hours after infection pulmonary cytokine and chemo- kine levels did not differ between PAR-1 KO and WT mice. At 48 hours, lung levels of TNF-a, IL-6 and IFN-g Figure 2 Lower bacterial outgrowth in protease-activated receptor-1 knockout mice in lungs and blood at 24 hours and in spleen and liver at 48 hours after induction of murine pneumococcal pneumonia. Bacterial outgrowth in (A) lung, (B) blood, (C) spleen and (D) liver 6, 24, and 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout (grey bars) mice. Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * indicates statistical significance as compared to wild-type (P <0.05, Mann-Whitney U test). Figure 2 Lower bacterial outgrowth in protease-activated receptor-1 knockout mice in lungs and blood at 24 hours and in spleen and liver at 48 hours after induction of murine pneumococcal pneumonia. Bacterial outgrowth in (A) lung, (B) blood, (C) spleen and (D) liver 6, 24, and 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout (grey bars) mice. Inflammatory response mortality as compared to WT mice (P <0.05). Median sur- vival time was 2 days and 21 hours in PAR-1 KO mice as compared to 2 days and 12 hours in WT mice. Moreover, at 2 days and 17 hours after infection, 64% of PAR-1 KO mice was still alive, while only 21% of WT mice had survived until that time point. To investigate the impact of PAR-1 on lung pathology, we determined histopathology scores of lung tissue slides obtained 24 and 48 hours after infection. Pneumococcal pneumonia was associated with pulmonary inflammation and damage as evidenced by the occurrence of bronchitis, interstitial inflammation, edema and endothelialitis. Mean histopathological scores were lower in PAR-1 KO mice at both 24 and 48 hours after infection (Figure 3A to 3C). To obtain insight in the role of PAR-1 in neutro- phil recruitment to the primary site of infection, we per- formed Ly-6G staining on lung sections at 24 and 48 hours after infection. While there were no significant differences at 24 hours after infection, PAR-1 KO mice showed significantly lower neutrophil numbers in lung tissue later on, as evidenced by lower Ly-6G positivity at 48 hours after infection (Figure 4A to 4C). Histology and immunohistochemistry The right lung was fixed in 10% formalin/PBS for 24 hours and embedded in paraffin. Sections of 5 µm were cut, stained with hematoxylin and eosin (H & E) and analyzed by a pathologist who was blinded for groups as described [14]. To score lung inflammation and damage, the entire section was analyzed with respect to the following para- meters: bronchitis, interstitial inflammation, edema, endothelialitis, pleuritis and thrombus formation. Each parameter was graded on a scale of 0 to 4. The total histo- pathological score was expressed as the sum of the scores. Granulocyte staining was performed using fluorescein isothiocyanate-labeled anti-mouse Ly-6G monoclonal antibody (mAb) (Pharmingen, San Diego, CA, USA) as Figure 1 Protease-activated receptor-1 accelerates mortality in murine pneumococcal pneumonia. Survival of wild-type (dashed line, open symbols) and protease-activated receptor-1 knockout (solid line, closed symbols) mice in murine pneumococcal pneumonia (14 mice per group). * indicates statistical significance as compared to wild-type (P <0.05, log rank test). Figure 1 Protease-activated receptor-1 accelerates mortality in murine pneumococcal pneumonia. Survival of wild-type (dashed line, open symbols) and protease-activated receptor-1 knockout (solid line, closed symbols) mice in murine pneumococcal pneumonia (14 mice per group). * indicates statistical significance as compared to wild-type (P <0.05, log rank test). Page 3 of 8 Page 3 of 8 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Inflammatory response mortality as compared to WT mice (P <0.05). Median sur- vival time was 2 days and 21 hours in PAR-1 KO mice as compared to 2 days and 12 hours in WT mice. Moreover, at 2 days and 17 hours after infection, 64% of PAR-1 KO mice was still alive, while only 21% of WT mice had survived until that time point. Bacterial outgrowth Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * indicates statistical significance as compared to wild-type (P <0.05, Mann-Whitney U test). Page 4 of 8 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 A. B. 15 ore C. 0 5 10 * ** 48h 24h lung histopathology sco Figure 3 Lower histopathology scores in protease-activated receptor-1 knockout mice in murine pneumococcal pneumonia. Representative microphotographs of hematoxylin and eosin stained lung sections, 48 hours after induction of pneumococcal pneumonia in (A) wild-type and (B) protease-activated receptor-1 knockout mice (100 times original magnification). (C) Total pathology scores 24 and 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout mice (grey bars). Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * and ** indicate statistical significance as compared to wild-type (P <0.05 and P <0.01 respectively, Mann-Whitney U test). B. A. B. A. A. 15 ore C. 0 5 10 * ** 48h 24h lung histopathology sco C. Figure 3 Lower histopathology scores in protease-activated receptor-1 knockout mice in murine pneumococcal pneumonia. Representative microphotographs of hematoxylin and eosin stained lung sections, 48 hours after induction of pneumococcal pneumonia in (A) wild-type and (B) protease-activated receptor-1 knockout mice (100 times original magnification). (C) Total pathology scores 24 and 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout mice (grey bars). Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * and ** indicate statistical significance as compared to wild-type (P <0.05 and P <0.01 respectively, Mann-Whitney U test). were considerably higher in PAR-1 KO mice as com- pared to WT mice (P <0.01 to P <0.05), whereas pul- monary IL-10, MCP-1 and MIP-2 concentrations did not differ between groups. IL-12 remained undetectable in lung homogenates at all time points. sepsis PARs are considered to play a pivotal role in the crosstalk between coagulation and inflammation [4]. Since data on the role of PAR-1 in severe infection are sparse and the function of PAR-1 in bacterial pneumonia and sepsis to date is unknown, we here investigated the involvement of PAR-1 in the host response to pneumo- coccal pneumonia. Bacterial outgrowth We show that PAR-1 hampers anti- bacterial defense, which is associated with more lung damage, more lung neutrophil influx and more systemic inflammation, altogether resulting in a higher mortality. To investigate the role of PAR-1 in the systemic inflammatory response, we determined levels of the above mentioned cytokines in plasma (Table 2). At 6 hours after infection, cytokine levels were below detec- tion (data not shown). At 24 hours after infection, PAR-1 KO mice had substantially lower plasma levels of TNF-a and MCP-1 (P <0.001) and a trend toward lower IL-6 concentrations (P = 0.08) when compared with WT mice. These differences had subsided at 48 hours. IL-10, IL-12 and IFN-g levels stayed below detection throughout the course of the disease (data not shown). Previous studies examined the role of PAR-1 in endo- toxemia and abdominal sepsis induced by CLP, revealing partially contradicting results [10,11,17]. Our finding that PAR-1 deficiency improves survival early in severe mur- ine pneumococcal pneumonia is in accordance with data by Niessen et al, who, using a PAR-1 antagonist, showed that functional PAR-1 reduces survival in polymicrobial sepsis induced by CLP, a finding which was associated with dendritic cell-mediated sustainment of proinflam- matory and procoagulant mechanisms [11]. These authors also showed that PAR-1 KO mice had a better survival in a 90% lethal dose (LD90) model of endotoxin- induced toxicity [11], a finding that differed from an Discussion S. pneumoniae is a major cause of morbidity and mortal- ity in humans and antibiotic resistance in this pathogen is increasing, which urges the need to study the host defense mechanisms that influence the outcome of pneu- mococcal pneumonia and sepsis [1]. In pneumonia and Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Page 5 of 8 7.5 y) VOLGT A. B. 0.0 2.5 5.0 7.5 * 48h 24h lung L6-G (% positivity C. Figure 4 Reduced pulmonary neutrophil influx in protease-activated receptor-1 knockout mice in later stage of murine pneumococcal pneumonia. Representative slides of lung Ly-6G staining (brown) 24 and 48 hours after induction of pneumococcal pneumonia in (A) wild-type and (B) protease-activated receptor-1 knockout mice (200 times original magnification). (C) Quantitation of pulmonary Ly-6G 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout mice (grey bars). Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * indicates statistical significance as compared to wild-type (P <0.05, Mann-Whitney U test). VOLGT A. B. B. A. 7.5 y) 0.0 2.5 5.0 7.5 * 48h 24h lung L6-G (% positivity C. C. Figure 4 Reduced pulmonary neutrophil influx in protease-activated receptor-1 knockout mice in later stage of murine pneumococcal pneumonia. Representative slides of lung Ly-6G staining (brown) 24 and 48 hours after induction of pneumococcal pneumonia in (A) wild-type and (B) protease-activated receptor-1 knockout mice (200 times original magnification). (C) Quantitation of pulmonary Ly-6G 48 hours after induction of pneumococcal pneumonia in wild-type (open bars) and protease-activated receptor-1 knockout mice (grey bars). Data are expressed as box-and-whisker diagrams depicting the smallest observation, lower quartile, median, upper quartile and largest observation (eight mice per group). * indicates statistical significance as compared to wild-type (P <0.05, Mann-Whitney U test). [11]. Additional studies using lower infectious doses are warranted to establish whether PAR-1 deficiency impacts on survival in less severe pneumonia. Kaneider et al reported an unaltered survival of PAR-1 KO mice relative to WT mice in CLP-induced sepsis [10]. However, they also showed that early treatment with a PAR-1 antagonist (at t = 0) did improve survival in CLP, whereas adminis- tration of a PAR-1 agonist at a later time point (t = 4 hours) also conveyed a survival benefit [10]. From their [11]. Data are medians (interquartile ranges). TNF-a, tumor necrosis factor-a; IL, interleukin; MCP-1, monocyte chemotactic protein-1; MIP-2, macrophage inflammatory protein-2. B.D., below detection determined. * and indicate statistical significance compared to wild-type (WT) (P <0.05 and P <0.01 respectively, Mann-Whitney U test). Discussion Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Page 6 of 8 Table 2 Plasma cytokine and chemokine levels in wild-type (WT) and protease-activated receptor-1 (PAR-1 KO) mice 24 and 48 hours after induction of pneumococcal pneumonia. T = 24 T = 48 WT n = 8 PAR-1 KO n = 8 WT n = 8 PAR-1 KO n = 8 TNF-a (pg/ml) 12.2 (11.4-15.4) 7.00 (4.10-9.00) 13.7 (7.80-30.6) 17.6 (14.1-22.7) IL-6 (pg/ml) 51.0 (8.58-92.9) 8.10 (2.50-49.8) 58.9 (49.8-210) 51.5 (41.2-69.7) MCP-1 (ng/ml) 129 (48.6-152) 23.9 (18.7-41.6) 75.0 (64.7-386) 151 (68.2-191) Data are medians (interquartile ranges). TNF-a, tumor necrosis factor-a; IL-6, interleukin-6; MCP-1, monocyte chemotactic protein-1. indicates statistical significance compared to wild-type (WT) (P <0.01 respectively, Mann Whitney U test). Table 2 Plasma cytokine and chemokine levels in wild-type (WT) and protease-activated receptor 24 and 48 hours after induction of pneumococcal pneumonia. plasmin, trypsin, cathepsin G, elastase, chymase, and, as mentioned, MMP-1 [6,18], and multiple cell types present in the lung express PAR-1, including macro- phages, mast cells, fibroblasts and airway smooth muscle cells [6]. Hence, the net effect of PAR-1 activation depends on the cell types and proteases present during various stages of the infection. This may also explain the partially contradictory results obtained on the role of PAR-1 in CLP-induced abdominal sepsis. Of note, however, in accordance with our current findings regarding lung pathology and neutrophil recruitment after infection with S. pneumoniae, PAR-1 was reported to participate in the acute lung inflammation elicited by intrapulmonary instil- lation of bleomycin, as reflected by reduced inflammatory cell influx in PAR-1 KO mice [22]. This [22] and other studies [23,24] have further implicated PAR-1 as a proin- flammatory receptor in acute as well as chronic lung injury. It was therefore unexpected that PAR-1 KO mice displayed higher concentrations of the proinflammatory cytokines TNF-a, IL-6 and IFN-g in lung tissue during pneumonia. We can only speculate on the mechanism: theoretically, the increase in cytokine generation in PAR-1 KO mice could be the result of diminished anti-inflamma- tory endothelial protein C receptor (EPCR-)APC effects via PAR-1 or to compensatory increases of other proin- flammatory signaling pathways such as upregulation of PAR-2 to PAR-4 outweighing the loss of PAR-1 as a proinflammatory signaling pathway. Further studies are needed to dissect the exact mechanisms and cell types at play mediating PAR-1 effects after infection by S. pneumoniae. Discussion studies these investigators concluded that PAR-1 is detri- mental in early phases of sepsis but beneficial in later phases, which could explain the absence of a net survival benefit in PAR-1 KO mice in their studies [10]. A very recent study identified matrix metalloproteinase (MMP)- 1a as a PAR-1 agonist in mice; blockade of MMP-1a activity protected against CLP-induced lethality in WT but not in PAR-1 KO mice, suggesting that MMP-1 acti- vation of PAR-1 contributes to an adverse outcome of polymicrobial abdominal sepsis [18]. Clearly, the studies on the role of PAR-1 endotoxic shock and CLP-induced sepsis are not fully consistent. We did not evaluate the effects of pharmacologic blockade of PAR-1 in pneumo- coccal pneumonia; such studies could reveal potential time-dependent effects of PAR-1 inhibition and the pos- sible impact of therapeutic PAR-1 blockade in the con- text of concurrent antibiotic treatment. The survival advantage of PAR-1 KO mice in our study corresponded with lower bacterial loads at various stages of the infection. In addition, PAR-1 KO mice displayed lower lung pathology scores and a reduced number of neutrophils in lung tissue. The mechanisms underlying these differences remain to be elucidated. Understanding the role of PAR-1 signaling in infection is difficult due to the multiple and in part opposite effects ascribed to this receptor. Indeed, although APC and thrombin can both activate PAR-1, APC affects the vascular endothelium in a way that clearly is distinct from thrombin signaling. Specifically, APC can exert anti-inflammatory, anti-apop- totic and vasculoprotective signals in endothelial cells via PAR-1, processes in which the endothelial protein C receptor plays a pivotal role [7,19], whereas thrombin induces vascular hyperpermeability via PAR-1 [20]. To make things more complex, activation of PAR-1 by low doses of thrombin can (like APC) result in a barrier pro- tective effect [8], whereas a very recent investigation pro- vided evidence that activated coagulation factor VII (FVIIa) can exert a barrier protective effect in endothelial cells via activation of PAR-1 [21]. Moreover, PAR-1 can be activated by proteases other than FVIIa, thrombin and APC, including activated coagulation factor × (FXa), Discussion Additional studies using lower infectious doses are warranted to establish whether PAR-1 deficiency impacts on survival in less severe pneumonia. Kaneider et al reported an unaltered survival of PAR-1 KO mice relative to WT mice in CLP-induced sepsis [10]. However, they also showed that early treatment with a PAR-1 antagonist (at t = 0) did improve survival in CLP, whereas adminis- tration of a PAR-1 agonist at a later time point (t = 4 hours) also conveyed a survival benefit [10]. From their earlier study demonstrating an unaltered mortality of PAR-1 KO mice after a high-dose endotoxin challenge [17]. In contrast to the studies performed by Niessen and colleagues [11], the survival benefit of PAR-1 KO mice in our study was only temporary. This does not necessarily mean there is no effect of PAR-1 deficiency in later stages of the disease but could be related to the fact that our model of severe pneumococcal pneumonia is an LD100 model as opposed to the models used by Niessen et al Table 1 Pulmonary cytokine and chemokine levels in wild-type (WT) and protease-activated receptor-1 (PAR-1 KO) mice 6, 24 and 48 hours after induction of pneumococcal pneumonia. T = 6 T = 24 T = 48 WT n = 8 PAR-1 KO n = 8 WT n = 8 PAR-1 KO n = 8 WT n = 8 PAR-1 KO n = 8 TNF-a (pg/ml) 4.45 (3.28-7.10) 5.70 (5.20-7.20) 10.5 (1.30-24.0) 13.0 (3.50-56.1) 27.3 (15.9-40.6) 175 (43.7-215)** IL-6 (pg/ml) 7.95 (2.50-11.3) 11.6 (10.0-14.7) 227 (23.0-1418) 366 (15.5-1460) 217 (147-382) 770 (397-1642)* IL-10 (pg/ml) B.D. B.D. 44.3 (34.2-66.6) 34.7 (30.5-39.3) 16.9 (10.0-24.8) 25.1 (16.3-169) IFN-g (pg/ml) B.D. B.D. 11.7 (9.30-17.4) 13.7 (3.10-34.0) 4.80 (2.90-8.85) 16.7 (11.3-23.2)* MCP-1 (ng/ml) 0.15 (0.13-0.29) 0.19 (0.12-0.28) 8.38 (4.30-9.93) 5.28 (2.48-7.28) 3.29 (2.88-6.34) 5.69 (2.68-10.1) MIP-2 (ng/ml) B.D. B.D. 6.89 (3.81-15.9) 12.6 (0.20-37.6) 1.95 (0.77-17.3) 1.14 (1.01-5.05) Data are medians (interquartile ranges). TNF-a, tumor necrosis factor-a; IL, interleukin; MCP-1, monocyte chemotactic protein-1; MIP-2, macrophage inflammatory protein-2. B.D., below detection determined. * and ** indicate statistical significance compared to wild-type (WT) (P <0.05 and P <0.01 respectively, Mann-Whitney U test). y cytokine and chemokine levels in wild-type (WT) and protease-activated receptor-1 (PAR-1 KO) hours after induction of pneumococcal pneumonia. ble 1 Pulmonary cytokine and chemokine levels in wild-type (WT) and protease-activated recepto ce 6, 24 and 48 hours after induction of pneumococcal pneumonia. Schouten et al. Abbreviations APC acti ated p 8. Bae JS, Kim YU, Park MK, Rezaie AR: Concentration dependent dual effect of thrombin in endothelial cells via Par-1 and Pi3 kinase. J Cell Physiol 2009, 219:744-751. APC: activated protein C; CAP: community-acquired pneumonia; CBA: cytometric bead array; CFU: colony forming units; CLP: cecal ligation and puncture; EPCR: endothelial protein C receptor; ELISA: enzyme-linked immunosorbent assay; FVIIa: activated coagulation factor VII; FXa: activated coagulation factor X; H & E: hematoxylin & eosin; IFN: interferon; IL: interleukin; KO: knockout; LD90: 90% lethal dose; mAb: monoclonal antibody; MCP: monocyte chemoattractant protein; MIP: macrophage inflammatory protein; MMP: matrix metalloproteinase; PAR: protease-activated receptor; PBS: phosphate-buffered saline; S. pneumonia: Streptococcus pneumonia; TNF: tumor necrosis factor; WT: wild type. 9. Ludeman MJ, Kataoka H, Srinivasan Y, Esmon NL, Esmon CT, Coughlin SR: PAR1 cleavage and signaling in response to activated protein C and thrombin. J Biol Chem 2005, 280:13122-13128. 9. Ludeman MJ, Kataoka H, Srinivasan Y, Esmon NL, Esmon CT, Coughlin SR: PAR1 cleavage and signaling in response to activated protein C and thrombin. J Biol Chem 2005, 280:13122-13128. 10. Kaneider NC, Leger AJ, Agarwal A, Nguyen N, Perides G, Derian C, Covic L, Kuliopulos A: ’Role reversal’ for the receptor PAR1 in sepsis-induced vascular damage. Nat Immunol 2007, 8:1303-1312. 10. Kaneider NC, Leger AJ, Agarwal A, Nguyen N, Perides G, Derian C, Covic L, Kuliopulos A: ’Role reversal’ for the receptor PAR1 in sepsis-induced vascular damage. Nat Immunol 2007, 8:1303-1312. PBS: phosphate-buffered saline; S. pneumonia: Streptococcus pneumonia; TNF: tumor necrosis factor; WT: wild type. 11. Niessen F, Schaffner F, Furlan-Freguia C, Pawlinski R, Bhattacharjee G, Chun J, Derian CK, Andrade-Gordon P, Rosen H, Ruf W: Dendritic cell PAR1-S1P3 signalling couples coagulation and inflammation. Nature 2008, 452:654-658. References 1. Niederman MS, Mandell LA, Anzueto A, Bass JB, Broughton WA, Campbell GD, Dean N, File T, Fine MJ, Gross PA, Martinez F, Marrie TJ, Plouffe JF, Ramirez J, Sarosi GA, Torres A, Wilson R, Yu VL: Guidelines for the management of adults with community-acquired pneumonia - diagnosis, assessment of severity, antimicrobial therapy, and prevention. Am J Respir Crit Care Med 2001, 163:1730-1754. Acknowledgements The authors thank Marieke ten Brink and Joost Daalhuisen for their technical assistance during the animal experiments, Regina de Beer for performing histopathological and immunohistochemical stainings and Daan de Boer and Arie Hoogendijk for performing Ly-6G quantitation. The authors thank the Dutch Thrombosis Foundation for supporting this study. 16. Schouten M, van der Sluijs KF, Gerlitz B, Grinnell BW, Roelofs JJ, Levi MM, van ‘t Veer C, van der Poll T: Activated protein C ameliorates coagulopathy but does not influence outcome in lethal H1N1 influenza: a controlled laboratory study. Crit Care 2010, 14:R65. 17. Pawlinski R, Pedersen B, Schabbauer G, Tencati M, Holscher T, Boisvert W, Andrade-Gordon P, Frank RD, Mackman N: Role of tissue factor and protease-activated receptors in a mouse model of endotoxemia. Blood 2004, 103:1342-1347. The authors declare that they have no competing interests. The authors declare that they have no competing interests. 15. Rijneveld AW, Levi M, Florquin S, Speelman P, Carmeliet P, van der Poll T: Urokinase receptor is necessary for adequate host defense against pneumococcal pneumonia. J Immunol 2002, 168:3507-3511. Conclusions We show that in pneumococcal pneumonia, PAR-1 impairs the host defense response, as reflected by a reduced lethality, lower bacterial loads, lower lung histo- pathology scores and less pulmonary neutrophil influx in PAR-1 KO mice. Considering the complex role of PAR-1 in infection, related to the capacity of multiple proteases to activate PAR-1 resulting in differential cellular effects and the multiple cell types expressing PAR-1, this receptor Page 7 of 8 Page 7 of 8 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 at this moment does not represent a straightforward thera- peutic target in severe pneumonia and sepsis. Key messages • Protease activated receptor (PAR)-1 knock out (KO) mice have an improved survival as compared to wild-type (WT) mice in pneumococcal pneumonia. Am J Respir Crit Care Med 2001, 163:1730-1754. 2. Laterre PF, Garber G, Levy H, Wunderink R, Kinasewitz GT, Sollet JP, Maki DG, Bates B, Yan SCB, Dhainaut JF: Severe community-acquired pneumonia as a cause of severe sepsis: data from the PROWESS study. Crit Care Med 2005, 33:952-961. • PAR-1 KO mice have lower bacterial loads in lungs and blood at 24 hours and in spleen and liver at 48 hours after induction of pneumococcal pneumonia as com- pared to WT mice. 3. Bartlett JG, Dowell SF, Mandell LA, File TM Jr, Musher DM, Fine MJ: Practice guidelines for the management of community-acquired pneumonia in adults. Infectious Diseases Society of America. Clin Infect Dis 2000, 31:347-382. 4. Ossovskaya VS, Bunnett NW: Protease-activated receptors: contribution to physiology and disease. Physiological Reviews 2004, 84:579-621. • The favorable response in PAR-1 KO mice with regard to survival and bacterial outgrowth is accompanied by lower histopathology scores and less neutrophil influx in the lungs. 5. Sokolova E, Reiser G: A novel therapeutic target in various lung diseases: airway proteases and protease-activated receptors. Pharmacol Ther 2007, 115:70-83. 6. Peters T, Henry PJ: Protease-activated receptors and prostaglandins in inflammatory lung disease. Br J Pharmacol 2009, 158:1017-1033. 6. Peters T, Henry PJ: Protease-activated receptors and prostaglandins in inflammatory lung disease. Br J Pharmacol 2009, 158:1017-1033. g • Taken together, this study shows that PAR-1 hampers host defense in murine pneumococcal pneumonia. 7. Bae JS, Yang L, Manithody C, Rezaie AR: The ligand occupancy of endothelial protein C receptor switches the PAR-1-dependent signaling specificity of thrombin from a permeability-enhancing to a barrier- protective response in endothelial cells. Blood 2007, 110:3909-3916. Authors’ contributions MS participated in the design of the study, carried out the in vivo experiments and drafted the manuscript. CV participated in the design of the study, advised in laboratory matters and helped draft the manuscript. JJTHR performed pathology scoring, prepared part of the figures and helped draft the manuscript. ML helped to draft the manuscript and supervised the study. TP participated in the design of the study, supervised the study and helped draft the manuscript. All authors read and approved the final manuscript. MS participated in the design of the study, carried out the in vivo experiments and drafted the manuscript. CV participated in the design of the study, advised in laboratory matters and helped draft the manuscript. JJTHR performed pathology scoring, prepared part of the figures and helped draft the manuscript. ML helped to draft the manuscript and supervised the study. TP participated in the design of the study, supervised the study and 12. Connolly AJ, Ishihara H, Kahn ML, Farese RV Jr, Coughlin SR: Role of the thrombin receptor in development and evidence for a second receptor. Nature 1996, 381:516-519. 12. Connolly AJ, Ishihara H, Kahn ML, Farese RV Jr, Coughlin SR: Role of the thrombin receptor in development and evidence for a second receptor. Nature 1996, 381:516-519. 13. Rijneveld AW, Florquin S, Bresser P, Levi M, De Waard V, Lijnen R, van der Zee JS, Speelman P, Carmeliet P, van der Poll T: Plasminogen activator inhibitor type-1 deficiency does not influence the outcome of murine pneumococcal pneumonia. Blood 2003, 102:934-939. study. TP participated in the design of the study, supervised the study and helped draft the manuscript. All authors read and approved the final manuscript. 14. Schouten M, van ‘t Veer C, Roelofs JJ, Levi M, van der Poll T: Impact of the factor V Leiden mutation on the outcome of pneumococcal pneumonia: a controlled laboratory study. Crit Care 2010, 14:R145. 21. Sen P, Gopalakrishnan R, Kothari H, Keshava S, Clark CA, Esmon CT, Pendurthi UR, Rao LV: Factor VIIa bound to endothelial cell protein C doi:10.1186/cc11910 Cite this article as: Schouten et al.: Protease-activated receptor-1 impairs host defense in murine pneumococcal pneumonia: a controlled laboratory study. Critical Care 2012 16:R238. Author details 1 1Center for Experimental and Molecular Medicine (CEMM), Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. 2Center for Infection and Immunity Amsterdam (CINIMA), Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. 3Department of Pathology, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. 4Department of Internal Medicine, Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. 5Division of Infectious Diseases; Academic Medical Center, University of Amsterdam, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. 18. Tressel SL, Kaneider NC, Kasuda S, Foley C, Koukos G, Austin K, Agarwal A, Covic L, Opal SM, Kuliopulos A: A matrix metalloprotease-PAR1 system regulates vascular integrity, systemic inflammation and death in sepsis. EMBO Mol Med 2011, 3:370-384. 19. Danese S, Vetrano S, Zhang L, Ploplis VA, Castellino FJ: The protein C pathway in tissue inflammation and injury: pathogenic role and therapeutic implications. Blood 2010, 115:1121-1130. 20. Singleton PA, Moreno-Vinasco L, Sammani S, Wanderling SL, Moss J, Garcia JG: Attenuation of vascular permeability by methylnaltrexone: role of MOP-R and S1P3 transactivation. Am J Respir Cell Mol Biol 2007, 37:222-231. 21. Sen P, Gopalakrishnan R, Kothari H, Keshava S, Clark CA, Esmon CT, Pendurthi UR, Rao LV: Factor VIIa bound to endothelial cell protein C 21. Sen P, Gopalakrishnan R, Kothari H, Keshava S, Clark CA, Esmon CT, Pendurthi UR, Rao LV: Factor VIIa bound to endothelial cell protein C Received: 8 July 2012 Revised: 26 November 2012 Accepted: 14 December 2012 Published: 27 December 2012 Received: 8 July 2012 Revised: 26 November 2012 Accepted: 14 December 2012 Published: 27 December 2012 Page 8 of 8 Page 8 of 8 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 receptor activates protease activated receptor-1 and mediates cell signaling and barrier protection. Blood 2011, 117:3199-3208. 22. Howell DC, Johns RH, Lasky JA, Shan B, Scotton CJ, Laurent GJ, Chambers RC: Absence of proteinase-activated receptor-1 signaling affords protection from bleomycin-induced lung inflammation and fibrosis. Am J Pathol 2005, 166:1353-1365. 23. Jenkins RG, Su X, Su G, Scotton CJ, Camerer E, Laurent GJ, Davis GE, Chambers RC, Matthay MA, Sheppard D: Ligation of protease-activated receptor 1 enhances alpha(v)beta6 integrin-dependent TGF-beta activation and promotes acute lung injury. J Clin Invest 2006, 116:1606-1614. 24. Scotton CJ, Krupiczojc MA, Konigshoff M, Mercer PF, Lee YC, Kaminski N, Morser J, Post JM, Maher TM, Nicholson AG, Moffatt JD, Laurent GJ, Derian CK, Eickelberg O, Chambers RC: Increased local expression of coagulation factor × contributes to the fibrotic response in human and murine lung injury. J Clin Invest 2009, 119:2550-2563. doi:10.1186/cc11910 Cite this article as: Schouten et al.: Protease-activated receptor-1 impairs host defense in murine pneumococcal pneumonia: a controlled laboratory study. Critical Care 2012 16:R238. receptor activates protease activated receptor-1 and mediates cell signaling and barrier protection. Blood 2011, 117:3199-3208. 22. Howell DC, Johns RH, Lasky JA, Shan B, Scotton CJ, Laurent GJ, Chambers RC: Absence of proteinase-activated receptor-1 signaling affords protection from bleomycin-induced lung inflammation and fibrosis. Am J Pathol 2005, 166:1353-1365. 23. Jenkins RG, Su X, Su G, Scotton CJ, Camerer E, Laurent GJ, Davis GE, Chambers RC, Matthay MA, Sheppard D: Ligation of protease-activated receptor 1 enhances alpha(v)beta6 integrin-dependent TGF-beta activation and promotes acute lung injury. J Clin Invest 2006, 116:1606-1614. 24. Scotton CJ, Krupiczojc MA, Konigshoff M, Mercer PF, Lee YC, Kaminski N, Morser J, Post JM, Maher TM, Nicholson AG, Moffatt JD, Laurent GJ, Derian CK, Eickelberg O, Chambers RC: Increased local expression of coagulation factor × contributes to the fibrotic response in human and murine lung injury. J Clin Invest 2009, 119:2550-2563. doi:10.1186/cc11910 Cite this article as: Schouten et al.: Protease-activated receptor-1 impairs host defense in murine pneumococcal pneumonia: a controlled laboratory study. Critical Care 2012 16:R238. receptor activates protease activated receptor-1 and mediates cell signaling and barrier protection. Blood 2011, 117:3199-3208. g g p 22. Howell DC, Johns RH, Lasky JA, Shan B, Scotton CJ, Laurent GJ, Chambers RC: Absence of proteinase-activated receptor-1 signaling affords protection from bleomycin-induced lung inflammation and fibrosis. Am J Pathol 2005, 166:1353-1365. 23. Schouten et al. Critical Care 2012, 16:R238 http://ccforum.com/content/16/6/R238 Jenkins RG, Su X, Su G, Scotton CJ, Camerer E, Laurent GJ, Davis GE, Chambers RC, Matthay MA, Sheppard D: Ligation of protease-activated receptor 1 enhances alpha(v)beta6 integrin-dependent TGF-beta activation and promotes acute lung injury. J Clin Invest 2006, 116:1606-1614. 24. Scotton CJ, Krupiczojc MA, Konigshoff M, Mercer PF, Lee YC, Kaminski N, Morser J, Post JM, Maher TM, Nicholson AG, Moffatt JD, Laurent GJ, Derian CK, Eickelberg O, Chambers RC: Increased local expression of coagulation factor × contributes to the fibrotic response in human and murine lung injury. J Clin Invest 2009, 119:2550-2563. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
W2604586237.txt	https://rbeo.emnuvens.com.br/rbeo/article/download/74/pdf_1	pt		Cidades Artificiais: Um Estudo Exploratório do Projeto Urbano Porto Alegre 4D.	Revista Brasileira de Estudos Organizacionais	2,017	cc-by	8,973	REVISTA BRASILEIRA DE ESTUDOS ORGANIZACIONAIS SOCIEDADE BRASILEIRA DE ESTUDOS ORGANIZACIONAIS CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira 1 Resumo Uma cidade artificial é uma cidade projetada, criada do zero. Projetos contemporâneos de renovação urbana podem ser considerados projetos de cidade artificial pois o foco do projeto não está na população residente, ao contrário, desencadeiam processos de gentrificação e de atração de população externa. O objetivo deste artigo é, portanto, mostrar o caráter prático operacional dos projetos urbanos desta natureza descrevendo o Porto Alegre 4D. Se trata de uma política de organização espacial de uma parte da cidade. O argumento enfatiza como a vida da população do quarto distrito é desvalorizada. O projeto não é feito para aquele lugar e menos, ainda, para as pessoas que lá residem, mas sob a aparência de um desenho técnico beneficia grupos e setores econômicos. Palavras-chave: Projetos urbanos. Cidades artificiais. Gentrificação. Produção do espaço. ARTIFICIAL CITIES: AN EXPLORATORY STUDY OF THE URBAN PROJECT PORTO ALEGRE 4D Abstract An artificial city is a designed city, created from scratch. Contemporary projects of urban renewal can be considered an artificial city project because its focus is not on the resident population, on the contrary, they trigger processes of gentrification and attraction of external population. The objective of this article is, therefore, to show the operational-practical nature of urban projects of this kind, describing the Porto Alegre 4D project. It is a policy of spatial organization of the one part of the city. The 1 Mestre e Doutoranda em Planajamento Urbano e Regional – UFRGS. E.mail: arq.clarice@gmail.com Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira argument the ways in which the life of the population of the fourth district is devalued. The project is not for this place and, even less, for the people who live there, but under the appearance of a technical design, for the benefit of economic sectors and groups. Keywords: Urban Projects. Artificial Cities. Gentrification. Production Of Space. TÍTULO EM ESPANHOL Ciudades artificiales: un estudio exploratorio del Proyecto Urbano Porto Alegre 4D Resumen Una ciudad artificial es una ciudad diseñada, creada a partir de cero. Proyectos de renovación urbana contemporánea pueden ser consideradas como este tipo de proyecto con el foco en los procesos de gentrificación de disparo y la atracción de población extranjera. Por lo tanto, el propósito de este artículo es mostrar el carácter práctico-operativo de los proyectos urbanos de esta naturaleza, describiendo el proyecto Porto Alegre 4D. Se trata de una política de organización espacial para una parte de la ciudad. El argumento enfatiza cómo la vida de la populación del cuarto distrito es devaluada. El proyecto no es para aquel lugar y menos aún, para la gente que la reside, pero bajo la apariencia de un diseño técnico beneficia grupos y sectores económicos. Palabras-clave: Proyectos Urbanos. Ciudades Artificiales. Gentrificación. Producción Del Espacio. Introdução Uma cidade artificial2 é uma cidade projetada, criada do zero, que após concluída pode receber moradores que não tem origem naquele lugar. Uma pesquisa rápida no Google mostra uma profusão de cidades 2 Alexander (1966), defendendo o papel do planejamento urbano, faz a seguinte distinção: cidades que surgem espontaneamente e se desenvolvem ao longo dos anos são cidades naturais; cidades ou partes de cidades que são intencionalmente criadas por planejadores, são cidades artificiais. Mas, como diz o título do ensaio, cidades não são árvores. Esse é um claro exemplo do que Lefebvre (2008b) chama de ‘senhores do espaço’. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 78 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira artificiais futuristas, baseadas na alta tecnologia para recriar o meio ambiente artificial em um grande edifício apresentado como uma cidade autossustentável, flutuante, e que, supostamente, não causaria danos ao ecossistema. Seguindo a lógica de uma cidade criada do zero, Brasília a cidade modernista projetada pelos arquitetos Lúcio Costa e Oscar Niemeyer também pode ser chamada de uma cidade artificial. Porém, este artigo não trata nem de uma cidade do passado e nem de uma cidade do futuro, trata do presente, momento em que se presencia a tendência dos projetos de renovação urbana como “o caminho dourado para a sobrevivência urbana” (HARVEY, 2006, p. 175), baseado em atividades de consumo e acumulação capitalista. O objetivo deste artigo é mostrar o caráter-prático operacional dos projetos urbanos desta natureza. Para tanto, faz um estudo ainda exploratório do Masterplan Porto Alegre 4D, um projeto urbano concebido pelo setor privado em parceria com a Prefeitura Municipal de Porto Alegre (PMPA) e desenvolvido por uma equipe multidisciplinar da Universidade Federal do Rio Grande do Sul (UFRGS). Esta primeira exploração se baseou em entrevistas com envolvidos na elaboração do projeto, análise de documentos e artigos de revistas e jornais, e observação participante em eventos de apresentação e discussão do projeto. A escolha teórico-metodológica baseada na obra de Henri Lefebvre se justifica pois vivemos o fenômeno contemporâneo da “hipermercantilização da vida urbana” (BRENNER, MARCUSE e MAYER, 2012, p. 2). Fenômeno apontado, inicialmente como a mercantilização do espaço urbano, a partir dos anos 1960 e 1970 por Lefebvre (2001, 2008a), Castells (1977) e Harvey (2008). As formulações de Lefebvre estão subjacentes ou, mesmo, formalmente referidas por alguns dos autores mencionados. Além disso, constata-se que há uma tendência, no campo dos estudos críticos urbanos e dos estudos organizacionais, em recorrer à obra desse autor. Acredita-se que o retorno de Lefebvre decorre do avanço do sistema do capital, em busca de novos espaços de acumulação, sobre o espaço urbano. Portanto, se faz necessário aprofundar o estudo da obra de Lefebvre e verificar sua adequação, com as devidas mediações para analisar fenômenos contemporâneos. Nas partes que seguem serão desenvolvidos: (1) uma revisão dos conceitos principais da obra de Lefebvre acerca do espaço urbano e da produção do espaço; (2) descrição do projeto Porto Alegre 4D; (3) considerações finais com insights a partir da tríalética do espaço. É preciso reconhecer que os dois primeiros itens deste artigo estão muito separados. A intenção, até porque se trata de um estudo Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 79 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira exploratório e de um projeto urbano em fase inicial, não é aplicar o referencial lefebvriano para análise das informações sistematizadas na sequência. Sua presença serve, nesta versão do estudo, para que o leitor entenda o olhar que é dado ao projeto urbano, porque sua descrição tem o conteúdo que tem, e porque algumas interpretações estão ali colocadas. Ainda assim, espera-se contribuir na direção de uma maior aproximação entre os campos do planejamento urbano e dos estudos organizacionais, já que o tema da organização da cidade e da produção de políticas públicas referidas especificamente ao urbano se beneficia muito desta aproximação interdisciplinar. A produção do espaço urbano Para Lefebvre (2008b), o espaço não é produto, não é objeto, não é mercadoria, não é simplesmente um instrumento. Para o autor, a problemática do espaço vivido, em oposição ao espaço “neutro”, puramente epistemológico, é um aspecto essencial do conhecimento da realidade urbana. Em relação a essa questão, Lefebvre (2008b) explora diversas hipóteses. Na primeira, “o espaço é a forma pura”, conceito que exclui a ideologia, a interpretação e o não saber. O espaço seria, nessa ótica, capaz de articular “o social e o mental, o teórico e o prático, o ideal e o real” (LEFEBVRE, 2008b, p. 41). Entretanto, tal hipótese implica a liquidação do tempo histórico, além de pesar sobre a cientificidade abstrata e o saber absoluto. Ela tem profunda relação com o urbanismo modernista e sua crítica: momento de valorização absoluta do saber científico, em que arquitetos e urbanistas acreditavam mudar o social através da forma. Lefebvre os chamou de “senhores do espaço”. Ou seja, a forma concebida pelo técnico seria, depois, povoada pelos usuários que se comportariam de acordo com o recipiente em que estivessem envoltos. Nesse sentido, essa noção exclui o espaço histórico e vivido. Em outra hipótese, o espaço é social, um produto da sociedade, ele resulta do trabalho e da divisão do trabalho sendo, pois, o lugar geral dos objetos produzidos. O espaço é, então, uma objetivação do social. Confrontando tais formulações, Lefebvre (2008b, p. 44) conclui que o espaço não é nem ponto de partida, nem ponto de chegada, “mas um intermediário em todos os sentidos desse termo, ou seja, um modo e um instrumento, um meio e uma mediação”. O espaço é um instrumento político intencionalmente manipulado por um indivíduo ou um coletivo que detém poder sobre ele, e que o transforma conforme seus objetivos. “Nessa hipótese, a representação do espaço sempre serviria a uma estratégia, sendo ao mesmo tempo abstrata e concreta, pensada e desejada, isto é, projetada” (LEFEBVRE, 2008b, p. 44). O espaço é, então, carregado de ideologia e saber técnico. As representações do espaço servem, nesse sentido, para criar consenso e controlar as contradições inerentes ao modo Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 80 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira de produção capitalista (LEFEBVRE, 1991). Assim, o saber técnico legitimado para produzir a representação do espaço se transforma no poder de transformar ideias e sentimentos, ou seja, no poder de criar o consenso acerca do espaço. Essa dominação racional-funcional vincula-se à reprodução da força de trabalho pelo consumo, sendo o espaço “o meio e o modo, ao mesmo tempo, de organização do consumo na sociedade neocapitalista” (LEFEBVRE, 2008b, p. 46). Para Lefebvre (2008b, p. 46), “[...] as cidades são tão somente unidades de consumo correlatas às grandes unidades de produção”. A essa formulação se incorpora a teoria da alienação: o espaço instrumental se torna abstrato a partir de uma falsa representação de vocação estratégica, passando a ser “um lugar fora do tempo, da vida e da práxis”. O espaço assume a forma de mercadoria, em que, “no limite [,] não há mais exatamente ideologia, mas somente a falsa consciência, com os discursos que ela engendra”. Entretanto, o autor alerta que essa hipótese servia ao capitalismo concorrencial do século XIX, cujo enfoque recaia na reprodução dos meios de produção. Ao se renovar, o capitalismo incorporou a “reprodução das relações de produção” que se efetuam através da cotidianidade e dos momentos de lazer e cultura. Essa noção é complementar à última hipótese apresentada, mas leva a outra, mais complexa: “o espaço estaria essencialmente ligado à reprodução das relações (sociais) de produção” (LEFEBVRE, 2008b, p. 48). É o espaço da produção, ao mesmo tempo abstrato-concreto, homogêneo e desarticulado: Essa reprodução se realiza através de um esquema relativo à sociedade existente, cujo caráter essencial é ser conjunta-disjunta, dissociada, mantendo uma unidade, a do poder, na fragmentação. Este espaço homogêneofraturado não é somente o espaço global do planejamento ou o espaço parcelar do arquiteto e dos promotores imobiliários, é também o espaço das obras de arte, por exemplo, o do mobiliário e do design. É o esteticismo que unifica os fragmentos funcionais de um espaço deslocado realizando, assim, seu caráter homogêneo e fraturado (LEFEBVRE, 2008b, p. 49). Retoma-se, nesse ponto, a já citada noção de Harvey (2011) sobre a sedução pelo embelezamento e os símbolos de distinção social. Espaços homogêneos, com uma estética específica, criados em recortes, em partes fragmentadas no tecido urbano, tornam-se centros ou subcentros de consumo, polos de atração de uma determinada classe social, causando estranheza e constrangimento às demais classes e interferindo diretamente no poder do espaço, assim como na dinâmica da estrutura urbana. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 81 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Ao propor o método que possibilita uma análise da produção do espaço, o autor apontou para a necessidade de esclarecimento destes dois termos. A noção de espaço já foi aqui esclarecida (ele é um produto da sociedade que resulta do trabalho e da divisão do trabalho). Sobre o conceito de produção, Lefebvre (1991) segue as indicações de Marx na associação de dois significados inseparáveis. Em um deles, os homens produzem a própria vida, sua própria consciência e seu próprio mundo ao transformar a natureza. No outro, a produção está relacionada diretamente ao trabalho e a produção de um objeto, um produto. Esta abordagem relaciona a produção diretamente com a vida cotidiana. Assim produção, produto e trabalho são abstrações concretas que tornam possível as relações de produção. Mas, Lefebvre (1991) quer ir além, defendendo que é preciso definir de maneira mais rigorosa as relações dialéticas entre produção e produto, trabalho e produto, e natureza e produção. Logo, o autor esclarece: o trabalho é insubstituível, sendo um produto é algo que pode ser reproduzido tal e qual, repetidas vezes; a natureza cria, não produz, por isso, um espaço natural intocado não é o espaço produzido. Em síntese, a humanidade cria trabalho e produz objetos e, nesse processo, os seres humanos se produzem. Portanto, para Lefebvre (1991) se o conceito de produção transcende a oposição filosófica entre objeto e sujeito, como poderia a racionalidade inerente à produção ser definida? O próprio Lefebvre (1991, p. 69) responde: a produção “organiza uma sequência de ações com certa objetividade em vista (por exemplo, um objeto a ser produzido). Esta condição impõe uma dimensão temporal e outra espacial sobre as relações pelas quais os resultados são coextensivos”. Portanto, para Lefebvre (1991), no momento em que ocorre uma atividade orientada para certo objetivo, se estabelece a relação tridiática entre os elementos espaciais (o corpo), os materiais (pedra, madeira etc.) e os dispositivos (linguagem, ferramentas, agendas etc.). Esta relação é o fundamento do desenvolvimento denominado de tríade da produção do espaço. Nela, se relacionam dialeticamente: o espaço percebido (a prática do espaço); o concebido (a representação do espaço); e o vivido (os espaços de representação) - sendo imprescindível que os três sejam analisados articuladamente. Lefebvre (1991) ressalta a importância dos espaços públicos, aqueles sem o consumo organizado e direcionado, onde todos têm o direito de ir e vir, estar e permanecer, sem constrangimentos, onde tal triplicidade pode ser percebida sem interferências. Em A revolução urbana, Lefebvre (2008a) já esboçava o que chamou mais adiante de espaçologia - uma teoria unitária do espaço: [...] o conhecimento do fenômeno urbano só pode se constituir como ciência na e pela formação consciente de uma práxis urbana suplantando, com sua racionalidade própria, a práxis industrial atualmente realizada. É nesse Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 82 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira processo complexo que a análise recorta ‘objetos’, ou constrói ‘modelos’, todos provisórios, passíveis de revisão, de crítica. O que supõe, em primeiro lugar, a mencionada confrontação entre ideologia urbanística e a prática urbana dos grupos e classes sociais; em segundo lugar, a intervenção de forças sociais e políticas; em terceiro lugar a liberação das capacidades de invenção, sem excluir o utopismo que mais se aproxime do imaginário puro. (LEFEBVRE, 2008a, p. 125) A prática espacial (o espaço percebido) é a maneira dialética como sociedade e espaço se relacionam, por um lado a sociedade propõe o espaço, e por outro, o pressupõe, ou seja, o produz paulatinamente na medida em que o compreende e apropria. Lefebvre (1991, p. 38) questiona como seria a prática social na sociedade neocapitalista3. E responde: “a prática espacial é composta por uma associação muito próxima entre o espaço percebido, entre a realidade do dia-a-dia e a realidade urbana (as rotas e redes que conectam os lugares fora do trabalho, do lazer e da vida privada)”. A representação do espaço (o espaço concebido) é o espaço dos arquitetos, urbanistas, planejadores, engenheiros, ou de todos aqueles “que identificam o que é vivido e o que é percebido para o que é concebido” (LEFEBVRE, 1991, p. 38). É o espaço da tecnocracia, aonde se estabelecem os regimes discursivos. O espaço representacional (o espaço vivido) é o espaço como ele pode ser, plenamente vivido, momentos do presente: “um espaço que é dominado e por vezes experimentado passivamente, espaços em que a imaginação tende a transformar e se apropriar” (Lefebvre, 1991, p. 39). O espaço de representação é qualitativo, fluido e dinâmico, pois é um espaço vivo; é o espaço dos símbolos, das imagens. A produção do espaço inclui, portanto, a produção material, a produção de conhecimento, e a produção de significado. Essas dimensões formam uma trialética unitária e contraditória, como explica Schmid (2012). Para estabelecer as relações entre o percebido, o vivido e o concebido, algumas categorias de Lefebvre (1991, p.193) acerca da arquitetura do espaço são fundamentais. A primeira se relaciona ao espaço percebido e diz respeito a uma superimposição de certas relações sociais à estrutura e à forma urbana: (1) espaços acessíveis de uso comum – rotas ou caminhos, ruas, avenidas, em que o uso é pré-estabelecido seguindo certas regras de convívio; (2) limites e territórios proibidos – espaços nos quais o acesso é proibido relativamente (por vizinhos e amigos) ou 3 A expressão sociedade capitalista é usada por Lefebvre para referir-se ao que ficou conhecido, a partir da escola da regulação francesa como pós-fordismo. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 83 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira absolutamente (por vizinhos e inimigos); (3) locais de residência – permanentes ou temporários; (4) pontos de junção – lugares de passagem ou encontro, geralmente têm o acesso proibido com exceção de ocasiões especiais – são momentos de declaração de paz ou guerra, por exemplo. A segunda se relaciona ao espaço concebido e diz respeito ao espaço abstrato, Lefebvre (1991, p. 286-287) propõe a consideração de três formatos: (1) geométrico – tratado como ‘absoluto’, o espaço de referência, é o espaço euclidiano que se define por sua isotopia; (2) óptico ou visual – o que é meramente visto, reduzido a uma imagem que torna difícil ver, mas que “é falado mais e mais eloquentemente e escrito mais e mais copiosamente”; (3) fálico - é completamente preenchido com imagens ou objetos transicionais que simbolizam a força, a violência masculina, a brutalidade do poder político e seus meios (polícia, exército, burocracia). Mas, diz ele, “o espaço não é homogêneo, ele apenas tem uma homogeneidade como meta, como orientação [...]. Ele parece homogêneo, mas é multiforme. Seus formatos geométricos e visuais são complementares em suas antíteses”. Eles são “diferentes modos de atingir o mesmo resultado: a redução do ‘real’, por um lado, a um ‘plano’ existindo no vazio [...] e, por outro, à unidimensão de um espelho, de uma imagem, de puro espetáculo [...]. A terceira se relaciona ao espaço vivido e diz respeito às relações geralmente tratadas como psíquicas, mas que Lefebvre (1991, p. 186-187) prefere tratar como materiais, uma vez que surgem da conexão corpo/objeto e do espelho/ objeto. Por espelho, Lefebvre (1991) se refere ao processo de abstração que o objeto permite pelo reflexo do nosso corpo como um signo do que somos. Por meio desta relação disjunta entre forma e conteúdo, Lefebvre (1991) estabelece as seguintes classificações: (1) simetria (planos e eixos) - duplicação, refração, (as)simetria; (2) miragens e efeito miragem – reflexos da superfície versus profundidade, o revelado versus o oculto, o opaco versus o obscuro; (3) linguagem – mediação entre o implicador versus o implicado, ou entre o que confere valor versus o que tem valor, ou como mudança de direção através do discurso; (4) consciência - de si mesmo e do outro, do corpo e do ambiente abstrato dos outros e de tornar-se outro (alienação); (5) tempo – relação imediata entre repetição e diferenciação; (6) espaço – compreendendo as suas duplas determinações entre o imaginário e o real, o produzido e o produtor, o material e o social, o imediato e o mediato, o conectado e o separado. Por fim, para definir as propriedades dos espaços urbanos diferenciais (tempo-espaço), o autor introduz novos conceitos: (1) isotopia – um lugar e tudo que o cerca e que faz de tudo o “mesmo lugar”; (2) heterotopia – uma diferença que situa o lugar diferencial com relação ao lugar inicial como um outro lugar e que pode ir do contraste ao conflito ao se levar em consideração os ocupantes do lugar; (3) u-topia – o não lugar Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 84 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira que não tem lugar e busca um lugar para si. Nesse sentido, o u-tópico “não tem nada em comum com um imaginário abstrato”. Ele é real, “ele está no próprio coração do real, da realidade urbana que não pode existir sem esse fermento. [...] Esse é um lugar paradoxal aonde o paradoxo se torna o oposto do cotidiano” (LEFEBVRE, 2003, p. 38). Uma vez estabelecida a base teórico-metodológica da tríade da produção do espaço, a partir da qual foi construído o olhar sobre o projeto, a seguir se apresentam as informações sobre o Porto Alegre 4D. Porto Alegre 4D A região do 4º Distrito de Porto Alegre é caracterizada geograficamente por sua proximidade com o centro da cidade na sua extremidade sul e conexão com a região metropolitana na extremidade norte. A região é muito bem servida por vias de acesso e serviço de transporte coletivo (ônibus e trem metropolitano), uma vez que importante infraestrutura de conexão metropolitana e regional passa por ali. Essa oferta de infraestrutura de mobilidade tem relação com o desenvolvimento histórico do 4º Distrito. Desde o início do século XX, a região era ocupada por industrias do setor da fiação e metalurgia. No Plano Diretor de 1959 a região foi gravada como Distrito Industrial. Na época, o Plano seguia os preceitos modernistas da separação da habitação, lazer, trabalho e circulação por meio do zoneamento de atividades. Com o passar do tempo, assim como em outras cidades brasileiras, houve a saída da indústria para fora das cidades, causando uma reestruturação nos usos da região. Atualmente, a região apresenta baixa densidade populacional (Figura 1), baixa ocupação por residentes com renda acima de cinco salários mínimos (Figura 2), e uma variedade de tipologias tombadas ou inventariadas pelo setor de patrimônio histórico. Segundo estudos estudos da Prefeitura Municipal de Porto Alegre (PMPA), a região apresenta um grande potencial de renovação imobiliária (Figura 4). Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 85 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Figura 1: Densidade demográfica (hab/Km2) do 4º Distrito – Porto Alegre Fonte: IBGE (2010). Figura 2: Renda per capita acima de 5 salários mínimos no 4º Distrito – Porto Alegre Fonte: IBGE ( 2010) Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 86 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Figura 3: Potencial de renovação construtivo no 4º Distrito – Porto Alegre 87 Fonte: Masterplan 4D (2016) Os dados acima apresentados, associados ao discurso por parte dos envolvidos no projeto, indicam que a região do 4º Distrito seria um lugar sem vida. A mídia reforça essa suposição: Há pelo menos 30 anos o Quarto Distrito de Porto Alegre pede socorro. Após meio século de boom industrial, foi abandonado pelas fábricas, viu a degradação tomar conta de suas ruas e amargou o afastamento gradativo da vida urbana – agravado por alagamentos, insegurança e prostituição. Em meio a desacertos e desinteresse político, hoje iniciativas da sociedade civil tentam transformá-lo novamente em um espaço público vibrante (FELIN, 2014). Na mesma reportagem, os presidentes do Conselho de Arquitetos e Urbanistas (CAU-RS) e do Instituto dos Arquitetos do Brasil (IAB-RS) alertam para a necessidade de planos claros para a área. Segundo o presidente do CAU-RS, “a partir do momento em que a prefeitura não apresenta um plano específico para a região, ela fica ao sabor dos Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira interesses e precisa lidar caso a caso. Aí, às vezes, as obras são melhores para o investidor, às vezes melhores para a cidade” (FELIN, 2014). Importante notar que a matéria foi realizada pelo Jornal Zero Hora (vinculado ao Grupo RBS, filial da Rede Globo no Estado) quatro meses após a realização da Copa do Mundo em Porto Alegre, no momento de avaliação da experiência de grandes obras oportunizada pelo megaevento. Vale destacar que, por ocasião da realização da Copa do Mundo 2014, importantes obras de infraestrutura e equipamentos urbanos foram realizadas como a construção do novo estádio de futebol do Grêmio (na extremidade norte), o incremento da rede de fibra ótica, a duplicação de vias e o Grande Projeto Urbano, ainda em disputa, de Revitalização do Cais Mauá (na fronteira sul do 4º Distrito) 4. Os incrementos de infraestrutura associados à localização privilegiada (centro da cidade, aeroporto e rodovias interestaduais) favorecem a perspectiva de ‘revitalização’ da área baseada no desenvolvimento de atividades de inovação e no empreendedorismo associados ao desenho urbano. Sobre a disseminação da ideia de revitalização para legitimar o projeto Porto Alegre 4D, vale ressaltar essa não é uma região vazia e sem vida. A Figura 4 elucida esta questão - as cores mais escuras representam a maior concentração de domicílios particulares permanentes não ocupados. Em contraposição, a Figura 5 mostra em cores mais escuras a maior concentração de domicílios particulares permanentes ocupados. A leitura das duas imagens relacionando-as com a Figura 2 propicia a compreensão de que, sim, a área tem uma ocupação mais rarefeita e apresenta alguma parcela de lotes desocupados, mas não configura um vazio urbano. Figura 4: Domicílios particulares permanentes não ocupados Fonte: IBGE (2010) 4 A esse respeito ver: Oliveira, C. M.; Misoczky, M. C. Urban entrepreneurialism in FIFA World Cup host cities: the case of Porto Alegre. Organizações & Sociedade v. 23, p. 624-645, 2016. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 88 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Figura 5: Domicílios particulares permanentes ocupados Fonte: IBGE (2010) Percorrendo a região, por meio de espaços acessíveis ao uso comum como ruas e praças, percebe-se a diversidade do tipo de ocupação arquitetônica e da diversidade social. A região Norte dos Bairros Humaitá e Navegantes é ocupada por condomínios de classe média, todos cercados, por vilas urbanizadas e ocupações informais. Nesta parte do 4º Distrito é onde se localiza o estádio de futebol Arena do Grêmio e as torres de apartamentos de classe média. É possível perceber neste cenário limites e territórios proibidos, originando segregação sócioespacial, seja pelos muros dos condomínios fechados que contam com estrutura de lazer intramuros, seja pelo preconceito, medo, ou falta de vontade de circular em favelas, mesmo que urbanizadas. Neste caso, os muros são invisíveis e construídos cotidianamente. Seguindo em direção ao centro, o cenário é um pouco diferente, variando entre edificações de patrimônio histórico, galpões de indústria (alguns abandonados e outros em operação), habitações de classe média, cortiços, praças, terminais de ônibus, equipamentos públicos como escolas e abrigos da PMPA. A ocupação varia em áreas de tecido urbano mais ou menos consolidado, com mais ou menos comércio e com diferentes classes sociais. A parte do Bairro Floresta é mais consolidada, por sua proximidade com o centro e o Bairro Moinhos de Vento, de classe média-alta. Nesta fração do território observa-se o convívio entre moradores, comerciantes, profissionais do sexo e, mais recentemente, jovens empreendedores que se alinham com o conceito de economia criativa. Por sua proximidade com a malha urbana consolidada e pelos ainda a baixos custos de aluguel, a região se tornou um atrativo para este público. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 89 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira A partir desta transformação recente o projeto, que tem como lema transformar o 4º Distrito no lugar mais inovador da América Latina, começou a ser prospectado em meados de 2013, quando os participantes do grupo CITE (Cidadãos, Inovação, Tecnologia e Empreendedorismo5) organizaram uma viagem ao Vale do Silício. Na ocasião, foi uma comitiva de 30 pessoas incluindo o ex-Prefeito de Porto Alegre José Fortunati, os secretários de Governança Local e de Planejamento e os coordenadores do Gabinete de Inovação (Inovapoa) e de Gabinete Digital. Segundo o exPrefeito, Porto Alegre é hoje, entre as 50 maiores cidades do Brasil, a segunda em criatividade e uma das que mais investe na indústria criativa e da inovação. Nós queremos consolidar a cidade como uma referência na área. E a ida ao Vale do Silício significa trazer investimentos, trazer mais esperança para o desenvolvimento desse importante setor para a nossa cidade. Sem contar que as novas tecnologias aprimoram a prestação de serviços e quem ganha com isso é a população, é o cidadão (FERNANDES, 2013). A PMPA e o grupo CITE encontraram nas universidades parceiros interessados em compor o projeto, principalmente a Universidade Federal do Rio Grande do Sul (UFRGS)6 e a Pontifícia Universidade Católica (PUC-RS), que vislumbram a possibilidade de construir um parque tecnológico na área. Segundo um dos entrevistados envolvido no projeto, durante reuniões com o setor privado de tecnologia e inovação e universidades, as questões urbanas acabaram liderando muitas das discussões, uma vez que as empresas demonstravam a preocupação em investir em funcionários de alta qualidade, para em três ou quatro anos saírem em função de oportunidades de trabalho em outra cidade com mais atrativos urbanos. Diante desta questão, em 2015, uma comitiva composta pelos então Prefeito e Secretários da PMPA e representantes do setor privado (construção civil, educação e saúde, principalmente) fizeram uma viagem a Barcelona para conhecer a experiência do 22@Barcelona. Segundo o ex5 6 Ver notícia sobre a criação do CITE: http://www.baguete.com.br/noticias/13/05/2013/cite-quer-agitar-porto-alegre. Chama atenção como se apresentam: “grupo formado por jovens empreendedores em tecnologia, profissionais referência na área empresarial e acadêmica querem romper o marasmo de Porto Alegre e colocar a capital de volta no cenário internacional de investimentos em inovação”. Nada mais óbvio que definir uma área da cidade como sem vida, já que para esses ‘jovens’ (de fato, não tão jovens como se apresentam) toda a cidade é um marasmo. Nesta Universidade o projeto inclui o Núcleo de Tecnologia Urbana da Faculdade de Arquitetura e o Núcleo de Gestão da Inovação Tecnológica da Escola de Administração, entre outros. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 90 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Secretário Municipal da Fazenda Jorge Tonetto, que coordenava o grupo junto com o ex-Secretário de Urbanismo, Valter Nagelstein, “nesta primeira fase do projeto está sendo discutido todo o equipamento urbano necessário para dar uma ambiência de qualidade ao local e atrair investimentos de empresas” (FERRAZ, 2015). A mudança da coordenação do projeto para o ex-Secretário da Fazenda Jorge Tonetto representou um marco no desenvolvimento do projeto. Segundo um entrevistado, a Secretaria da Fazenda coordena o projeto pois é nela que se viabiliza economicamente a sua implementação. Ainda em julho de 2015 foram realizadas visitas técnicas em São Paulo e no Rio de Janeiro para conhecer as experiências das Operação Urbana Consolidada do Porto Maravilha e da venda de Certificados de Potencial Adicional de Construção (CEPACs). Em São Paulo as visitas foram na SP Urbanismo e no Banco do Brasil Investimentos. Foram criados, então, mecanismos de incentivo fiscal e isenção de impostos para que empresas se interessassem pela área. Dentre as leis de incentivos aprovadas ainda em 2015 estão: isenção do IPTU por cinco anos (propriedade, locação ou cessão), isenção do ITBI na aquisição de imóveis para empresas de base tecnológica ou inovadora que venham a se instalar na região; redução de 5% para 2% na alíquota do ISS para pesquisa e desenvolvimento nas áreas de tecnologias da saúde em toda a cidade. Além disso, a PMPA julgou oportuno utilizar o instrumento das Operações Urbanas Consorciadas (OUCs), à exemplo de São Paulo e Rio de Janeiro. Tendo em vista a venda de CEPACs para a região e a possiblidade da emissão de debêntures da dívida pública na bolsa de valores, a PMPA aprovou na Câmara de Vereadores o projeto de Lei que cria a Empresa de Gestão de Ativos do Município S.A. InvestPOA, a exemplo de São Paulo e Belo Horizonte. Para tanto, a PMPA coloca como garantia seu patrimônio, inclusive importantes equipamentos públicos e culturais da cidade. Embora a venda de CEPACs pareça ser a fórmula para o sucesso de uma OUC, experiências como a do Porto Maravilha no Rio de Janeiro demonstram que o setor privado não tem o interesse necessário para deslanchar uma operação deste porte. Neste caso, todos os CEPACs foram comprados pela Caixa Econômica Federal, ou seja, pelo setor público. No caso de Porto Alegre, segundo um dos entrevistados, mesmo com todo este aprimoramento nos instrumentos financeiros, a resposta do setor privado também não foi favorável, pelo menos enquanto a área mantivesse suas características de degradação física e de vulnerabilidade social. Assim, a posição dos empresários tanto da construção civil quanto do setor de tecnologia e informação corroboram para a necessidade do redesenho desta parte da cidade. Para tanto, em 2016, a PMPA contratou uma consultoria da UFRGS para a realização de um Masterplan para o 4D. Sob coordenação do arquiteto e professor da Faculdade de Arquitetura, Benamy Turkienicz, o projeto foi entregue à PMPA no final de 2016. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 91 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Durante o processo de elaboração do Masterplan uma série de workshops e palestras foram organizadas com diversos setores da sociedade, incluindo construção civil, empresas da área da saúde, as casas colaborativas associadas à economia criativa presentes na região, universidades, e workshops chamados de sociais, mas com uma presença inexpressiva da comunidade residente. Segundo o projeto disponibilizado pela PMPA os objetivos do Masterplan 4D (2016) são: reverter o processo de degradação do 4D, oportunizando investimentos econômicos e sociais; preservar, valorizar e qualificar o patrimônio construído; incorporar agenda ambiental na cidade; melhorar a mobilidade e os padrões de acessibilidade urbana; intensificar a urbanidade no 4D e garantir ambientes amigáveis para todas as faixas etárias; incentivar a população habitacional inclusiva. O projeto visa desenvolver áreas temáticas como clusters relacionados às áreas da saúde, ensino, tecnologia e indústria criativa. São incorporadas iniciativas já presentes no 4º Distrito como as casas colaborativas de economia criativa e a Zona de Inovação Sustentável em Porto Alegre ZISPOA) dirigida por Marc Weiss presidente do Global Urban Development (GUD) que está presente em Porto Alegre desde 2013. Weiss foi responsável pela implementação do NoMa7 em Washington, onde foi realizada a renovação de uma zona industrial a partir de atividades envolvendo mídia e tecnologia. Associado ao ZISPOA está o projeto de Living Lab, a exemplo do 22@Barcelona. No Masterplan 4D, uma rua é programada como via da inovação para abrigar as ações e estudos de novas tecnologias associadas aos problemas urbanos, na lógica do que vem sido denominado de Smart Cities. É importante destacar que o trabalho de Weiss é financiado pelo Banco Mundial e que ele trabalha diretamente com estudantes das universidades locais em atividades sempre associadas ao empreendedorismo e à inovação configurando-se, também, como uma estratégia de difusão de práticas neoliberais. No entanto, segundo o ZISPOA, o projeto é elaborado pela comunidade e conta com iniciativas como a Pulsar e Next Citizens. Outra estratégia do Masterplan se refere às chamadas quadras rápidas. Seriam quadras, ou até mesmo conjunto de quadras, que por pertencer a um mesmo proprietário ou por apresentar edificações sem uso de grande porte, teriam um alto potencial de transformação no curto prazo. A Figura 6 mostra os níveis deste potencial de transformação que 7 Weiss é professor adjunto da área de Assuntos Internacionais e Públicos da Universidade de Columbia e professor internacional visitante de Economia e Administração de Empresas da Unisinos. Em 1998, ele foi autor de um Plano de Desenvolvimento Econômico para Washington e, especificamente, para a área norte de Massachusetts Avenue (NoMa), apresentado como uma oportunidade potencial de desenvolvimento em uma área enferrujada (sic), ancorada por empresas de mídia e acessível por uma nova estação de metrô (ver: http://www.revistadigital.com.br/2014/07/america-latina-precisa-de-um-vale-do-silicio/.). Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 92 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira varia do mais rápido ao mais lento. Para as quadras rápidas são previstos Projetos de Intervenção Urbana (PIU), como são chamados no Masterplan, ou mesmo Grandes Projetos Urbanos (GPUs). São previstos pelo menos seis PIUs, tais como parque linear, shopping center, centro de cultura e centro administrativo. Os estudos vinculados ao Masterplan são cruciais para o desenvolvimento dos novos parâmetros urbanísticos e para a venda de CEPACS. Estão ainda em fase de definição os estudos de volumetria que exploram novos índices de aproveitamento e ocupação dos lotes, sempre desde uma perspectiva da quadra. São exploradas questões de insolação e ventilação, bem como relação com o patrimônio. Segundo um dos entrevistados, chegaram a estipular uma altura máxima de seis pavimentos em quarteirões que contivessem edificações de patrimônio histórico. Porém, esta possibilidade foi contestada por técnicos da Secretaria da Fazenda, uma vez que os índices “baixos” não seriam atrativos para o desenvolvimento destas quadras. Turkienicz, autor do Masterplan, questionado em apresentações do projeto, argumenta que não haverá mudança na densidade prevista hoje no Plano Diretor para a região, mas que novos índices serão estipulados para determinados pontos respeitando a média de densidade máxima. A Figura 6 mostra os estudos volumétricos e os parâmetros urbanísticos previstos no projeto. Figura 6: Simulações de ocupações de quadras Fonte: Masterplan 4D (2016). A única estratégia do Masterplan que se destina a tratar da população de alta vulnerabilidade social é estimular a produção de Habitação de Interesse Social (HIS) através de incentivos urbanos e fiscais. A responsabilidade da construção de HIS não seria do empreendedor, mas do Conselho Gestor que administrará os ativos e os recursos relacionados a OUC. O Conselho Gestor, assim como previsto no Estatuto da Cidade, terá competência de atribuir as prioridades e as ações necessárias para a implementação dos programas e projetos estabelecidos no Masterplan, assim como realizar contratos de administração para o cumprimento dos planos e metas estipulados. No Masterplan, este novo mecanismo de Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 93 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira gestão é interpretado como uma estratégia de promoção da gestão participativa no desenvolvimento do território, uma vez que o conselho seria formado por quatro ‘élices’: PMPA, setor privado, universidades e sociedade civil organizada. Cabe lembrar o caso do Porto Maravilha e o fato de que nenhuma das escolas e dos postos de saúde previstos foram construídos, pois os fundos destinados foram realocados para a construção do Museu do Amanhã. Para o 4º Distrito, os entrevistados que participaram da elaboração do projeto não julgam que isso seria um problema, afinal se em um primeiro momento os investimentos necessários são para os projetos âncora, isto não significa que no futuro não possam ser realocados em projetos de caráter mais social. Ainda no Masterplan 4D está prevista a localização de HIS de maneira pulverizada no território, a fim de evitar a criação de ‘ghetos’ como os hoje existentes. Se por um lado, no Masterplan é proposto um passo positivo com relação a esse problema; por outro, faltam mecanismos que evitem que esta estratégia sirva, na verdade, para a expulsão dos mais pobres da região. Aqui é preciso mencionar o projeto Porto Alegre Resiliente8, com financiamento da Fundação Rockfeller, que inclui no seu projeto estratégico, dentre outros programas, a revitalização do 4º Distrito e a agilização do título de propriedade às famílias de baixa renda. A combinação título de propriedade para HIS e construções de alta qualidade pulverizadas no território só tem um resultado já suficientemente concretizado em tantas cidades: a gentrificação rápida e eficaz. A articulação de um projeto urbano vai muito além do desenho arquitetônico, embora esta seja uma dimensão importante. Diferente da época do modernismo, a exemplo de Brasília, existe toda uma teia de articulações entre atores e instituições, e de instrumento de parcerias entre o setor público e o setor privado que vão além da forma urbana. Para a elaboração do projeto, a rede de atores que formam a coalisão para crescer, nos termos de Logan e Molotch (1993), pode ser facilmente percebida. Elites locais, políticos, grupos específicos das universidades e mídia corporativa formam esta coalisão. A população residente é colocada à margem deste processo, restringindo-se à presença em momentos e situações construídos para garantir a aparência de participação. A presença de algumas menções a ações que seriam para a população residente e o discurso da reversão do processo de degradação do 4D, oportunizando investimentos econômicos e sociais, quando detalhada mostra que o projeto desconsidera as necessidades desta população que, em última análise, não é o foco do projeto. Os tipos de investimento, o tipo de indústria e de usos não são os que se adequam à 8 Ver: http://portoalegreresiliente.org/. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 94 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira qualificação capacitação profissional da população local. Como dito em diversas apresentações do projeto, a intenção é atrair uma população que hoje não tem interesse em morar ali, mas que, com os devidos atrativos de urbanidade, cultura e economia teriam vontade de ali se estabelecer. Este fato é evidenciado não somente pelas declaradas intenções do projeto, mas pode ser percebido também no desenvolvimento do Masterplan. Abstrações que separam forma e conteúdo são recorrentes. As apresentações do projeto levam em consideração apenas a forma e a estrutura urbana.Em termos lefebvrianos, o espaço é tratado como geométrico, isotópico. Dados sócio-espaciais são negligenciados sistematicamente. Quando as equipes proponentes do projeto são confrontadas, por exemplo, com perguntas sobre habitação de interesse social, não sabem informar qual é a necessidade em termos quantitativos, menos ainda qualitativos. Entretanto, isto não impede o discurso de que o estoque de habitação social poderia servir não somente a população do território em questão, como também a de seu entorno. O discurso é reforçado por apresentações visuais, de cenários futuros. Em termos lefebvrianos, ele se demostra também fálico, representando a brutalidade com o que o Estado e o mercado pretendem se impor sobre os moradores da região. Essas questões ficam evidentes em um espaço de junção, segundo categorização dos espaços vividos por Lefebvre (1991). Segundo o autor, somente nos espaços públicos é que a tríade da produção do espaço pode ser percebida na sua plenitude. Foi possível percebê-la no evento realizado no Espaço Cultural Vila Flores9 no dia 13 de setembro de 2016, com o nome Encontro Cidade: uma construção coletiva. O Vila Flores pode ser categorizado como um espaço de junção. Embora não seja público, ele é aberto ao público sistematicamente pela diversidade de atividades que abriga. É um ponto de encontro entre pessoas não só da região, mas da população da cidade em geral. No evento promovido pelo Vila Flores o projeto foi apresentado pelo arquiteto coordenador do projeto urbano. No debate que se iniciou logo após a apresentação, moradores da região e de Porto Alegre confrontaram o projeto. Representantes do Movimento Nacional de Luta pela Moradia (MNLM), assistentes sociais da PMPA, coletivos de economia criativa estabelecidos na região, moradores, arquitetos e urbanistas, pesquisadores e militantes partidários estavam presentes. Todos defenderam que as imagens que apresentavam o projeto não expressam e nem contemplam os moradores da região. Primeiro, pelo uso da simetria. A leitura de apenas duas situações opostas, ou seja, assimétricas da região como um todo, são uma racionalização que oculta o conteúdo do lugar. Segundo, por uma questão de linguagem. As 9 Ver: https://vilaflores.wordpress.com/about/. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 95 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira simulações em 3D expressam o que Lefebvre (2008a) denomina de espaço homogêneo e fraturado, disjunto. Homogêneo porque pasteuriza, é artificial, não é daquele lugar, é de qualquer lugar do mundo. Fraturado e disjunto porque não representa a linguagem de quem ali pertence e, por consequência, ao quebrar o sentimento de pertencimento do lugar, expulsa a população inclusive simbolicamente. O fenômeno da gentrificação não se dá apenas pelo aumento do custo de vida (moradia, alimentação, conveniência, transporte etc.), mas pela exclusão do sentimento de pertencimento ao lugar. Segundo moradores presentes no evento, o projeto não se relaciona com a história do bairro, não a preserva nem valoriza. A relação harmônica entre o antigo e o novo perde para a geração de mais valia via capital imobiliário. Pode-se dizer que os espaços culturais previstos, como as pequenas bibliotecas públicas (a exemplo das feitas em Medellín) e a Praça das Artes, que contariam com espaço para arte a capacitação profissional seriam respostas aos problemas sociais apresentados na região. Embora a difusão da cultura seja importante para todas as classes sociais, tal estratégia se alinha à noção do uso da cultura para promover o consumo a partir da imagem da cidade. Como visto a exemplo do Museu Guggenheim em Bilbao e do Museu do Amanhã no Rio de Janeiro. Considerações Finais A análise de um projeto urbano deve ir muito além da compreensão da dimensão da arquitetura em grande escala, sendo uma complexa relação de atores e processos a ele relacionado. Sendo assim, para garantir a harmonia social na produção de intervenções em territórios marcados pela contradição fundante do sistema do capital, é preciso ideologias bem sucedidas, como Lefebvre (2014) indica, ao referirse a ilusões efetivas. Ou seja, ideias que têm implicações práticooperacionais concretas. No caso dos projetos urbanos, a ideia de que uma área precisa ser revitalizada se impõe como o indispensável carreador da cientificidade abstrata. Ainda que a crítica de Lefebvre (2008b) tenha sido feita para o urbanismo modernista, ela propicia inspiração para a crítica ao urbanismo neoliberal contemporâneo, já que arquitetos e urbanistas, agora com a colaboração dos gestores da inovação (‘senhores do espaço’ associados aos ‘senhores da inovação’), continuam reproduzindo a lógica da forma concebida pelos técnicos que exclui, intencionalmente, o espaço histórico e vivido. Estes ‘senhores’ tomam o espaço como instrumento a ser intencionalmente manipulado conforme objetivos político-econômicos. As representações que produzem tentam criar consensos, transformar ideias e sentimentos (inclusive dos que vivem aquele espaço) e reorganizar o espaço como mercadoria a ser consumida e para o consumo. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 96 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira Assim se dissemina, por exemplo, a crença de que uma área precise ser revitalizada. Smith (2006) se refere à revitalização/renovação/regeneração urbana como um eufemismo alinhado à gentrificação generalizada como estratégia de planejamento. O eufemismo reside em fazer crer que os espaços delimitados para um determinado projeto urbano não têm vida e, portanto, precisam ser revitalizados. Claro, ter vida significa contribuir para processos cada vez mais intensivos de acumulação. A concentração de população de baixa renda na região do 4º Distrito exemplifica esta situação, que pode ser percebida igualmente na área de outros projetos urbanos como a do Porto Maravilha. Assim, a gentrificação generalizada, nos termos de Smith (2006), alimenta a necessidade de gerar valor e, sobretudo, de valorização do valor associado aos processos cíclicos de produção-consumodistribuição-circulação. Para que isso ocorra sem perturbações, são necessárias representações do espaço que disseminem e criem consenso em torno da necessidade de revitalização urbana e, ao mesmo tempo, consolidam sua expressão contemporânea: o projeto urbano. Percebe-se a transformação do papel do planejamento, de corretor dos problemas da cidade para empreendedor, ou seja, um instrumento que reforça o valor de troca e o consequente acesso desigual da população ao meio urbano. O arquiteto responsável pelo Masterplan 4D explicou em um debate sobre o projeto que “a cidade é uma mercadoria que precisa ser desenhada” (TURKIENICZ, 2016). A cidade é meio de produção essencial para a reprodução das relações sociais de produção. Para garantir que esse meio se concretize efetivamente, os recortes urbanos que fragmentam o tecido do percebido e do vivido são indispensáveis. O Projeto Urbano Porto Alegre 4D vem se mostrando, ainda que em fases iniciais, como um espaço de potencial conflito entre a prática espacial, a associação entre o espaço percebido, a realidade do dia a dia e a realidade urbana, daqueles que fazem do quarto distrito um território de vida e trabalho e a representação do espaço dos tecnocratas (aqui se usa livremente o termo cunhado por Lefebvre para se referir aso profissionais técnicos relacionados ao urbanismo para incluir, para além desses, os empreendedores, especuladores, acadêmicos, administradores públicos – a ‘coalisão para crescer’) que estabelecem os regimes discursivos que transformam o vivido e o percebido no concebido. A superimposição dos interesses destes ‘senhores’ está se produzindo através dos três formatos indicados por Lefebvre (1991): geométrico, visual e fálico. No entanto, a meta da homogeneidade, a redução do real a um plano, ao unidimensional, encontra, ainda que de maneira frágil no momento da escrita deste texto, a contraposição de lugares de encontro e de resistência, a introdução do paradoxal, a possibilidade da heterotopia e, quem sabe, da u-topia. Referências Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 97 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira ALEXANDER, Christopher. A city is not a tree. Design, n. 206, p. 46-55, 1966. BRENNER, Neil; MARCUSE, Peter; MAYER, Margit. Cities for people not for profit: an introduction. In: BRENNER, Neil; MARCUSE, Peter; MAYER, Margit (Eds.). Cities for People, Not for Profit: Critical Urban Theory and the Right to the City. New York: Routledge. 2012. p. 1-10. CASTELLS, Manuel. The Urban Revolution: A Marxist approach. London: Edward Arnold, 1977. FELIN, Bruno. Quarto Distrito espera há pelo menos 30 anos por uma revitalização. Jornal Zero Hora, Porto Alegre, 15 nov. 2014. Disponível em: (http://zh.clicrbs.com.br/rs/porto-alegre/noticia/2014/11/quartodistrito-espera-ha-pelo-menos-30-anos-por-uma-revitalizacao4643635.html), acesso em 29 jan. 2017. FERNANDES, Melina. Fortunati parte em missão oficial ao Vale do Silício. Notícias do Portal da PMPA, Porto Alegre, 21 mai. 2013. Disponível em: (http://www2.portoalegre.rs.gov.br/portal_pmpa_novo/default.php?p_ noticia=160745&FORTUNATI+PARTE+EM+MISSAO+OFICIAL+AO+VA LE+DO+SILICIO), acesso em 30 nov. 2016. FERRAZ, Fernanda. Parques tecnológicos de Barcelona são modelo para o 4º Distrito. Notícias do Portal da PMPA, Porto Alegre, 22 jun. 2015 Disponível em: (http://www2.portoalegre.rs.gov.br/portal_pmpa_novo/default.php?p_ noticia=178597&PARQUES+TECNOLOGICOS+DE+BARCELONA+SAO+ MODELO+PARA+O+4O+DISTRITO), acesso em 30 nov. 2016. IBGE. Censo Demográfico 2010. Disponível em: (http://www.ibge.gov.br/home/estatistica/populacao/censo2000/), acesso em 30 jun. 2015. MASTERPLAN 4D. 4D Distrito de Inovação de Porto Alegre. Porto Alegre, 21 de Dezembro de 2016. Disponível em: (http://4distrito.portoalegre.rs.gov.br/sites/default/files/Masterplan_4D istrito.pdf), acesso em 13 de fev. 2017. LEFEBVRE, Henri. The production of space. Oxford: Blackwell, 1991. _________. A revolução urbana. Belo Horizonte: Editora UFMG, 2008a. _________. Espaço e política. Belo Horizonte: Editora UFMG, 2008b. Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 98 CIDADES ARTIFICIAIS: UM ESTUDO EXPLORATÓRIO DO PROJETO URBANO PORTO ALEGRE 4D Clarice M. de Oliveira LOGAN, John R; MOLOTCH, Harvey L. The City as a Growth Machine. In: FAINSTEIN, S. Susan; CAMPBELL, Scott. Readings in Urban Theory. Oxford: Blackwells, 1993. HARVEY, David. Condição Pós-Moderna: uma pesquisa sobre as origens da mudança cultural. São Paulo: Loyola, 2011. ______. A produção capitalista do espaço. São Paulo: Annablume Editora, 2006. ______. Social Justice and the city. Athens: The University of Georgia Press, 2008. SCHMID, Christian. Henri Lefebvre, the right to the city, and the new metropolitan mainstream. In: BRENNER, Neil; MARCUSE, Peter; MAYER, Margit (Eds.). Cities for people, not for profit: critical urban theory and the right to the city. New York: Routledge, 2012. p. 42-62. SMITH, Neil. A gentrificação generalizada: de uma anomalia local à “regeneração” urbana como estratégia urbana global. In: BIDOUZACHARIASEN, Catherine (Coord.) De volta à cidade: dos processos de gentrificação às políticas públicas de “revitalização” dos centros urbanos. São Paulo: Annablume Editora, 2006. p. 59-87. TURKIENICZ, Benamy. Debate Ricardo Fayos e Benamy Turkienicz. Organização UrbsNova: Escolha de Engenharia da UFRGS, Porto Alegre 22 de dez 2016. Gravado. Submetido em: 15/12/2016 Aprovado em: 10/02/2017 Revista Brasileira de Estudos Organizacionais  v. 3. n. 1, p. 77-99, jun.2016, eISSN: 2447-4851 Doi 10.21583/2447-4851.rbeo.2016.v3n1.74 Sociedade Brasileira de Estudos Organizacionais 99
https://openalex.org/W4283583582	https://ojs.iptjournal.org/index.php/ipt/article/download/39/38	Russian	null	Взрывобезопасный каталитический синтез гидроксиламинсульфата	Promyšlennye processy i tehnologii	2,022	cc-by	5,435	Аннотация Проанализированы условия синтеза гидроксиламина при взаимодействии Н2 и NO в водном растворе серной кислоты в присутствии катализатора «платина на электрогра- фите» в каскаде трехфазных реакторов смешения. Исследованы факторы, обеспечива- ющие взрывобезопасность протекания процесса: давление, соотношение Н2 : NO и со- держание N2O в газообразных продуктах синтеза. Описана конструкция реакторов, обеспечивающая взрывобезопасность процесса. Предложен алгоритм управления этим процессом. Приведены данные о составе газовой фазы на входе в ректоры синтеза гид- роксиламинсульфата (ГАС), состав газообразных продуктов синтеза ГАС, условия и результаты синтеза ГАС в промышленных условиях. Предлагаемые решения по управ- лению каталитическим синтезом гидроксиламинсульфата позволяют значительно упростить управление за счет создания простого и понятного алгоритма и, одновре- менно с поддержанием взрывобезопасной концентрации NO в синтез-газе, анализиро- вать содержание N2O в газообразных продуктах синтеза, оценивать стабильность ра- боты реакторного блока, необходимость замены катализатора и необходимость оста- нова реакторного блока. Дополнительный ввод необходимого количества оксида азота (II) непосредственно в реакторы каскада без контакта с синтез-газом и водородом повышает взрывобезопасность процесса. Ключевые слова каталитический синтез гидроксиламинсульфата, взрывобезопасность процесса, алго- ритм управления процессом, конструкция реакторов 2NO + H2 + H2SO4 → (NH2OH)2∙H2SO4 (1) 2NO + H2 + H2SO4 → (NH2OH)2∙H2SO4 12.03.2022 12.03.2022 Поступила в редакцию УДК 547.238 УДК 547.238 DOI: 10.37816/2713-0789-2022-2-2(4)-38-49 А В А М В Д б С В А Взрывобезопасный каталитический синтез гидроксиламинсульфата А. В. Артемов,1, М. В. Дюбанов, С. В. Ардамаков** * Национальный исследовательский центр «Курчатовский институт», Москва, Россия Межведомственный центр аналитических исследований в области физики, химии и биологии при Президиуме Российской академии наук, Москва, Россия * ООО «Волгатехноол», Тольятти, Россия Взрывобезопасный каталитический синтез гидроксиламинсульфата А. В. Артемов,1, М. В. Дюбанов, С. В. Ардамаков** * Национальный исследовательский центр «Курчатовский институт», Москва, Россия Межведомственный центр аналитических исследований в области физики, химии и биологии при Президиуме Российской академии наук, Москва, Россия * ООО «Волгатехноол», Тольятти, Россия 1 Для переписки: Email: arsenyart@icloud.ru Введение (1) Гидроксиламинсульфат (ГАС) — (NH2OH)2∙H2SO4, один из основных реаген- тов в производстве капролактама. Синтез ГАС основан на каталитическом восста- новлении NO водородом в серной кислоте: В промышленности процесс обычно проводят при температуре 35–45оС, избы- точном давлении 0.45–0.60 атм. и объем- ном соотношении Н2 : NO, равном (1.7–1.8) : 1. В качестве катализатора 1 Для переписки: Email: arsenyart@icloud.ru Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 38 Взрывобезопасный каталитический синтез гидроксиламинсульфата используют платину, нанесенную на мел- кодисперсный графит («платина на элек- трографите»). Процесс проводят в каскаде трехфазных реакторов смешения (обычно 4–6 реакторов в каскаде). Синтез-газ (NO + H2) получают смешиванием NO и H2 в смесителе при указанных соотноше- ниях и давлении и направляют в реакторы каскада [1]. Водный 18–25% раствор сер- ной кислоты и суспензию катализатора в этом растворе вводят в каждый реактор каскада таким образом, чтобы стабилизи- ровать концентрацию катализатора в каждом реакторе на уровне 30–50 г/л [2]. Такая организация процесса позволяет получать ГАС с максимальным содержанием 275 г/л в конечном продукте. чтобы в реакторах это соотношение дости- гало оптимальной величины (1.7–1.8) : 1, как это предусмотрено технологическим регламентом получения ГАС; 2. увеличить концентрацию серной кислоты до 26–45% на всех этапах техно- логического цикла производства; 2. увеличить концентрацию серной кислоты до 26–45% на всех этапах техно- логического цикла производства; 3. при проведении процесса в области избыточного давления 0.7–3.0 атм. необхо- димо контролировать содержание водо- рода в синтез-газе, чтобы предотвратить взрывоопасную ситуацию на стадиях полу- чения синтез-газа в смесителе и подачи синтез-газа в реакторы каскада. В соответствии с основной реакцией получения ГАС (1) увеличение выхода ГАС возможно при увеличении давления в системе и увеличении концентрации сер- ной кислоты. Однако, увеличение давле- ния на стадии приготовления синтез-газа (NO + H2) может привести к взрывоопас- ному режиму ведения процесса, как это показано на рисунке 1. Область технологи- ческих параметров (H, S, P), обеспечиваю- щая взрывобезопасное протекание про- цесса, на рисунке 1 обозначена как А. Для безопасного ведения процесса получения ГАС в области избыточного давления 0.7–3.0 атм. необходимо изменить состав синтез-газа и увеличить в нем содержание водорода до Н=0.73 (S=2.70). Эта область технологических параметров обозначена на рисунке 1 как Б. Такой состав синтез- газа неприемлем из-за низкого содержа- ния в нем NO, что неизбежно приведет к снижению выхода ГАС. Для устранения отрицательного влияния низкого содержа- ния NO в синтез-газе при проведении про- цесса в области избыточного давления 0.7–3.0 атм. необходимо: Рисунок 1 — Граница взрывоопасной области ведения процесса получения ГАС. Введение S — объемное отношение Н2 : NO в синтез- газе; Н — объемная доля водорода в синтез- газе. Область А — безопасная область ведения процесса, реализуемая в настоящее время. Область Б — безопасная область ведения процесса при избыточном давлении 0.7–3.0 атм. Рисунок 1 — Граница взрывоопасной области ведения процесса получения ГАС. S — объемное отношение Н2 : NO в синтез- газе; Н — объемная доля водорода в синтез- газе. Область А — безопасная область ведения процесса, реализуемая в настоящее время. Область Б — безопасная область ведения процесса при избыточном давлении 0.7–3.0 атм. Зависимость Н = j(Р), приведенная на рисунке 1 и показывающая границу взрывоопасной области в интервале значе- ний Р = 0.25–3.0 атм., может быть с Зависимость Н = j(Р), приведенная на рисунке 1 и показывающая границу взрывоопасной области в интервале значе- ний Р = 0.25–3.0 атм., может быть с Зависимость Н = j(Р), приведенная на рисунке 1 и показывающая границу взрывоопасной области в интервале значе- ний Р = 0.25–3.0 атм., может быть с 1. дополнительно вводить NO в каж- дый реактор каскада в таком количестве, Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 39 А. В. Артемов, М. В. Дюбанов, С. В. Ардамаков водного раствора 26–45 % серной кислоты в каждый из реакторов каскада. Осу- ществляют регулированную (см. поз. 7, 15, 16, 17, 18, 19 на рисунке 2) подачу суспен- зии катализатора в водном растворе 26–45 % кислоты в каждый из реакторов каскада. Осуществляют дополнительную регулированную (см. поз. 20, 21, 22, 23, 24, 25 на рисунке 2) подачу оксида азота (II) в каждый из реакторов каскада для полу- чения объемного соотношения водорода и оксида азота (II), равного (1.7–1.8) : 1. высокой точностью преобразована к ли- нейному виду H = 0.66 + 0.074 ∙ lnP, а условием взрывобезопасного ведения процесса является соотношение высокой точностью преобразована к ли- нейному виду H = 0.66 + 0.074 ∙ lnP, а условием взрывобезопасного ведения процесса является соотношение H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС H ≧ 0.66 + 0.074 ∙ lnP. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС При избыточном давлении 0.7–3.0 атм. и при концентрации серной кислоты 26–45 % масс. синтез ГАС осуществляют в соответствии с принципиальной схемой, приведенной на рисунке 2. Процесс прово- дят следующим образом. Смесь оксида азота (II) и водорода через распредели- тельное устройство подают в каждый из реакторов каскада (1–6). Реакционная смесь самотеком перетекает из одного ре- актора каскада в другой. Процесс прово- дят при избыточном давлении 0.7–3.0 атм. в присутствии катализатора «платина на электрографите». Соотношение смеси ок- сида азота (II) и водорода, подаваемых в нижнюю часть каждого из реакторов кас- када, регулируют таким образом, чтобы для объемной доли водорода (Н) в синтез- газе выполнялось условие Дополнительные подачи осуществ- ляют таким образом, чтобы концентрация катализатора в каждом реакторе каскада была 30–50 г/л. Синтез ГАС проводят при интенсивном перемешивании жалюзийной мешалкой. Верхняя зона реактора снаб- жена жалюзийным пеногасителем для ликвидации образования пено-газового слоя. Отходящие (хвостовые) газы стадии синтеза ГАС направляют на сжигание. Предложенные взрывобезопасные условия синтеза ГАС были реализованы на промышленной установке [3], значи- тельно расширяют область возможных значений технологических параметров процесса (сравнить области А и Б на ри- сунке 1) и позволяют увеличить выход ГАС на 12–25% (таблица 1). H ≧ 0,66 + 0,074 ∙ lnP. Осуществляют регулированную (см. поз. 8, 10, 11, 12, 13, 14 на рисунке 2) подачу H ≧ 0,66 + 0,074 ∙ lnP. Осуществляют регулированную (см. поз. 8, 10, 11, 12, 13, 14 на рисунке 2) подачу H ≧ 0,66 + 0,074 ∙ lnP. Рисунок 2 — Принципиальная технологическая схема получения ГАС. Рисунок 2 — Принципиальная технологическая схема получения ГАС. Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 40 Взрывобезопасный каталитический синтез гидроксиламинсульфата Таблица 1 — Условия и результаты синтеза ГАС в промышленных условиях Темпе- ратура, оС Избыточное давление, атм Концен- трация Катализа- тора в ре- акторах каскада, г/л Мольное отноше- ние H2 : NO (S) Кон- центра- ция H2SO4, % масс. Концен- трация H2SO4 в суспензии катализа- тора. % масс. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Объем- ная доля Н2 в син- тез-газе (Н) 0.66 + 0,071 4×inP Вы- ход ГАС, кг/ч 36 3.0 32 1.75:1 26 26 0.740 0.738 3856 37 2.9 41 1.70:1 43 43 0.736 0.736 3962 41 0.7 38 1.80:1 30 30 0.640 0.635 4011 42 2.8 43 1.72:1 38 38 0.735 0.734 3585 38 2.6 33 1.75:1 29 29 0.730 0.728 3619 39 3.0 42 1.72:1 45 45 0.735 0.738 3781 40 0.9 30 1.71:1 41 41 0.660 0.652 3917 42 1.7 50 1.73:1 28 28 0.700 0.698 3825 38 1.4 44 1.78:1 32 32 0.690 0.684 3976 39 2.1 36 1.79:1 40 40 0.715 0.713 4001 41 3.0 45 1.80:1 44 44 0.735 0.738 4012 41 2.2 47 1.75:1 43 43 0.720 0.716 3916 41 2.3 48 1.75:1 37 37 0.725 0.719 3882 42 2.4 36 1.78:1 27 27 0.725 0.723 3716 осуществляется за счет внутреннего змее- вика (поверхность теплообмена 55 м2) и внешней рубашки (поверхность теплооб- мена 36 м2). Реактор рассчитан на рабочее давление до 4 атм. Предельное рабочее давление в рубашке 6 атм. Реактор имеет размеры (мм): внутренний диаметр 3600; ширина 4320; высота с приводом 8320. Ре- актор имеет устройства ввода (штуцеры) серной кислоты, расположенные в средней и верхней частях реактора [1]. Реконстру- ированный барботер состоит из двух зон, разделенных между собой непроницаемой перегородкой, имеющей горизонтальное расположение и позволяющей осуществ- лять раздельный ввод водорода и оксида азота (II) в реакционную зону, причем ввод водорода осуществляют из зоны, рас- положенной ниже перегородки. Конструк- ция нижней части реактора вместе с бар- ботером приведена на рисунке 3. Другим путем снизить или полностью исключить возможность взрывоопасной ситуации при синтезе ГАС является изме- нение конструкции реактора и использова- ние раздельного ввода водорода и оксида азота (II). Это изменение конструкции можно осуществить двумя путями. В первом случае реконструкцию осу- ществляют без существенного изменения основной конструкции трехфазного реак- тора — реконструкция касается только барботера для раздельного подвода компо- нентов синтез-газа [4]. Трехфазный реак- тор смешения имеет объем 50 м3 и снабжен жалюзийной мешалкой. Частота вращения мешалки 110 об/мин, мощность привода 110 кВт, диаметр мешалки 1.25 м. Реактор состоит из эллиптического днища, эллип- тической крышки и цилиндрического кор- пуса, соединенных фланцами. На крышке, днище и корпусе монтируются устройства, необходимые для функционирования реак- тора. Теплообмен в реакторе 41 Промышленные процессы и технологии. 2022. Т. 2. № 2(4) А. В. Артемов, М. В. Дюбанов, С. В. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Ардамаков Рисунок 4 — Трехфазный реактор смешения для синтеза ГАС гидрированием оксида азота (II) в растворе серной кислоты с раздельным вводом газообразных реагентов в реакционную зону. Рисунок 3 — Схема ввода газообразных реагентов в трехфазный реактор смешения для получения гидроксиламинсульфата при горизонтальном расположении перегородки. 1 — нижнее днище корпуса реактора; 2 — нижняя зона газораспределительного устройства (барботера); 3 — труба для ввода катализатора; 4 — мешалка; 5 — патрубок для ввода водорода; 6 — патрубок для ввода оксида азота (II); 7 — перегородка; 8 — верхняя зона газораспределительного устройства (барботера). Рисунок 4 — Трехфазный реактор смешения для синтеза ГАС гидрированием оксида азота (II) в растворе серной кислоты с раздельным вводом газообразных реагентов в реакционную зону. Такая последовательность ввода газо- вых компонентов (ввод водорода располо- жен ниже ввода оксида азота (II)) обуслов- лена физико-химическими свойствами га- зов и механизмом протекания этой катали- тической реакции [5, 6], в соответствии с которым начальной стадией является вза- имодействие водорода с активным цен- тром катализатора с образованием адсор- бированного комплекса: Сборку реактора проводят следующим образом. На эллиптическое днище (3) с присоединенными (приваренными) к нему штуцером ввода жидких продуктов (27) и технологическим штуцером (28) помещают полую тарелку (5) в сборе с заглушкой (29) и трубой малого диаметра (8). Всю кон- струкцию (3, 27, 28, 5, 29, 8) закрепляют с помощью фланцевого соединения (4). По- лая тарелка имеет сквозные отверстия (7) для изолированного прохода суспензии ка- тализатора в водном растворе серной кис- лоты и продуктов реакции и отверстия (6) для ввода водорода в реакционный объем. Труба малого диаметра (8) в нижней части над полой тарелкой (5) имеет кольцевой упор (9), который с внешней стороны имеет резьбу для присоединения к нему трубы среднего диаметра (10). Верхняя часть трубы малого диаметра (8) имеет резьбу для присоединения к ней с Kt + H2 « [Kt…H2] с последующим взаимодействием этого комплекса с оксидом азота (II): [Kt…H2] + NO « [Kt…H2…NO]. [Kt…H2] + NO « [Kt…H2…NO]. Во втором случае для раздельного ввода Н2 и NO значительной реконструк- ции подвергают сам реактор [7]. Схема ре- конструированного реактора приведена на рисунке 4. Реактор имеет цилиндрический корпус (1) и эллиптическую крышку (2). Снаружи корпуса (1) имеется рубашка в виде змеевика из полутруб для охлажде- ния реакционного объема (на рисунке 4 не показана). Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 42 Взрывобезопасный каталитический синтез гидроксиламинсульфата помощью устройства (18) штуцера для подвода водорода (17). H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС образом, чтобы обеспечить свободное вра- щение трубы большого диаметра (14) с за- крепленной на ней мешалкой (15). К устройству (21) с помощью фланцево-резь- бового соединения (20) подсоединяют устройство (18), имеющее штуцер для ввода водорода (17). Демонтаж реактора проводят в обратном порядке. Трубу среднего диаметра (10) закреп- ляют с помощью резьбового соединения, имеющегося в нижней части трубы (10), с кольцевым упором (9). Труба среднего диаметра имеет в нижней части кольцевой упор (12), а в верхней части опору (23). Между кольцевыми упорами (9) и (12) труба среднего диаметра (10) имеет равно- мерно распределенные отверстия (11) для ввода оксида азота (II) в реакционный объем. Верхняя часть трубы среднего диа- метра (10) имеет резьбу для последующего присоединения к ней с помощью устрой- ства (21) и фланцево-резьбового соедине- ния (20) штуцера (19) для подвода оксида азота (II). Реактор работает следующим образом (см. рисунок 4). В реактор через штуцер (27) вводят суспензию катализатора в вод- ном растворе серной кислоты и продуктов реакции. Эта суспензия через сквозные от- верстия (7) поступает в реакционный объем, расположенный выше полой та- релки (5). В этот же реакционный объем через отверстия (6), расположенные в верхней части полой тарелки (5), посту- пает водород, а через отверстия (11), рас- положенные в нижней части трубы сред- него диаметра (10), поступает оксид азота (II). Суспензия, водород и оксид азота (II) интенсивно перемешиваются в реакционном объеме с помощью мешалки (15). Пенно-газовый слой разрушается с помощью пеногасителя (16) и газообраз- ные продукты выводятся из реактора с по- мощью штуцера (25). Жидкие продукты реакции выводятся из реактора с помощью штуцера (26) и направляются во внешнюю циркуляционную систему (на рисунке 4 не показана), где охлаждаются с помощью дополнительного охлаждающего устрой- ства (на рисунке 4 не показан). Часть жид- ких продуктов реакции рециркулируют с помощью штуцера (27) в реактор, а остав- шуюся часть направляют либо в другой реактор каскада, либо выводят в качестве продукта реакции. На кольцевой упор (12) трубы сред- него диаметра (10) устанавливают опор- ный подшипник (13). На опорный подшип- ник (13) надевают трубу большого диа- метра (14), на которой закреплена ме- шалка (15). На рисунке 4 показана одна мешалка; на трубе большого диаметра мо- гут быть закреплены несколько мешалок одного или разных типов. После чего с по- мощью одного из фланцевых соединений (4) устанавливают цилиндрический корпус (1), на котором смонтирован тарельчатый пеногаситель (16) и штуцер (26) для вы- вода жидких продуктов. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Далее, с помо- щью другого фланцевого соединения (4) устанавливают эллиптическую крышку (2), на которой смонтирован штуцер (25) вывода газообразных продуктов. Затем на трубу большого диаметра (14) жестко за- крепляют элемент привода перемешиваю- щего устройства (24) и сверху между тру- бами среднего (10) и большого (14) диа- метра устанавливают подшипник (22) и за- крепляют его с помощью устройства (21), например, с помощью резьбового соедине- ния (на рисунке 4 не показано). Закрепле- ние подшипника (22) проводят таким Преимущества конструкции реактора, приведенной на рисунке 4, заключаются в следующем: 1) раздельная подача в реакционный объем водорода и оксида азота (II) 43 Промышленные процессы и технологии. 2022. Т. 2. № 2(4) А. В. Артемов, М. В. Дюбанов, С. В. Ардамаков реактор (конструкции барботера и трех- фазного реактора смешения). Вместе с этим и состав продуктов реакции, прежде всего повышенное содержание оксида азота (I), может быть причиной взрыво- опасности процесса. Поэтому процесс син- теза ГАС ведут обычно при содержании N2O в газообразных продуктах реакции в интервале 0.10–0.11 мольных долей (м.д.). При увеличении содержания N2O до зна- чения большего или равного 0.16 м.д,. ре- акторный блок отключают. Повышенное содержание N2O в газообразных продук- тах реакции возникает в основном из-за протекания побочной реакции: исключает вероятность возникновения взрывоопасной ситуации; 2) включение в систему циркуляции дополнительного охлаждающего устрой- ства позволяет надежно регулировать тем- пературный режим процесса и снижает до минимума вероятность протекания побоч- ных процессов; 3) интенсивное перемешивание реак- ционной массы достигается за счет исполь- зования внутреннего перемешивания и внешней циркуляционной системы, что позволяет проводить процесс в области, близкой к кинетической; 4) высокая надежность конструкции реактора обеспечивается наличием трех фиксирующих и опорных элементов для внутренних конструкций реактора: фикси- рование полой тарелки во фланце, соеди- няющем эллиптическое днище и цилин- дрический корпус; опорные кольца на тру- бах малого и среднего диаметров; 2NO + H2 = N2O + H2O, 2NO + H2 = N2O + H2O, что связано со снижением активности ка- тализатора. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Поэтому управление процес- сом каталитического синтеза ГАС, обеспе- чивающее взрывобезопасность процесса как по составу синтез-газа (NO + H2), так и по содержанию N2O в газообразных про- дуктах реакции, и проведение самого про- цесса состоит из следующих основных ста- дий [8]: 5) наличие в конструкции реактора пеногасителей снижает вероятность обра- зования пенно-газового слоя и выноса ка- тализатора; 1) при заданном избыточном давле- нии синтеза ГАС проводят расчет содер- жания водорода в синтез-газе, соответ- ствующего предельному содержанию водо- рода в синтез-газе, обеспечивающего взры- вобезопасное ведение процесса и готовят синтез-газ с этим содержанием водорода; 6) простота сборки и демонтажа реак- тора; 7) в реакционном объеме реактора, расположенном выше полой тарелки, от- сутствуют какие-либо конструкционные элементы, способствующие образованию застойных зон, что увеличивает эффектив- ность действия катализатора; 2) проводят сравнение рассчитанного содержания водорода с заданным техноло- гическим регламентом интервалом допу- стимого содержания водорода в синтез- газе; 8) раздельный и перекрестный ввод реагирующих предварительно диспергиро- ванных газов в реакционный объем способ- ствует увеличению скорости реакции и ее протеканию в области, близкой к кинети- ческой. 3) если содержание водорода в синтез- газе удовлетворяет и условию взрывобез- опасности и условию синтеза ГАС, то при- готовленный синтез-газ направляют в ре- акторы каскада; Обсужденные выше методы взрыво- безопасного каталитического синтеза ГАС касаются анализа входных параметров: давления, состава синтез-газа и раздель- ного ввода компонентов синтез-газа в 4) если содержание водорода в синтез- газе удовлетворяет только условию взры- вобезопасности и не удовлетворяет Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 44 Взрывобезопасный каталитический синтез гидроксиламинсульфата условию синтеза ГАС, то помимо направ- ления синтез-газа в реакторы каскада, независимо от ввода синтез-газа, по выде- ленной линии направляют дополнительное количество оксида азота (II), такое, чтобы суммарное количество водорода в реакто- рах каскада удовлетворяло условию син- теза ГАС; случае объемное отношение Н2 : NO в син- тез-газе больше 2 и выходит за рамки до- пустимого отношения для синтеза ГАС (Н2 : NO = (1.7–1.8):1). Поэтому по выде- ленной линии непосредственно в реакторы каскада подают дополнительное количе- ство NO. После этого содержание водо- рода в синтез-газе снижается до 0.64 м.д., что соответствует условию синтеза ГАС (Н2 : NO = 1.78, рисунок 1). 5) проводят постоянное определение содержания N2O в газообразных продук- тах синтеза ГАС с сохранением результа- тов определений; Таблица 2 — Состав газовой фазы на входе в реакторы синтеза ГАС. Ризб = 2 атм. Точка от- бора газо- вой фазы Ед. изм. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Ардамаков Таблица 4 — Состав газообразных продуктов синтеза ГАС (Х), мольные доли (м.д.) Номер п/п Компоненты газовой фазы NO H2 N2 N2O CH4 H2O Сумма (1) 0.14 0.54 0.10 0.09 0.07 0.06 1.00 (2) 0.14 0.53 0.10 0.10 0.07 0.06 1.00 (3) 0.14 0.52 0.10 0.11 0.08 0.05 1.00 (4) 0.14 0.51 0.09 0.12 0.08 0.06 1.00 а 4 — Состав газообразных продуктов синтеза ГАС (Х), мольные доли (м.д.) 3) дополнительно вводить необходи- мое количество оксида азота (II) непосред- ственно в реакторы каскада без контакта с синтез-газом и водородом, что повышает взрывобезопасность процесса; Как видно из приведенных в таблице 3 данных, величина ХN2O имеет не только положительный градиент: 3) дополнительно вводить необходи- мое количество оксида азота (II) непосред- ственно в реакторы каскада без контакта с синтез-газом и водородом, что повышает взрывобезопасность процесса; р д X(2) N2O – X(1) N2O = 0.09 – 0.08 = + 0.01; X(3) N2O – X(2) N2O = 0.08 – 0.09 = – 0.01; X(4) N2O – X(3) N2O = 0.09 – 0.08 = + 0.01, что удовлетворяет условиям ведения син- теза ГАС. X(2) N2O – X(1) N2O = 0.09 – 0.08 = + 0.01; X(3) N2O – X(2) N2O = 0.08 – 0.09 = – 0.01; X(4) N2O – X(3) N2O = 0.09 – 0.08 = + 0.01, 4) использовать эти технические ре- шения для совершенствования технологии других каталитических процессов взаимо- действия двух газообразных реагентов, об- разующих взрывоопасные смеси (напри- мер, Н2 + О2 или Н2 + СО) при проведе- нии процесса в трехфазных реакторах сме- шения («сларри-реакторах»). что удовлетворяет условиям ведения син- теза ГАС. В другом случае (таблица 4) при опре- делении содержания N2O в газообразных продуктах синтеза ГАС в серии из после- довательных испытаний обнаружен только положительный градиент: X(2) N2O – X(1) N2O = 0.10 – 0.09 = + 0.01; X(3) N2O – X(2) N2O = 0.11 – 0.10 = + 0.01; X(4) N2O – X(3) N2O = 0.12 – 0.11 = + 0.01. X(2) N2O – X(1) N2O = 0.10 – 0.09 = + 0.01; X(3) N2O – X(2) N2O = 0.11 – 0.10 = + 0.01; X(4) N2O – X(3) N2O = 0.12 – 0.11 = + 0.01. H ≧ 0.66 + 0.074 ∙ lnP. Решения по управлению каталитиче- ским синтезом ГАС Компоненты газовой фазы NO H2 Сумма В емкости- смесителе при условии взрывобез- опасной кон- центрации водорода м.д. 0.29 0.71 1.00 кмоль 24.78 79.02 103.8 кг 734.40 158.04 892.44 Дополнитель- ное количе- ство оксида азота (II) кг 321.11 0.11 321.22 Суммарное количество газовой фазы в каждом ре- акторе кас- када кг 1055.56 158.15 1213.71 кмоль 35.19 79.08 114.27 м.д. 0.36 0.64 1.00 6) если содержание N2O больше или равно 0.16 м.д., то проводят остановку ре- акторного блока из-за превышения содер- жания N2O предельного значения; 7) если содержания N2O меньше 0.16 м.д., то проводят расчет градиента из- менения концентрации N2O в реакционных газах по результатам по крайней мере че- тырех последовательных измерений содер- жания N2O. Если градиент устойчиво по- ложительный, то проводят замену катали- затора и корректировку режима процесса, начиная с расчета содержания водорода в синтез-газе при установившемся в про- цессе давлении. Результаты использования предлагае- мого алгоритма приведены ниже. В таб- лице 2 приведен состав газовой фазы (без учета примесных компонентов) на входе в реактор синтеза ГАС при избыточном дав- лении 2 атм. Одновременно проводят многократное определение содержания N2O в газообраз- ных продуктах синтеза ГАС. Результаты этих определений приведены в таблице 3. Как видно из таблицы 2 в емкости-сме- сителе при условии взрывобезопасной кон- центрации Н2 его предельное (наимень- шее) значение равно 0.71 м.д. В этом Таблица 3 — Состав газообразных продуктов синтеза ГАС (Х), мольные доли (м.д.) Номер п/п Компоненты газовой фазы NO H2 N2 N2O CH4 H2O Сумма (1) 0.13 0.54 0.10 0.08 0.08 0.07 1.00 (2) 0.12 0.56 0.09 0.09 0.07 0.07 1.00 (3) 0.10 0.60 0.09 0.08 0.07 0.06 1.00 (4) 0.13 0.53 0.10 0.09 0.08 0.07 1.00 Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 45 А. В. Артемов, М. В. Дюбанов, С. В. СПИСОК ЛИТЕРАТУРЫ [1]. В. И. Герасименко, А. А. Огар- ков, С. В. Ардамаков, И. В. Лукьянов, А. В. Артемов. Технологические аспекты синтеза гидроксиламинсульфата. Россий- ский химический журнал, 2006, Т. 50, №3, с.64-71. [1]. В. И. Герасименко, А. А. Огар- ков, С. В. Ардамаков, И. В. Лукьянов, А. В. Артемов. Технологические аспекты синтеза гидроксиламинсульфата. Россий- ский химический журнал, 2006, Т. 50, №3, с.64-71. В этих условиях проводят корректировку режима процесса, начиная с расчета со- держания водорода в синтез-газе при уста- новившемся в процессе давлении и прово- дят замену катализатора. Выводы Промышленные процессы и технологии. 2022. Т. 2. № 2(4) Выводы [2]. В. И. Герасименко, С. В. Ардама- ков, И. В. Лукьянов. Патент РФ 2690931, МПК С01В21/14. Способ получения гид- роксиламинсульфата. Предлагаемые решения по управле- нию каталитическим синтезом ГАС позво- ляют: 1) значительно упростить управление за счет создания простого и понятного ал- горитма; [3]. А. К. Чернышев, В. И. Гераси- менко, Б. А. Сокол и др. Капролактам: свойства, производство, применение. – Москва: Междунар. науч.-инормац. ком- пания "Инфохим", 2016, Т. 1, С. 396–402 2) одновременно с поддержанием взрывобезопасной концентрации NO в син- тез-газе анализировать содержание N2O в газообразных продуктах синтеза, оцени- вать стабильность работы реакторного блока, необходимость замены катализа- тора и необходимость останова реактор- ного блока; [4]. К. В. Мукалин. Реактор синтеза гидроксиламинсульфата. Катализ в про- мышленности, 2007, №6, С.25-29 [5]. В. Л. Лопатин, Н. В. Кулькова, М. И. Темкин. О механизме гидрирования [5]. В. Л. Лопатин, Н. В. Кулькова, М. И. Темкин. О механизме гидрирования Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 46 Взрывобезопасный каталитический синтез гидроксиламинсульфата окиси азота. Кинетика и катализ, 1982, №4, С. 863–867. гидроксиламинсульфата. Известия Са- марского научного центра Российской ака- демии наук, 2007, Специальный выпуск, С. 179–182. окиси азота. Кинетика и катализ, 1982, №4, С. 863–867. [6]. Н. Н. Саводник, Н. В. Кулькова, Д. М. Дохолов, В. Л. Лопатин, М. И. Тем- кин. Кинетика каталитического синтеза гидроксиламина из окиси азота и водо- рода. Кинетика и катализ, 1972, №6, С. 1520–1526. [8]. С. В. Ардамаков, А. В. Гераси- менко, И. В. Лукьянов Патент РФ 2702575 МПК С01В 21/14 Способ управления про- цессом синтеза гидроксиламинсульфата, опубл. 08.10.2019, БИ № 28. [7]. К. В. Мукалин, В. П. Щукин. Оп- тимизация технологической схемы синтеза Артемов Арсений Валерьевич — доктор химических наук, профессор, ведущий научный сотрудник Национального исследовательского центра «Курчатовский инсти- тут» (Российская Федерация, 123182, Москва, пл. Академика Курчатова, 1). Дюбанов Михаил Викторович — начальник отдела материалов Межведомствен- ного центра аналитических исследований в области физики, химии и биологии при Пре- зидиуме Российской академии наук (Российская Федерация, 117342, ул. Профсоюзная, д.65, стр.6). Дюбанов Михаил Викторович — начальник отдела материалов Межведомствен- ного центра аналитических исследований в области физики, химии и биологии при Пре- зидиуме Российской академии наук (Российская Федерация, 117342, ул. Профсоюзная, д.65, стр.6). Ардамаков Сергей Витальевич — директор ООО «Волгатехноол» (Российская Федерация, 445007, Тольятти, ул. Новозаводская, 6) Ардамаков Сергей Витальевич — директор ООО «Волгатехноол» (Российская Федерация, 445007, Тольятти, ул. Новозаводская, 6) 47 47 А. В. Артемов, М. В. Дюбанов, С. В. Ардамаков Explosion-proof catalytic synthesis of hydroxylamine sulfate A. V. Artemov,1, M. V. Dubanov, S. V. Ardamakov** * National Research Center «Kurchatov Institute», Moscow, Russia Interdepartmental Center for Analytical Research in Physics, Chemistry and Biology at the Presidium of the Russian Academy of Sciences, Moscow, Russia * LLC «Volgatekhnool», Togliatty, Russia * National Research Center «Kurchatov Institute», Moscow, Russia Interdepartmental Center for Analytical Research in Physics, Chemistry and Biology at the Presidium of the Russian Academy of Sciences, Moscow, Russia * LLC «Volgatekhnool», Togliatty, Russia Keywords hydroxylamine sulfate catalytic synthesis, process explosion safety, process control algo- rithm, reactor design. Abstract The conditions for the synthesis of hydroxylamine by reaction of hydrogen and nitric oxide in aqueous sulfuric acid in a cascade reactor system composed of three-phase mixing reactors in the presence of a platinum-on-electrographite catalyst, have been analyzed. The factors that provide the explosion safety of the process have been studied, such as pressure, H2 : NO ratio, and N2O content in the gaseous synthesis products. The design of the reactors, which enables the explosion safety of the process, is described. An algorithm for the process control is proposed. Data on the composition of the gas phase mixture entering the reactors for hydroxylamine sulfate (HS) synthesis, the composition of the gaseous products of HS syn- thesis, along with the industrial scale HS synthesis conditions and outcomes are presented. The proposed solution for the control the catalytic synthesis of hydroxylamine sulfate makes it possible to a great extent simplify the process management by creating a simple and understandable algorithm together with maintaining an explosion-proof NO concentration in the synthesis gas. It also provides the opportunity to analyze the N2O content in the gaseous synthesis products, to evaluate the operation stability of the reactor unit and the necessity of the catalyst replacement and shutdown of the reactor system when required. Additional introduction of the required amount of nitric oxide directly into the cascade reactors without the contact with synthesis gas and hydrogen further improves the explosion safety of the entire process. 1 Corresponding author: Email: arsenyart@icloud.ru Khim. Zh.], 2006, vol. 50, №3, pp. 64-71. (In Russ.) Промышленные процессы и технологии. 2022. Т. 2. № 2(4) REFERENCES Khim. Zh.], 2006, vol. 50, №3, pp. 64-71. (In Russ.) [1]. Gerasimenko V. I, Ogarkov A. A., Ardamakov S. V., Luk’yanov I. V., Arte- mov A. V. Tekhnologicheskiye aspekty sin- teza gidroksilaminsul'fata [Technological As- pects of the Synthesis of Hydroxylamine] Rossiyskiy khimicheskiy zhurnal [Ross. [2]. Gerasimenko V. I, Ogarkov A. A., Ardamakov S. V., Luk’yanov I. V. Method for obtaining hydroxylamine sulfate. Pat. 14329 Ukraine (In Russ.) 1 Corresponding author: Email: arsenyart@icloud.ru Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 48 Взрывобезопасный каталитический синтез гидроксиламинсульфата [3]. Chernyshev A. K., Gerasimenko V. I., Sokol B. A. [et al.] Kaprolaktam: svoystva, proizvodstvo, primeneniye [Capro- lactam: properties, production, application] Moscow, Infokhim Publ., 2016, Vol. 1, pp. 396–402 (In Russ.) gidroksilamina iz okisi azota i vodoroda [Ki- netics of catalytic synthesis of hydroxylamine from nitric oxide and hydrogen]. Kinetika i kataliz [Kinetics and Catalysis], 1972, №6, pp. 1520–1526. [7]. Mukalin K. V., Schukin V. P. Opti- mizatsiya tekhnologicheskoy skhemy sinteza gidroksilaminsul'fata [Optimization of tech- nological scheme for synthesis of hydroxyla- mine sulfate]. Izvestiya Samarskogo nauch- nogo tsentra Rossiyskoy Federatsii akademii nauk [Proceedings of Samara Scientific Cen- ter of Russian Academy of Sciences], 2007, Spec. issue, pp. 179–182. (In Russ.) [7]. Mukalin K. V., Schukin V. P. Opti- mizatsiya tekhnologicheskoy skhemy sinteza gidroksilaminsul'fata [Optimization of tech- nological scheme for synthesis of hydroxyla- mine sulfate]. Izvestiya Samarskogo nauch- nogo tsentra Rossiyskoy Federatsii akademii nauk [Proceedings of Samara Scientific Cen- ter of Russian Academy of Sciences], 2007, Spec. issue, pp. 179–182. (In Russ.) [4]. Mukalin K. V. Reaktor sinteza gidroksilaminsul'fata [Hydroxylamine sulfate synthesis reactor] Kataliz v promyshlennosti [Catalysis in Industry], 2007, № 6, pp. 25–29 (In Russ.) [5]. Lopatin V. L., Kul’kova N. V., Temkin M. I. O mekhanizme gidrirovaniya okisi azota [On mechanism of nitric oxide hy- drogenation]. Kinetika i kataliz [Kinetics and Catalysis], 1982, №4, pp. 863–867. (In Russ.) [8]. Ardamakov S.V., Gerasimenko V.I Luk’yanov I.V. Method for controlling hy- droxylamine sulfate synthesis process. Pat. 2702575, Russian Federation, 2019. (In Russ.) [8]. Ardamakov S.V., Gerasimenko V.I Luk’yanov I.V. Method for controlling hy- droxylamine sulfate synthesis process. Pat. 2702575, Russian Federation, 2019. (In Russ.) [6]. Savodnik N. N., Kul’kova N. V., Dokholov D. M., Lopatin V. L., Tem- kin M. I. Kinetika kataliticheskogo sinteza Artemov A. V. — Dr. Sc. (Eng.), Professor, Leading Researcher, National Research Center "Kurchatov Institute" (Academician Kurchatov Square, 1, Moscow, 123182, Russian Federation) Dubanov M. V. Просьба ссылаться на эту статью следующим образом: Взрывобезопасный каталитический синтез гидроксиламинсульфата / А. В. Артемов, М. В. Дюбанов, С. В. Ардамаков // Промышленные процессы и технологии. 2022. Т. 2. № 2(4). С. 38 – 49. DOI: 10.37816/2713-0789-2022-2-2(4)-38-49 REFERENCES — Head Of Materials Department, Interdepartmental Center for Ana- lytical Research in Physics, Chemistry and Biology at the Presidium of the Russian Academy of Sciences (Profsoyuznaya st., 65, building 6, Moscow, 117342, Russian Federation) Dubanov M. V. — Head Of Materials Department, Interdepartmental Center for Ana- lytical Research in Physics, Chemistry and Biology at the Presidium of the Russian Academy of Sciences (Profsoyuznaya st., 65, building 6, Moscow, 117342, Russian Federation) Ardamakov S. V. — director of LLC "Volgatekhnool" (Novozavodskaya str., 6, Togli- atty, 445007, Russian Federation) Ardamakov S. V. — director of LLC "Volgatekhnool" (Novozavodskaya str., 6, Togli- atty, 445007, Russian Federation) Please cite this article as: Artemov A. V., Dubanov M. V., Ardamakov S. V. Explosion-proof catalytic synthesis of hydroxylamine sulfate. Industrial processes ang Technologies, 2022, vol. 2, no. 2(4), pp. 38 – 49 DOI: 10.37816/2713-0789-2022-2-2(4)-38-49 Artemov A. V., Dubanov M. V., Ardamakov S. V. Explosion-proof catalytic synthesis of hydroxylamine sulfate. Industrial processes ang Technologies, 2022, vol. 2, no. 2(4), pp. 38 – 49 DOI: 10.37816/2713-0789-2022-2-2(4)-38-49 Промышленные процессы и технологии. 2022. Т. 2. № 2(4) 49 Промышленные процессы и технологии. 2022. Т. 2. № 2(4)
https://openalex.org/W2804189069	http://www.jisem-journal.com/download/predicting-software-flaws-with-low-complexity-models-based-on-static-analysis-data.pdf	English	null	Predicting Software Flaws with Low Complexity Models based on Static Analysis Data	Journal of information systems engineering & management	2,018	cc-by	5,706	ABSTRACT Due to the constant evolution of technology, each day brings new programming languages, development paradigms, and ways of evaluating processes. This is no different with source code metrics, where there is always new metric classes. To use a software metric to support decisions, it is necessary to understand how to perform the metric collection, calculation, interpretation, and analysis. The tasks of collecting and calculating source code metrics are most often automated, but how should we monitor them during the software development cycle? Our research aims to assist the software engineer to monitor metrics of vulnerability threats present in the source code through a reference prediction model, considering that real world software have non-functional security requirements, which implies the need to know how to monitor these requirements during the software development cycle. As a first result, this paper presents an empirical study on the evolution of the Linux project. Based on static analysis data, we propose low complexity models to study flaws in the Linux source code. About 391 versions of the project were analyzed by mining the official Linux repository using an approach that can be reproduced to perform similar studies. Our results show that it is possible to predict the number of warnings triggered by a static analyzer for a given software project revision as long as the software is continuously monitored. Keywords: source code static analysis, source code metrics, common weakness enumeration, prediction, linux Journal of Information Systems Engineering & Management, 2018, 3(2), 17 ISSN: 2468-4376 Lucas Kanashiro 1, Athos Ribeiro 1, David Silva 2, Paulo Meirelles 1, 3, Antonio Terceiro 4 Lucas Kanashiro 1, Athos Ribeiro 1, David Silva 2, Paulo Meirelles 1, 3, Antonio Terceiro 4 1 FLOSS Competence Center (CCSL), University of São Paulo (USP), São Paulo, BRAZIL 2 UnB Faculty in Gama (FGA), University of Brasília (UnB), Brasília, BRAZIL 3 Department of Health Informatics, Federal University of São Paulo (UNIFESP), São Paulo, BRAZIL 4 QA Services Team, Linaro Limited, Curitiba, BRAZIL Published: April 14, 2018 Published: April 14, 2018 Copyright © 2018 by Author/s and Licensed by Lectito BV, Netherlands. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. *Corresponding Author: paulormm@ime.usp.br Citation: Kanashiro, L., Ribeiro, A., Silva, D., Meirelles, P. and Terceiro, A. (2018). Predicting Software Flaws with Low Complexity Models based on Static Analysis Data. Journal of Information Systems Engineering & Management, 3(2), 17. https://doi.org/10.20897/jisem.201817 Citation: Kanashiro, L., Ribeiro, A., Silva, D., Meirelles, P. and Terceiro, A. (2018). Predicting Software Flaws with Low Complexity Models based on Static Analysis Data. Journal of Information Systems Engineering & Management, 3(2), 17. https://doi.org/10.20897/jisem.201817 RELATED WORKS Evans and Larochelle (2002) states that although security vulnerabilities are well understood, it is not a common practice to include techniques to detect or avoid them in the development processes and suggests that instead of solely relying on programmers, tools to prevent and detect software vulnerabilities should be built and integrated in software development cycles. It is also known that static analysis helps programmers detect software vulnerabilities before they reach production (Evans and Larochelle, 2002) in a cheaper way than solely conducting manual inspections (Johnson et al., 2013), thus the interest in static analyzers. Zheng et al. (2006) analyzed the effectiveness of security oriented static analyzers using tests and the number of failures reported by customers as parameters. They concluded that static analyzers are effective tools to find flaws in source code. However, a solution on how to monitor such flaws is not presented. Our research aims to find solutions for this issue. Pan et al. (2006) proposes a method to predict if a file or function in a given project is buggy or bug-free based on the correlation between code quality metrics, like size, complexity and coupling, and software bugs. The authors used static analyzers to extract data from 76 revisions of Latex2rtf and 887 revisions of Apache HTTP so they could use the 10-fold cross-validation method with a machine learning algorithm to perform their predictions. The prediction precision was over 80% for files and over 70% for functions. The authors declared that the costs to run the static analysis to generate the method inputs were too high for large projects. In this paper we are testing low cost functions to predict some flaws, but running the analyses is still needed. In another work, Penta et al. (2008) performed an analysis of the evolution of vulnerabilities detected by static analyzers in three different Free Software applications: Samba, Squid, and Horde. The authors analyzed how the number of vulnerabilities varies over time. Since they analyzed different development versions (not only releases), they could analyze aspects of the development process, like bug fixing efforts right before a release. The focus of this work was to understand how long vulnerabilities tend to remain in the system by modeling the vulnerabilities decay time with a statistical distribution and comparing the decay time of different classes of vulnerabilities. Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data Although no standard to define software flaws is widely accepted, some organizations define and classify them. The most accepted form is the one defined by MITRE, where software common flaws are cataloged as CWEs1 (Common Weakness Enumeration). This catalog establishes a unique language to describe known software flaws, like divisions by zero and buffer overflows. There are several static analyzers that can detect flaws (Black, 2001). With the support of such tools, this work presents a method to define low complexity models for the amount of flaws found in a software through a case study of two different types of common weaknesses detected in Linux: Use of Uninitialized Variable and NULL pointer Dereference. Thus, we have obtained low complexity polynomial functions to describe the number of these flaws present in Linux over time. To show our ideas and results in this context, the remainder of this paper is organized as follows: the next section describes several related works investigated in this research. The third section shows the research design, discussing our research question as well as presenting the data collection and analysis approach. The fourth section presents the candidate low complexity models to predict flaws in Linux. The fifth section discusses the results, selecting and applying the candidate low complexity models. The last section concludes the paper, highlighting its main contributions and pointing to possibilities for future work. RELATED WORKS Our research is based on the hypothesis that to effectively take action on such bug fixing effort predictions, monitoring the number of flaws or the flaws decay time should be done continuously, during the software development cycle. This work is a first step towards such automation. First, we present a model to predict the number of flaws present in a specific version of the Linux kernel. Then, we show that such model must be continuously updated to be effective, i.e., the greater the difference between the last version used to train the model and the version used for the predictions, the less effective are the predictions made. 1 cwe.mitre.org Q1 – Is it possible to define low complexity models to predict flaws in the Linux source code? INTRODUCTION Source code static analysis is a good means to provide inputs to support software quality assurance. For instance, these inputs may be software structural attributes such as object-oriented metrics as presented by Chidamber and Kemerer (2002). Attributes related to the possible behaviors of the software at execution time (Ernst, 2005), which includes analyses of behaviors that may lead to security issues, could also serve as input provided for quality assurance (Ferzund et al., 2009; Misra and Bhavsar, 2003; Nagappan et al., 2006). From the source code static analysis point of view, a vulnerability is the result of one or more flaws in software requirements, design, implementation, or deployment (Black et al., 2007). Vulnerability correction can be very expensive in later stages of a software development cycle, hence the importance of finding and correcting flaws in early stages, before they expose actual vulnerabilities. Copyright © 2018 by Author/s and Licensed by Lectito BV, Netherlands. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Data Collection Although it is common to refer to the whole operating system as Linux, the name refers to the kernel. Its development has been active for over 20 years and it is written in the C programming language. The Linux source code showed itself adequate for this study due to its size and the fact that it is widely used around the world, which led us to believe there is a higher likelihood of finding flaws in it. A total of 391 versions of the kernel were analyzed. These version were downloaded from the project's official Git repositories2. By the time of our first analyses, the versions between 2.6.11 and 3.9 were available. After that, we have tested the candidate models in three recent kernel versions: 4.0, 4.5, and 4.9. The static analyzer selected was Cppcheck. According to the description in its homepage3, it aims to be a sound (Black, 2001) static analyzer, since its main goal is to find real bugs in the source code, not generating false positives. The main characteristic of sound static analyzers is the low rate of false positives. The Cppcheck output is composed of a location (source code line) and a message, describing a problem (warning or error). We ran Cppcheck 1.67 in each version of Linux and the corresponding output files for each of the analyzed versions are publicly available at github.com/lucaskanashiro/linux-analysis. RESEARCH DESIGN To guide the development of a reproducible method to define low complexity functions that enable the longitudinal study of software flaws, we addressed the following research question: Q1 – Is it possible to define low complexity models to predict flaws in the Linux source code? 1 cwe.mitre.org © 2018 by Author/s 2 / 9 2 / 9 Journal of Information Systems Engineering & Management, 3(2), 17 Figure 1. CWE 476 dereferences evolution Figure 1. CWE 476 dereferences evolution With low complexity functions, we are able to automate the continuous monitoring of flaws in a simple way, without the need to build complex models, which would be expensive to maintain. If we can easily infer the number of flaws predicted in a software as well, improvements in the software development cycle based on it become possible. p We could, for example, make better decisions on when to increase the efforts to fix bugs or refactor code. 2 git.kernel.org 3 cppcheck.sourceforge.net 4 Bash, Blender, FFmpeg, Firefox, Gstreamer, Inetutils, OpenSSH, OpenSSL, Python2.7, and Ruby-2.1 cppcheck.sourceforge.net 4 Bash, Blender, FFmpeg, Firefox, Gstreamer, Inetutils, OpenSSH, OpenSSL, Python2.7, and Ruby-2.1 3 cppcheck.sourceforge.net 4 Bash, Blender, FFmpeg, Firefox, Gstreamer, Inetutils, OpenSSH, OpenSSL, Python2.7, and Ruby-2.1 Data Analysis The risk of Linux versions with more modules to stand out compared to other versions was identified upon analyzing the absolute number of flaws found. Thus, the number of flaws was normalized based on the number of modules in each version, resulting in the flaw rate per module that is our dependent variable. To analyze the influence of the project growth, we used the total number of modules as our independent variable. In this context, a kernel module is one C programming language source code file (header files are not counted for this matter). Since there are several known flaws, we have conducted a previous study with 10 popular Free Software projects4 from different domains to find the most recurrent flaws and select them to perform the analyses for this work. Since this paper proposes to find one model per flaw analyzed, we selected the two most frequent flaws found in those projects: NULL pointer dereference and use of uninitialized variable. Both are cataloged by MITRE and as Common Weakness Enumerations, the former as CWE476 and the latter as CWE457. 3 / 9 © 2018 by Author/s Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data Figure 2. CWE 457 dereferences evolution Figure 3. 4-Plot for CWE 476 – Number of NULL pointer Figure 2. CWE 457 dereferences evolution Figure 3. 4-Plot for CWE 476 – Number of NULL pointer Figure 3. 4-Plot for CWE 476 – Number of NULL pointer As we can see in Figures 1 and 2, in the course of Linux evolution, the total number of flaws dereferences does not always increase. It increases until a certain point and then there is a tendency towards stabilization, which is, most likely, the result of the Linux development process evolution. To understand the behavior of the data we performed some analyses, such as correlation matrix between the flaws and the number of modules, using the Pearson correlation coefficient (Rossman, 1996). Also, to confirm that the data can not be represented by a linear function, we used the 4-plot technique developed by NIST5. This technique facilitates the understanding of the data distribution and its randomness. 5 itl.nist.gov/div898/handbook/eda/section3/eda3332.htm Data Analysis For this study, identifying these characteristics was necessary to obtain information for the definition of a model which satisfies the studied data set, or that could, at least, eliminate unwanted options. For Linux, the data does not follow a normal distribution, but a long-tailed distribution, so that it is not constant and varies widely, as we can see in Figures 3 and 4. Both analyses showed that the flaw rates were not strongly correlated and they did not fit a normal distribution, which corroborates with the fact that exponential and weibull distributions can model the behavior of defect occurrence in a wide range of software (Penta, 2002; Jones, 1991; Li et al., 2004; Wood, 1996). With this in mind we discarded the possibility of using a linear model. Since our proposal is based on low complexity models and linear functions do not fit the data set, we investigated polynomial models. Initially, an identification of outliers (Hawkins, 1980) work was done for the definition of the polynomial models through a technique presented by Tukey (1977), using a box plot. The outliers found were removed to improve the candidate model accuracy. After that, parametric and non-parametric models © 2018 by Author/s 4 / 9 Journal of Information Systems Engineering & Management, 3(2), 17 Figure 4. 4-Plot for CWE 457 – Use of uninitialized variable were built. The non-parametric model is a black box method from which it is not possible to obtain a mathematical function. In this study, it is used as a base for the definition of a parametric model. As a result, we have a function that receives the total number of modules in the project as input and returns the flaw rate as output. We also have used the Locally Weighted Regression (LOESS) non parametric method to provide a smooth Figure 4. 4-Plot for CWE 457 – Use of uninitialized variable Figure 5. CWE 476 data models prediction curve Figure 4. 4-Plot for CWE 457 – Use of uninitialized variable Fi 5 CWE 476 d d l di i Figure 5. CWE 476 data models prediction curve were built. The non-parametric model is a black box method from which it is not possible to obtain a mathematical function. In this study, it is used as a base for the definition of a parametric model. Data Analysis As a result, we have a function that receives the total number of modules in the project as input and returns the flaw rate as output. were built. The non-parametric model is a black box method from which it is not possible to obtain a mathematical function. In this study, it is used as a base for the definition of a parametric model. As a result, we have a function that receives the total number of modules in the project as input and returns the flaw rate as output. We also have used the Locally Weighted Regression (LOESS) non-parametric method to provide a smooth surface using a non-parametric regression (Cleveland and Devlin, 1988). In short, the LOESS method applies several minor regressions to the data set as well as it can guide the definition of a possible parametric model. Therefore, we have applied polynomial regressions to obtain the parametric models definition for this study.Posteriorly, to validate the obtained models, we used the K-Fold cross-validation technique, which consists in dividing the data set in K groups, where one of these groups is used to test the model and the other groups are used to train the model (Picard and Cook, 1984). With this kind of cross-validation, we can test the model with values from different versions of the kernel, obtaining the model prediction error. We use the mean squared error, which guarantees the model reliability, given that this error has a low value. After the validation, the model errors were compared and we could chose the best model, getting to low complexity (polynomial) functions capable of predicting the Linux source code flaw rates proposed in this paper. Finally, we have tested the candidate models collecting new data from recent Linux versions (such as 4.0, 4.5, and 4.9) and comparing them to the predicted values from our low complexity polynomial functions LOW COMPLEXITY MODELS DEFINITION To achieve a lower error, we could use a polynomial function with a higher degree. However, to build a suitable model, we need to avoid overfitting, the extreme adjustment of the model to our data set, which would disturb the extrapolation for future inputs. Our approach based on a non-parametric model as reference helps avoiding a possible overfitting of the built model, providing a more flexible model. p g p g For the models definition, we identified and removed the outliers in the data set, defined the non-parametric model with the LOESS method and, finally, defined two parametric models through polynomial regression: one 5 / 9 © 2018 by Author/s © 2018 by Author/s Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data quadratic polynomial and one cubic polynomial. In Figures 5 and 6, we present the “real” flaw rate values and flaws evolution curves (”cubic” “quadratic” and “LOESS”) in the built models Figure 6. CWE 457 data models prediction curve Table 1. Error Obtained Through cross-validation Model CWE Mean squared error Quadratic 476 0.000000958 Cubic 476 0.000000862 Quadratic 457 0.000000137 Cubic 457 0.000000112 Figure 6. CWE 457 data models prediction curve Figure 6. CWE 457 data models prediction curve Figure 6. CWE 457 data models prediction curve Table 1. Error Obtained Through cross-validation Model CWE Mean squared error Quadratic 476 0.000000958 Cubic 476 0.000000862 Quadratic 457 0.000000137 Cubic 457 0.000000112 quadratic polynomial and one cubic polynomial. In Figures 5 and 6, we present the “real” flaw rate values and flaws evolution curves (”cubic”, “quadratic” and “LOESS”) in the built models. quadratic polynomial and one cubic polynomial. In Figures 5 and 6, we present the “real” flaw rate values and flaws evolution curves (”cubic”, “quadratic” and “LOESS”) in the built models. ( q Figure 5 shows that the built cubic model is the one with the best approximation for the reference model curve (LOESS model). It is especially good for data extrapolation on data sets higher or lower than the data interval used in this study. On the edges, the cubic model has a better approximation from the reference model, while the quadratic model tends to diverge from it. 6 These numbers can be verified by comparing the values in Table I, dividing the quadratic error values by the cubic ones LOW COMPLEXITY MODELS DEFINITION Still, although all the models presented in Figure 6 have a good approximation for the reference model curve, the cubic model fits the reference model better than the quadratic one, both in the minimum and maximum limits. Finally, to compare the models, we performed a cross-validation with the K-fold method to analyze the performance of both models in a data set that was not used for training those models. In short, we used a ten-fold cross-validation (K = 10) for this study, which means dividing the sample in ten groups for training and testing (Kohavi, 1995). Table 1 presents the mean squared error (MSE), associated to each of the models obtained through the cross-validation. For instance, the MSE measures the quality of a prediction model, so its values closer to zero are better, and they always are non-negatives. Therefore, the low error (MSE < 0.000001), obtained from the presented functions, guarantees the models reliability. For both CWEs, the cubic model stood out. The error associated with the quadratic model for CWE476 was approximately 11.14% higher than the cubic model in “real” analysis situations, while the error associated with the quadratic model for CWE457 was approximately 22.32% higher than the cubic model6. Therefore, in both cases, the cubic models are better for predicting the Linux flaw rate evolution, since they show a better approximation from the reference models when it comes to the extrapolation of the analyzed data boundaries. RESULTS For each CWE studied, two models were defined and one was selected. Thus, it is possible to use low complexity functions, with low error, to define a model to predict flaws in Linux, answering our research question. This is feasible when we can obtain a meaningful data set, as it was the case for Linux, where we used 391 different versions of the project. In short, comparing and selecting the models for each CWE, we have Equations 1 and 2 representing the selected cubic models: tax_CWE476(modules) = (1.911224 ∗ 10 -15) ∗ modules 3 − (1.72028 ∗ 10 -19) ∗ modules 2 + (4.857479 ∗ 10 -6) ∗ modules − (0.03460173) (1 tax_CWE476(modules) = (1.911224 ∗ 10 -15) ∗ modules 3 d − (1.72028 ∗ 10 -19) ∗ modules 2 (1) + (4.857479 ∗ 10 -6) ∗ modules − (0.03460173) © 2018 by Author/s 6 / 9 Journal of Information Systems Engineering & Management, 3(2), 17 Journal of Information Systems Engineering & Management, 3(2), 17 Journal of Information Systems Engineering & Management, 3(2), 17 tax_CWE457(modules) = (−6.466983 ∗ 10 -16) ∗ modules 3 + (5.603787 ∗ 10 -11) ∗ modules 2 − (1.639652 ∗ 10 -16) ∗ modules + (0.02287291) (2) I b ifi d h h ffi i f h i 1 d 2 l l i h fl l l Table 2. Some flaw rates and predictions Version tax_CWE476 tax CWE457 Modules linux-v2.6.11 0.005735325 0.008992424 14123 linux-v2.6.39 0.008927095 - 30245 linux-v3.16-rc3 - 0.006577706 37095 linux-v3.9-rc8 0.006460368 0.005663884 33899 linux-v4.0.0 0.006435337 0.006036167 37793 linux-v4.5.0 0.006829370 0.005503281 40209 linux-v4.9.0 0.007679160 0.004861321 42234 Table 3. Comparison between model and real value CWE 476 Version model_tax_CWE476 real_ tax_CWE476 linux-v4.0 0.006435337 0.006085783 linux-v4.5 0.006829370 0.005968813 linux-v4.9 0.007679160 0.005729981 Table 4. Comparison between model and real value CWE 457 Version model_tax_CWE457 real_ tax_CWE457 linux-v4.0 0.006036167 0.006562062 linux-v4.5 0.005503281 0.006615434 linux-v4.9 0.004861321 0.006179855 tax_CWE457(modules) = (−6.466983 ∗ 10 -16) ∗ modules 3 + (5.603787 ∗ 10 -11) ∗ modules 2 − (1.639652 ∗ 10 -16) ∗ modules + (0.02287291) tax_CWE457(modules) = (−6.466983 ∗ 10 -16) ∗ modules 3 + (5.603787 ∗ 10 -11) ∗ modules 2 − (1.639652 ∗ 10 -16) ∗ modules + (0.02287291) (2) (2) It can be verified that the coefficients of the equations 1 and 2 are low values, since the flaw rates are also low. It is uncommon to find buggy functions in Linux (Ferreira et al., 2016). RESULTS So, when this already low value is divided by the number of modules, which tends to increase over time, we get very low values for the flaw rates. Table 2 presents the defined models with the flaw rates in some known points, as first collected version (2.6.11), last analyzed version (3.9), the version with the higher flaw rates (2.6.39 for CWE476 and 3.16 for CWE457), as well as, recent Linux versions for prediction points (4.0, 4.5, and 4.9). In the one hand, the flaw rate for CWE476 (NULL pointer dereference) peaked in version 2.6.39 and then started to decrease, but as soon as the number of modules starts to increase, it tends to increase, reaching a new peak and finally starting to decrease again. In the other hand, the flaw rate for CWE457 (Uninitialized variables) tends to decrease, as shown in Table 2. We can observe that the flaws always oscillate with a few peaks, but the values tend to decrease. The values for CWE476 emphasize that fact, since the flaw rate on the second peak is lower than the previous one. After aplying the models to kernel versions later than 3.9, we compared the real flaw rate values and the values estimated by our model. We obtained the real values by dividing the number of occurrences of CWE476 and CWE457 by the number of modules of the respective kernel versions. The real values are shown in Tables 3 and 4. By analyzing the real values and the values proposed by the models for Linux versions greater than 3.9, we can see that the further the latest version used to train the model is from the version analyzed, the greater the error on its predictions. This leads to the need of updating the model for a certain interval of new Linux releases, making it possible to reduce the error, improving the precision of the statistical model predictions. An interesting means to achieve such updates is by performing continuous static analysis on software repositories (the kernel repository in this specific context) and automate the model updating tasks. In the next Section we discuss our final remarks and our plans for future work, including the development of a platform to perform continuous static analysis on software repositories and make the analyses available. © 2018 by Author/s FINAL REMARKS By identifying frequent flaws in source code, such as NULL pointer dereferences and uses of uninitialized variables (CWE476 and CWE457, respectively), a series of exploratory analyses were carried out and we defined low complexity models to predict the mentioned flaws. We observed, from Cppcheck reports, that source code flaws are not easily detected in Linux, which gave us a small number of flaws per module in the performed analysis. To test the defined models, we also analyzed Linux versions that were more recent than the ones analyzed to define the selected functions, using the presented models to predict the flaw rates, and comparing the obtained values with the results of static analysis. The results showed that it is possible to predict the number of flaws 7 / 9 © 2018 by Author/s Kanashiro et al. / Predicting Software Flaws with Low Complexity Models based on Static Analysis Data triggered by the static analyzer for a given version of the kernel, as long as the revisions used for training the model are not distant from the version we want to generate predictions for. It is important to say that this may not be true for other software, since it depends on how much is changed between versions. But the opposite also holds: the model may work better for software with smaller deltas between versions and this should be investigated in future work. In this study, we did not consider the possibility of the used static analyzer to report false alarms (Landi, 1992) (false positives), increasing the flaw rate of this work with fake data (not real flaws), or false negatives, omitting flaws from our research, which would decrease our flaw rates. Even with this limitation, since the used sample is large, statistically, the proposed models are valid. They could also be applied with other sources for the flaws, like manually mining repositories for bug fixes or searching mailing lists for confirmed bug reports, cases where most flaws, if not all of them, are positives. We did not follow this approach because we intend to scale this research using an automated approach. In our analysis, we could observe that the presented rates of flaws per module have peaks during the evolution of Linux's source code, that represent phases of source code's growth, where flaws were added, and refactoring phases, where flaws were fixed. ACKNOWLEDGEMENTS National Council for the Improvement of Higher Education (CAPES) for supporting Lucas Kanashiro and Athos Ribeiro during the production of this paper. Also, Fabio Kon (University of São Paulo) for reviewing and providing important feedback on this paper. FINAL REMARKS Thus, a next step for this research would be applying the analysis of other factors that influence the variation of source code flaws, such as the number of developers who are working with quality assurance. This would make it possible to define multivariate models to aid software engineers on making decisions around the software development cycle.Moreover, from the defined low complexity models analyses, we want to identify some aspects of the development process as a next step, besides studying the flaws history in the course of different releases. For instance, the models could be used to analyze the development cycle in specific moments, such as new features development (increase in number of flaws) and moments of stabilization, refactoring, and bug fixing (decrease in number of flaws) through Linux development. The reason to build less complex models is the fact that they will require less effort to perform future analysis. In the future, we want to automate all these analyses and perform continuous monitoring of flaws in source code. In fact, our next step involves developing a platform to perform static analysis on different software repositories with different static analyzers. This will provide the data needed to perform studies like the present research and take decisions to improve the software development cycle and the software quality itself. We also intend to open the collected data, which might be useful for other researchers and software developers. REFERENCES Black, P. E. (2001). Static analyzers in software engineering. Black, P. E., Kass, M. and Koo, M. (2007). Source code security analysis tool functional specification version 1.0. In Special Publication (NIST SP) 500-268, NIST, pp. 5. https://doi.org/10.6028/NIST.SP.500-268 Chidamber, S. and Kemerer, C. (2002). A metrics suite for object oriented design. MIT, Cambridge, MA, USA. Chidamber, S. and Kemerer, C. (2002). A metrics suite for object oriented design. MIT, Cambridge, MA, USA. Cleveland, W. S. and Devlin, S. J. (1988). Locally weighted regression: An approach to regression analysis by local fittting. Chidamber, S. and Kemerer, C. (2002). A metrics suite for object oriented design. MIT, Cambridge, MA, USA. Cleveland W S and Devlin S J (1988) Locally weighted regression: An approach to regression analysis by local g Ernst, M. D. (2005). Static and dynamic analysis: synergy and duality. Evans, D. and Larochelle, D. (2002). Improving security using extensible lightweight static analysis. IEEE Software, 19(1), 42–51. https://doi.org/10.1109/52.976940 p g Ferreira, G., Malik, M. M., Kästner, C., Pfeffer, J. and Apel, S. (2016). Do #ifdefs influence the occurrence of vulnerabilities? An empirical study of the linux kernel. CoRR, vol. abs/1605.07032. [Online]. Available at: http://arxiv.org/abs/1605.07032. https://doi.org/10.1145/2934466.2934467 Ferzund, J., Ahsan, S. and Wotawa, F. (2009). Empirical evaluation of hunk metrics as bug predictors. Hawkins, D. M. (1980). Identification of outliers. Londres, Chapman and Hall. https://doi.org/10.1007/978-94-015- 3994-4 Ferzund, J., Ahsan, S. and Wotawa, F. (2009). Empirical evaluation of hunk metrics as bug predictors. Hawkins D M (1980) Identification of outliers Londres Chapman and Hall https://doi org/10 1007/978 94 015 , J , , , ( ) p g p Hawkins, D. M. (1980). Identification of outliers. Londres, Chapman and Hall. https://doi.org/10.100 3994-4 Johnson, B., Song, Y., Murphy-Hill, E. and Bowdidge, R. (2013). Why don’t software developers use static analysis tools to find bugs?. In 2013 35th International Conference on Software Engineering (ICSE), pp. 672–681. https://doi.org/10.1109/ICSE.2013.6606613 p g Jones, W. D. (1991). Reliability models for very large software systems in industry. In Software Reliability Engineering, International Symposium on. IEEE, pp. 35–42. https://doi.org/10.1109/ISSRE.1991.145351 8 / 9 © 2018 by Author/s Journal of Information Systems Engineering & Management, 3(2), 17 R. (1995). A study of cross-validation and bootstrap for accuracy estimation and model selection. andi, W. (1992). Undecidability of static analysis. ACM Letters on Programming Languages and Systems 1(4), 323–337. https://doi.org/10.1145/161494.161501 Li, P. L., Shaw, M., Herbsleb, J., Ray, B. and Santhanam, P. (2004). REFERENCES Empirical evaluation of defect projection models for widely-deployed production software systems. ACM SIGSOFT Software Engineering Notes, 29(6), 263– 272. https://doi.org/10.1145/1041685.1029930 p g Misra, S. and Bhavsar, V. (2003). Relationships between selected software measures and latent bug-density: Guidelines for improving quality. p g q y Nagappan, N., Ball, T. and Zeller, A. (2006). Mining metrics to predict component failures. Pan, K., Kim, S. and E. J. W. Jr. (2006). Bug Classification Using Program Slicing Metrics. In 2006 Sixth IEEE International Workshop on Source Code Analysis and Manipulation, pp. 31–42. https://doi.org/10.1109/SCAM.2006.6 Penta, M. D., Cerulo, L. and Aversano, L. (2008). The Evolution and Decay of Statically Detected Source Code Vulnerabilities. In 2008 Eighth IEEE International Working Conference on Source Code Analysis and Manipulation, pp. 101–110. https://doi.org/10.1109/SCAM.2008.20 Picard, R. R. and Cook, R. D. (1984). Cross-validation of regression models. Journal of the American Statistical Association, 79(387), 575–583, 1984. https://doi.org/10.1080/01621459.1984.10478083 Picard, R. R. and Cook, R. D. (1984). Cross-validation of regression models. Journal of the American Statistical Association, 79(387), 575–583, 1984. https://doi.org/10.1080/01621459.1984.10478083 ( ) p g Rossman, A. J. (1996). Workshop statistics: Discovery with data. Tukey, J. W. (1977). Exploratory data analysis. y, J ( ) p y y Wood, A. (1996). Predicting software reliability. Computer, 29(11), 69–77. https://doi.org/10.1109/2.544240 Wood, A. (1996). Predicting software reliability. Computer, 29(11), 69–77. https://doi.org/10.1109/2.544240 , ( ) g y p , ( ), p // g/ / Zheng, J., Williams, L., Nagappan, N., Snipes, W., Hudepohl, J. P. and Vouk, M. (2006). On the value of static analysis for fault detection in software. Zheng, J., Williams, L., Nagappan, N., Snipes, W., Hudepohl, J. P. and Vouk, M. (2006). On the value of static analysis for fault detection in software. © 2018 by Author/s 9 / 9
https://openalex.org/W4299490849	https://zenodo.org/records/3445765/files/1.pdf	English	null	Effect of exchange rate, inflation rate and interest rate on Balance of Payment (BOP): a research from India and Bangladesh	Zenodo (CERN European Organization for Nuclear Research)	2,017	cc-by	4,317	Effect of exchange rate, inflation rate and interest rate on Balance of Payment (BOP): a research from India and Bangladesh Author’s Details: (1)Kamran najaf ,(2) Samana Salim Author’s Details: (1)Kamran najaf ,(2) Samana Salim Masters in Business Administration-Taylor's University, Lakeside Campus, Jalan Taylor's, Subang J Keywords Exchange Rate, Interest Rate, Inflation Rate, Balance of Payments (BOP). Abstract Exchange rate remains an important element in international macroeconomics; its impact is evidence various currency disasters in many countries and gets attractions for the economists around the worlds. Exchange rate has been proved its nature in evaluating the state financial situation. It is a proportional research of Balance of Payment (BOP) of two nations: Bangladesh and India. This text mentioned that BOP be supposed to be constant. It must not be in debit (Deficit) and in credit (surplus) as in both situations, it will be adverse. Thus this study evaluate the variation in three important rates of any country that is 1) interest rate,2)inflation rate, and 3) foreign exchange rate on BOP. Outcome describes that inflation and exchange rate will have favorable impact mean while interest rate will have adverse impact on BOP in both countries. 1. Introduction The economists try to overcome these issues.  The economists deploy the methods of for the development of exports and betterment of the BOP and also cut down the  The domestic crisis of the country has a great effect on BOP. The economists try to overcome these issues.  The domestic crisis of the country has a great effect on BOP. The economists try to overcome these issues.  The economists deploy the methods of for the development of exports and betterment of the BOP and also cut down the exchange rate.  The economists deploy the methods of for the development of exports and betterment of the BOP and also cut down the exchange rate.  International Monetary Fund (IMF) gives loan either short term or long term to developing countries to float their economy, but it will have a negative impact such as, repayments schedule which is made in installments, it drive the currency of the country toward decline which result in financial crisis.  International Monetary Fund (IMF) gives loan either short term or long term to developing countries to float their economy, but it will have a negative impact such as, repayments schedule which is made in installments, it drive the currency of the country toward decline which result in financial crisis. y y  Tendency from rural areas to urban areas drive to more growth in industries and increase the export proportion which helps BOP in turn of surplus.  Tendency from rural areas to urban areas drive to more growth in industries and increase the export proportion which helps BOP in turn of surplus.  Foreign investment (FI) also has a major impact on the BOP. Inflow of FI will result in surplus and outflow will create the deficit.  Foreign investment (FI) also has a major impact on the BOP. Inflow of FI will result in surplus and outflow will create the deficit.  Corruption is also major impact. Due to it, foreign investor reluctant to invest in country and it will result in negative impact on the BOP.  Corruption is also major impact. Due to it, foreign investor reluctant to invest in country and it will result in negative impact on the BOP. Vol-6-Issue-7 July-2017 ISSN (2304-7151) Vol-6-Issue-7 July-2017 ISSN (2304-7151) American Based Research Journal 1. Introduction Balance of payments (BOP) provides the detail knowledge of the country demand and supply of currency. BOP must remain constant, it should not be either deficit or surplus as in both situations it is not favorable for a country. BOP could also utilize to determine the economy of the country in the worldwide economic rivalry. Furthermore, the impact of ups and down in balance of trade will be result in evaluating of balance of payments (BOP) of country. BOP introduces methodology which helps to determine the payments made and receipts during certain period of time of the country from other countries in private and public sectors. Consequently, equation of BOP is receipts for example exports, attendance of the foreigner for tourism which make payments in country, remittance from oversee local resident, gift and grant got from different countries in the form of cash, or capital assets and payments made for import of different items and instruments, purchase of oversee services, and gifts made to other countries. To balance this equation (receipts and payments), from last couple year’s government’s show more interest to correctly perform the economic strategy to determine export to different countries, and cut down imports at similar time.Whereas, in those countries which having the stability in their policies, this will be achievable easily in long term macroeconomic policies of their balance of payments goals. In opposition, many developing nations having with poor application of that strategy will result in adverse growth. In this research, we are matching the BOP of Bangladesh and India by using the many variables for example: exchange rate, inflation rate and interest rate. Since several years, BOP of both the countries has been in deficit for a longer time. Because of the downfall in the currency of the country’s capital account tends to be in surplus If the capital account remains surplus it has a major impact on the BOP of a country which is mention as follow  The domestic crisis of the country has a great effect on BOP. The economists try to ov  The domestic crisis of the country has a great effect on BOP. The economists try to overcome these issues.  The economists deploy the methods of for the development of exports and betterment of the BOP and also cut down the exchange rate.  The domestic crisis of the country has a great effect on BOP.  Foreign trade also plays a vital role to maintain BOP of the country.  Foreign trade also plays a vital role to maintain BOP of the country.  From the research it was concluded that efficient fiscal policy utilized to boost level of saving. 2. Methodology  From the research it was concluded that efficient fiscal policy utilized to boost level of saving. 2. Methodology 2.4. Inflation Rate Inflation rate knows as general rise in price which can be evaluate by general purchasing power of normal personal. Inflation evaluates by comparing the same sets of things at different time (past and present).Increase in inflation rate will have a negative impact on economy. Inflation will have a huge impact on the stander of living of normal person. They have to pay more to buy the same things or services as compare with the previous purchased. Inflow of cash should be increase as compare to increase in inflation otherwise it would be harder to survive for the normal personal 2.3. Foreign Exchange Rate Exchange rate can be defined as such unit currency of a country that is paid to buy unit of the other country currency. Factors which impact on the exchange rate the most which are as such:  Mark up rate  Inflation factor  Political instability of the country  Internal Factors of a country (sectarian issue, races, terrorism, civil war etc.)  Internal Factors of a country (sectarian issue, races, terrorism, civil war etc ernal Factors of a country (sectarian issue, race  Governance Quality of a country Exchange rate influences a lot on the BOP. Exchange rate of Bangladesh and India is very low which has adverse impact on BOP. From this study, we come to conclusion that exchange rate play a vital role in the balance of payments. Balance of payments is the list of transactions with the other country which demonstrate the receipts and payments with other country for the certain period of time. 1. Introduction  In this new tenure, multinational bank loan facilities increase in the economy because of deposits, and inflow of FI and it will result in positive impact on BOP because to inflow of c  In this new tenure, multinational bank loan facilities increase in the economy because of increase in mark up on deposits, and inflow of FI and it will result in positive impact on BOP because to inflow of cash in the country. The interest rate is low in those countries because of type of Debt Structure of debt(Long term or short term or it p , p p y  The interest rate is low in those countries because of type of Debt, Structure of debt(Long term or short term or it repayment schedule),and strategy for the holding the foreign reserve.  Exchange rate policies have a major impact on the exports of the country. Other factors which influence the most of exports of the country that is Black Market which greater or lesser the demand and supply of dollar http://www.abrj.org 2.1. Balance of Payments The BOP shows the receipts and payments which is book of tractions with the world. If a country in adverse impact on the economy. The BOP shows the receipts and payments which is book of tractions with the world. If a country inflation rate increase, it will have adverse impact on the economy. Vol-6-Issue-7 July-2017 ISSN (2304-7151) Table 2. ANOVA for Bangladesh Source DF SS MS F Regression 3 1770284 442463 25.69 (0.000) Error 22 415573 18175 Total 25 2185857 The high value of F statistic and extremely low p value (given in parenthesis below F value) support the regression model. To support the regression model, very low P value (given in parenthesis right side by F value) and F statistic 3.2. Regression Analysis of INDIA 3.1. Regression Analysis of Bangladesh For this study, I collect data from International Financial Statistics (IFS) for inflation, exchange rate, BOP, and interest rate.By evaluating the regression between BOP, and three rates interest rate, foreign rate, and inflation rate, and I came to conclusion which is mentioned as below: BOP =35, inflation rate +25.4, interest rate -35.9, foreign exchange rate +97.10 Table 1. Regression Analysis of Bangladesh Predictor Coef St. Dev T value P Constant 34.7 160.0 0.11 0.720 Inflation rate 25.423 8.326 2.02 0.007 Interest rate -35.86 15.07 -1.47 0.036 Foreign exchange rate 97.10 2.358 6.10 0.000 BOP =35, inflation rate +25.4, interest rate -35.9, foreign exchange rate +97.10 Table 1. Regression Analysis of Bangladesh Predictor Coef St. Dev T value P Constant 34.7 160.0 0.11 0.720 Inflation rate 25.423 8.326 2.02 0.007 Interest rate -35.86 15.07 -1.47 0.036 Foreign exchange rate 97.10 2.358 6.10 0.000 Page 2 American Based Research Journal Vol-6-Issue-7 July-2017 ISSN (2304-7151) S = 134.4, R2=81.0%, R2dj =77.7% R2 =71.0% demonstrates a good relationship between independent variables (inflation rate, foreign exchange rate, and interest rate) and dependent variable (BOP).Coefficient of inflation is 25.423,foreign exchange=97.21 which shows positive relationship and have a major influence on BOP. Although interest rate has negative value but has a great influence on BOP. Nil hypothesis value should be rejected for independent variables. The outcome from the ANONA(Fisher Model) for Bangladesh shows in the Table2 Vol-6-Issue-7 July-2017 ISSN (2304-7151) To support the regression model, very low P value (given in parenthesis below F value) and F statistic. 3.2. Regression Analysis of INDIA To evaluate the regression among BOP and interest rate, exchange rate, inflation rate following equation is create which is mentioned below: BOP = -537, Inflation rate+ 142, interest rate -191,foreign exchange +97.1 R2=78.21% demonstrates that the model is good enough. The coefficient of inflation rate and exchange rate are 133.40 and 97.20 retrospectively which have optimistic impact on BOP and it has a great impact on the BOP. The outcome from the ANOVA(Fisher Model) for India also are demonstrate in table no 4. Variable Coef St Dev T value P BOP -572.3 983.6 -0.44 0.473 Inflation rate 131.20 77.11 2.54 0.086 Interest rate -197. 1104.3 -1.76 0.066 Foreign Exchange rate 98.20 13.131 5.81 0.000 S = 836.7, R2= 78.2%, R2 (adj) = 74.3% Table 4. ANOVA for India Source DF SS MS F Regression 3 54200325 13550088 18.42 (0.000) Variable Coef St Dev T value P BOP -572.3 983.6 -0.44 0.473 Inflation rate 131.20 77.11 2.54 0.086 Interest rate -197. 1104.3 -1.76 0.066 Foreign Exchange rate 98.20 13.131 5.81 0.000 S = 836.7, R2= 78.2%, R2 (adj) = 74.3% Table 4. ANOVA for India Source DF SS MS F Regression 3 54200325 13550088 18.42 (0.000) Variable Coef St Dev T value P BOP -572.3 983.6 -0.44 0.473 Inflation rate 131.20 77.11 2.54 0.086 Interest rate -197. 1104.3 -1.76 0.066 Foreign Exchange rate 98.20 13.131 5.81 0.000 S = 836.7, R2= 78.2%, R2 (adj) = 74.3% Table 4. ANOVA for India Source DF SS MS F Regression 3 54200325 13550088 18.42 (0.000) Variable Coef St Dev T value P BOP -572.3 983.6 -0.44 0.473 Inflation rate 131.20 77.11 2.54 0.086 Interest rate -197. 1104.3 -1.76 0.066 Foreign Exchange rate 98.20 13.131 5.81 0.000 S = 836.7, R2= 78.2%, R2 (adj) = 74.3% Table 4. ANOVA for India Source DF SS MS F Regression 3 54200325 13550088 18.42 (0.000) http://www.abrj.org American Based Research Journal Vol-6-Issue-7 July-2017 ISSN (2304-7151) Error 27 16054778 698036 Total 30 702550103 To support the regression model, very low P value (given in parenthesis below F value) and F statistic. American Based Research Journal Vol-6-Issue-7 July-2017 ISSN (2304-7151) Error 27 16054778 698036 Total 30 702550103 Figure 3. Foreign exchange rate of Bangladesh Figure 3. Foreign exchange rate of Bangladesh 3.4. Analysis of India’s BOP Over Last Years American Based Research Journal Figure2. Interest rate of Bangladesh Figure2. Interest rate of Bangladesh The below graph mentions the worth of Bangladeshi currency unit demised The below graph mentions the worth of Bangladeshi currency unit demised in previous years as compare to US Dollar. In 2009, the Bangladeshi currency value impaired to 71.21 as compare to US Dollar due to instability of political and inflation increase by 13%.Value of coefficient is 10.218 that shows exchange g p g y in previous years as compare to US Dollar. In 2009, the Bangladeshi currency value g rate have strong correlation with BOP. If there is 1% modification in value will cause 10.218% impact on BOP. The graph mentions there is no correlation exists in value term of foreign exchange r modification in value will cause 10.218% impact on BOP. The graph mentions there is no correlation exists in value term of foreign exchange rate. modification in value will cause 10.218% impact on BOP. The graph mentions there is no correlation exists in value term of foreign exchange rate. 0 20 40 60 80 1981 1991 2001 2011 Year wise Dollar Rate 0 20 40 60 80 1981 1991 2001 2011 Year wise Dollar Rate Year wise Dollar Rate 3.4. Analysis of India’s BOP Over Last Years The inflation was at top in 1991 that was 13.8% from the previous years and remained stable in 2003 and 2004 that was 3.8% and again rise in 2008 but that rising rate was lesser than Bangladesh’s inflation rate. The value of coefficient 132.30 demonstrates that inflation rate have direct correlation with the BOP of India. Inflation does have great influence on BOP of India as compare to Bangladesh. The graph mentions that there was no correlation consists in value of inflation. http://www.abrj.org Page 5 Figure 4. Inflation rate of India The interest rate remained steady from 1983 to 1987 and greater than before in 1987 and after that again it was steady in 2000.The interest rate increased in 1997 and dropped at its peak value of 13.45 and then decline yet again in year 2005 and following this it happened to raised again in 2009 that paying attention in additional to extra FDI as compare to Bangladesh‘s 0 10 20 1981 1991 2001 2011 0 10 20 1981 1991 2001 2011 Figure 4. Inflation rate of India 3.3. Analysis of Bangladesh’s BOP over Last Couple of Years Due to instability of political and domestic circumstance of the country these factors influenced a lot on the BOP, inflation was low in 1986 to 1997 and it getting higher from 1998 to 2007 and reach at the top then decline in between 1998 to 1999 and inflation rate in 2004 was at the value 4.32% which is as low in last 25 years. The rate was changed in 2007 due to political instability in country and disorder position of law and order. The BOP was greatly influenced by coefficient of inflation which is rate=25.423%.it shows positive influence on the BOP. There is change in the value of inflation by 1% because of change in the value of BOP and it will result in 25.423%.The graph demonstrates there is no correlation consists of in value of inflation. Figure 1. Inflation rate of Bangladesh 0 20 1981 1991 2001 2011 Inflation rate of Bangladesh Figure 1. Inflation rate of Bangladesh Impact of interest rate on BOP: If Inflation increases, then the interest rate will also rise. The higher interest rate will attract the stakeholders to invest in such country. The graph mentions the value of interest rate decrease from 1981 to 1989 due to decrease in inflation rate. In 1999 the value of interest rate is high that was 13.44% and then it declined up to 2006 and after it started to increase and in 2009 it was 8.4% and the foreign direct investment value increased from 2006 to 2007. The coefficient of interest rate=-38.71 mentions its adverse effect on the BOP. 2% modification in value of interest rate will impact in the decline in value of BOP by 38.71%. The graph mention there is no correlation consisted of in value of interest in figure 2. http://www.abrj.org Page 4 0 10 20 1981 1991 2001 2011 YEAR WISE INTEREST RATE 0 10 20 1981 1991 2001 2011 YEAR WISE INTEREST RATE YEAR WISE INTEREST RATE http://www.abrj.org Figure 5. Interest rate of India Figure 5. Interest rate of India The value of Indian currency unit decline as compare to US Dollar but not as much as the Bangladeshi currency in past couple of years. The value of Indian currency unit is 42.41 at 2007 the depreciated value was lesser than Bangladeshi currency at similar year is 60.52 as compare to US Dollar. The value of coefficient 97.2 mentioned positive correlation with BOP. The graph mentioned that there is no correlation consists in value of foreign exchange rate. /www.abrj.org Page 6 Figure 6. Foreign exchange rate of India 4. Conclusion BOP gives us knowledge in depth regarding country’s currency supply and demand. BOP plays a vital role in economy of a country. After this research we came to conclude that inflation and foreign exchange rate will have positive impact on the BOP of both the countries and interest rate will results in negative impact on them. It was also evaluated from this research that supervision of both countries need assist to resolve the issues for shakiness of BOP. For such Countries like Bangladesh and India, it is very important to having proper evaluating systems attached with appropriate macroeconomic strategies. Such strategies should be introduced which attracted inflow of cash from the foreigners e.g. remittance etc. Moreover, these developed monitoring at different strategic levels (Government or public sector) (e.g. 0 50 100 1981 1991 2001 2011 YEARS WISE DOLLAR RATE YEARS WISE DOLLAR RATE YEARS WISE DOLLAR RATE Figure 6. Foreign exchange rate of India 4 Conclusion 0 50 100 1981 1991 2001 2011 YEARS WISE DOLLAR RATE 2001 Figure 6. Foreign exchange rate of India Figure 4. Inflation rate of India The interest rate remained steady from 1983 to 1987 and greater than before in 1987 and after that again it was steady in 2000.The interest rate increased in 1997 and dropped at its peak value of 13.45 and then decline yet again in year 2005 and following this it happened to raised again in 2009 that paying attention in additional to extra FDI as compare to Bangladesh‘s http://www.abrj.org Vol-6-Issue-7 July-2017 ISSN (2304-7151) American Based Research Journal American Based Research Journal Vol-6-Issue-7 July-2017 ISSN (2304-7151) financial system. The value of coefficient was -197.2 that demonstrated negative correlation with the BOP, that mentioned 2% raised in interest rate value would demised in value of BOP by 197.2. The graph demonstrated that there was no relationship consists in value of interest. Figure 5. Interest rate of India The value of Indian currency unit decline as compare to US Dollar but not as much as the Bangladeshi currency in past couple of years. The value of Indian currency unit is 42.41 at 2007 the depreciated value was lesser than Bangladeshi currency at similar year is 60.52 as compare to US Dollar. The value of coefficient 97.2 mentioned positive correlation with BOP. The graph mentioned that there is no correlation consists in value of foreign exchange rate. 0 5 10 15 20 25 1981 1991 2001 2011 YEAR WISE INTEREST RATE 0 50 100 1981 1991 2001 2011 YEARS WISE DOLLAR RATE financial system. The value of coefficient was -197.2 that demonstrated negative correlation with the BOP, that mentioned 2% raised in interest rate value would demised in value of BOP by 197.2. The graph demonstrated that there was no relationship consists in value of interest. 0 5 10 15 20 25 1981 1991 2001 2011 YEAR WISE INTEREST RATE References [1]Calvo, G. A. "Staggered prices in a utility-maximizing framework," Journal of monetary Economics, 12(3), 383-398(1983) [1]Calvo, G. A. "Staggered prices in a utility-maximizing framework," Journal of monetary Economics, 12(3), 383-398(1983) [1]Calvo, [2] De La Cruz Martínez, J. "Mexico’s balance of payments and exchange rates: a cointegration analysis," The North American Journal of Economics and Finance, 10(2), 401-421 (1999) [2] De La Cruz Martínez, J. "Mexico’s balance of payments and exchange rates: a cointegration analysis," The North American Journal of Economics and Finance, 10(2), 401-421 (1999) [2] De La Cruz Martínez, J. "Mexico’s balance of payments and exchange rates: a cointegration analysis," The North American Journal of Economics and Finance, 10(2), 401-421 (1999) [3] Atkinson, G., & Hamilton, K. "International trade and the ecological balances of payments," Resources Policy, 28(1), 27-37 (2002) [3] Atkinson, G., & Hamilton, K. "International trade and the ecological balances of payments," Resources Policy, 28(1), 27-37 (2002) [4] Thanh, N. N., & Kalirajan, K. "Can devaluation be effective in improving the balance of payments in Vietnam?" Journal of Policy Modeling, 28(4), 467-476 (2006) [4] Thanh, N. N., & Kalirajan, K. "Can devaluation be effective in improving the balance of payments in Vietnam?" Journal of Policy Modeling, 28(4), 467-476 (2006) [5] Thacker, S. C. "The high politics of IMF lending," World Politics, 52(01), 38-75 (1999) [5] Thacker, S. C. "The high politics of IMF lending," World Politics, 52(01), 38-75 (1999) [6] Choe, S. C. "The impacts of globalization on the urban spatial-economic system in Korea. In Globalization and Urban Development, 59-78 Springer Berlin Heidelberg (2005) [7] Hejazi, W., & Pauly, P. "Motivations for FDI and domestic capital formation" Journal of International Business Studies, 34(3), 282-289 (2003) [8] Habib, M., & Zurawicki, L. "Corruption and foreign direct investment." Journal of international business studies, 291- 307 (2002) [9] Weller, C. E., & Scher, M. "Multinational bank credit in less industrialized economies" Journal of International Business Studies, 32(4), 833-851 (2001) [10] Globerman, S., & Shapiro, D. "Governance infrastructure and US foreign direct investment," Journal of International Business Studies, 34(1), 19-39 (2003) [11] Jha, R. "Downward rigidity of Indian interest rates," Economic and Political Weekly, 469-474 ( [12] Schnitzer, M. "Debt v. foreign direct investment: the impact of sovereign risk on the structure of international capital flows,"Economica, 69(273), 41-67 (2002) [13] Eichengreen, B., Rhee, Y., & Tong, H. American Based Research Journal planning and development department, Research department, central banks (state banks), Ministries of economies, foreign affairs commerce, trade, and industries, Departments of statistics and budgets) can plays management body role for the imposing the economic policies. Similarly, provide reasonable atmosphere for foreigner to invest in the country on consistent basis such as stability of political situation, law and order positions, incentive for tourism in country etc. This will help in achieving balance of payments targets and by having steady exchange rates , this will help in longer period of time. 4. Conclusion BOP gives us knowledge in depth regarding country’s currency supply and demand. BOP plays a vital role in economy of a country. After this research we came to conclude that inflation and foreign exchange rate will have positive impact on the BOP of both the countries and interest rate will results in negative impact on them. It was also evaluated from this research that supervision of both countries need assist to resolve the issues for shakiness of BOP. For such Countries like Bangladesh and India, it is very important to having proper evaluating systems attached with appropriate macroeconomic strategies. Such strategies should be introduced which attracted inflow of cash from the foreigners e.g. remittance etc. Moreover, these developed monitoring at different strategic levels (Government or public sector) (e.g. http://www.abrj.org Page 6 http://www.abrj.org References "China and the exports of other Asian countries," Review of World Economics, 143(2),201-226 (2007) [14] Acharyya, R. "Exchange Rate Policy and Black Market Premium on Foreign Exchange: Theory and Evidence.,"Economic and Political Weekly, (2001) [15] Nagaishi, M. "Stock market development and economic growth: dubious relationship," Economic and Political Weekly, (2004-2012 ) [16] Raipuria, K. "Foreign Trade Eludes Economic Crystal Ball Gazers," .Economic and Political Weekly, 2308-2317 (1998) [16] Raipuria, K. "Foreign Trade Eludes Economic Crystal Ball Gazers," .Economic and Political Weekly, 2308-2317 (1998) [17] Lopez, J. H., Schmidt-Hebbel, K., & Servén, L. "How effective is fiscal policy in raising national saving?" Review of Economics and Statistics, 82(2), 226-238 (2000) [17] Lopez, J. H., Schmidt-Hebbel, K., & Servén, L. "How effective is fiscal policy in raising national saving?" Review of Economics and Statistics, 82(2), 226-238 (2000) http://www.abrj.org Page 7 Page 7
https://openalex.org/W2801032454	https://publications.goettingen-research-online.de/bitstream/2/77811/2/journal.pone.0197329.pdf	English	null	Specific serum and CSF microRNA profiles distinguish sporadic behavioural variant of frontotemporal dementia compared with Alzheimer patients and cognitively healthy controls	PloS one	2,018	cc-by	14,167	Abstract Editor: Junming Yue, University of Tennessee Health Science Center, UNITED STATES Editor: Junming Yue, University of Tennessee Health Science Center, UNITED STATES Received: February 14, 2018 Accepted: April 29, 2018 Published: May 10, 2018 Editor: Junming Yue, University of Tennessee Health Science Center, UNITED STATES Received: February 14, 2018 Accepted: April 29, 2018 Published: May 10, 2018 Copyright: © 2018 Denk et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original h d di d Information on circulating miRNAs in frontotemporal lobar degeneration is very limited and conflicting results have complicated an interpretation in Alzheimer’s disease thus far. In the present study we I) collected samples from multiple clinical centers across Germany, II) defined 3 homogenous patient groups with high sample sizes (bvFTD n = 48, AD n = 48 and cognitively healthy controls n = 44), III) compared expression levels in both CSF and serum samples and IV) detected a limited set of miRNAs by using a MIQE compliant protocol based on SYBR-green miRCURY assays that have proven reliable to generate reproducible results. We included several quality controls that identified and reduced technical variation to increase the reliability of our data. We showed that the expression levels of circulating miRNAs measured in CSF did not correlate with levels in serum. Using cluster analysis we found expression pattern in serum that, in part, reflects the genomic organization and affilia- tion to a specific miRNA family and that were specifically altered in bvFTD, AD, and control groups. Applying factor analysis we identified a 3-factor model characterized by a miRNA signature that explained 80% of the variance classifying healthy controls with 97%, bvFTD with 77% and AD with 72% accuracy. MANOVA confirmed signals like miR-320a and miR- 26b-5p at BH corrected significance that contributed most to discriminate bvFTD cases with 96% sensitivity and 90% specificity and AD cases with 89% sensitivity and specificity Copyright: © 2018 Denk et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. RESEARCH ARTICLE OPEN ACCESS Citation: Denk J, Oberhauser F, Kornhuber J, Wiltfang J, Fassbender K, Schroeter ML, et al. (2018) Specific serum and CSF microRNA profiles distinguish sporadic behavioural variant of frontotemporal dementia compared with Alzheimer patients and cognitively healthy controls. PLoS ONE 13(5): e0197329. https://doi.org/10.1371/ journal.pone.0197329 ¶ Membership of the FTLDc study group is provided in the Acknowledgments. * jahn@uke.de Specific serum and CSF microRNA profiles distinguish sporadic behavioural variant of frontotemporal dementia compared with Alzheimer patients and cognitively healthy controls Johannes Denk1, Felix Oberhauser1, Johannes Kornhuber2, Jens Wiltfang3, Klaus Fassbender4, Matthias L. Schroeter5, Alexander E. Volk6, Janine Diehl-Schmid7, Johannes Prudlo8, Adrian Danek9, Bernhard Landwehrmeyer10, Martin Lauer11, Markus Otto10, Holger Jahn1,12, for the FTLDc study group¶ a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 Department of Psychiatry and Psychotherapy, University Medical Center Hamburg-Eppendorf, Hamburg, Germany, 2 Department of Psychiatry and Psychotherapy, Friedrich-Alexander-University of Erlangen- Nuremberg, Erlangen, Germany, 3 Department of Psychiatry and Psychotherapy, University Medical Center Goettingen, Goettingen, Germany, 4 Department of Neurology, Saarland University, Homburg, Germany, 5 Clinic for Cognitive Neurology, University Clinic Leipzig and Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany, 6 Institute of Human Genetics, University Medical Center Hamburg- Eppendorf, Hamburg, Germany, 7 Department of Psychiatry, Technical University of Munich, Munich, Germany, 8 Department of Neurology, University of Rostock, Rostock, Germany, 9 Department of Neurology, Ludwig-Maximilians-University, Munich, Germany, 10 Department of Neurology, University of Ulm, Ulm, Germany, 11 Department of Psychiatry and Psychotherapy, University of Wuerzburg, Wuerzburg, Germany, 12 AMEOS Klinikum, Heiligenhafen, Heiligenhafen, Germany ¶ Membership of the FTLDc study group is provided in the Acknowledgments. jahn@uke.de Introduction The role of microRNAs (miRNAs) in neurodegenerative disorders has gained growing interest in the field [1]. This is due to matured technologies, which now enable the reliable detection of miRNAs in body fluids [2]. MiRNAs are small non-coding RNA that selectively bind different messengerRNAs (mRNA) to downregulate its translation into proteins. These ~22 nucleotides long molecules control fundamental biological processes such as neurogenesis, immune responses and aging and are critical to cellular expression homeostasis [3]. In addition, the transport of miRNAs in extracellular vesicles such as exosomes secreted by neurons and glia plays a key role in intercellular communication and neuroinflammation [4]. The fact that miR- NAs are released as circulating miRNAs into the bloodstream not only enables detection, but deregulated signals may reflect neurodegenerative conditions that occur in AD [1, 5]. Thus, miRNAs are easily accessible in minimally invasive body fluids such as serum and cerebrospi- nal fluid (CSF) [6] and are also known to remain largely stable towards freeze-thawing cycles and RNAse activity [7, 8]. Detection by RT-qPCR is accurate and inexpensive and potentially easier to integrate into clinical routine than assays detecting proteins. A deeper investigation of the role of miRNAs may also foster our understanding of neurodegenerative diseases and lead to new therapeutic approaches. In contrast to AD, frontotemporal lobar degeneration (FTLD) is a neurodegenerative disor- der for which information on circulating miRNAs is very limited. Our knowledge on the miR- Nome is still scarce. FTLD is characterized by a progressive degeneration of the frontal and anterior temporal lobes causing pathological changes in behaviour and language. About 10– 30% of FTLD cases have a known genetic predisposition. Mutations in genes like C9orf72, GRN (progranulin), MAPT (microtubule associated protein tau) or TBK-1 (encoding TANK- binding kinase 1) can cause FTLD [9–12]. Autosomal dominant inherited cases often occur clustered in families [13]. The large majority of sporadic FTLD cases is, however, of unknown etiology, albeit genetic alterations are to be expected there too [14]. The behavioural variant (bvFTD) accounts for more than half of the cases [15] and is histopathologically described by distinct inclusion bodies either comprised of Tau (FTLD-TAU) [16] or ubiquitinated TDP-43 (FTLD-TDP) [17]. Despite considerable efforts, in-vivo biomarkers are not yet available for FTLD. We hypothesize that miRNAs in serum and CSF may serve as biomarkers differentiating patients with bvFTD, AD and cognitively healthy controls. Serum and CSF miRNA profiles distinguish bvFTD, AD and controls compared to healthy controls, respectively. Correlation analysis revealed that miRNAs from the 3-factor model also correlated with levels of protein biomarker amyloid-beta1-42 and phosphorylated neurofilament heavy chain, indicating their potential role in the monitoring of progressive neuronal degeneration. Our data show that miRNAs can be reproducibly mea- sured in serum and CSF without pre-amplification and that serum includes higher expressed signals that demonstrate an overall better ability to classify bvFTD, AD and healthy controls compared to signals detected in CSF. the state Baden-Wu¨rttemberg (D.3830), the state Baden-Wu¨rttemberg (D.3830), Boehringer Ingelheim Ulm University BioCenter (D.5009) and the Thierry Latran Foundation. The work of A.E.V was funded by the Deutsche Forschungsgemeinschaft (DFG, VO 2028/1-1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: MO gave invited talks for Teva, Virion and Lilly and scientific advice for Axon and Neuroalliance. This does not alter our adherence to PLOS ONE policies on sharing data and materials. Abstract Funding: The study was supported by grants from the German Federal Ministry of Education and Research (project: FTLDc 01GI1007A), the JPND network PreFrontAls (01ED1512), the EU (FAIR-PARK II 633190), the German Research Foundation/DFG (SFB1279), the foundation of PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 1 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Introduction Hence, we designed a study follow- ing guidelines such as the “minimum information for publication of quantitative real-time PCR experiments” (MIQE) [18] by using SYBR-green based LNA assays (Exiqon A/S), which have proven reliable to measure miRNAs in body fluids compared to other technologies. We profiled two customized miRNA panels each including n = 96 comparable assays in 48 bvFTD, 48 AD and 44 control CSF and serum samples to identify possible biomarker pattern. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 2 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Patient data Our study population included n = 140 CSF and n = 131 serum samples of bvFTD (48 CSF / 48 serum), AD (48 CSF / 47 serum) and cognitively healthy control cases (HC) (44 CSF / 38 serum). For each case a pair of CSF and serum material was available except for n = 9 serum samples. All samples were provided by the German consortium for frontotemporal lobar degeneration (FTLDc), which were collected from 10 academic centers across Germany [19]. All bvFTD patients met standard diagnostic criteria according to Rascovsky et al. (2011) [20]. Alzheimer’s disease (AD) was diagnosed according to criteria from the National Institute of Neurological and Communicative Diseases and Stroke (NINCDS)–Alzheimer’s Disease and Related Disorders Association [21]. Participants were assessed with extensive diagnostic tools including physical and neurological examination, clinical laboratory testing and analysis of CSF, genetic screening of C9orf72, GRN, MAPT for pathogenic mutations. CSF was obtained by lumbar puncture in a sitting position according to standard procedures [22]. 4 ml CSF was collected into a polypropylene test tube for routine diagnosis as well as for further studies. CSF was free of blood contaminations and tested for hemoglobin. CSF was centrifuged (1600 g, RT, 10 min) and frozen within 30 min after the puncture at -80˚C until use. Blood was drawn in 7.5 ml S-Monovette Serum Gel Z tubes (Sarstedt, Germany), incubated for 10 min at RT, cen- trifuged (2000 g, 4˚C, 10 min) and frozen within 30–40 min after blood sampling at -80˚C until use. The CSF and serum samples were at no time thawed/refrozen. Genetic analyses DNA was available from 61 participants. Screening of C9orf72 for pathogenic repeat expansion was performed by amplicon length analysis and repeat-primed PCR was conducted in 61 sam- ples. In case of a suspected expansion, southern blot analysis was performed for confirmation (Akimoto et al., 2014) [23]. In samples from participants with a familial history of a neurode- generative disease, all exons and flanking intronic regions in MAPT and GRN were screened by Sanger sequencing (details available upon request). The sequencing results were compared to reference sequences (GenBank entry NM_005910, NM_002087) using the SEQUENCE pilot software (JSI medical systems). Ethics statement Collection and analysis of samples were approved by the local Ethics Committees of Depart- ments of Neurology and Psychiatry participating in the German FTLD Consortium, a quality- controlled, monitored, multicenter initiative (Ulm approval number 20/10) [19]. All partici- pants of the registry gave their written informed consent for all investigations and their data were fully anonymized prior to any entries in the database. The data we worked with were fully anonymized. All investigations were carried out according to international Good Labora- tory Practice (GLP) and Good Clinical Practice (GCP) standard. RT-qPCR Total RNA was reverse transcribed using the Universal cDNA synthesis kit II, (Exiqon, P/N 203301). Briefly, cDNA was synthesized using 8 μl of total RNA isolated from CSF or serum including 1 μl of a spike-in mix containing synthetic UniSp6 and cel-39-3p (Exiqon, P/N 203203) using a poly-T primer with a 3’ degenerate anchor and a 5’ universal tag in a 20 μl reaction. Reverse transcription was performed at 42˚C for 60 min and 95˚C for 5 min on a qTower 2.2 (Jena Analytik). 10 μl cDNA was diluted 50x with RNase-free water and added to 500 μl 2x PCR Master mix according to the protocol for the miRCURY LNA Universal RT microRNA PCR System (Exiqon A/S). 10 μl of each sample was assayed on a custom Pick- &-Mix microRNA PCR panel containing pre-spotted LNA primers (S1 Table). Both the CSF and the serum panel each consisted of n = 96 abundant and highly expressed miRNAs. N = 76 miRNAs were identical on both panels, but n = 20 different. All miRNAs were selected based on a preliminary pilot study [1] and a comparable study by Mooney et al. [25], that both used OpenArray technology and pre-amplification. A no-template control (NTC) of RNase-free water was co-purified and profiled like the samples to measure background. Amplification was performed on a qTower 2.2 (Jena Analytik) using the following thermal cycling protocol: 95˚C for 10 min, 45 amplification cycles at 95˚C for 10 s and at 60˚C for 60 s (fluorescent reading), followed by a melting curve analysis. Laboratory markers The concentration of albumin, IgG, IgA and IgM in CSF and sera was determined as described earlier (Jesse et al., 2011) [24]. Samples were analysed in the neurochemical lab at the depart- ment of Neurology in Ulm. Commercially available ELISA kits were used for Neurofilament light chain (NfL) (IBL, Hamburg, Germany), phosphorylated Neurofilament heavy chain (pNfH) (Biovendor, Heidelberg, Germany), Progranulin (Biovendor, Heidelberg, Germany), PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 3 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Tau, pTau and amyloid-beta1-42 (Fujirebio, Hannover, Germany) according to the manufac- turer’s instructions. Tau, pTau and amyloid-beta1-42 (Fujirebio, Hannover, Germany) according to the manufac- turer’s instructions. RNA isolation Total RNA, including miRNA, was purified from 250 μl cell-free CSF and 200 μl serum sam- ples using the miRNeasy Serum/Plasma Kit (Qiagen, P/N 217184). The protocol was applied according to the manufacturer’s recommendations with the following modifications. QIAzol Lysis Reagent mixture containing 1.25 μg/ml MS2 bacteriophage RNA (Roche Applied Sci- ence) and 1 μl RNA spike-in template mixture containing synthetic UniSp2, UniSp4, and UniSp5 (Exiqon, P/N 203203) was added to all samples. The total RNA was not treated with DNase and eluted with 22 μl RNAse-free water and stored at -80˚C until use. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Pre-processing, quality control and normalisation We adapted our RT-qPCR protocol to the MIQE guidelines in order to provide sufficient experimental details to increase the reliability of our data (S2 Table) [18]. Raw unprocessed fluo- rescence values were exported from the qTower 2.2 and analysed with the “qpcR” package [26]. Cq values were calculated by using the second derivative maximum after fitting a 5-parameter sigmoidal model to the fluorescence data [27]. In addition, qPCR runs with kinetic outliers were identified and removed [28]. Furthermore, all signals showing < 80% valid data after applying the following QC criteria were excluded from analysis: (1) Signals with Cq 37 and dCq 3 (CtNTC−CtmiR-x) were considered background, (2) all assays were inspected for distinct melting curves and (3) serum samples with a dCq 7 (CqmiR-23a −miR-451) were considered at high risk of hemolysis [29]. NormFinder and GeNorm were used to identify potential reference miRNAs using GenEx [30, 31]. CSF samples were normalised to miR-101-3p, miR-320a and miR-99a-5p. Serum sam- ples were normalised to let-7i-5p, miR-23a-3p, miR-23b-3p and miR-30e-5p. Relative expres- sion was calculated as follows: dCt = Ct (Ctmean RefmiR −CtmiR) and ddCt = mean dCtAD or bvFTD − mean dCtHC. Higher values indicate higher expression. Only signals with a ddCt \|0.58\|, which corresponds to a fold change of \|1.5\|, were considered as differentially regulated. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 4 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Absolute quantification Synthetic oligos that match mature miRNA sequences (S3 Table) were ordered from IDT (Inte- grated DNA Technologies) to prepare an initial pool of oligos each with a concentration of 33 nM. Standard curves were generated by preparing 4-fold serial dilutions (n = 7) each at a vol- ume of 80 μl containing 0.8 μg/μl MS2 carrier RNA (Roche) following RT-qPCR as described earlier. Standards were measured as cDNA duplicates and PCR triplicates and assayed on a cus- tom Pick-&-Mix microRNA PCR Panel containing n = 96 pre-spotted LNA primers (S3 Table). Background was measured with a no template control including water. Standard curves were obtained by fitting a regression line to the samples with known concentrations. Statistics Two-tailed unpaired Mann-Whitney t-tests at a significance level of 5% were used to deter- mine statistical differences between two groups and chi square tests for dichotomous variables to examine demographic group differences. One-way MANOVA was applied to identify statis- tical differences between more than two groups, respectively. P < 0.05 (two sided) was consid- ered as statistically significant. The false discovery rate was controlled by using the Benjamini– Hochberg procedure when conducting multiple comparisons. Correlations were determined using Pearson r for normally distributed data or the Spearman test when data were not nor- mally distributed at a 5% significance level. To evaluate diagnostic accuracy we used receiver operating curves and discriminant analysis. Unsupervised hierarchical clustering and differen- tial correlations were calculated by using the DiffCorr v0.4.1 package for R [32]. Exploratory Factor Analysis (EFA) permits examination of how unmeasured latent vari- ables (factors) summarize patterns of correlations found in the measured relationships between miRNA expression levels. The following criteria for the factorability of a correlation were used: 1) correlation of all variables with at least 0.3 with at least one other variable, 2) the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy for each variable must be 7.5 and Bartlett’s test of sphericity should be significant, 3) the diagonals of the anti-image correla- tion should be above 0.5, 4) the communalities above 0.3 and 5) only miRNAs based on factor loadings \|0.5\| were considered significant in contributing to the respective factor. A scree plot assisted of how many factors should be extracted. The eigenvalues indicate how much of the variance is explained by each factor. Calculation and graphics were done using XLStat (v19.4), GraphPad (v7.03) and SPSS (v24) software and the R 3.4.2 statistical programming language (R Development Core 2008). PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Table 1. Demographic and descriptive parameters for study population. Characteristic1 / samples HC (CSF n = 44) (serum n = 38)2 bvFTD (CSF n = 48) (serum n = 48) AD (CSF n = 48) (serum n = 47)2 p-value (ANOVA / x2) AD vs HC AD vs bvFTD bvFTD vs HC Age (years) 64 ± 11.3 65 ± 9.2 65 ± 9.3 ns ns ns Sex (male/female) 20/24 30/18 22/26 ns ns ns MMSE nm 24 ± 4 21 ± 5.3 na < .05 na Tau [pg/ml] 317.6 ± 118.8 433.4 ± 414.7 738 ± 288.9 < .0001 < .0001 ns pTau [pg/ml] 35.1 ± 6.7 59.4 ± 37.2 96 ± 38.2 < .0001 < .0001 < .05 Aβ1–42 [pg/ml] 1031.5 ± 272.2 872.1 ± 338 513 ± 160.6 < .0001 < .0001 < .05 NfL [pg/ml] 1449.1 ± 940.2 2706.4 ± 1816.2 2157.6 ± 1026.4 ns ns < .05 pNfH [pg/ml] 307.6 ± 151.8 464.56 ± 253.85 540.33 ± 244.96 < .05 ns ns Tested / not tested for mutation in C9orf72, MAPT, GRN (n) na 41 / 7 11 / 37 11 / 37 20 / 28 11 / 37 11 / 37 na na na Tested negative for mutation in C9orf72/MAPT/GRN nm 41 11 11 20 11 11 na na na 1 Data represent mean ± SD analysed by 1-way ANOVA (age and CSF biomarker) and t-test (MMSE). Gender ratio was analysed by χ2. 2 Reduced number of serum samples did not lead to a significant change in demographical data. AD = Alzheimer’s disease, bvFTD = behavioural variant frontotemporal dementia, HC = healthy controls, na = not applicable, nm = not measured. 1 Data represent mean ± SD analysed by 1-way ANOVA (age and CSF biomarker) and t-test (MMSE). Gender ratio was analysed by χ2. 2 Reduced number of serum samples did not lead to a significant change in demographical data. AD = Alzheimer’s disease, bvFTD = behavioural variant frontotemporal dementia, HC = healthy controls, na = not applicable, nm = not measured. https://doi.org/10.1371/journal.pone.0197329.t001 MAPT and GRN were identified in the tested AD (n = 11) and bvFTD (n = 11) cases. To address possible center-effects, we compared the baseline miRNA expression data of the study sites (S1 Fig). Characteristics of study population As summarized in Table 1, no significant difference was observed in the distribution of age and gender. The bvFTD group had higher MMSE scores compared to the AD group (p<0.05). Classical protein biomarkers Tau and pTau were increased in CSF in the AD patients com- pared to the bvFTD and control cases (p<0.0001). Conversely, amyloid-beta1-42 levels were decreased in both the AD and bvFTD group compared to the controls (p<0.0001). We also measured levels of the neurofilament light and the phosphorylated heavy chain but detected large variation in CSF. NfL levels were observed significantly higher (p<0.05) in bvFTD cases and pNFH levels were higher (p<0.05) in AD cases compared to controls. Since we were pri- marily interested in the sporadic, non-genetic bvFTD, we examined our samples for mutations in the genes C9orf72, GRN and MAPT. A total of 41 of the 48 bvFTD and 20 of the 48 AD cases were tested negative for the most prominent gene C9orf72. Further, no mutations in the genes PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 5 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Expression levels of circulating miRNAs in serum do not correlate with levels in CSF Following quality control and the removal of low expressed signals, we were able to compare 34 of the 76 identical CSF and serum miRNA assays in 131 samples. Taking into account all subjects, an average correlation of r = 0.077 (p = ns) indicated no association of miRNA expression levels in serum compared to expression levels in CSF. If we considered expression levels of individual miRNAs, we observed that some of these correlated significantly between serum and CSF, however, only with weak (r<0.3) associations. This finding applied to all three subgroups. Taking into account the mean values of the respective subgroups, the control (r = 0.051, p = ns), bvFTD (r = 0.077, p = ns) and AD (r = 0.057, p = ns) group displayed no significant associations. However, expression levels of a few miRNAs showed significant corre- lations of moderate extent in the respective subgroups. This applied to both miR-19b-3p (r = 0.37, p = 0.023) and miR-25-3p (r = 0.45, p = 0.011) in the control group. Significant posi- tive correlations of miRNA expression levels in serum with expression levels in CSF were also found for miR-143-3p (r = 0.34, p = 0.023), miR-29b-3p (r = 0.33, p = 0.025) and miR-29c-3p (r = 0.44, p = 0.002) in the bvFTD group. Interestingly, miR-24-3p showed a significant nega- tive correlation, r = -0.33, p = 0.026. In the AD group, serum expression levels of let-7f-5p (r = 0.32, p = 0.032), miR-100 (r = 0.36, p = 0.017) and miR-143-3p (r = 0.32, p = 0.047) corre- lated positive with levels in CSF. In turn, we observed a significant negative correlation of miR-197-3p (r = -0.42, p = 0.005) and miR-30a-5p (r = -0.38, p = 0.012). Quality controls indicate low technical variation and consistent detection of miRNA levels in CSF and serum With a mean Cq of 33.61 ± 2.2, the expression of circulating miRNAs in CSF was > 26-fold lower compared to serum with a mean Cq of 28.89 ± 3.66. This was also reflected by the increased number of positive signals detected for each miRNA in our serum data. In total, 77 miRNAs displayed 80% positive signals per miRNA in serum while this only applied to 55 miRNAs in CSF. RNA isolation was monitored by using a subset of synthetic miRNAs that simulate high (UniSp2), medium (UniSp4) and low (UniSp5) expressed signals to measure and compare extraction efficiency [29]. cDNA synthesis was controlled by UniSp6 and cel- miR-39. Our results demonstrate constant extraction efficiency across all samples with accept- able intra-assay variation as well as constant efficiency of the reverse transcription step with no signs of inhibition (S2A Fig). However, two serum samples were removed due to suspicious expression of spikes. We also checked the degree of hemolysis in our serum samples. The obtained data were not affected by erythrocyte miRNA contamination. Only a minority of serum samples indicated a possible but no high risk of hemolysis (S2B Fig). Using Normfinder and GeNorm, we were able to identify a subset of miRNAs for CSF and serum that consider- ably reduced technical variation compared to un-normalized data (S2C and S2D Fig). Further- more, standard curves based on synthetic miRNA oligos were generated for a subset of assays to assess PCR efficiency and to estimate copy numbers. For the selected set of miRNA assays, PCR efficiencies were within an acceptable range of E = 0.95–0.98, which demonstrated low variation and good assay performance (S3 Fig). We also measured CSF and serum samples iso- lated at different days to test the inter-assay variance of our protocol. SYBR-green miRNA PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 6 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls assays displayed consistent expression with low inter-assay variance for serum (R2 = 0.99, F (1,70) = 11078, p < .0001) and CSF (R2 = 0.97, F(1,74) = 2215, p < .0001) showing a trend towards increasing day-day variation at higher expression levels as seen in the CSF data (S4 Fig). PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Cluster analysis of serum expression levels reveals miRNA families and genomic clusters altered in bvFTD and AD samples As a first approach to understand the interaction of miRNAs, hierarchical clustering was applied to the serum dataset containing 131 samples measured for the expression of 96 miR- NAs. To reduce the noise, we excluded miRNAs with low expression values (detailed in Mate- rial and Methods). Using average linkage bottom up clustering a total of 7 clusters in the control, 6 clusters in the bvFTD and 7 clusters in the AD group were detected (Fig 1). These clusters each contain 2–24 miRNAs (Fig 1). Notably, many cluster contained miRNAs with similar sequences (miRNA families) as well as miRNAs of a polycistronic unit that share a common promotor (referred to as genomic cluster, http://www.mirbase.org/, <10 kb). Geno- mic clusters containing only members of a particular family (homo-clusters) as well as clusters containing miRNAs with different seed sequences (hetero-clusters) were present. A good example for co-expression related to similar sequences is Cluster 7 in the control group that contains hsa-miR-27a and hsa-miR-27b or hsa-miR-148a and hsa-miR-148b in Cluster 4. In contrast, the genomic cluster let-7e/miR-99b/125a in Cluster 3 is an example of a hetero-clus- ter. In addition, we identified at least three Modules (dotted lines) in each condition. Modules are clusters of highly interconnected miRNAs that are characterized by dense interactions. Interestingly, the modules differed not only in terms of number and composition of miRNAs between the conditions but Modules 2 and 3 were also identified as densely connected groups in the bvFTD and AD group, which was not observed in the control group. As expected, we identified several pair-wise differential correlations in each condition with a trend of increased correlations in the bvFTD group. This was supported by the fact that some families such as PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 7 / 23 Factor analysis reveals miRNA structure in serum, which allows the classification of bvFTD, AD and control cases The variety of differential correlations in the individual conditions made an interpretation dif- ficult. This led us to consider all samples simultaneously using factor analysis to reduce the number of manifest variables to a few hypothetical variables that could be associated with the diagnoses of our patient samples. The goal of factor analysis is to detect a small set of factors that elucidate as much of the variance of the output variables as possible. The factorability of 73 miRNAs in serum was initially examined. Applying several well-rec- ognised criteria [33–35], 29 miRNAs were tested, which led to an acceptable subject-to-item ratio of ~ 1:5. A total of 44 variables were eliminated because they either failed to meet the minimum criteria or did not contribute to a simple factor structure by primary factor loadings of <\|0.5\|. All items in this analysis had primary loadings over .5. Some of the variables showed cross-loadings, but most showed strong primary loadings and were therefore not excluded. Looking at the scree plot of the serum data, eigenvalues showed that in each case 34.97%, 19.56%, 12.73% and accordingly a total of 67.26% of the variation were explained by the first three factors (S5 Fig). The fourth factor also had an eigenvalue > 1, but explained only a small part of the total variability with 4.72%. As a result, the three-factor solution was preferred to the four-factor model as depicted in Fig 2. Fig 2A shows how the initial miRNAs are correlated with the three factors. We could observe that Factor 1 (green) positively correlated with hsa- let-7g-5p, -miR-101-3p, -106a-5p, -106b-5p, -18b-5p, -20a-5p, -26b-5p, -29b-3p, -301a-3p, -30b-5p and -27a-3p and negatively correlated with hsa-miR-1246, -146a-5p, -30d-5p, -miR- 320a and -320b. In contrast. Factor 2 (red) positively correlated with hsa-let-7d-5p, -let-7f-5p and -miR-98-5p and negatively correlated with hsa-miR-15b-3p, -16-5p, -32-5p, -451a, -532- 5p and -19b-3p whereas Factor 3 (brown) negatively correlated with hsa-let-7c-5p, -let-7e-5p, -miR-22-3p and miR-29c-3p. The factor loading matrix for this final solution is presented in S6 Table. Interestingly, the use of the coordinates of the observations of the two-factor model already resulted in a significant discrimination of cognitively healthy controls from the bvFTD and AD group (Fig 2B). More interestingly, the structure also visually indicated a trend towards a grouping of AD and bvFTD cases, however, with a certain number of false positive signals, respectively (Fig 2B). Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Fig 1. Unsupervised hierarchal clustering of miRNA levels measured in serum of healthy control, bvFTD and AD samples. Using the DiffCorr package, the genes were grouped according to their expression patterns in each subtypes (cognitively healthy controls, bvFTD and AD) using the cluster.molecule function. We used (1 −correlation coefficient) as a distance measure (the cutoff value was a coefficient of 0.6) based on the cutree function. We then visualized the module network using the get.eigen.molecule and get.eigen.molecule.graph functions. MiRNAs that share similar seed sequences (miRNA families) are coloured. MiRNAs that are co-transcribed as a polycistronic unit (http://www.mirbase.org/, < 10 kb) are listed in S5 Table. AD = Alzheimer’s disease, bvFTD = behavioural variant frontotemporal dementia, M = Module. htt //d i /10 1371/j l 0197329 001 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Fig 1. Unsupervised hierarchal clustering of miRNA levels measured in serum of healthy control, bvFTD and AD samples. Using the DiffCorr package, the genes were grouped according to their expression patterns in each subtypes (cognitively healthy controls, bvFTD and AD) using the cluster.molecule function. We used (1 −correlation coefficient) as a distance measure (the cutoff value was a coefficient of 0.6) based on the cutree function. We then visualized the module network using the get.eigen.molecule and get.eigen.molecule.graph functions. MiRNAs that share similar seed sequences (miRNA families) are coloured. MiRNAs that are co-transcribed as a polycistronic unit (http://www.mirbase.org/, < 10 kb) are listed in S5 Table. AD = Alzheimer’s disease, bvFTD = behavioural variant frontotemporal dementia, M = Module. https://doi.org/10.1371/journal.pone.0197329.g001 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 8 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls mir-17 or let-7 or clusters such as miR-29b/29a and miR-29b/29c appeared more intermixed in the bvFTD condition. S4 Table shows the top 20 significantly differential co-expressed miR- NAs (FDR<0.05) in bvFTD cases compared to cognitively healthy controls. Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Fig 2. Factor and discriminant analysis of miRNA signals detected in serum. We have used factor and discriminant analysis to identify the variables that can best characterize and classify the bvFTD, AD and healthy control samples in our serum cohort. The following results are shown: (A) factor model of n = 29 serum miRNAs that load on the Factors 1–3 with factor loadings \|0.5\|, (B) observation plot showing coordinates of the observations resulting from the two-factor model (Factors 1–2), and result of the discriminant analysis using a re-factoring 2-factor model based on the n = 29 miRNAs from the original 3-factor model (Fig 2A) using (C) a training (n = 78) and validation (n = 53) set and (D) all cases (n = 131). Fig 2. Factor and discriminant analysis of miRNA signals detected in serum. We have used factor and discriminant analysis to identify the variables that can best characterize and classify the bvFTD, AD and healthy control samples in our serum cohort. The following results are shown: (A) factor model of n = 29 serum miRNAs that load on the Factors 1–3 with factor loadings \|0.5\|, (B) observation plot showing coordinates of the observations resulting from the two-factor model (Factors 1–2), and result of the discriminant analysis using a re-factoring 2-factor model based on the n = 29 miRNAs from the original 3-factor model (Fig 2A) using (C) a training (n = 78) and validation (n = 53) set and (D) all cases (n = 131). https://doi.org/10.1371/journal.pone.0197329.g002 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Factor analysis reveals miRNA structure in serum, which allows the classification of bvFTD, AD and control cases To improve the differentiation of the samples, the miRNAs of the third factor, which also contributed with 12.73% to the total variance, were taken into account in a subsequent discriminant analysis as shown in Fig 2C. After the data was split into a training- (n = 78) and validation (n = 53) set, 84.62% and respectively 60% of the total cases were successfully classified based on the analysis (Fig 2C). If the individual subgroups were considered, 100% of the controls, 70% of the bvFTD and 86% of the AD cases of the training set were correctly classified (Fig 2C). On the other hand, the validation set showed that 71% of controls, 67% of bvFTD and 44% of AD cases were correctly assigned (Fig 2C). Considering the entire data set (n = 131 serum samples), the controls were correctly predicted with 97%, the bvFTD cases with 77% and the AD cases with 72% accuracy (Fig 2D). Looking at the CSF data, n = 15 miRNAs were suitable for factor analysis. This resulted in an acceptable subject-to-item ratio of ~ 1:9. The scree plot indicated that the first three factors had eigenvalues >1, which explained 26.33%, 24.25% and 9.29% of the variance, respectively (S5 Fig). Since we did not observe a trend towards a valuable separation of our samples using the CSF miRNA factor model, we further concentrated on our serum data. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 9 / 23 Expression analysis identifies significantly de-regulated miRNAs between bvFTD, AD and control samples Next, we tested whether the miRNAs of the 3-factor model as well as other miRNAs in our CSF and serum data displayed significantly different expression levels and to what extent these signals contributed to the classification of our samples by calculating corresponding AUC val- ues. Using MANOVA, we identified a number of miRNAs, all of which were differentially expressed over a ddCt of \|0.58\| (S7 Table). Except for miR-30d-5p and miR-101-3p, all signals of Factor 1 were significantly de-regulated. In contrast, only miR-22-3p of Factor 3 was found as differentially expressed but none of the signals from Factor 2. In our CSF data, we identified a total of 10 miRNAs with significantly different expression levels (Fig 3A). The best classifier was miR-125a-5p that discriminated bvFTD cases with 72% sensitivity and 81% specificity as well as AD cases with 74% sensitivity and 82% specificity from our controls (Fig 3B). In contrast, we also observed miRNAs with significantly different expression levels between AD and bvFTD patients (Fig 3A), which showed more moderate classification performance. With an AUC of 0.73 (95% CI, 0.623–0.832), miR-30a-5p yielded PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 10 / 23 In the second expression pattern, miRNAs were either significantly up- or downregulated both in bvFTD and AD patients com- pared to controls (Fig 3G and 3J). For example, miR-301a-3p demonstrated an AUC of 0.96 (95% CI, 0.918–0.996) and classified bvFTD cases with high sensitivity (96%) and specificity (84%), whereas miR-26b-5p showed and AUC of 0.97 (95% CI, 0.940–0.999) and classified AD cases with 89% sensitivity and specificity with respect to our control group (Fig 3H and 3I). Accordingly, miR-320a was the best classifier against controls observed as collectively down reg- ulated in bvFTD and AD. ROC analysis resulted in an AUC of 0.96 (95% CI, 0.909–1.003), which classified bvFTD cases with 96% sensitivity and 90% specificity and an AUC of 0.90 (95% CI, 0.835–0.969) to predict AD cases with 83% sensitivity and 90% specificity (Fig 3K and 3I). We also examined the respective groups on a gender-specific basis. As expected, most sig- nals showed the same differentially expressed levels in both men and women. Interestingly, in our serum data, we identified signals with increased classification performance to classify bvFTD from AD in the male but not in the female cohort compared to the original analysis that considered both males and females. Signals miR-103a-3p (AUC = 0.80), miR-106a-5p (AUC = 0.80) and miR-1246 (AUC = 0.85) demonstrated increased sensitivities (70–77%) and specificities (75–85%) to classify bvFTD from AD cases. This trend was also overserved in our CSF data but less significant in terms of classification performance. We also examined the respective groups on a gender-specific basis. As expected, most sig- nals showed the same differentially expressed levels in both men and women. Interestingly, in our serum data, we identified signals with increased classification performance to classify bvFTD from AD in the male but not in the female cohort compared to the original analysis that considered both males and females. Signals miR-103a-3p (AUC = 0.80), miR-106a-5p (AUC = 0.80) and miR-1246 (AUC = 0.85) demonstrated increased sensitivities (70–77%) and specificities (75–85%) to classify bvFTD from AD cases. This trend was also overserved in our CSF data but less significant in terms of classification performance. MiRNAs from factor model in serum correlate with levels of amyloid-beta1- 42 and neurofilaments light chain detected in CSF As described previously, we were unable to observe a global relationship between miRNA expression levels in serum and CSF. Instead, only a weak trend towards associations of single signals was observed. However, we were also interested in whether the expression levels of miRNAs correlated with those of classical protein biomarkers in the CSF. As a result, we found a large number of significant correlations of our serum miRNAs with these biomarkers. The most interesting finding was that individual miRNAs that correlated with Factor 1 predomi- nantly correlated with amyloid-beta1-42 (Fig 4A) whereas miRNAs that correlated with Factor 2 primarily correlated with pNfH (Fig 4B). For example, miR-320a (r = 0.54, p = 0.028) showed a significantly positive correlation, whereas miR-18b-5p (r = -0.42, p = 0.042) dis- played a significantly negative correlation with CSF amyloid-beta1-42 levels. Both signals also positively correlated with Factor 1 (Fig 2A). Another example are let-7d (r = -0.52, p<0.001), let-7f (r = -0.48, p<0.001) and miR-98-5p (r = -0.44, p = 0.005), all of which show significantly negative correlations with CSF pNFH levels (Fig 4B). These signals were also found to posi- tively correlate with Factor 2 as identified by factor analysis (Fig 2A). Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Fig 3. Differentially expressed miRNAs in bvFTD, AD and cognitively healthy control cases detected in CSF and serum. Expression levels of n = 96 circulating miRNAs were measured in CSF (n = 140) and serum (n = 131) samples from bvFTD (n = 48/48) and AD patients (n = 48/47) as well as healthy controls (n = 44/38) and compared using MANOVA and ROC curves. Displayed are signals with significantly different expression levels after multiple comparisons detected in (A) CSF: group comparisons of AD, bvFTD and healthy controls with (B-C) corresponding ROC curves and (D) serum: group comparisons of AD, bvFTD and healthy controls (up- and downregulated miRNAs) with (E-F) corresponding ROC curves, (G) serum: group comparisons of AD, bvFTD and healthy controls (only upregulated miRNAs) with (H-I) corresponding ROC curves and (J) serum: group comparisons of AD, bvFTD and healthy controls (only downregulated miRNAs) with (K-L) corresponding ROC curves. Expression ratio: ddCt = mean dCtAD or bvFTD −mean dCtHC. Dotted lines indicate ddCt cut-off of \|0.58\|. Error bars indicate mean ± SEM. BH = Benjamini-Hochberg. https://doi.org/10.1371/journal.pone.0197329.g003 Fig 3. Differentially expressed miRNAs in bvFTD, AD and cognitively healthy control cases detected in CSF and serum. Expression levels of n = 96 circulating miRNAs were measured in CSF (n = 140) and serum (n = 131) samples from bvFTD (n = 48/48) and AD patients (n = 48/47) as well as healthy controls (n = 44/38) and compared using MANOVA and ROC curves. Displayed are signals with significantly different expression levels after multiple comparisons detected in (A) CSF: group comparisons of AD, bvFTD and healthy controls with (B-C) corresponding ROC curves and (D) serum: group comparisons of AD, bvFTD and healthy controls (up- and downregulated miRNAs) with (E-F) corresponding ROC curves, (G) serum: group comparisons of AD, bvFTD and healthy controls (only upregulated miRNAs) with (H-I) corresponding ROC curves and (J) serum: group comparisons of AD, bvFTD and healthy controls (only downregulated miRNAs) with (K-L) corresponding ROC curves. Expression ratio: ddCt = mean dCtAD or bvFTD −mean dCtHC. Dotted lines indicate ddCt cut-off of \|0.58\|. Error bars indicate mean ± SEM. BH = Benjamini-Hochberg. Fig 3. Differentially expressed miRNAs in bvFTD, AD and cognitively healthy control cases detected in CSF and se PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 11 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls the best classification by separating bvFTD from AD cases with 78% sensitivity and 68% speci- ficity (Fig 3C). In our serum data, a total of n = 31 miRNAs were identified with significantly different expression levels between our subgroups. Two basic expression patterns could be observed. One was characterized by signals that were differentially regulated in the bvFTD or AD group compared to the controls (Fig 3D). Here, bvFTD cases separated best from controls with 77% sensitivity and 72% specificity by miR-27a with an AUC of 0.86 (95% CI, 0.775–0.935) (Fig 3E). In contrast, miR-20a-5p demonstrated an AUC of 0.85 (95% CI, 0.768–0.938) and the highest specificity of 92%, whereas miR-29b-3p displayed an AUC of 0.83 (95% CI, 0.738–0.931) by sep- arating AD cases from controls with 93% sensitivity (Fig 3F). In the second expression pattern, miRNAs were either significantly up- or downregulated both in bvFTD and AD patients com- pared to controls (Fig 3G and 3J). For example, miR-301a-3p demonstrated an AUC of 0.96 (95% CI, 0.918–0.996) and classified bvFTD cases with high sensitivity (96%) and specificity (84%), whereas miR-26b-5p showed and AUC of 0.97 (95% CI, 0.940–0.999) and classified AD cases with 89% sensitivity and specificity with respect to our control group (Fig 3H and 3I). Accordingly, miR-320a was the best classifier against controls observed as collectively down reg- ulated in bvFTD and AD. ROC analysis resulted in an AUC of 0.96 (95% CI, 0.909–1.003), which classified bvFTD cases with 96% sensitivity and 90% specificity and an AUC of 0.90 (95% CI, 0.835–0.969) to predict AD cases with 83% sensitivity and 90% specificity (Fig 3K and 3I). In our serum data, a total of n = 31 miRNAs were identified with significantly different expression levels between our subgroups. Two basic expression patterns could be observed. One was characterized by signals that were differentially regulated in the bvFTD or AD group compared to the controls (Fig 3D). Here, bvFTD cases separated best from controls with 77% sensitivity and 72% specificity by miR-27a with an AUC of 0.86 (95% CI, 0.775–0.935) (Fig 3E). In contrast, miR-20a-5p demonstrated an AUC of 0.85 (95% CI, 0.768–0.938) and the highest specificity of 92%, whereas miR-29b-3p displayed an AUC of 0.83 (95% CI, 0.738–0.931) by sep- arating AD cases from controls with 93% sensitivity (Fig 3F). Discussion In the present study we I) collected samples from multiple clinical centers across Germany, II) defined 3 homogenous patient groups with increased sample sizes (bvFTD n = 48, AD n = 48 12 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls 71/journal.pone.0197329 May 10, 2018 13 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 13 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 13 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Fig 4. Correlations of miRNA expression levels in serum with CSF protein biomarkers. Depicted are normalized expression levels dCt = Ct (Ctmean RefmiR −CtmiR) of (A) miRNAs from the original 3-factor model that correlated with Factor 1 vs CSF levels of amylod-beta1-42 in the control group and (B) miRNAs from the original 3-factor model that correlated with Factor 2 vs CSF levels of pNfH in the bvFTD group. pNfH = phosphorylated neurofilament heavy chain. htt //d i /10 1371/j l 0197329 004 Fig 4. Correlations of miRNA expression levels in serum with CSF protein biomarkers. Depicted are normalized expression levels dCt = Ct (Ctmean RefmiR −CtmiR) of (A) miRNAs from the original 3-factor model that correlated with Factor 1 vs CSF levels of amylod-beta1-42 in the control group and (B) miRNAs from the original 3-factor model that correlated with Factor 2 vs CSF levels of pNfH in the bvFTD group. pNfH = phosphorylated neurofilament heavy chain. Fig 4. Correlations of miRNA expression levels in serum with CSF protein biomarkers. Depicted are normalized expression levels dCt = Ct (Ctmean RefmiR −CtmiR) of (A) miRNAs from the original 3-factor model that correlated with Factor 1 vs CSF levels of amylod-beta1-42 in the control group and (B) miRNAs from the original 3-factor model that correlated with Factor 2 vs CSF levels of pNfH in the bvFTD group. pNfH = phosphorylated neurofilament heavy chain. https://doi.org/10.1371/journal.pone.0197329.g004 https://doi.org/10.1371/journal.pone.0197329.g004 and cognitively healthy controls n = 44), III) compared expression levels in both CSF and serum samples and IV) focused on a limited set of miRNAs. Many studies use RT-qPCR to search for circulating miRNA biomarkers but do not consider the MIQE guidelines, do not control for hemolysis in serum or plasma samples, or use an inappropriate normalization method. Discussion We included appropriate quality control procedures that identified and reduced known (pre-) analytical sources of variation (S1–S4 Figs) and adapted our protocol to the MIQE guidelines (S2 Table) to increase the experimental transparency and reliability of our data. One major finding was that we did not observe a strong association of miRNA expression levels throughout the cohort or subgroups between our CSF and serum samples. This is consis- tent with results from a comparable study by Wang et al. and Freischmidt et al., who also observed a poor association of miRNA levels in CSF and serum [36, 37]. Wang et al., however, showed an increased correlation of CSF miRNA levels with miRNAs from the Choroid Plexus [37]. In our case single miRNAs such as miR-29b and miR-29c indicated a small trend for a correlation (R2 = 10–20%) between CSF and serum. However, these miRNAs are known to control fibrinogen production, which is a basic pathway both in the brain and the periphery [38]. In summary, the results point to autonomous networks that may independently respond to neurodegenerative processes and show little similarity in cognitively healthy controls. How- ever, one has to keep in mind that low CSF levels point to low input into CSF and that levels might dilute out during circulation and after blood-brain-barrier transport. This might be another reason why potential changes in CSF cannot necessarily be seen in serum. The main goal was to identify miRNA signatures to classify our bvFTD, AD and cognitively healthy control cases using multivariate statistics. To increase reliability, we focused on the serum data, which contained higher quality signals compared to the CSF data. One approach towards this aim was to use co-expression analysis. The first step included unsupervised cluster analysis to analyse possible interactions within each subgroup. An interesting observation was that the cluster analysis, which based on cell-free serum expression levels, grouped many sig- nals according to their affiliation with a polycistronic unit (e.g. miR-106a/-b cluster) or a par- ticular miRNA family (e.g. mir-17, let-7, mir-15 or mir-320 family). These (co-)expression patterns are known to occur within tissue cells [39, 40], different blood cell types [41] or whole blood [42]. The result therefore suggested that the profile of circulating miRNAs in serum appears to be more similar to the cell than expected but also point to a more unspecific release into the circulation. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls Factor 1 correlated with miR-106b-5p and miR-20a-5p, which in turn belong to the mir- 106b~25 and mir-17 cluster. Both were found as significantly upregulated in bvFTD and AD, which is in line with results from Cheng et al, who investigated and validated miRNA levels isolated from serum exosomes using RNA seq [44]. In addition, both signals demonstrated a good to excellent diagnostic accuracy to distinguish AD cases from controls. Interestingly, these miRNAs were also shown to regulate APP in vitro and in neuronal cell lines [45]. In addi- tion, we identified a differential correlation pattern for the mir-30b/30d cluster. We observed that miR-30b-5p was positively and miR-30d-5p negatively correlated with Factor 1 and we further confirmed significantly upregulated expression levels in bvFTD and AD compared to our controls. Notably, miR-30b-5p was found to correlate with amyloid plaque density by a study of Burgos et al. [5]. In addition, we found miR-30d-5p (R2 = 22%) to positively correlate with CSF amyloid-beta1-42 levels in our cognitively healthy controls. This also applied to miR- 320a (R2 = 30%), which is predicted to downregulate APP [46], whereas its family member miR-320b was found as a possible regulator of human-specific neural development [47]. We could also confirm both signals as significantly down-regulated in bvFTD and AD compared to our cognitively healthy controls. Overall, miR-320a and miR-320b showed a good to excel- lent diagnostic accuracy to correctly separate bvFTD and AD cases from controls, with miR- 30b-5p ranking only AD cases correctly. Factor 2 correlated positively with the let-7f/7d and let-7f/mir-98 clusters, whose members belong to the let-7 family, but we could not observe significantly different expression levels. However, let-7d and miR-98 were found as significantly down-regulated in AD compared to controls by Burgos et al. [5]. Interestingly, miR-98-5p was also found to act as a target for AD by regulating the production of beta-amyloid through modulating SNX6 Expression [48]. Another study showed, that inhibition of miR-98 in N2a/APP cells up-regulated the IGF-1 protein level and suppressed Aβ production [49]. Interestingly, we found all members of the let-7f/7d (R2 = 23–27%) and let-7f/mir-98 (R2 = 19%) cluster to negatively correlate with CSF pNfH levels in the bvFTD group. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Discussion Interestingly, Leidinger and colleagues identified similar miRNA clusters and families as specifically enriched in blood cell populations positive for different CD mark- ers, which supports this idea [41]. However, the extent to which members of a miRNA family (independent transcription units) displayed similar expression patterns in serum was unex- pected because the mechanisms responsible for interchromosomally-coordinated co-expres- sion are not yet fully understood and are the subject of current research [43]. In the second step we calculated differential correlations to better compare the individual profiles between our subgroups. However, the increased number of differential correlations made the interpre- tation difficult and prompted us to use factor analysis, which considered samples of all sub- groups. The initially 73 serum miRNAs were reduced by ~40% to 29 signals, which in turn resulted in a three-factor miRNA model. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 14 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls with Factor 2 and Factor 3 showed an overall lower diagnostic potential compared to those correlated with Factor 1. Except miR-29b-3p (AUC = 0.83) for the classification of AD cases. In summary, all signals from the three-factor model explained in total > 67% of the vari- ance. We therefore calculated a subsequent discriminant analysis to evaluate the performance of this model to classify our subgroups. As expected from the structure, the miRNA signature was able to identify cognitively healthy controls with 97% accuracy. The result outperforms other blood-based assays such as the detection of amyloid levels [55] and performs at least as well [56–58], if not better [59–61] compared to miRNAs in blood that have been identified in other studies. More interestingly, the same signature was able to classify bvFTD cases with 77% and AD cases with 72% accuracy and can at least in part compete with results from tradi- tional protein based tests in CSF [62]. This is mostly due to the fact that the majority of miR- NAs that were either up- or downregulated in AD compared to our cognitively healthy controls displayed a similar expression pattern in bvFTD patients. We believe that these rather unspecific signals are mainly due to the underlying neurodegeneration observed in AD and bvFTD. There were basically no signals on our panel that showed an opposing expression except for those that have been identified in our gender specific analysis as mentioned earlier. However, this should be the focus of further studies as specific signals would improve diagno- sis by helping to exclude other dementias. Concerning our CSF data, we identified individual miRNAs that showed significantly dif- ferent expression levels and displayed diagnostic potential. For example miR-125a-5p that dis- criminated cognitively healthy controls from AD with good (AUC = 0.84) or miR-30a-5p that classified bvFTD from AD cases with moderate (AUC = 0.73) accuracy. This is in line with Cogswell, who also identified these signals as significantly upregulated in AD [63]. However, due to quality control we lost some signals in our CSF data and factor analysis did not reveal a miRNA signature with high discriminatory value. In summary, our data show that our circulating miRNA profile in CSF was not comparable to that in serum and that serum miRNAs were better detectable compared to those in CSF. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 This also applied to miR-532 (R2 = 34%) and the mir-15 fam- ily members’ miR-15b-3p (R2 = 37%) and miR-16-5p (R2 = 37%) that negatively correlated with Factor 2. miR-15b-3p displayed significantly upregulated expression in AD and miR-16- 5p significantly downregulated expression levels in bvFTD and AD compared to controls, however, each below a ddCt of .58. In addition, both miRNAs correlated significantly posi- tively with CSF pNfH levels in the bvFTD group. Neurofilaments are major proteins of neu- rons and are particularly concentrated in axons and detection in CSF provides information about the degree of axonal injury [50] and was found a biomarker for genetic frontotemporal dementia [51, 52]. Notably, Burgos et al. identified miR-16-5p to negatively correlate with Braak stage [5], which supports the idea of miR-16-5p as a potential marker of neuronal injury. Furthermore, the miR-15 family has also been shown to modulate Tau phosphorylation through ERK1 leading to neuronal death in Neuro2a cells and primary cortical neurones [53]. However, we did not observe significant correlations with CSF Tau or pTau levels in our groups. Factor 3 contained no more than 2 members of a specific miRNA family. With the excep- tion of the miR-29 family. In this case, miR-29c-3p correlated negatively with Factor 3, with the family member miR-29b-3p positively correlating with Factor 1. In contrast, miR-29a-3p did not correlate with any of the three factors but was found to negatively correlate with CSF Tau (r = -0.56) and pTau (r = -69) levels (data not shown). However, we only found expression levels of miR-29b-3p as significantly upregulated in AD compared to our controls. Interest- ingly, Hebert et al. identified the mir-29a/b-1 as significantly decreased in AD brain as well as its regulation of amyloid-beta levels by upregulated levels of BACE1 [54]. miRNAs correlating PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 15 / 23 In addition, circulating miRNAs in serum show a strong tendency to form clusters, either because of their genomic organization or because of homologies in their sequences. Co-expression analysis displayed differently co-expressed miRNAs between our subgroups. However, the identified associations were complex and difficult to interpret. In addition, possible associa- tions of other miRNAs could not be demonstrated, since we did not include related signals on our serum panel. We could also show that multivariate methods such as factor analysis can identify miRNA signatures in serum able to classify bvFTD, AD and control cases with accept- able diagnostic accuracy. Due to their genomic organization and transcriptional expression pattern we therefore think that it is more likely to identify a biomarker consisting of either members of a miRNA cluster or family compared to single miRNAs. Another interesting observation was that particularly miRNAs associated with our factor model also correlated with CSF amyloid-beta1-42 and phosphorylated neurofilament heavy chain levels either with our control- or bvFTD group. This suggested that de-regulated miRNAs of a family or cluster may possibly be able to monitor the neurodegenerative progression seen in AD or bvFTD. However, further evidence from cell culture experiments is necessary to better describe the functional associations of these miRNAs. Overall, when analysing miRNA clusters and fami- lies, it must be noted that it is based on current understanding and annotation and that this relationship will change over time whenever new miRNA species are identified. However, unsupervised clustering is based on what was actually measured, thus reduces this bias in part and is therefore well suited for the detection of networks. However, one shortcoming towards the identification of miRNA signatures using cluster and factor analysis is the limited number of signals in our study. The (factor-) structures identified by us result from a small part of the PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 16 / 23 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls miRNome that circulate in serum. However, there are other related (S5 Table) and novel miR- NAs that we did not measure but that may further increase complexity. The last miRBase update was 2014. Novel miRNAs have been identified by NGS, but their expression levels have not yet extensively been investigated using qPCR and are not yet available on PCR panels. An updated selection of abundant signals could thus provide novel candidates for miRNA profil- ing studies. A shortcoming towards our study group was that not all individual were tested for one of the disease causing mutations. However, the probability of carrying one of the tested mutations is generally low and the majority was tested negative so that a few possible mutation carriers should not bias the results. Furthermore, some variation at baseline miRNA expres- sion was observed across the different study sites. We think that (possible) centre effects should affect all miRNAs equally (due to a harmonized and standardised protocol for sampling and extraction) so that variation on a single miRNA basis should not occur. The variation may pos- sibly be the result of a numerically unequal distribution of the patient groups from the differ- ent centres. Multi-centre studies should therefore make sure that the proportion of patient groups from the various centres is as equal as possible. Considering the discussed points in fol- low up studies will greatly contribute to identify other relevant miRNAs to better understand the complex expression pattern overserved in body fluids, and may further improve the classi- fication performance to separate AD and bvFTD cases. Supporting information S1 Fig. Baseline expression across all miRNAs between different study sites. A main effect for baseline expression levels (across all miRNAs) between study sites was found in our serum (F(9, 121) = 5.21, p < 0.001) and CSF (F(9, 130) = 4.50, p < 0.001) data. Serum expression levels from the study site in Ulm (M = 29.17, SD = 1.06) were lower compared to Mu¨nchen-tu (M = 27.86, SD = 0.61), Homburg (M = 27.74, SD = 0.29), Erlangen (M = 27.9, SD = 0.46) and Hamburg (M = 28.3, SD = 0.93). In contrast, CSF expression levels from the study site in Go¨t- tingen (M = 32.42, SD = 1.23) were higher compared to Ulm (M = 33.59, SD = 0.65), Mu¨nchen- tu (M = 33.56, SD = 0.49) and Erlangen (M = 33.95, SD = 0.54). However, each difference was below the critical threshold of ddCt < \|0.58\| except for Go¨ttingen vs. Erlangen (ddCt = 0.62). (TIF) S2 Fig. Pre-analytical variation of circulating miRNA in CSF and serum samples. a) Box plot (whiskers: 2.5–97.5 percentile) of synthetic miRNAs display low technical variation with acceptable intra-assay variation of UniSp2: CqCSF 17.59 ± 0.37 and Cqserum 17.49 ± 0.49; UniSp4: CqCSF 24.92 ± 0.52 and Cqserum 24.42 ± 0.59 and with a trend of increasing variation towards the isolation of lower expressed transcripts like UniSp5: CqCSF 30.65 ± 0.92 and Cqserum 31.23 ± 0.51. UniSp6: CqCSF 18.63 ± 0.46 and Cqserum 17.59 ± 0.16 and cel-miR-39 Cqserum 24.07 ± 0.25 were used to monitor the cDNA synthesis reactions and indicated con- stant RT efficiency with no signs of inhibition. b) The hemolysis plot indicates expression ratios of constant miR-23a and red blood cell sensitive miR-451a to monitor serum samples for signs of cellular contamination or hemolysis. With a mean dCqmiR-23a –miR-451 = 4.05 ± 1.07, most of the serum samples did not display signs of hemolysis (dCq 5, yellow line). Only a few signals showed a dCq > 5 but none of the samples appeared at high risk of hemolysis (dCq 7, red line). c,d) The cumulative distribution plots display different miRNA normalisation strategies applied on the serum and CSF data. Normalization with reference miRNAs identified by NormFinder and GeNorm resembled normalisation to the global mean and considerably reduced technical variation compared to un-normalized data or data nor- malized to internal standards. Supporting information Top 20 list of differentially coexpressed miRNAs from the serum dataset. Shown are significantly pair-wise differential correlations (FDR < 0.05) of miRNA expression levels in serum between cognitively healthy controls and bvFTD cases using the comp.2.cc.fdr func- tion from the DiffCorr package. r = Pearson correlation coefficient, lfdr = local false discovery rate. (XLSX) S5 Table. Table of miRNAs used for hierarchal cluster analysis. Depicted are 1) miRNAs, 2) sequence information, 3) clustered miRNAs on genome (http://www.mirbase.org/, <10 kb), 4) paralogous miRNAs, 5) location on chromosome, 6) miRNA family and 7) correlation with Factor 1–3 from factor analysis. miRNAs in red were not included on our serum panel. (XLSX) S6 Table. Factor loadings and communalities based on a principal factor analysis for n = 29 miRNAs detected in n = 131 serum samples. Only miRNAs based on factor loadings \|0.5\| were considered significant in contributing to the respective factor. F = Factor. (XLSX) S4 Fig. Day-to-Day reproducibility. RNA from A) CSF and B) serum isolated with miRNeasy serum Kit on different days was measured on each corresponding 96-well panel. Interplate cal- ibrated Cq values were plotted against each other to assess the degree of inter-assay variability by linear regression. R2 = coefficient of determination. (TIF) S5 Fig. Scree plots of serum and CSF data. The scree plot shows the variance explained by each factor in a factor analysis and is used to assess the optimum number of factors to take into account for further analysis. (TIF) S1 Table. Raw Cq values of endogenous miRNAs (cut-off Cq < 37) detected in CSF and serum samples. Listed is the FOC (number of positive miRNAs), arithmetic mean of raw Cq values of each subgroup and the standard deviation, the miRBase accession and the miRNA sequence. bvFTD = behavioural variant frontotemporal dementia, Cq = quantification cycle, FOC = frequency of occurrence, HC = cognitively healthy controls, SD = standard deviation. (XLSX) S2 Table. MIQE checklist. Provided are necessary information recommended by the MIQE guidelines to increase experimental transparency of quantitative real-time PCR experiments. (XLSX) Supporting information (TIF) 17 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls S3 Fig. miRNA standard curves for SYBR green miRCURY PCR assays. Plotted are mean Cq values from n = 4 replicate standard curves vs the log2 copy numbers. Standard curves were generated for a subset of assays by using a dilution series of a pool of known input amounts of synthetic miRNA oligonucleotides corresponding to the target sequence of the assay. Red error bars depict mean Cq ± CI. R2 = coefficient of determination, E = PCR efficiency ± CI, CI = confidence interval. (TIF) S3 Fig. miRNA standard curves for SYBR green miRCURY PCR assays. Plotted are mean Cq values from n = 4 replicate standard curves vs the log2 copy numbers. Standard curves were generated for a subset of assays by using a dilution series of a pool of known input amounts of synthetic miRNA oligonucleotides corresponding to the target sequence of the assay. Red error bars depict mean Cq ± CI. R2 = coefficient of determination, E = PCR efficiency ± CI, CI = confidence interval. (TIF) S4 Fig. Day-to-Day reproducibility. RNA from A) CSF and B) serum isolated with miRNeasy serum Kit on different days was measured on each corresponding 96-well panel. Interplate cal ibrated Cq values were plotted against each other to assess the degree of inter-assay variability by linear regression. R2 = coefficient of determination. (TIF) S5 Fig. Scree plots of serum and CSF data. The scree plot shows the variance explained by each factor in a factor analysis and is used to assess the optimum number of factors to take into account for further analysis. (TIF) S1 Table. Raw Cq values of endogenous miRNAs (cut-off Cq < 37) detected in CSF and serum samples. Listed is the FOC (number of positive miRNAs), arithmetic mean of raw Cq values of each subgroup and the standard deviation, the miRBase accession and the miRNA sequence. bvFTD = behavioural variant frontotemporal dementia, Cq = quantification cycle, FOC = frequency of occurrence, HC = cognitively healthy controls, SD = standard deviation. (XLSX) S2 Table. MIQE checklist. Provided are necessary information recommended by the MIQE guidelines to increase experimental transparency of quantitative real-time PCR experiments. (XLSX) S3 Table. Sequence information of synthetic oligos used for absolute quantification. (XLSX) S4 Table. Acknowledgments We thank the following members of the FTLDc study group: Franziska Albrecht (Leipzig), Sarah Anderl-Straub (Ulm), So¨nke Arlt (Hamburg), Christine v. Arnim (Ulm), Svenja Busch- beck (Hamburg), Jan Kassubek (Ulm), Bernhard Landwehrmeyer (Ulm), Jan Lehmbeck (Hamburg), Albert Ludolph (Ulm), Katharina Schuemberg (Leipzig), Kai Schumacher (Ham- burg), Elisa Semler (Ulm), Ingo Uttner (Ulm). S3 Table. Sequence information of synthetic oligos used for absolute quantification. (XLSX) S4 Table. Top 20 list of differentially coexpressed miRNAs from the serum dataset. Shown are significantly pair-wise differential correlations (FDR < 0.05) of miRNA expression levels in serum between cognitively healthy controls and bvFTD cases using the comp.2.cc.fdr func- tion from the DiffCorr package. r = Pearson correlation coefficient, lfdr = local false discovery rate. (XLSX) S5 Table. Table of miRNAs used for hierarchal cluster analysis. Depicted are 1) miRNAs, 2) sequence information, 3) clustered miRNAs on genome (http://www.mirbase.org/, <10 kb), 4) paralogous miRNAs, 5) location on chromosome, 6) miRNA family and 7) correlation with Factor 1–3 from factor analysis. miRNAs in red were not included on our serum panel. (XLSX) S6 Table. Factor loadings and communalities based on a principal factor analysis for n = 29 miRNAs detected in n = 131 serum samples. Only miRNAs based on factor loadings \|0.5\| were considered significant in contributing to the respective factor. F = Factor. (XLSX) 18 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls S7 Table. Differentially expressed miRNAs detected in serum and CSF. Listed are miRNAs that displayed significantly different expression levels in either serum or CSF samples between cognitively healthy control, bvFTD and AD cases. Fold change is calculated as ddCt = dCtCtRef −CtmiRNA −dCtCtRef −CtmiRNA where larger values reflect higher abundance. AD = Alzheimer’s disease, bvFTD = behavioural variant frontotemporal dementia, HC = cognitively healthy con- trols, p = p-value. (XLSX) Project administration: Johannes Denk. Resources: Felix Oberhauser, Johannes Kornhuber, Jens Wiltfang, Klaus Fassbender, Matthias L. Schroeter, Janine Diehl-Schmid, Johannes Prudlo, Adrian Danek, Bernhard Landwehr- meyer, Martin Lauer, Markus Otto. Software: Johannes Denk. Supervision: Holger Jahn. Supervision: Holger Jahn. Visualization: Johannes Denk. Writing – original draft: Johannes Denk, Holger Jahn. Writing – original draft: Johannes Denk, Holger Jahn. Writing – review & editing: Johannes Denk, Felix Oberhauser, Johannes Kornhuber, Jens Wiltfang, Klaus Fassbender, Matthias L. Schroeter, Alexander E. Volk, Janine Diehl- Schmid, Johannes Prudlo, Adrian Danek, Bernhard Landwehrmeyer, Martin Lauer, Mar- kus Otto, Holger Jahn. Author Contributions Conceptualization: Johannes Denk, Holger Jahn. Conceptualization: Johannes Denk, Holger Jahn. Conceptualization: Johannes Denk, Holger Jahn. Data curation: Johannes Denk. Data curation: Johannes Denk. Data curation: Johannes Denk. Formal analysis: Johannes Denk. Investigation: Johannes Denk. Methodology: Johannes Denk, Alexander E. Volk. Project administration: Johannes Denk. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 3. Schmiedel JM, Klemm SL, Zheng Y, Sahay A, Bluthgen N, Marks DS, et al. Gene expression. Micro- RNA control of protein expression noise. Science (New York, NY). 2015; 348(6230):128–32. https://doi. org/10.1126/science.aaa1738 PMID: 25838385. 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. 5. Burgos K, Malenica I, Metpally R, Courtright A, Rakela B, Beach T, et al. Profiles of extracellular miRNA in cerebrospinal fluid and serum from patients with Alzheimer’s and Parkinson’s diseases correlate with disease status and features of pathology. PLoS One. 2014; 9(5):e94839. https://doi.org/10.1371/ journal.pone.0094839 PMID: 24797360; PubMed Central PMCID: PMCPMC4010405. 6. Weber JA, Baxter DH, Zhang S, Huang DY, Huang KH, Lee MJ, et al. The microRNA spectrum in 12 body fluids. Clin Chem. 2010; 56(11):1733–41. https://doi.org/10.1373/clinchem.2010.147405 PMID: 20847327. 7. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A. 2008; 105 (30):10513–8. https://doi.org/10.1073/pnas.0804549105 PMID: 18663219; PubMed Central PMCID: PMC2492472. 8. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res. 2008; 18(10):997–1006. https://doi. org/10.1038/cr.2008.282 PMID: 18766170. 9. Cruts M, Gijselinck I, Van Langenhove T, van der Zee J, Van Broeckhoven C. Current insights into the C9orf72 repeat expansion diseases of the FTLD/ALS spectrum. Trends Neurosci. 2013; 36(8):450–9. https://doi.org/10.1016/j.tins.2013.04.010 PMID: 23746459. 10. Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R, Lindholm C, et al. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature. 2006; 442 (7105):916–9. https://doi.org/10.1038/nature05016 PMID: 16862116. 11. Hutton M, Lendon CL, Rizzu P, Baker M, Froelich S, Houlden H, et al. Association of missense and 5’- splice-site mutations in tau with the inherited dementia FTDP-17. Nature. 1998; 393(6686):702–5. https://doi.org/10.1038/31508 PMID: 9641683. 12. Freischmidt A, Wieland T, Richter B, Ruf W, Schaeffer V, Muller K, et al. Haploinsufficiency of TBK1 causes familial ALS and fronto-temporal dementia. Nat Neurosci. 2015; 18(5):631–6. https://doi.org/10. 1038/nn.4000 PMID: 25803835. 13. Ghetti B, Oblak AL, Boeve BF, Johnson KA, Dickerson BC, Goedert M. Invited review: Frontotemporal dementia caused by microtubule-associated protein tau gene (MAPT) mutations: a chameleon for neu- ropathology and neuroimaging. Neuropathol Appl Neurobiol. 2015; 41(1):24–46. https://doi.org/10. 1111/nan.12213 PMID: 25556536; PubMed Central PMCID: PMCPMC4329416. 14. Blauwendraat C, Wilke C, Simon-Sanchez J, Jansen IE, Reifschneider A, Capell A, et al. The wide genetic landscape of clinical frontotemporal dementia: systematic combined sequencing of 121 conse- cutive subjects. Genet Med. 2017. https://doi.org/10.1038/gim.2017.102 PMID: 28749476. 15. Warren JD, Rohrer JD, Rossor MN. Clinical review. Frontotemporal dementia. BMJ (Clinical research ed). 2013; 347:f4827. Epub 2013/08/08. References 1. Denk J, Boelmans K, Siegismund C, Lassner D, Arlt S, Jahn H. MicroRNA Profiling of CSF Reveals Potential Biomarkers to Detect Alzheimer‘s Disease. PLoS One. 2015; 10(5):e0126423. https://doi.org/ 10.1371/journal.pone.0126423 PMID: 25992776; PubMed Central PMCID: PMCPMC4439119. 2. Mestdagh P, Hartmann N, Baeriswyl L, Andreasen D, Bernard N, Chen C, et al. Evaluation of quantita- tive miRNA expression platforms in the microRNA quality control (miRQC) study. Nat Methods. 2014; 11(8):809–15. https://doi.org/10.1038/nmeth.3014 PMID: 24973947. 3. Schmiedel JM, Klemm SL, Zheng Y, Sahay A, Bluthgen N, Marks DS, et al. Gene expression. Micro- RNA control of protein expression noise. Science (New York, NY). 2015; 348(6230):128–32. https://doi. org/10.1126/science.aaa1738 PMID: 25838385. 19 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. Ritz C, Spiess AN. qpcR: an R package for sigmoidal model selection in quantitative real-time polymer- ase chain reaction analysis. Bioinformatics. 2008; 24(13):1549–51. https://doi.org/10.1093/ bioinformatics/btn227 PMID: 18482995. 27. Spiess AN, Feig C, Ritz C. Highly accurate sigmoidal fitting of real-time PCR data by introducing a parameter for asymmetry. BMC Bioinformatics. 2008; 9:221. https://doi.org/10.1186/1471-2105-9-221 PMID: 18445269; PubMed Central PMCID: PMCPMC2386824. 28. Sisti D, Guescini M, Rocchi MB, Tibollo P, D’Atri M, Stocchi V. Shape based kinetic outlier detection in real-time PCR. BMC Bioinformatics. 2010; 11:186. https://doi.org/10.1186/1471-2105-11-186 PMID: 20385019; PubMed Central PMCID: PMCPMC2873533. 29. Blondal T, Jensby Nielsen S, Baker A, Andreasen D, Mouritzen P, Wrang Teilum M, et al. Assessing sample and miRNA profile quality in serum and plasma or other biofluids. Methods. 2013; 59(1):S1–6. https://doi.org/10.1016/j.ymeth.2012.09.015 PMID: 23036329. 30. Andersen CL, Jensen JL, Orntoft TF. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res. 2004; 64(15):5245–50. https://doi.org/10.1158/0008- 5472.CAN-04-0496 PMID: 15289330. 31. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, et al. Accurate normaliza- tion of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002; 3(7):Research0034. Epub 2002/08/20. PMID: 12184808; PubMed Central PMCID: PMCPmc126239. 32. Fukushima A. DiffCorr: an R package to analyze and visualize differential correlations in biological net- works. Gene. 2013; 518(1):209–14. https://doi.org/10.1016/j.gene.2012.11.028 PMID: 23246976. 33. Osborne JW. Best Practices in Exploratory Factor Analysis. Scotts Valley, CA: CreateSpace Indepen- dent Publishing; 2014. 34. MacCallum RC, Widaman KF, Zhang SB, Hong SH. Sample size in factor analysis. Psychol Methods. 1999; 4(1):84–99. https://doi.org/10.1037/1082-989x.4.1.84 PubMed PMID: WOS:000079084400007. 35. Henson RK, Roberts JK. Use of exploratory factor analysis in published research—Common errors and some comment on improved practice. Educ Psychol Meas. 2006; 66(3):393–416. https://doi.org/10. 1177/0013164405282485 PubMed PMID: WOS:000237628100003. 36. Freischmidt A, Muller K, Ludolph AC, Weishaupt JH. Systemic dysregulation of TDP-43 binding micro- RNAs in amyotrophic lateral sclerosis. Acta Neuropathol Commun. 2013; 1:42. https://doi.org/10.1186/ 2051-5960-1-42 PMID: 24252274; PubMed Central PMCID: PMCPMC3893596. 37. Wang WX, Fardo DW, Jicha GA, Nelson PT. A Customized Quantitative PCR MicroRNA Panel Pro- vides a Technically Robust Context for Studying Neurodegenerative Disease Biomarkers and Indicates a High Correlation Between Cerebrospinal Fluid and Choroid Plexus MicroRNA Expression. Mol Neuro- biol. 2016. https://doi.org/10.1007/s12035-016-0316-2 PMID: 27900678. 38. Fort A, Borel C, Migliavacca E, Antonarakis SE, Fish RJ, Neerman-Arbez M. 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. https://doi.org/10.1136/bmj.f4827 PMID: 23920254; PubMed Central PMCID: PMCPMC3735339. 16. Dickson DW, Kouri N, Murray ME, Josephs KA. Neuropathology of frontotemporal lobar degeneration- tau (FTLD-tau). J Mol Neurosci. 2011; 45(3):384–9. https://doi.org/10.1007/s12031-011-9589-0 PMID: 21720721; PubMed Central PMCID: PMCPMC3208128. 17. Neumann M, Sampathu DM, Kwong LK, Truax AC, Micsenyi MC, Chou TT, et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science (New York, NY). 2006; 314(5796):130–3. https://doi.org/10.1126/science.1134108 PMID: 17023659. 18. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: mini- mum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009; 55 (4):611–22. https://doi.org/10.1373/clinchem.2008.112797 PMID: 19246619. 19. Otto M, Ludolph AC, Landwehrmeyer B, Forstl H, Diehl-Schmid J, Neumann M, et al. [German consor- tium for frontotemporal lobar degeneration]. Nervenarzt. 2011; 82(8):1002–5. https://doi.org/10.1007/ s00115-011-3261-3 PMID: 21805118. 20. Rascovsky K, Hodges JR, Knopman D, Mendez MF, Kramer JH, Neuhaus J, et al. Sensitivity of revised diagnostic criteria for the behavioural variant of frontotemporal dementia. Brain. 2011; 134(Pt 9):2456– 77. https://doi.org/10.1093/brain/awr179 PMID: 21810890; PubMed Central PMCID: PMC3170532. 21. Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger-Gateau P, Cummings J, et al. Research cri- teria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. Lancet Neurol. 2007; 6(8):734–46. https://doi.org/10.1016/S1474-4422(07)70178-3 PMID: 17616482. 22. Lewczuk P, Kornhuber J, Wiltfang J. The German Competence Net Dementias: standard operating pro- cedures for the neurochemical dementia diagnostics. J Neural Transm. 2006; 113(8):1075–80. https:// doi.org/10.1007/s00702-006-0511-9 PMID: 16835693. 20 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls 23. Akimoto C, Volk AE, van Blitterswijk M, Van den Broeck M, Leblond CS, Lumbroso S, et al. A blinded international study on the reliability of genetic testing for GGGGCC-repeat expansions in C9orf72 reveals marked differences in results among 14 laboratories. Journal of medical genetics. 2014; 51 (6):419–24. https://doi.org/10.1136/jmedgenet-2014-102360 PMID: 24706941; PubMed Central PMCID: PMCPMC4033024. 24. Jesse S, Brettschneider J, Sussmuth SD, Landwehrmeyer BG, von Arnim CA, Ludolph AC, et al. Sum- mary of cerebrospinal fluid routine parameters in neurodegenerative diseases. J Neurol. 2011; 258 (6):1034–41. https://doi.org/10.1007/s00415-010-5876-x PMID: 21188408; PubMed Central PMCID: PMCPMC3101362. 25. Mooney C, Raoof R, El-Naggar H, Sanz-Rodriguez A, Jimenez-Mateos EM, Henshall DC. High Throughput qPCR Expression Profiling of Circulating MicroRNAs Reveals Minimal Sex- and Sample Timing-Related Variation in Plasma of Healthy Volunteers. PLoS One. 2015; 10(12):e0145316. https:// doi.org/10.1371/journal.pone.0145316 PMID: 26699132; PubMed Central PMCID: PMCPMC4689368. 26. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. Regulation of fibrinogen production by microRNAs. Blood. 2010; 116(14):2608–15. https://doi.org/10.1182/blood-2010-02- 268011 PMID: 20570858. 39. Ludwig N, Leidinger P, Becker K, Backes C, Fehlmann T, Pallasch C, et al. Distribution of miRNA expression across human tissues. Nucleic Acids Res. 2016; 44(8):3865–77. https://doi.org/10.1093/ nar/gkw116 PMID: 26921406; PubMed Central PMCID: PMCPMC4856985. 40. Sempere LF, Freemantle S, Pitha-Rowe I, Moss E, Dmitrovsky E, Ambros V. Expression profiling of mammalian microRNAs uncovers a subset of brain-expressed microRNAs with possible roles in murine and human neuronal differentiation. Genome Biology. 2004; 5(3). ARTN R13 https://doi.org/10.1186/ gb-2004-5-3-r13 PubMed PMID: WOS:000189345300007. 21 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls 41. Leidinger P, Backes C, Meder B, Meese E, Keller A. The human miRNA repertoire of different blood compounds. BMC Genomics. 2014; 15:474. https://doi.org/10.1186/1471-2164-15-474 PMID: 24928098; PubMed Central PMCID: PMCPMC4076980. 42. Staehler CF, Keller A, Leidinger P, Backes C, Chandran A, Wischhusen J, et al. Whole miRNome-wide differential co-expression of microRNAs. Genomics Proteomics Bioinformatics. 2012; 10(5):285–94. https://doi.org/10.1016/j.gpb.2012.08.003 PMID: 23200138; PubMed Central PMCID: PMCPMC5054199. 43. Schoenfelder S, Sexton T, Chakalova L, Cope NF, Horton A, Andrews S, et al. Preferential associations between co-regulated genes reveal a transcriptional interactome in erythroid cells. Nat Genet. 2010; 42 (1):53–61. https://doi.org/10.1038/ng.496 PMID: 20010836; PubMed Central PMCID: PMCPMC3237402. 44. Cheng L, Doecke JD, Sharples RA, Villemagne VL, Fowler CJ, Rembach A, et al. Prognostic serum miRNA biomarkers associated with Alzheimer’s disease shows concordance with neuropsychological and neuroimaging assessment. Mol Psychiatry. 2015; 20(10):1188–96. https://doi.org/10.1038/mp. 2014.127 PMID: 25349172. 45. Hebert SS, Horre K, Nicolai L, Bergmans B, Papadopoulou AS, Delacourte A, et al. MicroRNA regula- tion of Alzheimer’s Amyloid precursor protein expression. Neurobiol Dis. 2009; 33(3):422–8. Epub 2008/12/27. https://doi.org/10.1016/j.nbd.2008.11.009 PMID: 19110058. 46. Chandrasekaran S, Bonchev D. Network Topology Analysis of Post-Mortem Brain Microarrays Identi- fies More Alzheimer’s Related Genes and MicroRNAs and Points to Novel Routes for Fighting with the Disease. PLoS One. 2016; 11(1):e0144052. https://doi.org/10.1371/journal.pone.0144052 PMID: 26784894; PubMed Central PMCID: PMCPMC4718516. 47. Somel M, Liu X, Tang L, Yan Z, Hu H, Guo S, et al. MicroRNA-driven developmental remodeling in the brain distinguishes humans from other primates. PLoS biology. 2011; 9(12):e1001214. https://doi.org/ 10.1371/journal.pbio.1001214 PMID: 22162950; PubMed Central PMCID: PMCPMC3232219. 48. Li Q, Li X, Wang L, Zhang Y, Chen L. miR-98-5p Acts as a Target for Alzheimer’s Disease by Regulating Abeta Production Through Modulating SNX6 Expression. J Mol Neurosci. 2016; 60(4):413–20. https:// doi.org/10.1007/s12031-016-0815-7 PMID: 27541017. 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. 49. Hu YK, Wang X, Li L, Du YH, Ye HT, Li CY. MicroRNA-98 induces an Alzheimer’s disease-like distur- bance by targeting insulin-like growth factor 1. Neurosci Bull. 2013; 29(6):745–51. https://doi.org/10. 1007/s12264-013-1348-5 PMID: 23740209; PubMed Central PMCID: PMCPMC5561832. 50. Petzold A. Neurofilament phosphoforms: surrogate markers for axonal injury, degeneration and loss. J Neurol Sci. 2005; 233(1–2):183–98. https://doi.org/10.1016/j.jns.2005.03.015 PMID: 15896809. 51. Meeter LH, Dopper EG, Jiskoot LC, Sanchez-Valle R, Graff C, Benussi L, et al. Neurofilament light chain: a biomarker for genetic frontotemporal dementia. Ann Clin Transl Neurol. 2016; 3(8):623–36. https://doi.org/10.1002/acn3.325 PMID: 27606344; PubMed Central PMCID: PMCPMC4999594. 52. Lehmer C, Oeckl P, Weishaupt JH, Volk AE, Diehl-Schmid J, Schroeter ML, et al. Poly-GP in cerebro- spinal fluid links C9orf72-associated dipeptide repeat expression to the asymptomatic phase of ALS/ FTD. EMBO Mol Med. 2017; 9(7):859–68. https://doi.org/10.15252/emmm.201607486 PMID: 28408402; PubMed Central PMCID: PMCPMC5494528. 53. Hebert SS, Papadopoulou AS, Smith P, Galas MC, Planel E, Silahtaroglu AN, et al. Genetic ablation of Dicer in adult forebrain neurons results in abnormal tau hyperphosphorylation and neurodegeneration. Hum Mol Genet. 2010; 19(20):3959–69. https://doi.org/10.1093/hmg/ddq311 PMID: 20660113. 54. Hebert SS, Horre K, Nicolai L, Papadopoulou AS, Mandemakers W, Silahtaroglu AN, et al. Loss of microRNA cluster miR-29a/b-1 in sporadic Alzheimer’s disease correlates with increased BACE1/beta- secretase expression. Proc Natl Acad Sci U S A. 2008; 105(17):6415–20. https://doi.org/10.1073/pnas. 0710263105 PMID: 18434550; PubMed Central PMCID: PMC2359789. 55. Pesini P, Perez-Grijalba V, Monleon I, Boada M, Tarraga L, Martinez-Lage P, et al. Reliable Measure- ments of the beta-Amyloid Pool in Blood Could Help in the Early Diagnosis of AD. Int J Alzheimers Dis. 2012; 2012:604141. https://doi.org/10.1155/2012/604141 PMID: 22957297; PubMed Central PMCID: PMCPMC3431090. 56. Galimberti D, Villa C, Fenoglio C, Serpente M, Ghezzi L, Cioffi SMG, et al. Circulating miRNAs as Potential Biomarkers in Alzheimer’s Disease. Journal of Alzheimers Disease. 2014; 42(4):1261–7. https://doi.org/10.3233/jad-140756 PubMed PMID: WOS:000343763000015. 57. Tan L, Yu JT, Liu QY, Tan MS, Zhang W, Hu N, et al. Circulating miR-125b as a biomarker of Alzhei- mer’s disease. J Neurol Sci. 2014; 336(1–2):52–6. https://doi.org/10.1016/j.jns.2013.10.002 PMID: 24139697. 58. Kumar P, Dezso Z, MacKenzie C, Oestreicher J, Agoulnik S, Byrne M, et al. Circulating miRNA Bio- markers for Alzheimer’s Disease. Plos One. 2013; 8(7). https://doi.org/10.1371/journal.pone.0069807 PubMed PMID: WOS:000323369700081. 22 / 23 PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 Serum and CSF miRNA profiles distinguish bvFTD, AD and controls 59. Tan L, Yu J-T, Tan M-S, Liu Q-Y, Wang H-F, Zhang W, et al. PLOS ONE \| https://doi.org/10.1371/journal.pone.0197329 May 10, 2018 4. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci. 2014; 369(1652). https://doi.org/ 10.1098/rstb.2013.0516 PMID: 25135977; PubMed Central PMCID: PMCPMC4142037. Genome-Wide Serum microRNA Expres- sion Profiling Identifies Serum Biomarkers for Alzheimer’s Disease. Journal of Alzheimers Disease. 2014; 40(4):1017–27. https://doi.org/10.3233/jad-132144 PubMed PMID: WOS:000336262000019. 60. Dong H, Li J, Huang L, Chen X, Li D, Wang T, et al. Serum MicroRNA Profiles Serve as Novel Biomark- ers for the Diagnosis of Alzheimer’s Disease. Dis Markers. 2015; 2015:625659. https://doi.org/10.1155/ 2015/625659 PMID: 26078483; PubMed Central PMCID: PMCPMC4452867. 61. Satoh J, Kino Y, Niida S. MicroRNA-Seq Data Analysis Pipeline to Identify Blood Biomarkers for Alzhei- mer’s Disease from Public Data. Biomark Insights. 2015; 10:21–31. https://doi.org/10.4137/BMI. S25132 PMID: 25922570; PubMed Central PMCID: PMCPMC4401249. 62. Oeckl P, Steinacker P, Feneberg E, Otto M. Neurochemical biomarkers in the diagnosis of frontotem- poral lobar degeneration: an update. Journal of neurochemistry. 2016; 138 Suppl 1:184–92. https://doi. org/10.1111/jnc.13669 PMID: 27186717. 63. Cogswell JP, Ward J, Taylor IA, Waters M, Shi Y, Cannon B, et al. Identification of miRNA changes in Alzheimer’s disease brain and CSF yields putative biomarkers and insights into disease pathways. Journal of Alzheimer’s disease: JAD. 2008; 14(1):27–41. Epub 2008/06/06. PMID: 18525125. 23 / 23
https://openalex.org/W3172652159	https://discovery.ucl.ac.uk/id/eprint/10141804/1/10.21105.joss.03283.pdf	English	null	lenstronomy II: A gravitational lensing software ecosystem	Journal of open source software	2,021	cc-by	5,024	Summary lenstronomy is an Astropy-affiliated (Astropy Collaboration et al., 2018, 2013) Python pack- age for gravitational lensing simulations and analyses. lenstronomy was introduced by Birrer & Amara (2018) and is based on the linear basis set approach by Birrer et al. (2015). The user and developer base of lenstronomy has substantially grown since then, and the software has become an integral part of a wide range of recent analyses, such as measuring the Hubble constant with time-delay strong lensing or constraining the nature of dark matter from resolved and unresolved small scale lensing distortion statistics. The modular design has allowed the community to incorporate innovative new methods, as well as to develop enhanced software and wrappers with more specific aims on top of the lenstronomy API. Through community engagement and involvement, lenstronomy has become a foundation of an ecosystem of affiliated packages extending the original scope of the software and proving its robustness and applicability at the forefront of the strong gravitational lensing community in an open source and reproducible manner. tronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 283 1 Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 1 lenstronomy II: A gravitational lensing software ecosystem Simon Birrer1, 2, Anowar J. Shajib3, 4, Daniel Gilman5, Aymeric Galan6, Jelle Aalbers1, 2, Martin Millon6, Robert Morgan7, 8, Giulia Pagano9, Ji Won Park1, 2, Luca Teodori10, Nicolas Tessore11, Madison Ueland1, Lyne Van de Vyvere12, Sebastian Wagner-Carena1, 2, Ewoud Wempe13, Lilan Yang14, Xuheng Ding15, Thomas Schmidt4, Dominique Sluse12, Ming Zhang16, and Adam Amara17 Simon Birrer1, 2, Anowar J. Shajib3, 4, Daniel Gilman5, Aymeric Galan6, Jelle Aalbers1, 2, Martin Millon6, Robert Morgan7, 8, Giulia Pagano9, Ji Won Park1, 2, Luca Teodori10, Nicolas Tessore11, Madison Ueland1, Lyne Van de Vyvere12, Sebastian Wagner-Carena1, 2, Ewoud Wempe13, Lilan Yang14, Xuheng Ding15, Thomas Schmidt4, Dominique Sluse12, Ming Zhang16, and Adam Amara17 1 Kavli Institute for Particle Astrophysics and Cosmology and Department of Physics, Stanford University, Stanford, CA 94305, USA 2 SLAC National Accelerator Laboratory, Menlo Park, CA, 94025, USA 3 Department of Astronomy & Astrophysics, University of Chicago, Chicago, IL 60637, USA 4 Department of Physics and Astronomy, University of California, Los Angeles, CA 90095, USA 5 Department of Astronomy and Astrophysics, University of Toronto, 50 St. George Street, Toronto, ON, M5S 3H4, Canada 6 Institute of Physics, Laboratory of Astrophysics, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland 7 Physics Department, University of Wisconsin-Madison, 1150 University Avenue Madison, WI 53706, USA 8 Legacy Survey of Space and Time Corporation Data Science Fellowship Program, USA 9 Independent Researcher 10 Weizmann Institute, 234 Herzl Street, Rehovot, 7610001 Israel 11 Department of Physics and Astronomy, University College London, Gower Street, London, WC1E 6BT, UK 12 STAR Institute, Université de Liège, Quartier Agora - Allée du six Août, 19c, B-4000 Liège, Belgium 13 Kapteyn Astronomical Institute, University of Groningen, PO Box 800, 9700 AV Groningen, the Netherlands 14 School of Physics and Technology, Wuhan University, Wuhan 430072, China 15 Kavli IPMU (WPI), UTIAS, The University of Tokyo, Kashiwa, Chiba 277-8583, Japan 16 Xinjiang Astronomical Observatory, Chinese Academy of Sciences, 150 Science 1-Street, Urumqi 831001, China 17 Institute of Cosmology and Gravitation, University of Portsmouth, Portsmouth PO1 3FX, UK DOI: 10.21105/joss.03283 DOI: 10.21105/joss.03283 DOI: 10.21105/joss.03283 DOI: 10.21105/joss.03283 Software • Review • Repository • Archive Editor: Daniel S. Katz Reviewers: • @smsharma • @coljac Submitted: 03 May 2021 Published: 08 June 2021 License Authors of papers retain copyright and release the work under a Creative Commons Attribution 4.0 International License (CC BY 4.0). Background Gravitational lensing displaces the observed positions and distorts the shapes of apparent objects on the sky due to intervening inhomogeneous matter along the line of sight. Strong gravitational lensing describes the regime where the background source, such as a galaxy or quasar, is lensed by a massive foreground object, such as another galaxy or cluster of galaxies, to produce multiple images of the source in a highly distorted manner. The top row of Figure 1 illustrates such a process from the intrinsic galaxy to the data product at hand, including the lensing distortions, effects of the instrument, observational conditions, and noise. Analyses of strong gravitational lensing have provided a wealth of key insights into cosmol- ogy and astrophysics. For example, relative time delays of multiply imaged variable sources provided precision measurements on the expansion rate of the Universe (Birrer et al., 2020; Shajib et al., 2020; Wong et al., 2020). Small scale distortions in the lensing signal of resolved sources (Birrer, Amara, et al., 2017; Hezaveh et al., 2016; Vegetti et al., 2012) and unre- solved flux ratios (Gilman et al., 2020; Hsueh et al., 2020) constrain the nature of dark matter. Combined strong lensing and kinematic observables constrain the formation and evolution of galaxies (Shajib, Treu, et al., 2021; Sonnenfeld et al., 2015), and the lensing magnification effect provides an otherwise inaccessible angle on the early Universe (Cava et al., 2018; Zheng et al., 2012). License Figure 1: Illustration of the strong gravitational lensing phenomenology and the capabilities of lenstronomy in performing realistic simulations as well as reconstructing lensing and source properties from a given data set. Top row from left to right along the green arrow: A galaxy is lensed around a foreground massive object, becomes highly distorted, and has components appearing multiple times. Observations of this phenomena are limited in resolution (convolution), depending on the detector (pixelation), and are subject to noise. Bottom row from right to left along the red arrow: The inverse problem is solved with a linear basis set in the source morphology maximizing the likelihood of the model given the data. Figure 1: Illustration of the strong gravitational lensing phenomenology and the capabilities of lenstronomy in performing realistic simulations as well as reconstructing lensing and source properties from a given data set. Top row from left to right along the green arrow: A galaxy is lensed around a foreground massive object, becomes highly distorted, and has components appearing multiple times. Observations of this phenomena are limited in resolution (convolution), depending on the detector (pixelation), and are subject to noise. Bottom row from right to left along the red arrow: The inverse problem is solved with a linear basis set in the source morphology maximizing the likelihood of the model given the data. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 2 enstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 03283 different types will be growing by more than an order of magnitude (Collett, 2015; Oguri & Marshall, 2010). Such large samples can provide unprecedented statistical precision to stress- test our current understanding and exploit discovery potential. It is key that these demanding studies, at present and in the future, are conducted by reliable software and supported by reproducible and open-source analysis products to provide the most compelling and transparent evidence required to further our physical understanding. The primary design goal of lenstronomy is to facilitate scientific investigations into the outstanding and most pressing questions in the cosmology and astrophysics community. lens tronomy has been applied throughout its development to the most demanding modeling and inference problems in strong lensing and the software has evolved around the requirements of the scientific applications to facilitate robust analyses. The modular API of the original design of lenstronomy (Birrer & Amara, 2018) has accommodated the addition of new features. Code review processes in the development phase have led to additional benefits for the user community at large beyond the specific needs of the developer. lenstronomy provides reliable and well-tested specific functionalities, as well as top-level interfaces, which allow for adaptive and innovative usage in control by the scientific inves- tigator. Guidance for the user community is provided on multiple levels. First, source code is well documented and provided through readthedocs.org. Second, a set of jupyter note- books are provided in an extension repository. These notebooks demonstrate simplified ex- ample use cases, each notebook individually highlighting different specific functionalities of lenstronomy, including a starting guide notebook to introduce the modular design structure of lenstronomy. Third, end-to-end analysis pipelines of some of the published work are publicly available, providing ‘real-life’ examples at advanced levels. Statement of need Strong lensing studies have significantly enhanced, and sometimes challenged, our current fundamental understanding of the Universe. In the near future, with the onset of the next- generation ground and space-based wide and deep astronomical imaging (Rubin, Roman, Euclid observatories; Ivezić et al., 2019; Laureijs et al., 2011; Spergel et al., 2013) and interferometric (SKA; Dewdney et al., 2009) surveys, the number of discovered lenses of & Metcalf, 2015), and has found applications beyond its original aim due to the robust and high-standard design requirements. Ecosystem of affiliated packages lenstronomy has allowed the community to develop third-party analysis products and soft- ware products utilizing its core functionalities to provide more targeted and integrated software solutions for a wide range of scientific analyses. These open-source affiliated packages effec- tively create an ecosystem enhancing the capability of lenstronomy. They provide specified and tested solution for specific scientific investigations, such as plug-ins and direct imple- mentation for innovative source reconstruction algorithms (Galan et al., 2021; SLITronomy; Joseph et al., 2019), gravitational wave lensing computations (lensingGW; Pagano et al., 2020), automated pipelines for gravitational lensing reconstruction (dolphin; Shajib, Treu, et al., 2021), cluster source reconstruction and local perturbative lens modeling (lenstruction; Yang et al., 2020), enhancement in large-scale structure imaging survey simulations (DESC SLSprinkler; LSST Dark Energy Science Collaboration (LSST DESC) et al., 2021), rendering of sub-halos and line-of-sight halos (pyHalo; Gilman et al., 2020), galaxy morphology anal- ysis (galight; Ding et al., 2020), and hierarchical analyses to measure the Hubble constant (hierArc; Birrer et al., 2020). With the rise in popularity and the promises in dealing with ever complex data problems with fast deep-learning methods, dedicated tools for simulating large datasets for applying such methods to strong gravitational lensing (deeplenstronomy; Morgan et al., 2021), (baobab; Park et al., 2021), as well as end-to-end Bayesian Neural Network training and validation packages for Hubble constant measurements (h0rton; Park et al., 2021), and for a hierarchical analysis of galaxy-galaxy lenses (ovejero; Wagner-Carena et al., 2021) have been developed. The affiliated packages make best use of the lenstronomy modules without duplicating source code and make it possible to combine aspects of multiple affiliated packages in one single analysis. Track-record of applications lenstronomy has been applied in and contributed to more than 30 peer reviewed publications since its first public release in 2018. In particular, lenstronomy has been used to provide state-of-the-art measurements on real data sets, such as: (i) Hubble constant measurements from three quadruly lensed quasars with Hubble Space Telescope (HST) imaging (Birrer et al., 2019, 2016; Shajib et al., 2020), dynamical modeling in the hierarchical analysis by Birrer et al. (2020), and modeling of lensed supernovae (Mörtsell et al., 2020); (ii) inference of small scale dark matter properties from detailed studies of both, resolved imaging (Birrer, Amara, et al., 2017), and unresolved flux ratio statistics (Gilman et al., 2020); (iii) decomposition of quasar and host galaxy light in both, lensed and unlensed cases (Bennert et al., 2021; Ding et al., 2020); (iv) morphological studies of high-redshift sources in the cluster environment (Yang et al., 2021, 2020); (v) internal structure of galaxies (Shajib, Treu, et al., 2021; Shajib, Molina, et al., 2021); (vi) measurements of the weak lensing effect imprinted in Einstein rings (Birrer, Welschen, et al., 2017; Kuhn et al., 2021). Among the studies, some of them have applied a pipeline to uniformly analyse dozens of lenses of different types (Shajib et al., 2019; Shajib, Treu, et al., 2021; Shajib, Molina, et al., 2021), a milestone in moving towards utilizing thousands of lenses in the near future. Beyond analyzing data, many theoretical studies have been conducted using lenstronomy to investigate statistical robustness in present and anticipated future analyses (Birrer & Treu, 2019; Ding, Liao, et al., 2021; Li et al., 2021; Millon et al., 2020; Van de Vyvere et al., 2020), as well as to provide forecasts for anticipated future constraints for different science cases (Birrer & Treu, 2020; Çaǧan Şengül et al., 2020; Gilman et al., 2019). Particularly, three separate teams participated in the blind time-delay lens modeling challenge (Ding, Treu, et al., 2021) using lenstronomy. lenstronomy has seen a substantial development and incorporation of innovations and nu- merical recipes (Birrer, 2021; Galan et al., 2021; Joseph et al., 2019; Shajib, 2019; Tessore omy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 3 Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 3 Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 4 References Astropy Collaboration, Price-Whelan, A. M., Sipőcz, B. M., Günther, H. M., Lim, P. L., Crawford, S. M., Conseil, S., Shupe, D. L., Craig, M. W., Dencheva, N., Ginsburg, A., Vand erPlas, J. T., Bradley, L. D., Pérez-Suárez, D., de Val-Borro, M., Aldcroft, T. L., Cruz, K. L., Robitaille, T. P., Tollerud, E. J., … Astropy Contributors. (2018). The Astropy Project: Building an Open-science Project and Status of the v2.0 Core Package. The Astronomical Journal, 156(3), 123. https://doi.org/10.3847/1538-3881/aabc4f Astropy Collaboration, Robitaille, T. P., Tollerud, E. J., Greenfield, P., Droettboom, M., Bray, E., Aldcroft, T., Davis, M., Ginsburg, A., Price-Whelan, A. M., Kerzendorf, W. E., Conley, A., Crighton, N., Barbary, K., Muna, D., Ferguson, H., Grollier, F., Parikh, M. M., Nair, P. H., … Streicher, O. (2013). Astropy: A community Python package for astronomy. Astronomy & Astrophysics, 558, A33. https://doi.org/10.1051/0004-6361/201322068 Bennert, V. N., Treu, T., Ding, X., Stomberg, I., Birrer, S., Snyder, T., Malkan, M. A., Stephens, A. W., & Auger, M. W. (2021). A local baseline of the black hole mass scaling relations for active galaxies. IV. Correlations between MBH and host galaxy σ, stellar mass, and luminosity. arXiv e-Prints, arXiv:2101.10355. http://arxiv.org/abs/2101.10355 Birrer, S. (2021). Gravitational lensing formalism in a curved arc basis: A continuous descrip- tion of observables and degeneracies from the weak to the strong lensing regime. arXiv e-Prints, arXiv:2104.09522. http://arxiv.org/abs/2104.09522 Birrer, S., & Amara, A. (2018). lenstronomy: Multi-purpose gravitational lens modelling software package. Physics of the Dark Universe, 22, 189–201. https://doi.org/10.1016/ j.dark.2018.11.002 Birrer, S., Amara, A., & Refregier, A. (2017). Lensing substructure quantification in RXJ1131- 1231: a 2 keV lower bound on dark matter thermal relic mass. Journal of Cosmology and Astroparticle Physics, 2017(5), 037. https://doi.org/10.1088/1475-7516/2017/05/037 Birrer, S., Amara, A., & Refregier, A. (2016). The mass-sheet degeneracy and time-delay cosmography: analysis of the strong lens RXJ1131-1231. Journal of Cosmology and Astroparticle Physics, 2016(8), 020. https://doi.org/10.1088/1475-7516/2016/08/020 Birrer, S., Amara, A., & Refregier, A. (2015). Gravitational Lens Modeling with Basis Sets. The Astrophysical Journal, 813(2), 102. https://doi.org/10.1088/0004-637X/813/2/102 Birrer, S., Shajib, A. J., Galan, A., Millon, M., Treu, T., Agnello, A., Auger, M., Chen, G. C.-F., Christensen, L., Collett, T., Courbin, F., Fassnacht, C. D., Koopmans, L. V. E., Marshall, P. J., Park, J.-W., Rusu, C. E., Sluse, D., Spiniello, C., Suyu, S. H., … Van de Vyvere, L. (2020). TDCOSMO. IV. Hierarchical time-delay cosmography joint inference of the Hubble constant and galaxy density profiles. Astronomy & Astrophysics, 643, A165. Acknowledgements Support for this work was provided by the National Science Foundation through NSF AST- 1716527. AJS was supported by NASA through the STScI grant HST-GO-15320 and by a Dissertation Year Fellowship from the UCLA Graduate Division. This research was supported by the U.S. Department of Energy (DOE) Office of Science Distinguished Scientist Fellow Program. DG is supported by NASA HST-GO-15177. AG, MM LvdV, DS are supported by COSMICLENS: ERC grant agreement No 787886. LT is supported by International Helmholtz- Weizmann Research School for Multimessenger Astronomy. MU is supported by KIPAC and the Stanford Summer Research Program. XD is supported by NASA HST-GO-15115. TS is supported by NASA grant HST-GO-15320 and HST-GO-15652. MZ is supported by the II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 4 National Science Foundation of China. AA is supported by a Royal Society Wolfson Fellowship. We are grateful to the user community for valuable feedback and encouragement in continuing the development. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 5 References https://doi.org/10.1051/0004-6361/202038861 Birrer, S., & Treu, T. (2020). TDCOSMO V: strategies for precise and accurate measurements of the Hubble constant with strong lensing. arXiv e-Prints, arXiv:2008.06157. http: //arxiv.org/abs/2008.06157 Birrer, S., & Treu, T. (2019). Astrometric requirements for strong lensing time-delay cos- mography. Monthly Notices of the Royal Astronomical Society, 489(2), 2097–2103. https://doi.org/10.1093/mnras/stz2254 Birrer, S., Treu, T., Rusu, C. E., Bonvin, V., Fassnacht, C. D., Chan, J. H. H., Agnello, A., Shajib, A. J., Chen, G. C.-F., Auger, M., Courbin, F., Hilbert, S., Sluse, D., Suyu, S. H., Wong, K. C., Marshall, P., Lemaux, B. C., & Meylan, G. (2019). H0LiCOW - IX. Cosmographic analysis of the doubly imaged quasar SDSS 1206+4332 and a new measurement of the Hubble constant. Monthly Notices of the Royal Astronomical Society, 484(4), 4726–4753. https://doi.org/10.1093/mnras/stz200 Birrer, S., Welschen, C., Amara, A., & Refregier, A. (2017). Line-of-sight effects in strong lensing: putting theory into practice. Journal of Cosmology and Astroparticle Physics, 2017(4), 049. https://doi.org/10.1088/1475-7516/2017/04/049 Cava, A., Schaerer, D., Richard, J., Pérez-González, P. G., Dessauges-Zavadsky, M., Mayer, L., & Tamburello, V. (2018). The nature of giant clumps in distant galaxies probed by the anatomy of the cosmic snake. Nature Astronomy, 2, 76–82. https://doi.org/10.1038/ s41550-017-0295-x Collett, T. E. (2015). The Population of Galaxy-Galaxy Strong Lenses in Forthcoming Op- tical Imaging Surveys. The Astrophysical Journal, 811(1), 20. https://doi.org/10.1088/ 0004-637X/811/1/20 Çaǧan Şengül, A., Tsang, A., Diaz Rivero, A., Dvorkin, C., Zhu, H.-M., & Seljak, U. (2020). Quantifying the line-of-sight halo contribution to the dark matter convergence power spectrum from strong gravitational lenses. Physical Review D, 102(6), 063502. https://doi.org/10.1103/PhysRevD.102.063502 Dewdney, P. E., Hall, P. J., Schilizzi, R. T., & Lazio, T. J. L. W. (2009). The Square Kilometre Array. IEEE Proceedings, 97(8), 1482–1496. https://doi.org/10.1109/JPROC. 2009.2021005 Ding, X., Liao, K., Birrer, S., Shajib, A. J., Treu, T., & Yang, L. (2021). Improved time- delay lens modelling and H0 inference with transient sources. 504(4), 5621–5628. https: //doi.org/10.1093/mnras/stab1240 Ding, X., Silverman, J., Treu, T., Schulze, A., Schramm, M., Birrer, S., Park, D., Jahnke, K., Bennert, V. N., Kartaltepe, J. S., Koekemoer, A. M., Malkan, M. A., & Sanders, D. (2020). The Mass Relations between Supermassive Black Holes and Their Host Galaxies at 1 < z < 2 HST-WFC3. The Astrophysical Journal, 888(1), 37. https://doi.org/10. 3847/1538-4357/ab5b90 Ding, X., Treu, T., Birrer, S., Chen, G. C.-F., Coles, J., Denzel, P., Frigo, M., Galan, A., Marshall, P. J., Millon, M., More, A., Shajib, A. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 6 tronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 283 6 References J., Sluse, D., Tak, H., Xu, D., Auger, M. W., Bonvin, V., Chand, H., Courbin, F., … Williams, L. L. R. (2021). Time delay lens modelling challenge. Monthly Notices of the Royal Astronomical Society, 503(1), 1096–1123. https://doi.org/10.1093/mnras/stab484 Galan, A., Peel, A., Joseph, R., Courbin, F., & Starck, J.-L. (2021). SLITRONOMY: Towards a fully wavelet-based strong lensing inversion technique. Astronomy & Astrophysics, 647, A176. https://doi.org/10.1051/0004-6361/202039363 Gilman, D., Birrer, S., Nierenberg, A., Treu, T., Du, X., & Benson, A. (2020). Warm dark matter chills out: constraints on the halo mass function and the free-streaming length of dark matter with eight quadruple-image strong gravitational lenses. Monthly Notices of the Royal Astronomical Society, 491(4), 6077–6101. https://doi.org/10.1093/mnras/stz3480 Gilman, D., Birrer, S., Treu, T., Nierenberg, A., & Benson, A. (2019). Probing dark matter structure down to 107 solar masses: flux ratio statistics in gravitational lenses with line- of-sight haloes. Monthly Notices of the Royal Astronomical Society, 487(4), 5721–5738. https://doi.org/10.1093/mnras/stz1593 Hezaveh, Y. D., Dalal, N., Marrone, D. P., Mao, Y.-Y., Morningstar, W., Wen, D., Blandford, R. D., Carlstrom, J. E., Fassnacht, C. D., Holder, G. P., Kemball, A., Marshall, P. J., Murray, N., Perreault Levasseur, L., Vieira, J. D., & Wechsler, R. H. (2016). Detection of Lensing Substructure Using ALMA Observations of the Dusty Galaxy SDP.81. The Astrophysical Journal, 823(1), 37. https://doi.org/10.3847/0004-637X/823/1/37 Hsueh, J.-W., Enzi, W., Vegetti, S., Auger, M. W., Fassnacht, C. D., Despali, G., Koopmans, L. V. E., & McKean, J. P. (2020). SHARP - VII. New constraints on the dark matter free-streaming properties and substructure abundance from gravitationally lensed quasars. Monthly Notices of the Royal Astronomical Society, 492(2), 3047–3059. https://doi.org/ 10.1093/mnras/stz3177 Ivezić, Ž., Kahn, S. M., Tyson, J. A., Abel, B., Acosta, E., Allsman, R., Alonso, D., AlSayyad, Y., Anderson, S. F., Andrew, J., Angel, J. R. P., Angeli, G. Z., Ansari, R., Antilogus, P., Araujo, C., Armstrong, R., Arndt, K. T., Astier, P., Aubourg, É., … Zhan, H. (2019). LSST: From Science Drivers to Reference Design and Anticipated Data Products. The Astrophysical Journal, 873(2), 111. https://doi.org/10.3847/1538-4357/ab042c Joseph, R., Courbin, F., Starck, J.-L., & Birrer, S. (2019). Sparse Lens Inversion Technique (SLIT): lens and source separability from linear inversion of the source reconstruction problem. Astronomy & Astrophysics, 623, A14. https://doi.org/10.1051/0004-6361/ 201731042 Jullo, E., & Kneib, J.-P. (2009). Multiscale cluster lens mass mapping - I. Strong lensing modelling. 395(3), 1319–1332. https://doi.org/10.1111/j.1365-2966.2009.14654.x Keeton, C. R. (2011). GRAVLENS: Computational Methods for Gravitational Lensing (p. ascl:1102.003). Kuhn, F. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 7 tronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 283 7 References A., Birrer, S., Bruderer, C., Amara, A., & Refregier, A. (2021). Combining strong and weak lensing estimates in the Cosmos field. Journal of Cosmology and Astroparticle Physics, 2021(4), 010. https://doi.org/10.1088/1475-7516/2021/04/010 Laureijs, R., Amiaux, J., Arduini, S., Auguères, J.-L., Brinchmann, J., Cole, R., Cropper, M., Dabin, C., Duvet, L., Ealet, A., Garilli, B., Gondoin, P., Guzzo, L., Hoar, J., Hoekstra, H., Holmes, R., Kitching, T., Maciaszek, T., Mellier, Y., … Zucca, E. (2011). Euclid Definition Study Report. arXiv e-Prints, arXiv:1110.3193. http://arxiv.org/abs/1110.3193 Li, N., Becker, C., & Dye, S. (2021). The impact of line-of-sight structures on measuring H0 with strong lensing time-delays. Monthly Notices of the Royal Astronomical Society. https://doi.org/10.1093/mnras/stab984 Liesenborgs, J., De Rijcke, S., & Dejonghe, H. (2006). A genetic algorithm for the non- parametric inversion of strong lensing systems. Monthly Notices of the Royal Astronomical Society, 367(3), 1209–1216. https://doi.org/10.1111/j.1365-2966.2006.10040.x LSST Dark Energy Science Collaboration (LSST DESC), Abolfathi, B., Alonso, D., Armstrong, R., Aubourg, É., Awan, H., Babuji, Y. N., Bauer, F. E., Bean, R., Beckett, G., Biswas, R., Bogart, J. R., Boutigny, D., Chard, K., Chiang, J., Claver, C. F., Cohen-Tanugi, J., Combet, C., Connolly, A. J., … Zuntz, J. (2021). The LSST DESC DC2 Simulated Sky Survey. The Astrophysical Journal Supplement Series, 253(1), 31. https://doi.org/10. 3847/1538-4365/abd62c Millon, M., Galan, A., Courbin, F., Treu, T., Suyu, S. H., Ding, X., Birrer, S., Chen, G. C.-F., Shajib, A. J., Sluse, D., Wong, K. C., Agnello, A., Auger, M. W., Buckley-Geer, E. J., Chan, J. H. H., Collett, T., Fassnacht, C. D., Hilbert, S., Koopmans, L. V. E., … Van de Vyvere, L. (2020). TDCOSMO. I. An exploration of systematic uncertainties in the inference of H0 from time-delay cosmography. Astronomy & Astrophysics, 639, A101. https://doi.org/10.1051/0004-6361/201937351 Morgan, R., Nord, B., Birrer, S., Lin, J., & Poh, J. (2021). deeplenstronomy: A dataset simulation package for strong gravitational lensing. The Journal of Open Source Software, 6(58), 2854. https://doi.org/10.21105/joss.02854 Mörtsell, E., Johansson, J., Dhawan, S., Goobar, A., Amanullah, R., & Goldstein, D. A. (2020). Lens modelling of the strongly lensed Type Ia supernova iPTF16geu. Monthly Notices of the Royal Astronomical Society, 496(3), 3270–3280. https://doi.org/10.1093/ mnras/staa1600 Nightingale, J. W., Dye, S., & Massey, R. J. (2018). AutoLens: automated modeling of a strong lens’s light, mass, and source. MNRAS, 478(4), 4738–4784. https://doi.org/10. 1093/mnras/sty1264 Nightingale, J. References W., Hayes, R., Kelly, A., Amvrosiadis, A., Etherington, A., He, Q., Li, N., Cao, X., Frawley, J., Cole, S., Enia, A., Frenk, C., Harvey, D., Li, R., Massey, R., Negrello, M., & Robertson, A. (2021). PyAutoLens: Open-Source Strong Gravitational Lensing. The Journal of Open Source Software, 6(58), 2825. https://doi.org/10.21105/joss.02825 Oguri, M. (2010). glafic: Software Package for Analyzing Gravitational Lensing (p. ascl:1010.012). Oguri, M., & Marshall, P. J. (2010). Gravitationally lensed quasars and supernovae in future wide-field optical imaging surveys. Monthly Notices of the Royal Astronomical Society, 405(4), 2579–2593. https://doi.org/10.1111/j.1365-2966.2010.16639.x Pagano, G., Hannuksela, O. A., & Li, T. G. F. (2020). LENSINGGW: a PYTHON package for lensing of gravitational waves. Astronomy & Astrophysics, 643, A167. https://doi. org/10.1051/0004-6361/202038730 Park, J. W., Wagner-Carena, S., Birrer, S., Marshall, P. J., Lin, J. Y.-Y., Roodman, A., & LSST Dark Energy Science Collaboration. (2021). Large-scale Gravitational Lens Mod- eling with Bayesian Neural Networks for Accurate and Precise Inference of the Hubble Constant. The Astrophysical Journal, 910(1), 39. https://doi.org/10.3847/1538-4357/ abdfc4 Saha, P., & Williams, L. L. R. (2011). PixeLens: A Portable Modeler of Lensed Quasars (p. ascl:1102.007). Shajib, A. J. (2019). Unified lensing and kinematic analysis for any elliptical mass profile. Monthly Notices of the Royal Astronomical Society, 488(1), 1387–1400. https://doi.org/ 10.1093/mnras/stz1796 Shajib, A. J., Birrer, S., Treu, T., Agnello, A., Buckley-Geer, E. J., Chan, J. H. H., Christensen, L., Lemon, C., Lin, H., Millon, M., Poh, J., Rusu, C. E., Sluse, D., Spiniello, C., Chen, G. C.-F., Collett, T., Courbin, F., Fassnacht, C. D., Frieman, J., … Zhang, Y. (2020). STRIDES: a 3.9 per cent measurement of the Hubble constant from the strong lens system DES J0408-5354. Monthly Notices of the Royal Astronomical Society, 494(4), 6072–6102. https://doi.org/10.1093/mnras/staa828 Shajib, A. J., Birrer, S., Treu, T., Auger, M. W., Agnello, A., Anguita, T., Buckley-Geer, E. J., Chan, J. H. H., Collett, T. E., Courbin, F., Fassnacht, C. D., Frieman, J., Kayo, I., Lemon, C., Lin, H., Marshall, P. J., McMahon, R., More, A., Morgan, N. D., … Walker, A. R. (2019). Is every strong lens model unhappy in its own way? Uniform modelling of a sample of 13 quadruply+ imaged quasars. Monthly Notices of the Royal Astronomical Society, 483(4), 5649–5671. https://doi.org/10.1093/mnras/sty3397 Shajib, A. J., Molina, E., Agnello, A., Williams, P. R., Birrer, S., Treu, T., Fassnacht, C. D., Morishita, T., Abramson, L., Schechter, P. L., & Wisotzki, L. (2021). High-resolution imaging follow-up of doubly imaged quasars. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 8 References Monthly Notices of the Royal Astronomical Society, 503(2), 1557–1567. https://doi.org/10.1093/mnras/stab532 Shajib, A. J., Treu, T., Birrer, S., & Sonnenfeld, A. (2021). Dark matter haloes of massive elliptical galaxies at z ∼0.2 are well described by the Navarro-Frenk-White profile. Monthly Notices of the Royal Astronomical Society, 503(2), 2380–2405. https://doi.org/10.1093/ mnras/stab536 Sonnenfeld, A., Treu, T., Marshall, P. J., Suyu, S. H., Gavazzi, R., Auger, M. W., & Nipoti, C. (2015). The SL2S Galaxy-scale Lens Sample. V. Dark Matter Halos and Stellar IMF of Massive Early-type Galaxies Out to Redshift 0.8. The Astrophysical Journal, 800(2), 94. https://doi.org/10.1088/0004-637X/800/2/94 Spergel, D., Gehrels, N., Breckinridge, J., Donahue, M., Dressler, A., Gaudi, B. S., Greene, T., Guyon, O., Hirata, C., Kalirai, J., Kasdin, N. J., Moos, W., Perlmutter, S., Postman, M., Rauscher, B., Rhodes, J., Wang, Y., Weinberg, D., Centrella, J., … Shaklan, S. (2013). Wide-Field InfraRed Survey Telescope-Astrophysics Focused Telescope Assets WFIRST- AFTA Final Report. arXiv e-Prints, arXiv:1305.5422. http://arxiv.org/abs/1305.5422 Spilker, J. S., Marrone, D. P., Aravena, M., Béthermin, M., Bothwell, M. S., Carlstrom, J. E., Chapman, S. C., Crawford, T. M., de Breuck, C., Fassnacht, C. D., Gonzalez, A. H., Greve, T. R., Hezaveh, Y., Litke, K., Ma, J., Malkan, M., Rotermund, K. M., Strandet, M., Vieira, J. D., … Welikala, N. (2016). ALMA Imaging and Gravitational Lens Models of South Pole TelescopeSelected Dusty, Star-Forming Galaxies at High Redshifts. The Astrophysical Journal, 826, 112. https://doi.org/10.3847/0004-637X/826/2/112 Tessore, N., & Metcalf, R. B. (2015). The elliptical power law profile lens. Astronomy & Astrophysics, 580, A79. https://doi.org/10.1051/0004-6361/201526773 Van de Vyvere, L., Sluse, D., Mukherjee, S., Xu, D., & Birrer, S. (2020). The impact of mass map truncation on strong lensing simulations. Astronomy & Astrophysics, 644, A108. https://doi.org/10.1051/0004-6361/202038942 Vegetti, S., Lagattuta, D. J., McKean, J. P., Auger, M. W., Fassnacht, C. D., & Koopmans, L. V. E. (2012). Gravitational detection of a low-mass dark satellite galaxy at cosmological distance. Nature, 481(7381), 341–343. https://doi.org/10.1038/nature10669 Wagner-Carena, S., Park, J. W., Birrer, S., Marshall, P. J., Roodman, A., Wechsler, R. H., & LSST Dark Energy Science Collaboration. (2021). Hierarchical Inference with Bayesian Neural Networks: An Application to Strong Gravitational Lensing. The Astrophysical Journal, 909(2), 187. https://doi.org/10.3847/1538-4357/abdf59 Wong, K. C., Suyu, S. H., Chen, G. C.-F., Rusu, C. E., Millon, M., Sluse, D., Bonvin, V., Fassnacht, C. D., Taubenberger, S., Auger, M. W., Birrer, S., Chan, J. H. Birrer et al., (2021). lenstronomy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: //doi.org/10.21105/joss.03283 9 omy II: A gravitational lensing software ecosystem. Journal of Open Source Software, 6(62), 3283. https: 9 References H., Courbin, F., Hilbert, S., Tihhonova, O., Treu, T., Agnello, A., Ding, X., Jee, I., … Meylan, G. (2020). H0LiCOW XIII. A 2.4 per cent measurement of H0 from lensed quasars: 5.3σ tension between early- and late-Universe probes. Monthly Notices of the Royal Astronomical Society, 498(1), 1420–1439. https://doi.org/10.1093/mnras/stz3094 Yang, L., Birrer, S., & Treu, T. (2020). A versatile tool for cluster lensing source reconstruction - I. Methodology and illustration on sources in the Hubble Frontier Field Cluster MACS J0717.5+3745. Monthly Notices of the Royal Astronomical Society, 496(3), 2648–2662. https://doi.org/10.1093/mnras/staa1649 Yang, L., Roberts-Borsani, G., Treu, T., Birrer, S., Morishita, T., & Bradač, M. (2021). The evolution of the size-mass relation at z = 1-3 derived from the complete Hubble Frontier Fields data set. Monthly Notices of the Royal Astronomical Society, 501(1), 1028–1037. https://doi.org/10.1093/mnras/staa3713 Zheng, W., Postman, M., Zitrin, A., Moustakas, J., Shu, X., Jouvel, S., Høst, O., Molino, A., Bradley, L., Coe, D., Moustakas, L. A., Carrasco, M., Ford, H., Benı́tez, N., Lauer, T. R., Seitz, S., Bouwens, R., Koekemoer, A., Medezinski, E., … van der Wel, A. (2012). A magnified young galaxy from about 500 million years after the Big Bang. Nature, 489(7416), 406–408. https://doi.org/10.1038/nature11446
https://openalex.org/W2594683771	https://repository.ubn.ru.nl//bitstream/handle/2066/173222/173222.pdf	English	null	SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFβ-induced chondrogenic differentiation of bone marrow-derived mesenchymal stem cells	Scientific reports	2,017	cc-by	9,609	Version of the following full text: Publisher’s version Downloaded from: http://hdl.handle.net/2066/173222 Download date: 2024-10-24 Version of the following full text: Publisher’s version Downloaded from: http://hdl.handle.net/2066/173222 Download date: 2024-10-24 SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFbeta-induced chondrogenic differentiation of bone marrow-derived mesenchymal stem cells Kroon, L.M.G. de; Narcisi, R.; Akker, G.G. van den; Vitters, E.L.; Blaney Davidson, E.N.; Osch, G.J.V.M. van; Kraan, P.M. van der 2017, Article / Letter to editor (Scientific Reports, 7, (2017), article 43164) Doi link to publisher: https://doi.org/10.1038/srep43164 SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFβ-induced chondrogenic differentiation of bone marrow- derived mesenchymal stem cells Laurie M. G. de Kroon1,2, Roberto Narcisi2, Guus G. H. van den Akker1, Elly L. Vitters1, Esmeralda N. Blaney Davidson1, Gerjo J. V. M. van Osch2,3 & Peter M. van der Kraan1 received: 30 September 2016 accepted: 19 January 2017 Published: 27 February 2017 To improve cartilage formation by bone marrow-derived mesenchymal stem cells (BMSCs), the signaling mechanism governing chondrogenic differentiation requires better understanding. We previously showed that the transforming growth factor-β (TGFβ) receptor ALK5 is crucial for chondrogenesis induced by TGFβ. ALK5 phosphorylates SMAD2 and SMAD3 proteins, which then form complexes with SMAD4 to regulate gene transcription. By modulating the expression of SMAD2, SMAD3 and SMAD4 in human BMSCs, we investigated their role in TGFβ-induced chondrogenesis. Activation of TGFβ signaling, represented by SMAD2 phosphorylation, was decreased by SMAD2 knockdown and highly increased by SMAD2 overexpression. Moreover, TGFβ signaling via the alternative SMAD1/5/9 pathway was strongly decreased by SMAD4 knockdown. TGFβ-induced chondrogenesis of human BMSCs was strongly inhibited by SMAD4 knockdown and only mildly inhibited by SMAD2 knockdown. Remarkably, both knockdown and overexpression of SMAD3 blocked chondrogenic differentiation. Chondrogenesis appears to rely on a delicate balance in the amount of SMAD3 and SMAD4 as it was not enhanced by SMAD4 overexpression and was inhibited by SMAD3 overexpression. Furthermore, this study reveals that TGFβ-activated phosphorylation of SMAD2 and SMAD1/5/9 depends on the abundance of SMAD4. Overall, our findings suggest a more dominant role for SMAD3 and SMAD4 than SMAD2 in TGFβ-induced chondrogenesis of human BMSCs. Joint injuries frequently cause articular cartilage lesions that do not heal well in adults as articular cartilage has poor regenerative capacity. Since bone marrow-derived mesenchymal stem cells (BMSCs) can chondrogenically differentiate, they are promising for cell-based regeneration of damaged articular cartilage. Chondrogenic dif- ferentiation is potently induced by transforming growth factor-β (TGFβ)1–3. However, cartilage tissue formed by BMSC-derived chondrocytes does not completely resemble native articular cartilage as these cells tend to lose the chondrogenic phenotype due to hypertrophic differentiation4–8. Therefore, understanding how mediators of TGFβ signal transduction govern chondrogenesis will be crucial to improve cartilage regeneration by BMSCs. Binding of TGFβ to its type II serine/threonine receptor TGFBR2 activates the type I receptor TGFBR1, also termed activin receptor-like kinase 5 (ALK5), to intracellularly phosphorylate receptor-regulated SMAD proteins (R-SMADs). Activated R-SMADs form complexes with co-factor SMAD4 and these complexes translocate to the nucleus where they regulate gene transcription9. Previously, we demonstrated that either blocking the kinase activity or downregulating the expression of ALK5 in human BMSCs inhibits chondrogenesis6,10, indicating an important role for TGFβ signaling via ALK5 in chondrogenic differentiation of BMSCs. Note: Note: To cite this publication please use the final published version (if applicable). To cite this publication please use the final published version (if applicable). www.nature.com/scientificreports www.nature.com/scientificreports received: 30 September 2016 accepted: 19 January 2017 Published: 27 February 2017 Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Resultsfi Efficient knockdown and overexpression of SMAD2, SMAD3 and SMAD4. To study the role of SMAD2, SMAD3 and SMAD4 in TGFβ-induced SMAD phosphorylation and chondrogenesis of human BMSCs, their expression was modulated in human fetal BMSCs using short hairpin RNA (shRNA)-mediated knockdown and adenoviral (ad)-mediated overexpression.ifi ( ) p We confirmed efficient shRNA-mediated knockdown of SMAD2 (Fig. 1a,b), SMAD3 (Fig. 1c,d) and SMAD4 (Fig. 1e,f) at the mRNA and protein level. In addition, we observed that SMAD2 (Fig. 1a) and SMAD3 (Fig. 1c) mRNA levels, but not their protein levels (Fig. 1b,d), were increased in SMAD4-shRNA compared to Ctrl-shRNA. Although gene expression of SMAD3 (Fig. 1c) and SMAD4 (Fig. 1e) was similar between SMAD2-shRNA and Ctrl-shRNA, their protein expression (Fig. 1d,f) was decreased in SMAD2-shRNA. Next, we verified adenovi- ral overexpression of SMAD2 (Fig. 2a,b), SMAD3 (Fig. 2c,d) and SMAD4 (Fig. 2e,f) at gene and protein level. Notably, only SMAD4 protein (Fig. 2f) was slightly reduced in ad-SMAD2 compared to ad-LacZ. Thus, these data confirm efficient knockdown and overexpression of SMAD2, SMAD3 and SMAD4. TGFβ-activated phosphorylation of R-SMADs is altered by knockdown and overexpression of SMAD2, SMAD3 or SMAD4. TGFβ signaling appears to be a straightforward cascade in which the ALK5 receptor phosphorylates SMAD2 and SMAD3, which upon binding to SMAD4 translocate to the nucleus where they regulate gene transcription9. However, this system is more complex as multiple mechanisms have been dis- covered that control TGFβ signal transduction39. Since it remained unknown whether R-SMAD phosphorylation depends on the amount of SMAD2 and SMAD3 present, we investigated whether modulating the expression of SMAD2 and SMAD3 affected TGFβ-activated phosphorylation of R-SMADs. We only show the effects on phosphorylated SMAD2 (pSMAD2) due to difficult detection of pSMAD3. First, we verified that TGFβ induced pSMAD2 in Ctrl-shRNA cells (Fig. 3a) and in ad-LacZ cells compared to no stimulation (Fig. 3b). TGFβ-activated SMAD2 phosphorylation was reduced in the SMAD2-shRNA condition (Fig. 3a) and it was enhanced in ad-SMAD2 (Fig. 3b). No effect on pSMAD2 was observed with SMAD3-shRNA (Fig. 3a) or ad-SMAD3 (Fig. 3b). Moreover, to determine if co-factor SMAD4 is involved in TGFβ-activated R-SMAD phosphorylation, the expres- sion of SMAD4 was modulated. We found that pSMAD2 was lower in SMAD4-shRNA than in Ctrl-shRNA (Fig. 3a) and was similar between ad-SMAD4 and ad-LacZ (Fig. 3b). g g Next to SMAD2/3, TGFβ can activate SMAD1/5/9, and both these R-SMAD signaling pathways are important for chondrogenic induction6,10. SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFβ-induced chondrogenic differentiation of bone marrow- derived mesenchymal stem cells Laurie M. G. de Kroon1,2, Roberto Narcisi2, Guus G. H. van den Akker1, Elly L. Vitters1, Esmeralda N. Blaney Davidson1, Gerjo J. V. M. van Osch2,3 & Peter M. van der Kraan1 1Experimental Rheumatology, Department of Rheumatology, Radboud University Medical Center, Nijmegen, 6500 HB, The Netherlands. 2Department of Orthopedics, Erasmus MC University Medical Center, Rotterdam, 3015 CN, The Netherlands. 3Department of Otorhinolaryngology, Erasmus MC University Medical Center, Rotterdam, 3015 CN, The Netherlands. Correspondence and requests for materials should be addressed to P.M.V.K. (email: Peter. vanderKraan@radboudumc.nl) Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 1 www.nature.com/scientificreports/ ALK5 can activate two R-SMADs, SMAD2 and SMAD3, which share similar structures including a Mad-Homology 1 (MH1) domain, linker region and MH2 domain9,11. Despite their similarities, SMAD2 and SMAD3 have a differential role in TGFβ signal transduction12–15. Whereas Smad2 knockout mice die during gas- trulation16,17, Smad3 knockout mice develop cartilage and survive for several months after birth18–20. Moreover, SMAD3, but not SMAD2, is involved in enhanced transcriptional activity of SRY (Sex Determining Region Y)-Box 9 (SOX9); a master regulator of chondrogenesis, in human mesenchymal stem cells21,22. In contrast to SMAD3, SMAD2 lacks the MH1 domain. Therefore, SMAD2 cannot bind DNA without complex formation with SMAD4, which may explain the differential effects of SMAD2 and SMAD323–27. However, whether SMAD2 and SMAD3 have a different function during TGFβ-induced chondrogenic differentiation of human BMSCs remains largely unknown. Also the specific role of SMAD4 in chondrogenesis of human BMSCs has not been investigated. In mice, deletion of Smad4 causes early embryonic death28,29. Therefore, tissue-specific Smad4 knockout mice have been generated. Conditional deletion of Smad4 in the limb bud mesenchyme of mice leads to an absence of cartilage elements prefiguring the limb skeleton30,31, indicating that SMAD4 is required for cartilage formation. Furthermore, chondrocyte-specific Smad4 knockout mice exhibit dwarfism and impaired growth plate organiza- tion32. Similarly, Smad3 knockout mice have forelimb malformations, are smaller than wild-type mice and they develop spontaneous joint degeneration resembling the degenerative joint disease osteoarthritis in humans18,19. In humans, mutations in SMAD3 lead to skeletal anomalies and early-onset of osteoarthritis33–36. Moreover, SMAD3 is required for TGFβ-mediated suppression of hypertrophic differentiation of human articular chondrocytes37,38. Although SMAD3 does not appear directly required for embryonic cartilage and joint development, SMAD3 seems to be important for maintaining a stable cartilage phenotype by preventing cartilage degeneration. p g g p yp y p g g g Since SMAD2, SMAD3 and SMAD4 have been shown to differentially regulate TGFβ signaling and to have distinct functions during in vivo cartilage development and maintenance, we investigated their role in TGFβ-induced signaling and chondrogenesis of human BMSCs. SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFβ-induced chondrogenic differentiation of bone marrow- derived mesenchymal stem cells Laurie M. G. de Kroon1,2, Roberto Narcisi2, Guus G. H. van den Akker1, Elly L. Vitters1, Esmeralda N. Blaney Davidson1, Gerjo J. V. M. van Osch2,3 & Peter M. van der Kraan1 We knocked down and overexpressed either SMAD2, SMAD3 or SMAD4 in human fetal BMSCs and determined the effects on TGFβ-activated SMAD phos- phorylation and induction of chondrogenic differentiation. Resultsfi Moreover, the TGFβ receptor ALK5 is required for TGFβ-activated phosphoryl- ation of SMAD2/3 and SMAD1/5/940. Hence, we hypothesized there could be competition between SMAD2/3 and SMAD1/5/9 for phosphorylation by ALK5. We confirmed SMAD1/5/9 phosphorylation (pSMAD1/5/9) in response to TGFβ stimulation compared to no stimulation in Ctrl-shRNA (Fig. 3c) and ad-LacZ (Fig. 3d). TGFβ-induced pSMAD1/5/9 was higher in SMAD2-shRNA and SMAD3-shRNA than in Ctrl-shRNA (Fig. 3c). Although SMAD1/5/9 phosphorylation was similar between ad-SMAD2 and ad-LacZ, it was slightly reduced in ad-SMAD3 (Fig. 3d). Next, we studied the involvement of SMAD4 in TGFβ-activated SMAD1/5/9 phos- phorylation. We observed that in the SMAD4-shRNA condition TGFβ stimulation did not lead to induction of pSMAD1/5/9 (Fig. 3c), whereas TGFβ-induced pSMAD1/5/9 was comparable between ad-SMAD4 and ad-LacZ (Fig. 3d). ( g ) Taken together, these data demonstrate that activation of the SMAD2/3 and SMAD1/5/9 signaling path- ways by TGFβ was affected by modulating SMAD2, SMAD3 or SMAD4 expression, indicating that the levels of TGFβ-induced phosphorylated R-SMADs depend on the amount of R-SMAD2/3 and co-SMAD4. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 2 www.nature.com/scientificreports/ Figure 1. Short hairpin RNA-mediated knockdown of SMAD2, SMAD3 and SMAD4 expression. Following lentiviral transduction with either SMAD2-shRNA, SMAD3-shRNA, SMAD4-shRNA or non-mammalian shRNA control (Ctrl-shRNA), human fetal BMSCs were cultured in chondrogenic medium for 1 day. Short hairpin RNA-mediated knockdown was evaluated by determining gene (RT-qPCR; (a,c,e) and protein (Western blot; (b,d,f) expression of SMAD2 (a,b), SMAD3 (c,d) and SMAD4 (e,f). Protein levels were normalized to GAPDH and expressed as relative to Ctrl-shRNA. Gene expression was normalized to the mean CT value of RPS27a and TBP. Data are expressed as % relative to normalized mRNA levels in Ctrl-shRNA. Bars represent mean + S.D. of triplicate pellets from 2 experiments. p < 0.01; p < 0.001 compared to Ctrl-shRNA. Figure 1. Short hairpin RNA-mediated knockdown of SMAD2, SMAD3 and SMAD4 expression. Following lentiviral transduction with either SMAD2-shRNA, SMAD3-shRNA, SMAD4-shRNA or non-mammalian shRNA control (Ctrl-shRNA), human fetal BMSCs were cultured in chondrogenic medium for 1 day. Short hairpin RNA-mediated knockdown was evaluated by determining gene (RT-qPCR; (a,c,e) and protein (Western blot; (b,d,f) expression of SMAD2 (a,b), SMAD3 (c,d) and SMAD4 (e,f). Protein levels were normalized to GAPDH and expressed as relative to Ctrl-shRNA. Gene expression was normalized to the mean CT value of RPS27a and TBP. Data are expressed as % relative to normalized mRNA levels in Ctrl-shRNA. Bars represent mean + S.D. Resultsfi of triplicate pellets from 2 experiments. p < 0.01; p < 0.001 compared to Ctrl-shRNA. Knocking down either SMAD3 or SMAD4 strongly inhibits chondrogenesis. We determined the contribution of SMAD2, SMAD3 and SMAD4 during TGFβ-induced chondrogenic differentiation by knocking down the expression of these SMADs and culturing the BMSCs in pellets in chondrogenic medium containing TGFβ. After 1, 7 and 14 days we evaluated expression of chondrogenesis-specific genes ACAN, COL2A1 and SOX9, formation of cartilage matrix and the macroscopic appearance of the BMSC pellets In 1, 7 and 14 days-cultured BMSC pellets, ACAN (Fig. 4a), COL2A1 (Fig. 4b) and SOX9 (Fig. 4c) were sim- ilarly expressed in the SMAD2-shRNA and Ctrl-shRNA condition, except for a significant decrease in COL2A1 expression at day 1 in SMAD2-shRNA (p = 0.007). Although ACAN and SOX9 expression were mildly affected by SMAD3-shRNA, transcription of COL2A1 was significantly inhibited by SMAD3-shRNA at day 1 (p = 0.007), day 7 (p < 0.001) and day 14 (p = 0.029). Furthermore, a strong and significant inhibition of ACAN, COL2A1 and SOX9 expression was observed in SMAD4-shRNA at all time points (p < 0.01 for all genes and time points). Next, we analyzed cartilage matrix deposition by determining the presence of glycosaminoglycans (GAGs) and collagen type II protein. In addition, we measured the pellet size. Compared to control, TGFβ-induced dep- osition of GAGs was not inhibited by SMAD2-shRNA at day 14 (Fig. 4d), whereas it was significantly inhibited by SMAD3-shRNA (p = 0.010 at day 7; p < 0.001 at day 14). The same was observed after correcting the GAG content for the amount of DNA per pellet (Fig. 4e; Supplementary Fig. S1). Despite stimulation with TGFβ, no GAGs were formed in pellets of BMSCs transduced with SMAD4-shRNA (Fig. 4d,e). Consistent with this, analyz- ing the presence of GAGs and collagen type II protein in 14 days-cultured pellets by histology revealed that both cartilage components were similarly present in SMAD2-shRNA and Ctrl-shRNA, whereas they were reduced in SMAD3-shRNA and even absent in SMAD4-shRNA (Fig. 4f). These observations were reflected by the size of the pellets after 14 days of culturing. Compared to Ctrl-shRNA pellets, SMAD2-shRNA pellets had the same size, SMAD3-shRNA pellets were smaller and SMAD4-shRNA pellets were smallest (Fig. 4g). Next, we analyzed cartilage matrix deposition by determining the presence of glycosaminoglycans (GAGs) and collagen type II protein. In addition, we measured the pellet size. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Resultsfi Compared to control, TGFβ-induced dep- osition of GAGs was not inhibited by SMAD2-shRNA at day 14 (Fig. 4d), whereas it was significantly inhibited by SMAD3-shRNA (p = 0.010 at day 7; p < 0.001 at day 14). The same was observed after correcting the GAG content for the amount of DNA per pellet (Fig. 4e; Supplementary Fig. S1). Despite stimulation with TGFβ, no GAGs were formed in pellets of BMSCs transduced with SMAD4-shRNA (Fig. 4d,e). Consistent with this, analyz- ing the presence of GAGs and collagen type II protein in 14 days-cultured pellets by histology revealed that both cartilage components were similarly present in SMAD2-shRNA and Ctrl-shRNA, whereas they were reduced in SMAD3-shRNA and even absent in SMAD4-shRNA (Fig. 4f). These observations were reflected by the size of the pellets after 14 days of culturing. Compared to Ctrl-shRNA pellets, SMAD2-shRNA pellets had the same size, SMAD3-shRNA pellets were smaller and SMAD4-shRNA pellets were smallest (Fig. 4g). p p ( g g) Once BMSCs have differentiated into chondrocytes, they undergo hypertrophic maturation, which is characterized by enhanced expression of collagen type X α1 (COL10A1) and runt-related transcription factor Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 3 www.nature.com/scientificreports/ Figure 2. Adenoviral-mediated overexpression of SMAD2, SMAD3 and SMAD4. Human fetal BMSCs were transduced with adenovirus overexpressing SMAD2 (ad-SMAD2), ad-SMAD3, ad-SMAD4 or ad-LacZ as control, followed by culturing in chondrogenic medium for 1 day. To confirm overexpression, gene (RT-qPCR; (a,c,e) and protein (Western blot; (b,d,f) expression of SMAD2 (a,b), SMAD3 (c,d) and SMAD4 (e,f) were measured. Protein levels were normalized to GAPDH and expressed as relative to ad-LacZ. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of triplicate pellets from 2 experiments. p < 0.001 compared to ad-LacZ. Figure 2. Adenoviral-mediated overexpression of SMAD2, SMAD3 and SMAD4. Human fetal BMSCs were transduced with adenovirus overexpressing SMAD2 (ad-SMAD2), ad-SMAD3, ad-SMAD4 or ad-LacZ as control, followed by culturing in chondrogenic medium for 1 day. To confirm overexpression, gene (RT-qPCR; (a,c,e) and protein (Western blot; (b,d,f) expression of SMAD2 (a,b), SMAD3 (c,d) and SMAD4 (e,f) were measured. Protein levels were normalized to GAPDH and expressed as relative to ad-LacZ. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of triplicate pellets from 2 experiments. *p < 0.001 compared to ad-LacZ. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Resultsfi 2 (RUNX2)4–8. We observed that COL10A1 and RUNX2 mRNA levels were similar between SMAD2-shRNA, SMAD3-shRNA and Ctrl-shRNA in 7 and 14 days-cultured pellets (Supplementary Fig. S2a,b). In contrast, compared to Ctrl-shRNA, COL10A1 expression was lower in the SMAD4-shRNA than in the Ctrl-shRNA condition (Supplementary Fig. S2a), while SMAD4-shRNA had no significant effect on RUNX2 expression (Supplementary Fig. S2b).f Altogether, these results indicate that shRNA-mediated knockdown of SMAD2 had a minor inhibitory effect on TGFβ-induced chondrogenic differentiation of human BMSCs. However, SMAD3 knockdown strongly reduced cartilage deposition, and SMAD4 knockdown completely blocked chondrogenesis. SMAD3 overexpression results in a strong inhibition of chondrogenesis. Since knockdown SMAD3 overexpression results in a strong inhibition of chondrogenesis. Since knockdown of SMAD2, SMAD3 and SMAD4 inhibited chondrogenesis, we investigated whether overexpression of these SMADs might enhance TGFβ-induced chondrogenic differentiation of human BMSCs. f SMAD2, SMAD3 and SMAD4 inhibited chondrogenesis, we investigated whether overexpression of these MADs might enhance TGFβ-induced chondrogenic differentiation of human BMSCs. f Expression of ACAN (Fig. 5a), COL2A1 (Fig. 5b) and SOX9 (Fig. 5c) was comparable between the control con- dition (ad-LacZ) and ad-SMAD2, ad-SMAD3 and ad-SMAD4 after 1 day of pellet culturing. In 7 days-cultured pellets ACAN, COL2A1 and SOX9 were slightly lower expressed in ad-SMAD2 than in ad-LacZ (ACAN: p = 0.016; COL2A1: p < 0.001; SOX9: p = 0.047). However, after 14 days only ACAN expression was significantly decreased in ad-SMAD2 compared to ad-LacZ (p = 0.038). At all time points expression of ACAN, COL2A1 and SOX9 was significantly reduced by ad-SMAD3 (p < 0.05 for all genes), whereas their expression was not affected by ad-SMAD4.i y Deposition of GAGs (Fig. 5d) was significantly decreased in pellets of BMSCs transduced with ad-SMAD2 compared to control after 7 days (p = 0.009) and 14 days (p = 0.003). However, this significant reduction was not observed when the GAG content was corrected for the DNA content (Fig. 5e; Supplementary Fig. S3). Pellets of ad-SMAD3-transduced BMSCs had a significantly lower amount of GAGs per pellet (Fig. 5d: p < 0.001 at day 7; p = 0.006 at day 14) and per μg DNA (Fig. 5e; Supplementary Fig. S3; p = 0.007 at day 7; p < 0.001 at day 14) than the control condition. At all time points, GAG deposition (Fig. 5d,e) was similar between ad-SMAD4 and ad-LacZ. Histological analysis of 14 days-cultured pellets revealed that the abundance of GAGs and collagen type II protein (Fig. 5f) was similar between ad-SMAD2, ad-SMAD4 and ad-LacZ, whereas their abundance was strongly reduced in the ad-SMAD3 condition. In line with these observations, pellets of ad-SMAD2-transduced Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 4 www.nature.com/scientificreports/ Figure 3. TGFβ-activated phosphorylation of R-SMADs is affected by knockdown and overexpression of SMAD2, SMAD3 and SMAD4. The effects of shRNA-mediated knockdown and adenoviral-mediated overexpression of SMAD2, SMAD3 and SMAD4 on TGFβ-induced phosphorylation of SMAD2 (pSMAD2; (a,b) and pSMAD1/5/9 (c,d) were determined using Western blot. SMAD3 overexpression results in a strong inhibition of chondrogenesis. Since knockdown To confirm TGFβ-induced phosphorylation of SMAD2 and SMAD1/5/9, Ctrl-shRNA cells and ad-LacZ cells were either not stimulated (−) or stimulated (+) with TGFβ in chondrogenic medium. Protein levels were normalized to GAPDH and expressed as relative to unstimulated Ctrl-shRNA or ad-LacZ. Figure 3. TGFβ-activated phosphorylation of R-SMADs is affected by knockdown and overexpression of SMAD2, SMAD3 and SMAD4. The effects of shRNA-mediated knockdown and adenoviral-mediated overexpression of SMAD2, SMAD3 and SMAD4 on TGFβ-induced phosphorylation of SMAD2 (pSMAD2; (a,b) and pSMAD1/5/9 (c,d) were determined using Western blot. To confirm TGFβ-induced phosphorylation of SMAD2 and SMAD1/5/9, Ctrl-shRNA cells and ad-LacZ cells were either not stimulated (−) or stimulated (+) with TGFβ in chondrogenic medium. Protein levels were normalized to GAPDH and expressed as relative to unstimulated Ctrl-shRNA or ad-LacZ. BMSCs were slightly smaller than control pellets, whereas pellets of ad-SMAD3-transduced BMSCs were much smaller (Fig. 5g). The pellet size was comparable between the ad-SMAD4 and ad-LacZ condition (Fig. 5g). h In addition, we found that chondrocyte hypertrophy markers, COL10A1 and RUNX2, were sim- ilarly expressed between ad-LacZ and ad-SMAD4 after 7 and 14 days (Supplementary Fig. S4a,b). Although COL10A1 expression was lower in ad-SMAD2 than in ad-LacZ at day 7, the two conditions were compa- rable at day 14 (Supplementary Fig. S4a). At both time points, RUNX2 mRNA levels were similar between ad-SMAD2 and ad-LacZ (Supplementary Fig. S4b). Strikingly, in the ad-SMAD3 condition COL10A1 mRNA was lower (Supplementary Fig. S4a), whereas RUNX2 mRNA was higher than in the ad-LacZ condition (Supplementary Fig. S4b). pp y g Our data demonstrate that TGFβ-induced chondrogenesis was slightly inhibited by SMAD2 overexpression and not affected by SMAD4 overexpression, whereas it was strongly inhibited by SMAD3 overexpression. Discussionh The pro-chondrogenic effect of TGFβ on human BMSCs is well-known, however, the contribution of the down- stream signaling molecules SMAD2, SMAD3 and SMAD4 during chondrogenesis has been less well investigated. By using human fetal BMSCs as a model for TGFβ-induced chondrogenic differentiation, we demonstrate here that modulating SMAD2 expression had a minor effect on chondrogenesis. In contrast, knockdown as well as overexpression of SMAD3 strongly inhibited cartilage formation and SMAD4 knockdown completely blocked chondrogenesis.if To the best of our knowledge, the specific role of SMAD2 and SMAD3 during chondrogenic differentiation of human BMSCs has been studied in one study so far, which demonstrated that modulating SMAD2 expression does not affect chondrogenesis22. In contrast, we observed a slight reduction in cartilage formation when SMAD2 was knocked down or overexpressed. We also found that SMAD2 knockdown led to reduced protein levels of SMAD3 and SMAD4, whereas Furumatsu et al. did not report an effect of SMAD2 knockdown on SMAD3 and SMAD4 expression22. Since knockdown of either SMAD3 or SMAD4 inhibited chondrogenesis, reduced SMAD3 and SMAD4 abundance might explain why chondrogenesis was slightly inhibited by SMAD2 knockdown. Our Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 5 www.nature.com/scientificreports/ Figure 4. Chondrogenesis is mildly inhibited by SMAD2-shRNA, is strongly inhibited by SMAD3-shRNA and completely blocked by SMAD4-shRNA. Human fetal BMSCs were transduced (lentivirus) either with SMAD2-shRNA, SMAD3-shRNA, SMAD4-shRNA or Ctrl-shRNA and subsequently pellet-cultured for 1, 7 or 14 days in chondrogenic medium with TGFβ. The effect of shRNA-mediated knockdown of SMAD2, SMAD3 and SMAD4 on chondrogenesis was determined by RT-qPCR analysis of the chondrogenesis-specific genes; ACAN (a), COL2A1 (b) and SOX9 (c), deposition of glycosaminoglycans (GAGs) per pellet (d) and per μg DNA (e), histological examination of GAGs by Safranin O (Saf.O) staining and collagen type II by immunohistochemistry in sections of 14 days-cultured pellets (f), and macroscopic evaluation of 14 days- cultured pellets (g). In (f) representative images of consecutive pellet sections per condition are shown and the scale bar represents 500 μm. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of triplicate pellets from 2 experiments. p < 0.05; p < 0.01; p < 0.001 compared to Ctrl-shRNA. Figure 4. Chondrogenesis is mildly inhibited by SMAD2-shRNA, is strongly inhibited by SMAD3-shRNA and completely blocked by SMAD4-shRNA. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Discussionh Human fetal BMSCs were transduced (lentivirus) either with SMAD2-shRNA, SMAD3-shRNA, SMAD4-shRNA or Ctrl-shRNA and subsequently pellet-cultured for 1, 7 or 14 days in chondrogenic medium with TGFβ. The effect of shRNA-mediated knockdown of SMAD2, SMAD3 and SMAD4 on chondrogenesis was determined by RT-qPCR analysis of the chondrogenesis-specific genes; ACAN (a), COL2A1 (b) and SOX9 (c), deposition of glycosaminoglycans (GAGs) per pellet (d) and per μg DNA (e), histological examination of GAGs by Safranin O (Saf.O) staining and collagen type II by immunohistochemistry in sections of 14 days-cultured pellets (f), and macroscopic evaluation of 14 days- cultured pellets (g). In (f) representative images of consecutive pellet sections per condition are shown and the scale bar represents 500 μm. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of triplicate pellets from 2 experiments. p < 0.05; p < 0.01; p < 0.001 compared to Ctrl-shRNA. 6 Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 www.nature.com/scientificreports/ Figure 5. Chondrogenesis is strongly inhibited by SMAD3 overexpression, whereas it is mildly inhibited by SMAD2 overexpression and unaffected by SMAD4 overexpression. Human fetal BMSCs transduced either with adenoviral SMAD2 (ad-SMAD2), ad-SMAD3, ad-SMAD4 or ad-LacZ as control were pellet-cultured for 1, 7 or 14 days in chondrogenic medium with TGFβ. The effect of SMAD overexpression on chondrogenesis was determined by RT-qPCR analysis of the chondrogenesis-specific genes; ACAN (a), COL2A1 (b) and SOX9 (c), deposition of glycosaminoglycans (GAGs) per pellet (d) and per μg DNA (e), histological examination of GAGs by Safranin O (Saf.O) staining and collagen type II by immunohistochemistry in sections of 14 days- cultured pellets (f), and macroscopic evaluation of 14 days-cultured pellets (g). In (f) representative images of consecutive pellet sections per condition are shown and the scale bar represents 500 μm. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of t i li t ll t f 2 i t <0 05 <0 01 * <0 001 d t d L Z Figure 5. Chondrogenesis is strongly inhibited by SMAD3 overexpression, whereas it is mildly inhibited by SMAD2 overexpression and unaffected by SMAD4 overexpression. Human fetal BMSCs transduced either with adenoviral SMAD2 (ad-SMAD2), ad-SMAD3, ad-SMAD4 or ad-LacZ as control were pellet-cultured for 1, 7 or 14 days in chondrogenic medium with TGFβ. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Discussionh The effect of SMAD overexpression on chondrogenesis was determined by RT-qPCR analysis of the chondrogenesis-specific genes; ACAN (a), COL2A1 (b) and SOX9 (c), deposition of glycosaminoglycans (GAGs) per pellet (d) and per μg DNA (e), histological examination of GAGs by Safranin O (Saf.O) staining and collagen type II by immunohistochemistry in sections of 14 days- cultured pellets (f), and macroscopic evaluation of 14 days-cultured pellets (g). In (f) representative images of consecutive pellet sections per condition are shown and the scale bar represents 500 μm. Gene expression data are presented as −ΔCT compared to the mean CT value of RPS27a and TBP. Bars represent mean + S.D. of triplicate pellets from 2 experiments. p < 0.05; p < 0.01; **p < 0.001 compared to ad-LacZ. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 7 www.nature.com/scientificreports/ finding that TGFβ-induced chondrogenic differentiation of human BMSCs is more strongly inhibited by SMAD3 knockdown than by SMAD2 knockdown is in line with the aforementioned study22. While we employed contin- uous knockdown by viral transduction with shRNA, Furumatsu and co-workers employed transient knockdown by transfection with siRNA22. We, thus, verified that continuous knockdown of SMAD2 and SMAD3 had effects on chondrogenesis similar to transient knockdown. g Whereas the previous study has shown that SMAD3 overexpression accelerated chondrogenesis22, in our study SMAD3 overexpression blocked chondrogenic differentiation of human BMSCs. This discrepancy might be caused by a difference in chondrogenic media composition (details of medium components were not reported), cell source (fetal or adult origin, bone marrow or other tissue derived) or amount of overexpression (western blots of overexpressed SMAD3 were not shown). In addition, we found that the DNA content per pellet decreased over time in cells overexpressing SMAD3 (Supplementary Fig. S3). In multiple cell types, it has been shown that overexpression of SMAD3 leads to an inhibition of proliferation41–43. Importantly, proliferation of BMSCs early during chondrogenic differentiation is required for chondrogenesis44. Therefore, SMAD3 overexpression might have inhibited chondrogenesis, because it blocked proliferation. Overall, this indicates that constantly high levels of SMAD3 in BMSCs do not accelerate cartilage matrix formation. Since both knockdown and overexpression of SMAD3 inhibited chondrogenic differentiation, chondrogenesis seems to rely on well-balanced levels of SMAD3.hif gf , g y The present study, for the first time, investigated the role of SMAD4 during chondrogenic differentiation of human BMSCs. Discussionh Previous studies in mice have demonstrated that mesenchyme-specific deletion of Smad4 leads to an absence of the limb skeleton as a result of impaired mesenchymal condensation30,31; a process required for initiating chondrogenesis45. In our study, SMAD4 knockdown did not interfere with pellet formation, implying that human BMSCs can form pellets even with a low amount of SMAD4. Although BMSCs transduced with SMAD4-shRNA formed pellets, they did not undergo chondrogenic differentiation. This observation is in line with a previous study demonstrating that BMSCs isolated from Smad4 knockout mice show reduced expression of Sox930. Additionally, we found that SMAD4 knockdown inhibited TGFβ-induced SMAD1/5/9 phosphoryl- ation and expression of COL2A1; a direct transcriptional target of SOX946. In murine chondroprogenitor cells, complex formation between Smad4 and phosphorylated Smad1/5 proteins is required for transactivation of the Col2a1 promoter47. This might explain why chondrogenesis did not occur when we knocked down SMAD4. Furthermore, we previously demonstrated that the SMAD2/3 as well as the SMAD1/5/9 pathways are required for TGFβ-induced chondrogenesis of human BMSCs6,10. In the present study activation of both pathways was reduced by SMAD4 knockdown, thereby possibly explaining the absence of cartilage formation. Altogether, these findings underline the importance of SMAD4 during TGFβ-induced chondrogenic differentiation of human BMSCs.fh Following chondrogenesis, BMSC-derived chondrocytes display signs of hypertrophic differentiation. This is undesired for the formation of articular cartilage, as hypertrophic chondrocytes produce cartilage that will mineralize and ossify when implanted in vivo7. We found that modulating SMAD2 expression did not have an effect on expression of hypertrophic differentiation markers, neither did SMAD4 overexpression. SMAD3 overex- pression and SMAD4 knockdown did result in lower COL10A1 expression. In these conditions, however, next to hypertrophy markers, the markers of chondrogenesis were also much lower expressed than in the control condi- tion, following the principle that reduced chondrogenesis leads to reduced hypertrophic differentiation. Whereas COL10A1 expression was lower by SMAD3 overexpression, RUNX2 expression was slightly, but significantly, higher. This was surprising, because SMAD3 is required for repression of RUNX2 expression and chondrocyte hypertrophy38,48–51. Possibly, as a result of decreased chondrogenesis RUNX2 expression failed to go down during the early phase of chondrogenesis or RUNX2 expression levels might depend on the abundance of SMAD3, but this requires further investigation. Discussionh Overall, our results failed to show effects specifically on hypertrophic differen- tiation of BMSC-derived chondrocytes without effects on the induction of chondrogenesis.fff yf g Besides the effects on chondrogenic and hypertrophic differentiation, this study investigated the effect of modulating SMAD2, SMAD3 and SMAD4 expression on TGFβ signal transduction. After activation by TGFβ, the ALK5 receptor phosphorylates SMAD2/3 proteins9. We show that SMAD2 phosphorylation was reduced by knockdown of SMAD2, while it highly increased by SMAD2 overexpression. These findings suggests that TGFβ-activated phosphorylation depends on the number of R-SMADs present in the cytoplasm rather than, for instance, on the kinase activity or number of ALK5 receptors. , y p We previously showed that TGFβ does not only induce phosphorylation of SMAD2/3, but also of SMAD1/5/9 in BMSCs6,10,52. Although SMAD1/5/9 cannot be directly phosphorylated by the TGFβ receptor ALK5, TGFβ-activated SMAD1/5/9 phosphorylation requires the intracellular kinase domains of ALK540. Moreover, ALK5 is the only receptor through which TGFβ can induce phosphorylation of SMAD2/353,54. Consistent with a study showing enhanced pSMAD1/5/9 in SMAD3 knockout chondrocytes51, we demonstrate increased pSMAD1/5/9 when SMAD3 or SMAD2 were knocked down. This supports the idea that TGFβ-induced phos- phorylation of SMAD1/5/9 requires the ALK5 receptor and is independent of other receptors that can activate SMAD1/5/955,56. Although SMAD4 is an important co-factor for translocation of activated R-SMADs to the nucleus57, it has not been implicated in controlling the phosphorylation of R-SMADs. Patients with Myhre syndrome have mutations in SMAD4 that lead to decreased ubiquitination of SMAD4 protein, resulting in accumulation of SMAD458–61. In skin fibroblasts from these patients, pSMAD2/3 and pSMAD1/5/9 were enhanced compared to healthy con- trols58,61. Contrary to this, in our study SMAD4 overexpression had no effect on R-SMAD phosphorylation. On the other hand, SMAD4 knockdown reduced TGFβ-induced phosphorylation of SMAD2 and completely pre- vented SMAD1/5/9 phosphorylation. Based on these observations, we speculate that SMAD4 is required for R-SMAD phosphorylation, or prevents R-SMADs from de-phosphorylation62,63 or ubiquitination64,65, which adds a regulatory mechanism that controls TGFβ signal transduction. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 8 www.nature.com/scientificreports/ To conclude, this study reveals that TGFβ-activated phosphorylation of R-SMADs in BMSCs does not only depend on the levels of SMAD2 and SMAD3, but also on the presence of SMAD4. Moreover, our findings suggest that SMAD3 and SMAD4 are more important than SMAD2 for TGFβ-induced chondrogenic differentiation of human BMSCs. Methods C l i Culturing of human fetal bone marrow-derived mesenchymal stem cells. Human fetal BMSCs (#SCC7500, Lot#6890, ScienCell Research Laboratories, Carlsbad, CA, USA) were expanded in Mesenchymal Stem Cell Growth Medium (MSCGM™; Lonza, Basel, Switzerland) supplemented with 1% Penicillin- Streptomycin-Glutamine (Gibco, Carlsbad, CA, USA). Cells were cultured in a 37 °C-incubator with 5% CO2. After reaching 80% confluence, cells were passaged using 0.05% trypsin-EDTA (Gibco) and re-seeded (~6,000 cells/cm2) in MSCGM. After 4 passages, cells were stored in liquid nitrogen. Per experiment, fetal BMSCs were defrosted and expanded in MSCGM for another 2 or 3 passages. Induction of chondrogenic differentiation. BMSC pellets were obtained by centrifuging 200,000 cells at 300 × g for 8 minutes in polystyrene V-bottom tubes (Greiner Bio-One, Alphen a/d Rijn, Netherlands). Pellets were cultured for 1, 7 or 14 days in 0.5 mL of serum-free chondrogenic medium, which consisted of DMEM-high glucose-GlutaMAX, 1% Penicillin-Streptomycin-Glutamine (both from Gibco), 6.25 μg/mL Insulin, 6.25 μg/mL Transferrin, 6.25 ng/mL selenious acid 5.35 μg/mL linoleic acid, 1.25 mg/mL bovine serum albumin, 1.0 mg/mL sodium pyruvate, 0.4 mg/mL L-proline, 50 μg/mL sodium L-ascorbate, 10−7 M dexamethasone (all from Sigma-Aldrich), and 10 ng/mL TGFβ1 (Biolegend, San Diego, CA, USA). This medium was renewed 3 times per week. Short hairpin-mediated knockdown of SMAD2, SMAD3 or SMAD4. MISSION® TRC-Hs1.5 shRNA clones targeting SMAD2 (TRCN0000040036), SMAD3 (TRCN0000330056) or SMAD4 (TRCN0000010321), and Non-Mammalian shRNA control (SHC002) constructed in the pLKO.1-Puro plasmid vector were obtained from Sigma-Aldrich. Lentiviruses were packaged as described previously10. Briefly, HEK293T cells (ATCC, Manassas, VA, USA) were co-transduced with plasmids of Rev, Gag, Pol, VSV-G (Plasmid Factory, Bielefeld, Germany) and a pLKO.1-Puro plasmid vector by calcium phosphate precipitation in DMEM (Gibco) containing 10% FCS (Perbio Science, Erembodegem, Belgium), 0.01 mM cholesterol (Sigma-Aldrich) and 1% pyruvate (Gibco). Medium was renewed at day 1, 2 and 3 post-transduction, collecting the medium at day 2 and 3. Collected medium was filtered through a 0.45 μm filter and centrifuged at 134,350 × g for 2 hours (Sorvall WX80+, ThermoFisher Scientific). Lentivirus concentration was determined with the INNOTEST® HIV p24 Antigen assay (Fujirebio Europe, Gent, Belgium) and expressed as pg of p24/μL.l g p pg p μ At 20% confluence, BMSCs were infected with 1 pg p24 per cell in MSCGM (Lonza) supplemented with 100 μ g/mL protamine sulfate (Sigma-Aldrich) for 1 day. After culture-expanding infected cells for 2 days, pellets were prepared to induce chondrogenesis. Discussionh However, as cartilage formation was not enhanced by overexpression of SMAD4 and even inhib- ited by SMAD3 overexpression, induction of chondrogenic differentiation seems to rely on a delicate balance in the amount of SMAD3 and SMAD4. This also implies that continuously enhanced SMAD3 expression levels may not be a suitable strategy to improve chondrogenesis. Despite its exploratory nature, this study offers novel insights into the signaling mechanism governing the induction of chondrogenic differentiation of human BMSCs. Further research is required to delineate the function of SMAD4 in TGFβ-induced phosphorylation of R-SMADs and the mechanisms behind the detrimental effects of SMAD3 overexpression on chondrogenesis. Table 1. List of primers used for RT-qPCR. GAPDH (#G8795, Sigma-Aldrich), followed by 1 hour-incubation with HRP-linked antibody (1:1,500) against mouse-IgG (#P0260, Dako). Proteins were visualized by enhanced chemiluminescence using Prime Western Blotting Detection Reagent and a ImageQuant LAS4000 machine (GE Healthcare). Densitometry was performed using ImageJ software (release 1.46r; National Institute of Health, Bethesda, Maryland, USA). RNA extraction and gene expression analysis. Pellets (n = 3 per condition at each time point) were collected in TRIzol® (Sigma-Aldrich) and disrupted using MagNA Lyser instrument (Roche). After total RNA was isolated according to manufacturer’s protocol (Sigma-Aldrich), RNA samples were treated with DNAse (Invitrogen) to remove contaminating genomic DNA. RNA concentration and purity were measured using a NanoDrop® spectrophotometer (Isogen Life Science, Utrecht, the Netherlands). 0.5 μg RNA was reverse tran- scribed in cDNA with M-MLV Reverse Transcriptase (Invitrogen). Gene expression was measured by real-time Reverse Transcription Quantitative Polymerase Chain Reaction (RT-qPCR) on a StepOnePlus™ System using SYBR Green Master mix (Applied Biosystems) and the primers listed in Table 1. CT values were determined at a fixed threshold level of fluorescence and efficiency of all primers was between 90% and 110% (Table 1). Data were normalized to the mean CT value of Ribosomal protein 27a (RPS27A) and TATA-box binding protein (TBP). The following genes were used as markers of chondrogenesis; Aggrecan (ACAN), Collagen type IIα1 (COL2A1) and SRY (Sex Determining Region Y)-Box 9 (SOX9). In addition, markers of chondrocyte hypertrophy; collagen type 10 α1 (COL10A1) and runt-related transcription factor 2 (RUNX2), were measured. Histology. To macroscopically evaluate pellet size of 14-days cultured pellets, pictures were taken using a microscope (Wild M3B, Heerbrugg, Switzerland). Subsequently, pellets were fixed in 4% formalin for 14 hours, embedded in paraffin and sectioned (6 μm). Sections were stained with 0.1% aqueous Safranin O (Brunschwig Chemie, Amsterdam, the Netherlands), resulting in red-staining of negatively-charged GAGs. As counter-staining, 0.1% aqueous Fast Green (Brunschwig Chemie) was used to stain the cytoplasm blue/green. For immunohistochemical staining of collagen type II, sections were pre-treated with 1 mg/mL pronase and 10 mg/mL hyaluronidase (Sigma-Aldrich), and then incubated with 0.4 μg/mL antibody specific for collagen type II (#II-II6B3, Developmental Studies Hybridoma Bank, Iowa City, IA, USA) or 0.4 μg/mL mouse-IgG1 (#X0931, Dako, Glostrup, Denmark). Following incubation with alkaline phosphatase (AP)-conjugated secondary anti- body (1:50, #HK-321-UK, Biogenex, San Ramon, CA, USA), AP-activity was visualized (magenta color) by incu- bation with new-fuchsin substrate. Sections were counterstained with haematoxylin (purple) to visualize nuclei. Methods C l i 1 day after pellet preparation, the efficiency of shRNA-mediated knockdown of SMAD2, SMAD3 and SMAD4 was analyzed by RT-qPCR and Western blot. Adenoviral-mediated overexpression of SMAD2, SMAD3 or SMAD4. Adenoviruses for SMAD2, SMAD3 and SMAD4 were kindly provided by Dr. P. ten Dijke (Leiden University Medical Center, the Netherlands). Cells at 80% confluence were incubated for 3 hours with adenovirus (multiplicity of infection of 1 pfu/cell) overexpressing either SMAD2, SMAD3, SMAD4 or LacZ (control). Following transduction, cells were washed and centrifuged to obtain pellets for chondrogenic induction. Overexpression of SMAD2, SMAD3 and SMAD4 was verified at gene (RT-qPCR) and protein (Western blot) level in BMSCs cultured for 1 day in chon- drogenic medium. Western blot analysis of (phosphorylated) SMAD proteins. To evaluate the effects of SMAD knock- down and overexpression on activation of TGFβ signaling, cells transduced either with lentivirus (SMAD-shRNA) or adenovirus (SMAD overexpression) were seeded in chondrogenic medium without TGFβ. After 18 hours, cells were stimulated with 10 ng/mL TGFβ1 (Biolegend) for 1 hour and cell lysates were prepared as described below to determine pSMAD2 and pSMAD1/5/9 expression. p p p To determine protein abundance, cells were lysed (duplicate per condition) using lysis buffer (Cell-Signaling-Technology, Danvers, MA, USA) containing 1% protease inhibitor (Roche, Mannheim, Germany) and lysates were sonicated on ice. Protein concentration was determined using bicinchoninic acid assay. Duplicate samples per condition were pooled and 10 μg protein lysate was subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose membranes and overnight at 4 °C incubated with an antibody (1:1,000) recognizing SMAD2 (#40855, Abcam, Cambridge, UK), SMAD3 (#28379, Abcam), SMAD4 (#AF2097, R&D systems, Minneapolis, MN, USA), pSMAD2 (#3101 L, Cell Signaling Technology) or pSMAD1/5/9 (#9511 L, Cell Signaling Technology). Subsequently, membranes were incubated with HRP-linked antibody (1:1,500) against rabbit-IgG (#P0448, Dako, Glostrup, Denmark) or against goat-IgG (#P0449, Dako) for 1 hour at room temperature. To evaluate equal protein loading between conditions, GAPDH expression was determined. Membranes were overnight incubated with an antibody (1:20,000) recognizing Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 9 www.nature.com/scientificreports/ Gene Forward primer Reverse primer Product RPS27A TGGCTGTCCTGAAATATTATAAGGT CCCCAGCACCACATTCATCA 90 bp TBP GCTTCGGAGAGTTCTGGGATTG GCAGCAAACCGCTTGGGATTA 134 bp SMAD2 CCGACACACCGAGATCCTAAC AGGAGGTGGCGTTTCTGGAAT 127 bp SMAD3 CATCGAGCCCCAGAGCAATA GTGGTTCATCTGGTGGTCACT 88 bp SMAD4 CCAATCATCCTGCTCCTGAGT CCAGAAGGGTCCACGTATCC 130 bp ACAN GCCTGCGCTCCAATGACT ATGGAACACGATGCCTTTCAC 104 bp COL2A1 CACGTACACTGCCCTGAAGGA CGATAACAGTCTTGCCCCACTT 65 bp SOX9 TGGGCAAGCTCTGGAGACTT CCCGTTCTTCACCGACTTCCT 140 bp COL10A1 TTTTACGCTGAACGATACCAAATG CTGTGTCTTGGTGTTGGGTAGTG 66 bp RUNX2 GCAAGGTTCAACGATCTGAGA TTCCCGAGGTCCATCTACTG 141 bp References A short amino-acid sequence in MH1 domain is responsible for functional difference between Smad2 and Smad3. Oncogene 18, 1643–1648, doi: 10.1038/sj.onc.1202729 (1999). 6. Kim, J., Johnson, K., Chen, H. J., Carroll, S. & Laughon, A. Drosophila Mad binds to DNA and directly mediates activation o vestigial by Decapentaplegic. Nature 388, 304–308, doi: 10.1038/40906 (1997). g y p p g 7. Shi, Y. et al. Crystal structure of a Smad MH1 domain bound to DNA: insights on DNA binding in TGF-beta signaling. Cell 94 585–594, doi: 10.1016/S0092-8674(00)81600-1 (1998).h 27. Shi, Y. et al. Crystal structure of a Smad MH1 domain bound to DNA: insights on DNA bind 585–594, doi: 10.1016/S0092-8674(00)81600-1 (1998).h 28. Sirard, C. et al. The tumor suppressor gene Smad4/Dpc4 is required for gastrulation and later for anterior development of the mouse embryo. Genes & development 12, 107–119, doi: 10.1101/Gad.12.1.107 (1998).h y p 29. Yang, X., Li, C. L., Xu, X. L. & Deng, C. X. The tumor suppressor SMAD4/DPC4 is essential for epiblast proliferation and mesoderm induction in mice. P Natl Acad Sci USA 95, 3667–3672, doi: 10.1073/pnas.95.7.3667 (1998). y p 29. Yang, X., Li, C. L., Xu, X. L. & Deng, C. X. The tumor suppressor SMAD4/DPC4 is essential for ep induction in mice. P Natl Acad Sci USA 95, 3667–3672, doi: 10.1073/pnas.95.7.3667 (1998). p 0. Benazet, J. D. et al. Smad4 is required to induce digit ray primordia and to initiate the aggregation and differentiation o chondrogenic progenitors in mouse limb buds. Development 139, 4250–4260, doi: 10.1242/dev.084822 (2012). 30. Benazet, J. D. et al. Smad4 is required to induce digit ray primordia and to initiate the a chondrogenic progenitors in mouse limb buds. Development 139, 4250–4260, doi: 10.1242/dev 31. Lim, J. et al. BMP-Smad4 signaling is required for precartilaginous mesenchymal condensation independent of Sox9 in the mouse. Developmental biology 400, 132–138, doi: 10.1016/j.ydbio.2015.01.022 (2015). 2. Zhang, J. et al. Smad4 is required for the normal organization of the cartilage growth plate. Developmental biology 284, 311–322, doi 10.1016/j.ydbio.2005.05.036 (2005). j y 3. van de Laar, I. M. B. H. et al. Mutations in SMAD3 cause a syndromic form of aortic aneurysms and dissections with early-onse osteoarthritis. Nature genetics 43, 121–U165, doi: 10.1038/ng.744 (2011). 34. van de Laar, I. M. B. H. et al. Phenotypic spectrum of the SMAD3-related aneurysms–osteoarthritis syndrome. Journal of medical genetics 49, 47–57, doi: 10.1136/jmedgenet-2011-100382 (2012). 35. Valdes, A. M. et al. References References 1. Johnstone, B., Hering, T. M., Caplan, A. I., Goldberg, V. M. & Yoo, J. U. In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Experimental cell research 238, 265–272, doi: 10.1006/excr.1997.3858 (1998).f 1. Johnstone, B., Hering, T. M., Caplan, A. I., Goldberg, V. M. & Yoo, J. U. In vitro chondrogenesis of bone marrow-derived mesenchymal progenitor cells. Experimental cell research 238, 265–272, doi: 10.1006/excr.1997.3858 (1998).f 2. Mackay, A. M. et al. Chondrogenic differentiation of cultured human mesenchymal stem cells from marrow. Tissue engineering 4 415–428, doi: 10.1089/ten.1998.4.415 (1998). 3. Pittenger, M. F. et al. Multilineage potential of adult human mesenchymal stem cells. Science 284, 143–147, doi: 10.1126 science.284.5411.143 (1999). ( ) 4. Hunziker, E. B., Lippuner, K., Keel, M. J. & Shintani, N. An educational review of cartilage repair: precepts & practice–myths & misconceptions–progress & prospects. Osteoarthritis and cartilage/OARS, Osteoarthritis Research Society 23, 334–350, doi: 10.1016/j.joca.2014.12.011 (2015). j j 5. Yamasaki, S., Mera, H., Itokazu, M., Hashimoto, Y. & Wakitani, S. Cartilage Repair With Autologous Bone Marrow Mesenchymal Stem Cell Transplantation: Review of Preclinical and Clinical Studies. Cartilage 5, 196–202, doi: 10.1177/1947603514534681 (2014).f 6. Hellingman, C. A. et al. Smad signaling determines chondrogenic differentiation of bone-marrow-derived mesenchymal stem cells: inhibition of Smad1/5/8P prevents terminal differentiation and calcification. Tissue engineering. Part A 17, 1157–1167, doi: 10.1089/ten.TEA.2010.0043 (2011). 7. Pelttari, K. et al. Premature induction of hypertrophy during in vitro chondrogenesis of human mesenchymal stem cells correlates with calcification and vascular invasion after ectopic transplantation in SCID mice. Arthritis and rheumatism 54, 3254–3266, doi: 10.1002/art.22136 (2006). 8. Scotti, C. et al. Recapitulation of endochondral bone formation using human adult mesenchymal stem cells as a paradigm developmental engineering. Proc Natl Acad Sci USA 107, 7251–7256, doi: 10.1073/pnas.1000302107 (2010). g g 9. Heldin, C. H., Miyazono, K. & tenDijke, P. TGF-beta signalling from cell membrane to nucleus through SMAD proteins. Nature 390 465–471, doi: 10.1038/37284 (1997). 10. de Kroon, L. M. et al. Activin Receptor-Like Kinase Receptors ALK5 and ALK1 Are Both Required for TGFbeta-Induced Chondrogenic Differentiation of Human Bone Marrow-Derived Mesenchymal Stem Cells. PloS one 10, e0146124, doi: 10.1371/ journal.pone.0146124 (2015). j p 11. Massague, J., Seoane, J. & Wotton, D. Smad transcription factors. Genes & development 19, 2783–2810, doi: 10.1101/gad.1350705 (2005).f 12. Kretschmer, A. et al. Differential regulation of TGF-beta signaling through Smad2, Smad3 and Smad4. Oncogene 22, 6748–6763, doi: 10.1038/sj.onc.1206791 (2003). j 3. Yang, Y. C. et al. References Hierarchical model of gene regulation by transforming growth factor beta. Proc Natl Acad Sci USA 100 10269–10274, doi: 10.1073/pnas.1834070100 (2003).f p 14. Brown, K. A., Pietenpol, J. A. & Moses, H. L. A tale of two proteins: differential roles and regulation of Smad2 and Smad3 in TGF- beta signaling. Journal of cellular biochemistry 101, 9–33, doi: 10.1002/jcb.21255 (2007).ii 15. Piek, E. et al. Functional characterization of transforming growth factor beta signaling in Smad2- and Smad3-deficient fibroblasts. The Journal of biological chemistry 276, 19945–19953, doi: 10.1074/jbc.M102382200 (2001).t h f g y j 6. Nomura, M. & Li, E. Smad2 role in mesoderm formation, left-right patterning and craniofacial development. Nature 393, 786–790 doi: 10.1038/31693 (1998).f 7. Waldrip, W. R., Bikoff, E. K., Hoodless, P. A., Wrana, J. L. & Robertson, E. J. Smad2 signaling in extraembryonic tissues determine anterior-posterior polarity of the early mouse embryo. Cell 92, 797–808, doi: 10.1016/S0092-8674(00)81407-5 (1998). 18. Datto, M. B. et al. Targeted disruption of Smad3 reveals an essential role in transforming growth factor beta-mediated signal transduction. Molecular and cellular biology 19, 2495–2504, doi: 10.1128/MCB.19.4.2495 (1999).f 9. Zhu, Y., Richardson, J. A., Parada, L. F. & Graff, J. M. Smad3 mutant mice develop metastatic colorectal cancer. Cell 94, 703–714, doi 10.1016/S0092-8674(00)81730-4 (1998).i 20. Li, T. F. et al. Aberrant hypertrophy in Smad3-deficient murine chondrocytes is rescued by restoring transforming growth factor beta-activated kinase 1/activating transcription factor 2 signaling: a potential clinical implication for osteoarthritis. Arthritis and rheumatism 62, 2359–2369, doi: 10.1002/art.27537 (2010).h , , ( ) 21. Furumatsu, T., Ozaki, T. & Asahara, H. Smad3 activates the Sox9-dependent transcription on chromatin. The international journal of biochemistry & cell biology 41, 1198–1204, doi: 10.1016/j.biocel.2008.10.032 (2009). 22. Furumatsu, T., Tsuda, M., Taniguchi, N., Tajima, Y. & Asahara, H. Smad3 induces chondrogenesis through the activation of SOX9 via CREB-binding protein/p300 recruitment. The Journal of biological chemistry 280, 8343–8350, doi: 10.1074/jbc.M413913200 (2005). 23. Kawabata, M., Inoue, H., Hanyu, A., Imamura, T. & Miyazono, K. Smad proteins exist as monomers in vivo and undergo homo- and hetero-oligomerization upon activation by serine/threonine kinase receptors. The EMBO journal 17, 4056–4065, doi: 10.1093/ emboj/17.14.4056 (1998). j ( ) 4. Yingling, J. M. et al. Tumor suppressor Smad4 is a transforming growth factor beta-inducible DNA binding protein. Molecular and cellular biology 17, 7019–7028, doi: 10.1128/MCB.17.12.7019 (1997).f 5. Dennler, S., Huet, S. & Gauthier, J. M. Table 1. List of primers used for RT-qPCR. Glycosaminoglycans and DNA content measurements. Pellets cultured for 1, 7 or 14 days were digested overnight at 60 °C in 100 μL papain digestion buffer (pH = 6.4) containing 0.1% papain, 10 mM EDTA-disodium salt (both from Merck), 100 mM sodium acetate, and 5 mM L-Cysteine·HCL (both from Sigma-Aldrich). Following digestion, samples were diluted 1:8 in water to measure the GAG content or samples were diluted 1:6 in Tris-EDTA (TE) buffer to measure DNA content. The GAG content was measured by adding 200 μL dimethylmethylene blue (DMB) solution (0.05 mM DMB, 41 mM NaCl, 45 mM glycin and pH = 3.0) to 40 μL papain-digested sample (pre-diluted 1:8 in water) and absorbance was measured at 590 nm using an iMark Reader (Bio-Rad). To determine the DNA content, PicoGreen® stock solution (ThermoFisher Scientific) was diluted 1:200 in TE buffer (10 mM Tris-HCl, 1 mM EDTA and pH = 7.5) and 50 μL of this solution was mixed with 50 μL papain-digested sample (pre-diluted 1:6 in TE buffer). After 5 minutes dark incubation at room tem- perature, fluorescence was measured at 485/520 nm (excitation/emission) with a CLARIOstar (BMG Labtech, Offenburg, Germany) using DNA obtained from human HEK293T cells to set the standard curve. Data analysis. Data represent mean + standard deviation of 6 pellets (from two experiments with triplicate pellets per experiment) per condition per time point. Statistical analyses were performed using SPSS version 22 (IBM, Armonk, NY, USA). Normality and variance were verified with the Shapiro-Wilk test and Levene’s Test, respectively. Statistical differences between control and experimental conditions per time point were determined with two-tailed independent T-tests. Level of significance was set at P < 0.05. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 10 www.nature.com/scientificreports/ www.nature.com/scientificreports/ 7. Ionescu, A. M. et al. ATF-2 cooperates with Smad3 to mediate TGF-beta effects on chondrocyte maturation. Experimental cel research 288, 198–207, doi: 10.1016/S0014-4827(03)00181-2 (2003).f 38. Yang, X. et al. TGF-beta/Smad3 signals repress chondrocyte hypertrophic differentiation and are required for maintaining arti cartilage. The Journal of cell biology 153, 35–46, doi: 10.1083/jcb.153.1.35 (2001). gh f gy j ( ) 9. Massague, J. & Chen, Y. G. Controlling TGF-beta signaling. Genes & development 14, 627–644, doi: 10.1101/gad.14.6.627 (2000). 0. Goumans, M. J. et al. Activin receptor-like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Molecula cell 12, 817–828, doi: 10.1016/S1097-2765(03)00386-1 (2003). ( ) ( ) 1. Daly, A. C., Vizan, P. & Hill, C. S. Smad3 protein levels are modulated by Ras activity and during the cell cycle to dictate transforming th f t b t Th J l f bi l i l h i t 285 6489 6497 d i 10 1074/jb M109 043877 (2010) 41. Daly, A. C., Vizan, P. & Hill, C. S. Smad3 protein levels are modulated by Ras activity and during the cell cycle to dictate transforming growth factor-beta responses. The Journal of biological chemistry 285, 6489–6497, doi: 10.1074/jbc.M109.043877 (2010). g ph f g y j 2. Uemura, M. et al. Smad2 and Smad3 play different roles in rat hepatic stellate cell function and alpha-smooth muscle actin organization. Molecular biology of the cell 16, 4214–4224, doi: 10.1091/mbc.E05-02-0149 (2005). 2. Uemura, M. et al. Smad2 and Smad3 play different roles in rat hepatic stellate cell function and alpha smooth muscle actin organization. Molecular biology of the cell 16, 4214–4224, doi: 10.1091/mbc.E05-02-0149 (2005). g gy f 43. Sowa, H., Kaji, H., Yamaguchi, T., Sugimoto, T. & Chihara, K. Smad3 promotes alkaline phosphatase activity and mineraliza osteoblastic MC3T3-E1 cells. J Bone Miner Res 17, 1190–1199, doi: 10.1359/jbmr.2002.17.7.1190 (2002). j 44. Dexheimer, V., Frank, S. & Richter, W. Proliferation as a requirement for in vitro chondrogenes Stem cells and development 21, 2160–2169, doi: 10.1089/scd.2011.0670 (2012). 44. Dexheimer, V., Frank, S. & Richter, W. Proliferation as a requirement for in vitro chondrogenesis of human mesenchymal stem cells. Stem cells and development 21, 2160–2169, doi: 10.1089/scd.2011.0670 (2012). p 45. Hall, B. K. & Miyake, T. All for one and one for all: condensations and the initiation of skeletal development. BioEssays: news and reviews in molecular, cellular and developmental biology 22, 138–147, doi: 10.1002/(Sici)1521-1878(200002)22:2<138::Aid-Bies5> 3.0.Co;2-4 (2000). 46. www.nature.com/scientificreports/ Mutations at a single codon in Mad homology 2 domain of SMAD4 cause Myhre syndrome. Nature genetics 44 85–U118, doi: 10.1038/ng.1016 (2012). 59. Caputo, V. et al. A Restricted Spectrum of Mutations in the SMAD4 Tumor-Suppressor Gene Underlies Myhre Syndrome. American journal of human genetics 90, 161–169, doi: 10.1016/j.ajhg.2011.12.011 (2012).h j f g j j g 60. Lindor, N. M., Gunawardena, S. R. & Thibodeau, S. N. Mutations of SMAD4 account for both LAPS and Myhre syndro American Journal of Medical Genetics Part A 158A, 1520–1521, doi: 10.1002/ajmg.a.35374 (2012). 61. Piccolo, P. et al. SMAD4 mutations causing Myhre syndrome result in disorganization of extracellular matrix improved by losartan. European journal of human genetics: EJHG 22, 988–994, doi: 10.1038/ejhg.2013.283 (2014). 62. Lin, X. et al. PPM1A functions as a Smad phosphatase to terminate TGFbeta signaling. Cell 125, 915–928, doi: 10.10 cell.2006.03.044 (2006). 63. Wrighton, K. H., Lin, X. & Feng, X. H. Phospho-control of TGF-beta superfamily signaling. Cell research 19, 8–20, doi: 10.1 cr.2008.327 (2009). 64. Lin, X., Liang, M. & Feng, X. H. Smurf2 is a ubiquitin E3 ligase mediating proteasome-dependent degradation of Smad2 in transforming growth factor-beta signaling. Journal of Biological Chemistry 275, 36818–36822, doi: 10.1074/jbc.C000580200 (2000). 65. Zhang, Y., Chang, C., Gehling, D. J., Hemmati-Brivanlou, A. & Derynck, R. Regulation of Smad degradation and activity by Smurf2, an E3 ubiquitin ligase. Proc Natl Acad Sci USA 98, 974–979, doi: 10.1073/pnas.98.3.974 (2001). References Genetic Variation in the SMAD3 Gene Is Associated With Hip and Knee Osteoarthritis. Arthritis and rheumatism 62, 2347–2352, doi: 10.1002/art.27530 (2010). 6. Yao, J. Y. et al. Mutation analysis of the Smad3 gene in human osteoarthritis. European Journal of Human Genetics 11, 714–717, doi 10.1038/sj.ejhg.5201034 (2003). Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 11 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Lefebvre, V., Huang, W. D., Harley, V. R., Goodfellow, P. N. & deCrombrugghe, B. SOX9 is a potent activator of the chondrocyte- specific enhancer of the pro alpha 1(II) collagen gene. Molecular and cellular biology 17, 2336–2346, doi: 10.1128/MCB.17.4.2336 (1997).h 47. Hatakeyama, Y., Nguyen, J., Wang, X., Nuckolls, G. H. & Shum, L. Smad signaling in mesenchymal and chondroprogenitor cells. The Journal of bone and joint surgery. American volume 85-A Suppl 3, 13–18 (2003). 48. Alliston, T., Choy, L., Ducy, P., Karsenty, G. & Derynck, R. TGF-beta-induced repression of CBFA1 by Smad3 decreases cbfa1 and osteocalcin expression and inhibits osteoblast differentiation. The EMBO journal 20, 2254–2272, doi: 10.1093/emboj/20.9.2254 (2001). 49. Chen, C. G., Thuillier, D., Chin, E. N. & Alliston, T. Chondrocyte-intrinsic Smad3 represses Runx2-inducible matrix metalloproteinase 13 expression to maintain articular cartilage and prevent osteoarthritis. Arthritis and rheumatism 64, 3278–3289, doi: 10.1002/art.34566 (2012). 0. Ferguson, C. M. et al. Smad2 and 3 mediate transforming growth factor-beta1-induced inhibition of chondrocyte maturation Endocrinology 141, 4728–4735, doi: 10.1210/endo.141.12.7848 (2000).if gy 1. O’Keefe, R. J. et al. Smad3-deficient chondrocytes have enhanced BMP signaling and accelerated differentiation. J Bone Miner Re 21, 4–16, doi: 10.1359/Jbmr.050911 (2006). 52. Blaney Davidson, E. N. et al. Increase in ALK1/ALK5 ratio as a cause for elevated MMP-13 expression in osteoarthritis in humans and mice. J Immunol 182, 7937–7945, doi: 10.4049/jimmunol.0803991 (2009).h j 3. Nakao, A. et al. TGF-beta receptor-mediated signalling through Smad2, Smad3 and Smad4. The EMBO journal 16, 5353–5362, doi 10.1093/emboj/16.17.5353 (1997). j 4. ten Dijke, P. & Hill, C. S. New insights into TGF-beta-Smad signalling. Trends in biochemical sciences 29, 265–273, doi: 10.1016/j tibs.2004.03.008 (2004). ( ) 55. Liu, C. et al. Smads 2 and 3 are differentially activated by transforming growth factor-beta (TGF-beta) in quiescent and activated hepatic stellate cells. Constitutive nuclear localization of Smads in activated cells is TGF-beta-independent. The Journal of biological chemistry 278, 11721–11728, doi: 10.1074/jbc.M207728200 (2003). y j 56. Wrighton, K. H., Lin, X., Yu, P. B. & Feng, X. H. Transforming Growth Factor beta Can Stimulate Smad1 Phosphorylation Independently of Bone Morphogenic Protein Receptors. The Journal of biological chemistry 284, 9755–9763, doi: 10.1074/jbc. M809223200 (2009). 57. Lagna, G., Hata, A., Hemmati-Brivanlou, A. & Massague, J. Partnership between DPC4 and SMAD proteins in TGF-beta signa pathways. Nature 383, 832–836, doi: 10.1038/383832a0 (1996).f 8. Le Goff, C. et al. Acknowledgementsh g The authors would like to thank Prof. Dr. P. ten Dijke (Leiden University Medical Center, the Netherlands) for providing adenoviruses of LacZ, SMAD2, SMAD3 and SMAD4. This work was financially supported by the Dutch Arthritis Foundation (11-1-404). R.N. was funded by the VENI grant from STW (13659) and further supported by Translational Adult Stem Cell Research by ZonMw (116005009). Author Contributions L.M.G.d.K., R.N., G.G.H.v.d.A., E.N.B.D., G.J.V.M.v.O. and P.M.v.d.K conceived the experiments, L.M.G.d.K., G.G.H.v.d.A. and E.L.V. conducted the experiments, L.M.G.d.K., R.N., G.G.H.v.d.A., E.N.B.D., G.J.V.M.v.O. and P.M.v.d.K. analyzed the results. All authors reviewed the manuscript. L.M.G.d.K., R.N., G.G.H.v.d.A., E.N.B.D., G.J.V.M.v.O. and P.M.v.d.K conceived the experiments, L.M.G.d.K., G.G.H.v.d.A. and E.L.V. conducted the experiments, L.M.G.d.K., R.N., G.G.H.v.d.A., E.N.B.D., G.J.V.M.v.O. and P.M.v.d.K. analyzed the results. All authors reviewed the manuscript. Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 12 www.nature.com/scientificreports/ Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 Additional Information Supplementary information accompanies this paper at http://www.nature.com/srepi Competing financial interests: The authors declare no competing financial interests. How to cite this article: de Kroon, L. M. G. et al. SMAD3 and SMAD4 have a more dominant role than SMAD2 in TGFβ-induced chondrogenic differentiation of bone marrow-derived mesenchymal stem cells. Sci. Rep. 7, 43164; doi: 10.1038/srep43164 (2017). ublisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and nstitutional affiliations. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ © The Author(s) 2017 Scientific Reports \| 7:43164 \| DOI: 10.1038/srep43164 13
https://openalex.org/W3212015157	https://hal.univ-brest.fr/hal-03428256/document	English	null	Ultra‐wideband antenna system for in‐band full‐duplex applications	IET microwaves, antennas & propagation	2,021	cc-by	9,512	To cite this version: Hadi Hijazi, Marc Le Roy, Raafat Lababidi, Denis Le Jeune, André Pérennec. Ultra-wideband antenna system for in-band full-duplex applications. IET Microwaves Antennas and Propagation, 2021, pp.1- 13. 10.1049/mia2.12194. hal-03428256 O R I G I N A L R E S E A R C H PA P E R O R I G I N A L R E S E A R C H PA P E R This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. © 2021 The Authors. IET Microwaves, Antennas & Propagation published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology. Ultra‐wideband antenna system for in‐band full‐duplex applications Hadi Hijazi1,2 \| Marc Le Roy2 \| Raafat Lababidi1 \| Denis Le Jeune1 \| Andre Pérennec2 Hadi Hijazi1,2 \| Marc Le Roy2 \| Raafat Lababidi1 \| Denis Le Jeune1 \| Andre Pérennec2 1ENSTA Bretagne, Lab‐STICC, CNRS, Brest, France 2Univ Brest, Lab‐STICC, CNRS, Brest, France wileyonlinelibrary.com/journal/mia2 - 1 HAL Id: hal-03428256 https://hal.univ-brest.fr/hal-03428256v1 Submitted on 15 Nov 2021 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Received: 29 March 2021 - Revised: 14 August 2021 - Accepted: 16 August 2021 DOI: 10.1049/mia2.12194 IET Microwaves, Antennas & Propagation Revised: 14 August 2021 - Accepted: 16 August 2021 DOI: 10.1049/mia2.12194 Correspondence Correspondence Hadi Hijazi, ENSTA Bretagne, Lab‐STICC, CNRS, UMR 6285, F‐29200 Brest, France. Email: hadi.hijazi@ensta-bretagne.org K E Y W O R D S baluns, microstrip transitions, ultra wideband antennas, Vivaldi antennas IET Microw. Antennas Propag. 2021;1–13. Abstract In this study, we present an ultra‐wideband full‐duplex system constituted of a wideband 4‐element Vivaldi array and wideband microstrip‐to‐slotline baluns. The proposed system is characterised by its simplicity, high directivity, and high self‐interference cancellation levels over a wide frequency bandwidth. The system is fabricated using basic printed‐ circuit board (PCB) technology and can provide at least 50 dB of self‐interference cancellation over the bandwidth of operation, 4–40 GHz, with an average gain of 7.8 dBi. The system has a size of 8 cm 8 cm 9.4 cm and can be used as a high data rate link between two distant wireless nodes. To assess the merits of the proposed system and compare it to other published works, a new figure of merit (FoMWFD) dedicated to wide‐band full‐duplex antenna topologies is introduced in this study. Correspondence Hadi Hijazi, ENSTA Bretagne, Lab‐STICC, CNRS, UMR 6285, F‐29200 Brest, France. Email: hadi.hijazi@ensta-bretagne.org 1 \| INTRODUCTION RF Cancellation IF Cancellation BB Cancellation Digital Cancellation Feeding Network C D A A N L DAC PA Rx-M1 Rx-M2 Tx-M1 Tx-M2 Antenna(s) Rx Tx Digital Level Analog Level Antenna Level Targeted Cancellation: > 50 dB Targeted Cancellation: > 30 dB Targeted Cancellation: > 30 dB F I G U R E 1 General block diagram of an in‐band full‐duplex transceiver F I G U R E 1 General block diagram of an in‐band full‐duplex transceiver devices limits the maximum achievable level of cancellation. On the other hand, for bistatic systems, the basic technique is to increase the separation between the Tx and Rx antennas [19], however, this results in an increased system size. So, instead, the beams of the antennas can be directed somewhat in different directions to reduce the overlap between them [19], nonetheless, not all applications can tolerate to transmit and receive in different directions. Alternatively, using orthogonal polarisations for the Tx and Rx antennas [20] can increase the level of cancellation, but this requires the Tx antennas of one node to be aligned with the Rx antennas of the other node. Moreover, near‐field cancellation [17] showed the capability of achieving a decent level of cancellation with a simple system design and compactness. Also, circularly phased arrays [21–23], can provide a similar performance to near‐field cancellation with a quasi‐omnidirectional radiation pattern, however, it requires a more complicated feeding network. In addition to all the above, placing high impedance structures between the antennas were considered in [16], which can provide a high level of cancellation, however, the size of such structures can drastically increase the overall size of the system. Finally, based on the above techniques wide- band full‐duplex arrays were also demonstrated in [24–27]. Note that, one or more techniques can be used in conjunc- tion to achieve the maximum possible level of self‐ interference cancellation. direct path coupling, which can be predicted and measured. So, a copy of the transmitted signal is taken, and direct path losses and phase changes are applied to it, then it gets subtracted from the received signal [3]. At the analogue level, cancellation can take place at one of three stages: at the RF stage [6] before the amplifiers, then at the intermediate frequency (IF) stage [7], and at the base band (BB) stage [8] before the converters. 1 \| INTRODUCTION transmitted by the other node. Nominally, 110 dB of self‐ interference cancellation is required to be able to establish an in‐band full‐duplex communication [3]. Estimates of futuremobile trafficindicate that theglobal number of mobile subscriptions could be 13.8 billion in 2025 and 17.1 billion in 2030 [1]. The increase in the number of mobile users creates a spectrum congestion problem and elevates the pressure on the available mobile infrastructure to keep up with the increased demand for mobile services. These problems create the need for new approaches to find more spectrum resources by improving traditional spectrum sharing techniques and/or har- nessing new ones. Recently, in‐band full‐duplex (IBFD) tech- nology emerged as a promising solution for the spectrum congestion problem [2]. As opposed to the traditional out‐of‐ band full‐duplex (OBFD), which utilises orthogonal resources (frequency or time) to establish a full‐duplex communication, in‐ band full‐duplex allows the simultaneous transmission and reception between two communicating nodes at the same fre- quency and the same time slot. This can only be achieved by reducing the self‐interference signals that are coupled from the transmitter of one node to its own receiver below the noise floor level, allowing the receiver to receive the useful signal Figure 1 depicts a general block diagram of an in‐band full‐ duplex heterodyne transceiver where it can be seen that self‐ interference cancellation takes place at three stages or levels: at the antenna level, at the analogue level, and at the digital level. At the antenna level, the system can be monostatic where a shared antenna(s) is used to transmit and receive, or it can be bi‐static where separate antennas can be used. However, for monostatic systems a duplexing device is required to isolate the transmitted signal from the received signal, such as circulators [4] or hybrid transformers [5]. While, for bistatic systems, the transmit antennas can be placed or fed in a way to create null planes at the positions of the receive antennas which can reduce the coupling between them. On the other hand, both the analogue and digital cancel- lation circuitry benefit from the fact that the transmit and receive antennas are co‐located on the same board or platform, and therefore, the receiver section has knowledge of the originally transmitted signal and of the characteristics of the IET Microw. Antennas Propag. 2021;1–13. 2 - HIJAZI ET AL. 3.1 \| Principle of operation The proposed in‐band full‐duplex system is depicted in Figure 2 where the substrates were omitted for better clarity. The system is composed of four Vivaldi antennas which are placed on the perimeter of a circle of a fixed radius and sequentially rotated by 90° from one another, thus forming two diametrically opposite pairsof antennas, where eachpairis orthogonalto theother. One pair is used to transmit, and the other is used to receive. Then each pair of antennas is fed signals of equal amplitudes and 180° out‐of‐phase by using wideband microstrip‐to‐slotline baluns. To explain the principle of operation of the proposed system, two approaches are adopted [17]: an inter‐port approach that describes the power coupling between the input and output ports of theTx andRx baluns, anda fieldapproachthat describes the electric field behaviour in the far‐field region. Near‐field cancellation has been implemented in several previous works, mainly, with planar antennas operating at the low frequency side of the spectrum (below the K‐band), and over bandwidths not more than one or two octaves [17, 20, 28–30]. But, in this work, we aim to have a system that can operate up to the Ka‐band while covering most of the lower bands (starting from the C‐band). To achieve this goal, and to satisfy the previously mentioned specifications, we propose a system that consists of a 4‐element Vivaldi array and two microstrip‐to‐slotline baluns, as depicted in Figure 2. Although the individual devices of the system are planar, yet the system is three‐dimensional, which creates some vagueness about the operation of the near‐field cancellation technique, taking into consideration that it has only been demonstrated for planar systems. Thus, our work can be considered the first work to implement the near‐field cancellation technique over an extremely wide bandwidth with a three‐dimensional an- tenna system, and our contributions can be summarised as follows: Inter‐Port Approach: in order to develop this approach, we find that it is most suitable to work with power waves [31], where an denotes the power wave incident at port n and bn denotes the power wave reflected from that port, and accordingly, the total powers carried by each wave are equal to janj2 and jbnj2, respectively. The flow of power waves in the system is depicted in Figure 2b assuming ideal conditions where no mismatch exists between the system components. 1 \| INTRODUCTION Note that, it is necessary to suppress the self‐interference signals, at the antenna and analogue levels, by at least a certain amount (typically 60 dB [3]) to prevent saturating the analogue‐to‐ digital converter. The simultaneous transmit‐receive property of in‐band full‐duplex technology can enable the introduction of novel and efficient multiple access techniques [9], physical layer se- curity protocols [10], relaying solutions [11], and can reduce air interface delay [12]. Also, it might benefit radar systems [13], cognitive radios (CR) [14], and multiple‐input multiple output (MIMO) arrays [15]. So, considering their appealing advan- tages, it is highly desirable to implement wideband full‐duplex systems for civil, military, and space applications. Available publications on wideband full‐duplex systems indicate that achieving 50 dB of self‐interference cancellation level over a wide bandwidth is possible at the antenna level [16–18], however, the analogue and digital parts fall far behind with achieving wideband cancellation, mainly, due to circuitry bandwidth limitations and cost. Nevertheless, implementing a wideband full‐duplex antenna system remains very attractive as it can be used as a universal transmit‐receive system suitable for different standards, frequency bands, or bandwidths, and even for software‐defined radios (SDR) or cognitive radios (CR). In light of that, the scope of this work will be mainly focussed on wideband self‐interference cancellation techniques at the antenna level. In the next section, our system specifications, detailed contributions, and the potential applications of the proposed system will be presented. Then, the whole system with all the individual devices will be presented in Section 3, where its principle of operation will be explained. And in Section 4, the final system assembly, matching, self‐interference cancellation, and far‐field performance will be demonstrated both in sim- ulations and in measurements. Also, in this section, the new figure of merit will be introduced, and then the section will be closed with a table that compares this work to previous works in the literature, especially in terms of the newly introduced figure of merit. Several techniques were proposed to attain wideband self‐ interference cancellation at the antenna level. For monostatic systems, some implementations based on wideband circula- tors and hybrid transformers can be found in [4, 5], respec- tively. Nevertheless, the limited isolation of the duplexing - 3 3 HIJAZI ET AL. 2 \| SCOPE OF WORK 2 network, or wireless links between two remote base stations instead of using fibre optic. More broadly, the ‘sensing‐as‐ transmitting’ property, along with all other characteristics of in‐ band full‐duplex, can be valuable for any wideband multi‐ standard communication operation especially for cognitive radios which require continuously scanning a wide range of frequencies while transmitting. In this work, we aim to design an ultra‐wideband in‐band full‐ duplex system with at least 50 dB of self‐interference cancel- lation at the antenna level, that can transmit and receive in the same direction with a fairly high directivity. And as it was evident, from the published works, that bistatic techniques have more wideband potential than monostatic techniques, in particular, near‐field cancellation which can satisfy the required specifications if, at the same time, orthogonal polarisations are used for the transmit and receive antennas. And in addition to that, near‐field cancellation can generate the most compact full‐duplex systems as compared to all other bistatic tech- niques. However, it requires at least four antennas and two 180° out‐of‐phase power dividers (or baluns), and it can be sensitive to the baluns' amplitude and phase imbalances, and, also, to the antennas' placement.i 3.1 \| Principle of operation Now, the output power (Po) at the output of the Rx balun can be expressed in terms of power waves coming from the receive antennas 1 and 3 as follows: Achieving 50 dB of self‐interference cancellation, at the antenna level, over an extremely wide bandwidth (4– 40 GHz). Po ¼ jboj2 ¼ ao 0 þa00 oe jπ2 ¼ b1 þ b3e jπ2 ð1Þ Optimising the microstrip‐to‐slotline transition to operate over the desired bandwidth. The transition will be used to feed the antennas and it constitutes the basic building block of the used baluns. ð1Þ Also b1 and b3 can be expressed in terms of the power waves incident on the transmit antennas 2 and 4 as follows: Also b1 and b3 can be expressed in terms of the power waves incident on the transmit antennas 2 and 4 as follows: Proposing a way to align the antennas using a 3D‐printed support.i b1 ¼ S12a2 þ S14a4 ð2Þ b3 ¼ S32a2 þ S34a4 ð3Þ ð2Þ Introducing a new figure of merit, the wideband full‐duplex systems figure of merit (FoMWFD), to assess the perfor- mance of wideband in‐band full‐duplex systems, and it ac- counts for the achieved self‐interference cancellation level, the gain, the bandwidth, and the size of the system. ð3Þ Similarly, a2 and a4 can be written in terms of the power wave ai incident at the input of the Tx balun:fififi a2 ¼ bi 0 ¼ ffiffiffi 2 p 2 ai ð4Þ a4 ¼ b00 i ¼ ffiffiffi 2 p 2 aie jπ ð5Þ a2 ¼ bi 0 ¼ ffiffiffi 2 p 2 ai ð4Þfififi The proposed system can be considered for diverse ap- plications where capacity, low latency, and secrecy capabilities, offered by in‐band full‐duplex, are required. That may concern, for example, in cellular telecommunication systems, backhaul microwave links between the core network and the radio access ð4Þ a4 ¼ b00 i ¼ ffiffiffi 2 p 2 aie jπ ð5Þ ð5Þ 4 - HIJAZI ET AL. 4 (a) (b) F I G U R E 2 (a) An illustration of the proposed wideband in‐band full‐duplex system with 4‐element Vivaldi array and two microstrip‐to‐slotline baluns, and in the background a transparent view of the Vivaldi array showing the position and orientation of the microstrip feed lines. 3.1 \| Principle of operation So, the individual electric fields can be expressed as follows: Finally, by substituting (6) and (7) in (1) we can obtain the output power (Po) in terms of the input power (Pi): E → 1 ¼ E0ejπ y → ð9Þ E → 2 ¼ E0x → ð10Þ E → 3 ¼ E0y → ð11Þ E → 4 ¼ E0ejπx → ð12Þ E → 1 ¼ E0ejπ y → ð9Þ E → 2 ¼ E0x → ð10Þ E → 3 ¼ E0y → ð11Þ E → 4 ¼ E0ejπx → ð12Þ E → 1 ¼ E0ejπ y → ð9Þ E → 2 ¼ E0x → ð10Þ E → 3 ¼ E0y → ð11Þ E → 4 ¼ E0ejπx → ð12Þ ð9Þ Po ¼ ffiffiffi 2 p 2 ai S12 þ S14ejπ þ ffiffiffi 2 p 2 ai S32 þ S34ejπ ejπ 2 Po ¼ ffiffiffi 2 p 2 jaij2jS12 þ S14ejπ þ S32ejπ þ S34ej2πj2 Po ¼ 1 2PijS12 −S14 −S32 þ S34j2 ð8Þ ð10Þ ð11Þ ð12Þ ð8Þ Now, if the baluns are connected to the antennas, then the Tx and Rx electric fields can be described as follows: Now, if the baluns are connected to the antennas, then the Tx and Rx electric fields can be described as follows: Equation (8) implies that if S12 ¼ S14 and S32 ¼ S34, then the power coupled from the input port of the system to the output port will be equal to zero. This means that if the Rx antennas are placed along the perpendicular bisector of the Tx antennas, then infinite isolation is, theoretically, obtained between them. This conclusion holds only if the antennas are placed and aligned precisely and if the balun operates ideally, that is if the two output signals of the balun are of equal amplitudes and 180° out‐of‐phase. Nonetheless, in practice, there will be slight misplacements of antennas and imbalances in the amplitude and phase of the balun's output signals, thus the isolation between the E → Tx ¼ E → 2 þ E → 4ejπ ¼ 2E0x → ð13Þ E → Rx ¼ E → 1 þ E → 3ejπ ¼ 2E0ejπ y →¼ −2E0 y → ð14Þ ð13Þ ð14Þ Equation (13) implies that the electric fields of the two transmit antennas will combine constructively in the far‐field region and Equation (14) implies the same observation for the receive antennas. 3.1 \| Principle of operation (b) A schematic of the system showing antenna placement and numbering, flow of power waves, and the original electric fields radiated by each single antenna before connecting them to the baluns (a) ) (b) (a) F I G U R E 2 (a) An illustration of the proposed wideband in‐band full‐duplex system with 4‐element Vivaldi array and two microstrip‐to‐slotline baluns, and in the background a transparent view of the Vivaldi array showing the position and orientation of the microstrip feed lines. (b) A schematic of the system showing antenna placement and numbering, flow of power waves, and the original electric fields radiated by each single antenna before connecting them to the baluns By substituting (4) and (5) in (2) and (3) we obtain: By substituting (4) and (5) in (2) and (3) we obtain: input and the output ports of the system is expected to decrease. b1 ¼ ffiffiffi 2 p 2 S12ai þ ffiffiffi 2 p 2 S14aiejπ ¼ ffiffiffi 2 p 2 ai S12 þ S14ejπ ð6Þ b3 ¼ ffiffiffi 2 p 2 S32ai þ ffiffiffi 2 p 2 S34aiejπ ¼ ffiffiffi 2 p 2 ai S32 þ S34ejπ ð7Þ Field Approach: Assuming that the antennas are not con- nected to the baluns, then each individual antenna element radiates a linearly polarised electric field of magnitude E0 which is oriented in the opposite direction of the microstrip stub orientation, Figure 2b. 3.1 \| Principle of operation However, this conclusion only holds if each two opposite antennas are symmetrically rotated with respect to the centre of symmetry of the system, that is, their 5 HIJAZI ET AL. feeding microstrip lines are oriented in opposite directions (see the background image in Figure 2a), and if the antennas are fed signals that are 180° out‐of‐phase. The antenna rotation (or feeding line opposite orientation) condition is especially critical, otherwise, electric fields will combine destructively in the far‐field, which creates a far‐field null, and thus the system will not be able to transmit or receive in the direction of this null. feeding microstrip lines are oriented in opposite directions (see the background image in Figure 2a), and if the antennas are fed signals that are 180° out‐of‐phase. The antenna rotation (or feeding line opposite orientation) condition is especially critical, otherwise, electric fields will combine destructively in the far‐field, which creates a far‐field null, and thus the system will not be able to transmit or receive in the direction of this null. size of the system and the orientation of the radiation pattern. In addition to that, significant grating lobes can be observed in the far‐field radiation pattern if the separation between the opposite antennas is greater than a half‐ wavelength at a certain frequency. 3.2 \| Wideband balun design Note that if the system is not symmetric, that is, if the number of transmit and receive antennas is not the same or if they are fed differently, differences between the transmit and receive far‐field radiation patterns will emerge, thus it is extremely favourable to maintain high symmetry in the system. The antennas can be tangent or normal to the circle or could form any angle with the circle in general and can still achieve the same level of self‐ interference cancellation, however, this affects the total The initial design of the wideband balun used here was pro- posed by Rhode et al. [32] and it consists of three microstrip‐ to‐slotline transitions terminated with radial stubs as in Figure 3. The balun operates as follows: the input power is first injected in the middle microstrip line and is then coupled to the slotline at its centre. After that, in the slotline, the power will be divided equally between the left and right paths, and at the end of each path, the power will be coupled again to the (a) (b) (c) (d) F I G U R E 3 (a) Illustrative drawing of the microstrip‐to‐slotline balun with (b) the parameters of the design. (c) Simulated and measured input matching and insertion losses, and (d) amplitude imbalances (AI) and phase imbalances (PI) between the output ports. {L = 40 mm, W = 40 mm, Wm = 0.4 mm, Rm = 1.73 mm, θm = 165°, Ws = 0.1 mm, Ls = 8 mm, Rs = 1.93 mm, θs = 150°} (a) (b) (b) (a) (c) (d) (d) (c) F I G U R E 3 (a) Illustrative drawing of the microstrip‐to‐slotline balun with (b) the parameters of the design. (c) Simulated and measured input matching and insertion losses, and (d) amplitude imbalances (AI) and phase imbalances (PI) between the output ports. {L = 40 mm, W = 40 mm, Wm = 0.4 mm, Rm = 1.73 mm, θm = 165°, Ws = 0.1 mm, Ls = 8 mm, Rs = 1.93 mm, θs = 150°} 6 - 6 HIJAZI ET AL. substrate. This may come from a slight degradation of the substrate characteristics during the fabrication process which led to an increase in the value of its loss tangent (tanδ). corresponding output microstrip line. 3.2 \| Wideband balun design Since the two output microstrip lines are oriented in opposite directions, the two output signals will be inverted with respect to one another over a wide bandwidth. The used balun was designed and optimised, in CST Microwave Studio, based on the parametric analysis performed by us in [33], after that, two baluns were fabricated on a 203.2 μm thick RO4003C substrate (εr = 3.55), and they occupy an area of 40 mm 40 mm. Moreover Figure 3d shows that the amplitude and phase imbalances between the output ports of the balun are less than 1 dB and 7° respectively over the 4–40 GHz band, where more fluctuations in the measured results are observed as compared to the simulated results, which might also be attrib- uted to fabrication imperfections. In addition to that, poor levels of matching and isolation between output ports were observed (but not plotted here), which are considered normal as a result of the fact that a 3‐port device cannot be lossless, reciprocal, and matched at all ports at the same time. However, simulations proved that low or high levels of isolation between output ports of the balun do not affect the level of self‐interference cancel- lation obtained in the in‐band full‐duplex system. The measured S‐parameters in Figure 3c show that the balun is matched from 4 to 50 GHz, where a relatively good agreement with simulated results is observed below 30 GHz, while some unexpected jumps in the matching response start to appear beyond 30 GHz. This might probably come from the limitations of the high frequency cables and/or connectors used to conduct the measurements, nevertheless, none of these jumps go higher than −10 dB. Also, the measured insertion loss attains a value of −5 dB at 4 GHz and reaches a value of −11 dB at 40 GHz, however, 3 dB of the insertion loss come from the power division taking place and are not actually lost in the balun. The measured insertion loss drastically increases with frequency as compared to the simulated one, this means that in practice more power is being lost or absorbed by the 3.3 The Vivaldi antenna, Figure 4, is a uniplanar exponentially tapered slot antenna which is characterised by its wide F I G U R E 4 Illustrative drawing of the Vivaldi antenna {L = 70 mm, W = 40 mm, Lt ¼ 10 mm, Lv = 60 mm, Wv = 20 mm, Wm = 0.4 mm, Rm = 1.87 mm, θm = 80°, Ws = 0.1 mm, Rs = 1.85 mm, θs = 170°} F I G U R E 4 Illustrative drawing of the Vivaldi antenna {L = 70 mm W = 40 mm L ¼ F I G U R E 4 Illustrative drawing of the Vivaldi antenna {L = 70 mm, W = 40 mm, Lt ¼ 10 mm, Lv = 60 mm, Wv = 20 mm, Wm = 0.4 mm, Rm = 1.87 mm, θm = 80°, Ws = 0.1 mm, Rs = 1.85 mm, θs = 170°} (a) (b) F I G U R E 5 Simulated and measured (a) matching and (b) broadside maximum realised gain of the designed Vivaldi antenna - 7 - 7 HIJAZI ET AL. (a) (a) bandwidth, simplicity of its design, its low cost, and its highly directive far‐field radiation pattern. The antenna is formed of a microstrip‐to‐slotline transition acting as a feeding mech- anism, and an exponentially tapered slotline acting as a radiating element. The microstrip‐to‐slotline transition was designed, optimised, and used, previously, in the balun design with a matching bandwidth from below 4 GHz to beyond 40 GHz. So the same transition will be used to feed the Vivaldi antenna, however, note that, a microstrip radial stub with half the size of that used in the balun is adopted here. 3.3 The size reduction of the microstrip stub reduces the overlap with the slotline stub and, consequently, reduces the power leaked by the transition [33], which means that more power will be fed to the antenna, and this will enhance its efficiency.i (b) (b) On the other hand, the exponential taper can be defined by the following set of equations: y ¼ Aeαx þ B ð15Þ A ¼ 0:5ðW v −W sÞ eαL −eαLt ð16Þ B ¼ 0:5 W seαL −W veαLt eαL −eαLt ð17Þ y ¼ Aeαx þ B ð15Þ A ¼ 0:5ðW v −W sÞ eαL −eαLt ð16Þ B ¼ 0:5 W seαL −W veαLt eαL −eαLt ð17Þ ð17Þ where α is the growth rate of the exponential taper which can be tuned to obtain different antenna characteristics (band- width, beamwidth, and gain). In this design, α was set to 0.05 for maximum matching bandwidth. Also, in order for the radiating element to be matched in the same bandwidth as that of the transition, the length of the exponential taper (Lv) was set to approximately one wavelength at 4 GHz, and the aper- ture width (W v) was set to approximately a half‐wavelength at the central frequency (22 GHz). F I G U R E 5 Simulated and measured (a) matching and (b) broadside maximum realised gain of the designed Vivaldi antenna 4 \| SYSTEM ASSEMBLY AND PERFORMANCE q y ( ) The antennas were also fabricated on the same RO4003C substrate (εr = 3.55 and h = 203.2 μm) where the total size of each antenna element is 70 mm 40 mm. Figure 5a confirms that the antenna is matched, as expected, from 4 to 47 GHz. And Figure 5b shows that the maximum realised gain of the antenna is better than 5 dBi over the entire matching bandwidth (4–47 GHz) and better than 10 dBi between 7 and 40 GHz in measurements. The Vivaldi gain measurements were performed using three reference antennas to cover the entire matching bandwidth: the first antenna operates in the 2–18 GHz band, the second antenna operates in the 18–33 GHz band, and the third antenna covers the 33– 50 GHz band. The three measured gains are depicted in Figure 5b in three different colours: red, blue, and green respectively, while the simulated gain is plotted in black. It is observed that the measured gain is slightly shifted towards the lower frequency side as compared to the simulated gain, which results from fabrication imperfections, however, the shape of the measured curve is highly similar to the simulated curve. 4.1 \| System assembly To achieve self‐interference cancellation and ensure proper system performance, it is of great importance to preserve a high symmetry when assembling the system, that is, the antennas should be placed exactly at the same distance away from the centre of symmetry. While it is easy to place the antennas precisely and symmetrically in a simulation envi- ronment, yet, in practice it cannot be guaranteed. Moreover, it was noticed that the fabricated antennas tend to bend naturally due to the low thickness of the used substrate (203.2 μm), which also contributes to the asymmetry of the system. Also, the thickness of the substrate makes it very fragile and prone to fracturing, mainly, due to the tension exerted by the feeding cables and connectors. To resolve those mechanical issues, a 3D‐printed support is designed and fabricated to ensure precise antenna placement, to reduce antenna bending and to hold it fully erect, and to absorb mechanical tension from the feeding cables and connectors. 8 8 HIJAZI ET AL. design in Figure 6b. Moreover, in simulation, the phase‐ paired cables were modelled as straight pieces of coaxial lines (25 cm long) formed of aluminium inner and outer conductors and PTFE dielectric stuffing. In addition to that, the 2.92 mm connectors were not considered in simulations. And finally, note that the baluns that were considered in simulations are the simulated microstrip‐to‐slotline baluns and not some ideal baluns, which means that the signals entering the antennas are not ideal, that is they are not perfectly of equal amplitudes and 180° out‐of‐phase, but rather they are subject to the phase and amplitude imbalances of the simulated baluns. Figure 6b depicts the 3D‐printed support which was made of Polyvinyl‐Chloride (PVC, εr ≈3), and the different parts of the support were fixed together using PTFE screws (εr ≈2:1). The support is designed such that it only grasps the edges of the antennas and is kept sufficiently below the top of the antenna. This ensures that the support will not affect the antennas' matching or performance. But before installing the antennas inside the 3D‐printed support, Southwest 2.92 mm connectors were mounted on all devices, and four Keysight N5448B phase‐paired cables (5 ps skew) were used to connect the antennas to the baluns. System matching Figure 7a shows that the assembled in‐band full‐duplex system has a return loss better than 10 dB over an ultra‐ wide frequency band (3.5–49.35 GHz). The measured matching seems to be slightly better than the simulated one around some frequencies, and this can be attributed to the higher insertion losses of the fabricated baluns, and also to the power losses resulting from the phase‐paired cables and the 2.92 mm connectors. While these losses reflect positively on the system's matching, yet they will reflect negatively on the measured gain of the system, which is depicted in Figure 7c. 4.1 \| System assembly However, note here that the fabricated baluns were modified slightly to have their output ports at the top (as shown in Figure 6a), which makes it easier to connect them directly to the antennas as shown in Figure 6c and 6d. Finally, the measurements were carried out by using Rhode&Schwarz ZVA67 vector network analyser. 4.2 “SAB Chamber” is the measured coupl port of the Tx balun (Port A) and the balun (Port B) inside the anechoic ch “SAB Room” is the measured couplin port of the Tx balun (Port A) and the balun (Port B) in a normal room that h insulation and that contains many o instruments, etc.) which might cause The purpose of this measurement is to the one obtained in the anechoic cham reflections from the surrounding en effect on the performance of the syst Now, the measured level of self‐int for the whole system in the anechoic ch at 4 GHz and reaches about 70 dB a simulation it starts with 80 dB at 4 G maintain approximately the same value un The lower level of cancellation in measu as a result of fabrication imperfections, a in the 3D‐printed support, and the high phases and amplitudes of the baluns. Al system cancellation, which was obtain chamber, to the measured cancellatio orthogonal antennas 1 and 2, which is 3 50.5 at 40 GHz, we can notice that technique, with differential feeding, can extra 20 dB of cancellation, on averag cancellation obtained from cross‐pola comparing the system cancellation in the the system cancellation in a normal room the difference between the two is almo dicates that the reflections from the surr are less significant than the residual therefore, have a negligible effect on th mance of the proposed system. (a) (b) (c) F I G U R E 7 The proposed in‐band full‐duplex system (a) matching, (b) inter‐port coupling, and (c) broadside maximum realised gain (a) (a) “S21” is the coupling (or cross‐polarisation) between the two orthogonal antennas 1 and 2 in Figure 2b if they are excited individually without a balun. And the aim of this measure- ment is to see how much cancellation is initially obtained from the fact that the transmit and receive antennas are orthogonally polarised. “SAB Sim:” is the simulated coupling between the input port of the Tx balun (Port A) and the output port of the Rx balun (Port B). “SAB Chamber” is the measured coupling between the input port of the Tx balun (Port A) and the output port of the Rx balun (Port B) inside the anechoic chamber. .2.2 \| Self‐interference cancellation Figure 7b depicts the inter‐port coupling (or the insertion loss) between the different ports of the system. Note that self‐ interference cancellation is equal to the negative of the inter‐ port coupling (in dB). In fact, inter‐port coupling and self‐ interference cancellation express similar meanings in opposite senses. Inter‐port coupling tells how much of the power incident at the input port is transmitted to the output port, while self‐ interference cancellation tells by how much the power trans- mitted to the output port is weaker than the power incident at the input port, or inversely, it tells byhow muchthe powerincident at the input port is stronger than the power transmitted to the output port. The different plots in Figure 7b are explained below: 4.2.3 \| 4.2 (b) “SAB Room” is the measured coupling between the input port of the Tx balun (Port A) and the output port of the Rx balun (Port B) in a normal room that has no electromagnetic insulation and that contains many objects (chairs, tables, instruments, etc.) which might cause multipath reflections. The purpose of this measurement is to compare its result to the one obtained in the anechoic chamber, and to see if the reflections from the surrounding environment have any effect on the performance of the system. (b) Now, the measured level of self‐interference cancellation for the whole system in the anechoic chamber is about 50 dB at 4 GHz and reaches about 70 dB at 40 GHz, while in simulation it starts with 80 dB at 4 GHz and continues to maintain approximately the same value until it reaches 40 GHz. The lower level of cancellation in measurements was expected as a result of fabrication imperfections, antenna misplacement in the 3D‐printed support, and the higher imbalances in the phases and amplitudes of the baluns. Also, if we compare the system cancellation, which was obtained in the anechoic chamber, to the measured cancellation between the two orthogonal antennas 1 and 2, which is 30.5 dB at 4 GHz and 50.5 at 40 GHz, we can notice that the used cancellation technique, with differential feeding, can provide, at least, an extra 20 dB of cancellation, on average, on the top of the cancellation obtained from cross‐polarisation. Finally, by comparing the system cancellation in the anechoic chamber to the system cancellation in a normal room, we can notice that the difference between the two is almost negligible. This in- dicates that the reflections from the surrounding environment are less significant than the residual self‐interference, and therefore, have a negligible effect on the cancellation perfor- mance of the proposed system. (c) (c) F I G U R E 7 The proposed in‐band full‐duplex system (a) matching, (b) inter‐port coupling, and (c) broadside maximum realised gain 4.2 The simulated and measured characteristics of the system are depicted in Figure 7. Note that simulations were done taking into account the PTFE screws and the 3D‐printed support, which was modelled as a lossless dielectric material having εr ≈3. Also, simulations take into account the extra separa- tion distance between the antennas as per the fabricated F I G U R E 6 Pictures of (a) the fabricated balun, (b) antenna assembly inside the 3D‐printed support, (c) whole system assembly and (d) measurements inside the anechoic chamber F I G U R E 6 Pictures of (a) the fabricated balun, (b) antenna assembly inside the 3D‐printed support, (c) whole system assembly and (d) measurements inside the anechoic chamber F I G U R E 6 Pictures of (a) the fabricated balun, (b) antenna assembly inside the 3D‐printed support, (c) whole system assembly and (d) measurements inside the anechoic chamber F I G U R E 6 Pictures of (a) the fabricated balun, (b) antenna assembly inside the 3D‐printed support, (c) whole system ass the anechoic chamber the fabricated balun, (b) antenna assembly inside the 3D‐printed support, (c) whole system assembly and (d) measurements insid (a) (b) (c) F I G U R E 7 The proposed in‐band full‐duplex system (a) matching, (b) inter‐port coupling, and (c) broadside maximum realised gain HIJAZI ET AL. 9 HIJAZI ET AL. 4.2.2 \| Self‐interference cancellation Figure 7b depicts the inter‐port coupling (or the insertion loss) between the different ports of the system. Note that self‐ “S21” is the coupling (or cross‐polarisa orthogonal antennas 1 and 2 in Figure individually without a balun. And the ment is to see how much cancellatio from the fact that the transmit and orthogonally polarised. “SAB Sim:” is the simulated coupling b of the Tx balun (Port A) and the ou balun (Port B). \| Farfield performance The gain of the system is depicted in Figure 7c where it can be seen that there are some differences between the simulated and measured results. In fact, the simulated gain starts with a value of 4.5 dBi at 4 GHz and then starts to increase with frequency. It reaches a value of 10 dBi around 10 GHz and continues to maintain almost the same value throughout the rest of the bandwidth. Finally, it starts to drop again around 40 GHz. On the other hand, the measured gain starts with a value of 5 dBi 10 10 HIJAZI ET AL. at 4 GHz and then it continues to increase with frequency and reaches its peak value (about 10 dBi) around 10 GHz. From that point, the measured gain starts to drop with frequency until it hits 5 dBi at 40 GHz. The drop in the measured gain can be attributed to the higher insertion losses of the fabricated balun and also to the power losses resulting from the phase‐ paired cables and the 2.92 mm connectors, which were not considered in simulations. However the measured gain remains higher or equal to 5 dBi in the bandwidth of matching (4– 40 GHz), and the average gain in this bandwidth is 7.8 2.6 dBi. the system, though it cannot be avoided unless the separation distance between the antennas is reduced. However, for the presented Vivaldi array, this cannot be feasible because the separation distance is conditioned by the width of the single antenna element, which is usually equal or greater than a wavelength at the central frequency. This means, for such wideband system, whatever the antenna width is, there will always be a portion of the bandwidth (the higher frequency range) where the grating lobes are present. 4.3 In addition to that, Figure 8a, 8b, 8c and 8d depict some sample radiation pattern plots at 5, 15, 25, and 35 GHz, respectively, where good agreement between simulated and measured results is observed. The radiation pattern plots reveal that a directive main lobe is present in the yz‐plane, and that at least two grating lobes are present alongside the main lobe in the xz‐plane. And it seems that the number of grating lobes increases as we go higher in frequency. In fact, at least two grating lobes start to appear in the radiation pattern when the separation distance between the two opposite antennas is higher than a half‐wavelength at the frequency of measure- ment, and their number increases when the separation be- comes higher than a multiple of half‐wavelength. Similar results were observed in other publications on near‐field cancellation [17], and in publications on Vivaldi arrays [34]. This phenomenon can be considered as a major drawback of To assess the merits of the proposed system a new figure of merit, wideband full‐duplex figure of merit (FoMWFD), is introduced, and it accounts for the achieved self‐interference cancellation level, the gain, the bandwidth, and the normal- ised size of the system. Firstly, both the values of the self‐ interference cancellation and the gain of the system should be linear and not in dB, however, note that the dB value of the self‐interference cancellation should be positive before con- verting it to a linear scale. Secondly, concerning the size of the system, some systems have a two‐dimensional geometry, such as systems fabricated on printed‐circuit boards (PCBs) where the thickness of the substrate is almost negligible, while other systems possess a three‐dimensional geometry, which is the case for our system, also, some systems have their feeding (a) (b) (c) (d) F I G U R E 8 Radiation pattern plots at (a) 5 GHz, (b) 15 GHz, (c) 25 GHz, and (d) 35 GHz, respectively (b) (a) (a) (b) (d) (c) (c) (d) F I G U R E 8 Radiation pattern plots at (a) 5 GHz, (b) 15 GHz, (c) 25 GHz, and (d) 35 GHz, respectively 11 11 HIJAZI ET AL. network directly integrated on the same board or platform while other systems have their feeding networks built on separate boards and connected with cables. 4.3 These differences make it difficult to find a common and fair way to compare sizes of different systems, so here we propose a method which is based on three points: (a) we let the system be inscribed in a sphere of radius R, this radius will be used in the figure of merit to represent the size of the system, (b) if the feeding network is integrated with the antennas on the same board then it will be accounted for in the size calculation, and, otherwise, it will be disregarded, (c) after calculating the radius of the sphere it will be normalised by the wavelength at the centre frequency. Based on all the above, the figure of merit is expressed as follows: polarisations (singular or dual), number of antennas used, an- tenna efficiency and level of side lobes, and the performance of the feeding network. Finally, Table 1 compares this work to other in‐band full‐duplex antenna topologies previously published in the literature, especially in terms of bandwidth, level of self‐ interference cancellation, system gain and size, and the newly introduced figure of merit where it can be noticed that our proposed system achieves the highest score. Also, up to our knowledge, it works over the widest bandwidth (a decade bandwidth) and reaches the highest upper frequency (40 GHz) which is conditioned by a minimum gain level of 5 dBi, however, otherwise, the highest matching frequency can go up to 49 GHz. 5 FoMWFD ¼ log 10 SIC Gain FBW R=λc ð18Þ FBW ¼ fu −fl fc ð19Þ R ¼ 1 2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi l2 þ w2 þ h2 q ð20Þ FoMWFD ¼ log 10 SIC Gain FBW R=λc ð18Þ FBW ¼ fu −fl fc ð19Þ R ¼ 1 2 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi l2 þ w2 þ h2 q ð20Þ ð18Þ In this study, an ultra‐wideband full‐duplex system consisting of four Vivaldi antennas and two microstrip‐to‐slotline baluns is presented. The prototype was built on RO4003C substrate (εr ¼ 3:55, tan δ = 0.0027, h = 203.2 μm). The proposed system can achieve 50 dB of self‐interference cancella- tion at the antenna level and a gain better than 5 dBi over the 4–40 GHz frequency range with a highly directive radiation pattern, which proves that in‐band full‐duplex can be extended to ultra‐wideband operation. The proposed system is a general‐ purpose tool that can be used in various applications especially those that require high data‐rate links, or reconfigurability over a wide bandwidth. The performance of the proposed system can be further enhanced by considering the following perspectives: ð19Þ ð20Þ SIC: average self‐interference cancellation level. SIC: average self‐interference cancellation level. Gain: average system gain. FBW: fractional bandwidth. fu: upper frequency. fl: lower frequency. fl: lower frequency. fc: central frequency. fc: central frequency. ▪Due to the relatively high loss tangent of the used substrate and due to fabrication imperfections, the fabricated baluns suffer from high insertion losses, especially at high fre- quencies, which decreases the total gain of the system. To resolve these issues the baluns and antennas can be designed and fabricated on a different substrate with lower tangent loss, RT/Duroid 5880 (εr ¼ 2:2, tan δ = 0.0009) for example. Our primary simulations confirm that the current matching bandwidth (4–40 GHz) can still be maintained using the latter substrate with a thickness of 0.254 mm. ▪Due to the relatively high loss tangent of the used substrate and due to fabrication imperfections, the fabricated baluns suffer from high insertion losses, especially at high fre- quencies, which decreases the total gain of the system. To resolve these issues the baluns and antennas can be designed and fabricated on a different substrate with lower tangent loss, RT/Duroid 5880 (εr ¼ 2:2, tan δ = 0.0009) for example. 5 Our primary simulations confirm that the current matching bandwidth (4–40 GHz) can still be maintained using the latter substrate with a thickness of 0.254 mm. λc: wavelength at the central frequency. R: radius of the sphere circumscribing the full‐duplex system. l: length of the full‐duplex system. l: length of the full‐duplex system. w: width of the full‐duplex system. h: height of the full‐duplex system. The proposed figure of merit is an initial attempt to combine different parameters of in‐band full‐duplex antenna configura- tions or topologies in an expressive way to evaluate the perfor- mances of the proposed systems and compare them to each other. However, depending on the targeted application, it might be modified to incorporate other metrics related to the field of application, such as the type of polarisation, the number of ▪Also, integrating the baluns and antennas on the same board is currently under consideration, which allows us to get rid of the phase‐paired cables and the 2.92 mm connectors, and that reduces the losses in the system. T A B L E 1 A table comparing several wideband antenna systems for in‐band full‐duplex applications Reference Frequency range (GHz) FBW SIC (dB) Gain (dBi) l w h (cm £ cm £ cm) R (cm) FoMWFD [4]a 4–8 0.67 45 25 40 £ 40 £ 25.3 30.984 4.531 [22]a 0.5–2 1.2 45 5 20 £ 20 £ 12.5 15.462 4.77 [23]a 0.8–1.7 0.72 40 4 60 £ 60 £ 30 45 3.784 [16]a 6–19 1.04 60 12 38 £ 13 £ 19 22.215 4.651 This work 4–40 1.64 64 7.8 8 £ 8 £ 9.4 7.55 5.072 aSome parameters' values were extracted from figures or graphs. T A B L E 1 A table comparing several wideband antenna systems for in‐band full‐duplex applications 12 12 HIJAZI ET AL. ▪Moreover, the low substrate thickness is necessary for the wideband performance of the fabricated devices, however, this comes at the expense of the fragility of these devices. To encounter this problem, we designed a 3D‐printed support to hold the antennas and absorb any external mechanical tensions. However, as an alternative solution, one can in- crease the thickness of the substrate while sacrificing a portion of the obtained bandwidth. 6. Venkatakrishnan, S.B., Alwan, E.A., Volakis, J.L.: Wideband RF self‐ interference cancellation circuit for phased array simultaneous transmit and receive systems. IEEE Access. 6, 3425–3432 (2018) 7. Hadi Hijazi https://orcid.org/0000-0001-7337-452X Hadi Hijazi https://orcid.org/0000-0001-7337-452X Hadi Hijazi https://orcid.org/0000-0001-7337-452X 23. Etellisi, E.A., Elmansouri, M.A., Filipovic, D.: Broadband full‐duplex monostatic circular‐antenna arrays: circular arrays reaching simulta- neous transmit and receive operation. IEEE Antenn. Propag. Mag. 60(5), 62–77 (2018) 5 Van den Broek, D.J., Klumperink, E.A., Nauta, B.: An in‐band full‐ duplex radio receiver with a passive vector modulator downmixer for self‐interference cancellation. IEEE J. Solid. State Circ. 50(12), 3003–3014 (2015) 8. Liu, Y., et al.: Digitally assisted analog interference cancellation for in‐ band full‐duplex radios. IEEE Commun. Lett. 21(5), 1079–1082 (2017) 9. Amin, M., Hossain, M., Atiquzzaman, M.: In‐band full duplex wireless LANs: medium access control protocols, design issues and their chal- lenges. J. Inf. 11(4), 216 (2020) ▪In addition to the above, other than in‐band full‐duplex applications, the whole system can be used as one dual‐ polarised antenna array, and the way it is implemented, and fed can greatly enhance the level of cross‐polarisation over a very wide bandwidth. 10. Van Nguyen, B., Jung, H., Kim, K.: Physical layer security schemes for full‐ duplex cooperative systems: state of the art and beyond. IEEE Commun. Mag. 56(11), 131–137 (2018) 11. Liu, G., et al.: In‐band full‐duplex relaying: a survey, research issues and challenges. Commun. Surv. Tutorials, IEEE. 17(2), 500–524 (2015) ▪Finally, the whole system can be used to generate circularly polarised waves by modifying the feeding network. The circular polarisation can be obtained by feeding the four antennas signals having equal amplitudes and having phases of 0°, 90°, 180°, and 270°, consecutively. 12. Debaillie, B., et al.: In‐band full‐duplex transceiver technology for 5G mobile networks. In: ESSCIRC Conference 2015‐41st European Solid‐ State Circuits Conference (ESSCIRC), pp. 84–87. IEEE (2015) 13. Stove, A.G.: Linear FMCW radar techniques. IEE Proceedings F (Radar and Signal Processing), vol. 139, pp. 343–350. IET (1992) 14. Darabkh, K.A., et al.: A–Z overview of the in‐band full‐duplex cognitive radio networks. Comput. Commun. 145, 66–95 (2019) PERMISSION TO REPRODUCE MATERIALS FROM OTHER SOURCES None. 19. Everett, E., Sahai, A., Sabharwal, A.: Passive self‐interference suppres- sion for full‐duplex infrastructure nodes. IEEE Trans. Wireless Com- mun. 13(2), 680–694 (2014) 20. Ha, J., et al.: Monostatic Co‐ polarized full‐duplex antenna with left‐or right‐hand circular polarization. IEEE Trans. Antenn. Propag. 65(10), 5103–5111 (2017) DATA AVAILABILITY STATEMENTi 21. Moulder, W.F., Perry, B.T., Herd, J.S.: Wideband Antenna array for simultaneous transmit and receive (STAR) applications. In: 2014 IEEE Antennas and Propagation Society International Symposium (APSURSI), pp. 243–244. IEEE (2014) The data that support the findings of this study are available from the corresponding author upon reasonable request. ORCID 22. Lian, R., et al.: A high‐isolation, ultra‐wideband simultaneous transmit and receive antenna with monopole‐like radiation characteristics. IEEE Trans. Antenn. Propag. 66(2), 1002–1007 (2017) CONFLICT OF INTEREST 18. Prasannakumar, P.V., et al.: Wideband quasi‐monostatic simultaneous transmit and receive reflector antenna. IEEE Trans. Antenn. Propag. 68(4), 2630–2637 (2019) The authors declare that there is no conflict of interest. ACKNOWLEDGEMENTS This project has been supported by The French Directorate General of Armaments (DGA), the European Regional Development Fund (ERDF) of the European Union, the Brittany Region (France), the Departmental Council of Fini- stère and Brest Métropole as part of the Cyber‐SSI project within the framework of the Brittany 2015–2020 State‐Region Contract (CPER). 15. Du, X., et al.: MIMO broadcast channel with continuous feedback using full‐duplex radios. In: 2014 48th Asilomar Conference on Signals, Sys- tems and Computers, pp. 1701–1705. IEEE (2014) 16. Prasannakumar, P.V., Elmansouri, M.A., Filipovic, D.S.: Wideband decoupling techniques for dual‐polarized Bi‐static simultaneous transmit and receive antenna subsystem. IEEE Trans. Antenn. Propag. 65(10), 4991–5001 (2017) 17. Wu, J., Li, M., Behdad, N.: A wideband, unidirectional circularly polarized antenna for full‐duplex applications. IEEE Trans. Antenn. Propag. 66(3), 1559–1563 (2018) REFERENCES 1. IT Union: IMT traffic estimates for the years 2020 to 2030. Report ITU (2015) 1. IT Union: IMT traffic estimates for the years 2020 to 2030. Report ITU (2015) 2. Sabharwal, A., et al.: In‐band full‐duplex wireless: challenges and op- portunities. IEEE J. Sel. Area Commun. 32(9), 1637–1652 (2014) 24. Alwan, E.A., Hovsepian, A., Volakis, J.L.: Ultra‐wideband dual polari- zation arrays with collocated elements for high isolation simultaneous transmit and receive systems. In: 2016 IEEE International Symposium on Phased Array Systems and Technology (PAST), pp. 1–3. IEEE (2016) 3. Bharadia, D., McMilin, E., Katti, S.: Full duplex radios. In: Proceedings of the ACM SIGCOMM 2013 Conference on SIGCOMM, pp. 375–386 (2013) 3. Bharadia, D., McMilin, E., Katti, S.: Full duplex radios. In: Proceedings of the ACM SIGCOMM 2013 Conference on SIGCOMM, pp. 375–386 (2013) 25. Hovsepian, A., Alwan, E.A., Volakis, J.L.: Wideband scanning array of spiral antennas for simultaneous transmit and receive (STAR). In: 2017 IEEE International Symposium on Antennas and Propagation and USNC/URSI National Radio Science Meeting, pp. 487–488. IEEE (2017) 4. Prasannakumar, P.V., Elmansouri, M.A., Filipovic, D.S.: Broadband monostatic simultaneous transmit and receive reflector antenna system. In: 2017 IEEE International Symposium on Antennas and Propagation and USNC/URSI National Radio Science Meeting, pp. 553–554. IEEE (2017) 4. Prasannakumar, P.V., Elmansouri, M.A., Filipovic, D.S.: Broadband monostatic simultaneous transmit and receive reflector antenna system. In: 2017 IEEE International Symposium on Antennas and Propagation and USNC/URSI National Radio Science Meeting, pp. 553–554. IEEE (2017) 26. Zhong, J., Alwan, E.A., Volakis, J.L.: Ultra‐wideband dual‐linear polarized phased array with 60° scanning for simultaneous transmit and receive systems. In: 2017 International Workshop on Antenna Technology: Small Antennas, Innovative Structures, and Applications (iWAT), pp. 140–141. IEEE (2017) 5. Etellisi, E.A., Elmansouri, M.A., Filipovic, D.S.: Wideband monostatic co‐ polarized co‐channel simultaneous transmit and receive broadside circular array antenna. IEEE Trans. Antenn. Propag. 67(2), 843–852 (2018) HIJAZI ET AL. 13 International Microwave and RF Conference (IMaRC), pp. 400–404. IEEE (2015) 27. Elmansouri, M.A., Kee, A.J., Filipovic, D.S.: Wideband antenna array for simultaneous transmit and receive (STAR) applications. IEEE Antenn. Wireless Propag. Lett. 16, 1277–1280 (2016) 33. Hijazi, H., et al.: 4‐40 GHz in‐phase/180° out‐of‐phase power dividers with enhanced isolation. In: 14th European Conference on Antennas and Propagation–EuCAP 2020, pp. 1–4 (2020) 28. Etellisi, E.A., Elmansouri, M.A., Filipovic, D.S.: Wideband monostatic simultaneous transmit and receive (STAR) antenna. IEEE Trans. Antenn. Propag. 64(1), 6–15 (2015) 34. REFERENCES Chan, K.K.M., Tan, A.E.C., Rambabu, K.: Decade bandwidth circularly polarized antenna array. IEEE Trans. Antenn. Propag. 61(11), 5435–5443 (2013) 29. Etellisi, E.A., Elmansouri, M.A., Filipovic, D.S.: Wideband monostatic spiral array for full‐duplex applications. In: 2017 IEEE International Symposium on Antennas and Propagation and USNC/URSI National Radio Science Meeting, pp. 1101–1102. IEEE (2017) 30. Filipovic, D.S., Elmansouri, M., Etellisi, E.A.: On wideband simultaneous transmit and receive (STAR) with a single aperture. In: 2016 IEEE In- ternational Symposium on Antennas and Propagation (APSURSI), pp. 1075–1076. IEEE (2016) How to cite this article: Hijazi, H., et al.: Ultra‐ wideband antenna system for in‐band full‐duplex applications. IET Microw. Antennas Propag. 1–13 (2021). https://doi.org/10.1049/mia2.12194 How to cite this article: Hijazi, H., et al.: Ultra‐ wideband antenna system for in‐band full‐duplex applications. IET Microw. Antennas Propag. 1–13 (2021). https://doi.org/10.1049/mia2.12194 31. Kurokawa, K.: Power waves and the scattering matrix. IEEE Trans. Microw. Theor. Tech. 13(2), 194–202 (1965) 32. Rohde, U., et al.: Ultra wide band balun/180° power divider using microstrip‐slotline‐microstrip transition. In: 2015 IEEE MTT‐S
W4386357242.txt	https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.24063975.v1	en		PAMeT-SNN: Predictive Associative Memory for Multiple Time Series based on Spiking Neural Networks with Case Studies in Economics and Finance	null	2,023	cc-by	8,532	PAMeT-SNN: Predictive Associative Memory for Multiple Time Series based on Spiking Neural Networks with Case Studies in Economics and Finance Iman AbouHassan 1 , Nikola Kasabov 2 , Tanmay Bankar 1 , Rishabh Garg 1 , and Basabdatta Sen Bhattacharya 1 1 Posted on 1 Sep 2020 — CC-BY 4.0 — https://doi.org/10.36227/techrxiv.24063975.v1 — e-Prints posted on TechRxiv are preliminary reports that are not peer reviewed. They should not b... 2 Affiliation not available Auckland University of Technology October 31, 2023 Abstract This paper offers for the first time a novel method for creating Predictive Associative Memories of Time series (PAMeT), where a full set of time series variables is used to create a machine learning predictive model (global learning), and afterward, only a few temporal variables are used at a shorter time to recall the model on new data (local recall). Inspired by human brain processes, PAMeT-SNN leverages the brain-inspired SNN NeuCube and the concept of spatio-temporal associative memories (STAM). PAMeT-SNN is a 4-dimensional spatio-temporal structure. First, it encodes time series data into spike sequences that reflects on the changes in the data over time and then maps the temporal variables into the SNN using temporal similarity to define the spatial locations of the variables in the SNN. It then learns temporal associations between a full set of time series variables, thereby memorizing these temporal associations as spatio-temporal connections between neurons. These connections are activated when only part of the time series data is used to recall the model on new data. The proposed groundbreaking method is exemplified with PAMeT-SNN for predictive modeling on two distinct case study time series data sets: trade dynamics and commodity prices. In these case studies, the method effectively captures the intricate data dynamics, enabling accurate forecasting of future values using a minimal set of variables. The method has the potential to be applied in diverse domains. 1 > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1 PAMeT-SNN: Predictive Associative Memory for Multiple Time Series based on Spiking Neural Networks with Case Studies in Economics and Finance Iman AbouHassan, Nikola K. Kasabov, Life Fellow IEEE, Tanmay Bankar, Rishabh Garg, Basabdatta Sen Bhattacharya, Senior Member IEEE Abstract—This paper offers for the first time a novel method for creating Predictive Associative Memories of Time series (PAMeT), where a full set of time series variables is used to create a machine learning predictive model (global learning), and afterward, only a few temporal variables are used at a shorter time to recall the model on new data (local recall). Inspired by human brain processes, PAMeT-SNN leverages the braininspired SNN NeuCube and the concept of spatio-temporal associative memories (STAM). PAMeT-SNN is a 4-dimensional spatio-temporal structure. First, it encodes time series data into spike sequences that reflects on the changes in the data over time and then maps the temporal variables into the SNN using temporal similarity to define the spatial locations of the variables in the SNN. It then learns temporal associations between a full set of time series variables, thereby memorizing these temporal associations as spatio-temporal connections between neurons. These connections are activated when only part of the time series data is used to recall the model on new data. The proposed groundbreaking method is exemplified with PAMeT-SNN for predictive modeling on two distinct case study time series data sets: trade dynamics and commodity prices. In these case studies, the method effectively captures the intricate data dynamics, enabling accurate forecasting of future values using a minimal set of variables. The method has the potential to be applied in diverse domains. Impact Statement —As time series data are widely used in economics, finance, health, engineering, and environment, the proposed method offers a new trend in AI, where a PAMeT-SNN model is created globally on a large number of variables/data and recalled and updated locally on smaller subsets of variables/data. I. AbouHassan* (iabouhassan@tu-sofia.bg) is a Ph.D. candidate at the Technical University of Sofia, Bulgaria, and holds a senior position at Lebanon’s Central Bank, Beirut, Lebanon (iabouhassan@bdl.gov.lb). N.K. Kasabov* (nkasabov@aut.ac.nz) is a Professor with the School of Engineering, Computer and Mathematical Science, Auckland University of Technology, AUT WZ building, St. Paul, Auckland,1010, New Zealand. He is also Chair Professor with the Intelligent Systems Research Centre at University of Ulster UK, and Visiting Professor with IICT Bulgarian Academy of Sciences and with Dalian University, China. He is Honorary Professor of the University of Auckland, NZ and Teesside University UK. Tanmay Bankar (tanmaybankar@gmail.com) is an undergraduate student at BITS Pilani, K. K. Birla Goa, India. Rishabh Garg (rishabhgargdps@gmail.com) is an undergraduate student at BITS Pilani, K. K. Birla Goa, India. Basabdatta Sen Bhattacharya (basabdattab@goa.bits-pilani.ac.in) is an Associate Professor at BITS Pilani, K. K. Birla Goa, India. Index Terms— Multiple time series, Predictive associative memory for multiple time series, Associative Memory, Spiking neural networks, NeuCube, Imports, Commodity, Economic and Financial time series, Forecasting, STAM, PAMeT-SNN. I. INTRODUCTION I n today's globalized world, predictive modeling of time series data, such as economic, environmental, etc. is crucial for world prosperity. Many of the time series data are of multiple variables that interact with each other over time. The challenge is how to use a large scale of available data to create a machine learning model that can be recalled at different times with a smaller number of available variables measured at shorter times, still achieving good predictive accuracy. The paper proposes for the first time a predictive associative memory for time series based on spiking neural networks (PAMeT-SNN). The model utilizes the concept of spatiotemporal associative memory [7]. The proposed PAMeT-SNN addresses some of the limitations of conventional neural network methods to handle non-linear and volatile time series data, and offers a novel comprehensive approach to predict time series with reduced input features without compromising prediction accuracy and explainability, thus contributing to the advancement of explainable AI and incremental learning systems. The case studies in economic forecasting also constitute an unique contribution. Associative memory (AM) is defined as the ability to learn and remember relationships between items and to recall these associations using partial information. Several AM for time series have been developed using statistical methods [35-38]. Although associative memory is a fundamental concept in brain information processing, the mechanisms underpinning it are still unknown [10]. The human brain learns incrementally from a lot of information and performs spatio-temporal associations between unrelated patterns to robustly store this knowledge in space and time [39,40,41]. This structured knowledge between events and their spatiotemporal attributes is retrieved to make future decisions, when only part of the information is available, either due to limited > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < sources, limited time to make the decision, or both. Resembling the human brain is one of the most important goals in computational intelligence. AM models based on neural networks have been already proposed in the literature. The Hopfield network [2] and the bi-directional AM [42] are recurrent neural networks that can be used for pattern recognition. However, they deal with vector-based data and not with spatio-temporal or temporal data. A Hetero-Associative Memory Network (HAM) to model AM using spiking neural networks proved its effectiveness in learning and recalling associations between input and output patterns using simulations [3]. An auto-associative memory based on a spiking neural network (SNN) uses FPGA architectures' vast connection resources to mimic biological brain networks' axons [4]. SNN-based models to encode different memories using different subsets of encoding neurons with temporal codes is proposed in [5,39,41]. Simulation results show that synaptic modification of connections between input layers and hidden layers allows for hetero-associative memory, and recurrent connections between hidden layers allow for auto-associative memory. A network of spiking neurons with AM capability is used to build a dynamic pattern recognition system [6]. AM of spatio-temporal data [7] allows the model to be trained on all temporal variables and recalled on part of them in both time or space. The paper is organized as follows: Section 2 presents the background knowledge of SNN, NeuCube [11] and STAM [7]. Section 3 introduces the proposed PAMeT-SNN framework, which functional diagram is shown in Fig.1. Section 4 presents an experimental study using the method on trade time series data, and section 5 applies the method on a second case study time series data. Section 6 is the conclusion. II. SPIKING NEURAL NETWORKS (SNN), THE NEUCUBE BRAIN INSPIRED SNN ARCHITECTURE AND STAM A. Spiking Neuron Models In SNN, information is represented and processed as sequences of spikes, i.e., binary events at specified times. SNN departs the traditional neural network learning algorithms by the fact that they can learn “time” in their connections. There are several types of spiking neurons and SNN introduced so far [8]-[17]. The Leaky integrate-and-fire (LIF) neuron model is adopted in many SNN systems (for a review, see [13]). It is a simple RC-circuit with a potential leakage, characterized by low computational cost. The neuron will emit a spike when the accumulated input voltage reaches a threshold and then reset to a resting state. The membrane potential u(t) of the neuron is shown in (1). 𝜏𝑚 𝜕𝑢 𝜕𝑡 = 𝑅. 𝐼(𝑡) − 𝑢(𝑡) (1) where, R is resistance, I(t) membrane current, u(t) membrane potential, and 𝜏𝑚 =𝑅. 𝐶 neuron’s membrane time constant. 2 B. Spiking Learning Rules 1) Spike-Timing Dependent Plasticity (STDP) The Spike-Timing Dependent Plasticity (STDP) learning rule is implemented [8] where the firing order of connected neurons determines the synaptic weight. If a pre-synaptic neuron fires first, before a post-synaptic neuron, then longterm potentiation (LTP) is established to strengthen the synapse weight. However, if a post-synaptic neuron activates first, a long-term depression (LTD) occurs, weakening the connection weight. Thus, the modification of the synapse (increase or decrease in weights) can be defined as a function of the firing times of pre-synaptic and post-synaptic neurons based on the difference between tpre and tpost . STDP learning creates long-term memory by changing connection weights to form LTP or LTD. Once data is learned, the SNNcube retains the connections as long-term memory. Then, if just part of the new information is inputted, a chain of activities would fire in the SNNcube based on established connections [9]. Thus, the NeuCube can be explored for learning long spatio-temporal patterns and utilized as associative memory and as a predictive system for event prediction when only some initial new input data is presented. The STDP function is defined as in (2,3). 𝑡𝑝𝑟𝑒 −𝑡𝑝𝑜𝑠𝑡 +𝑎 . 𝑒 𝜏+ if tpre < tpost ∆𝜔 = { (2,3) 𝑡𝑝𝑜𝑠𝑡 −𝑡𝑝𝑟𝑒 −𝑎 . 𝑒 𝜏− if tpre > tpost where ∆ω is the change in weights; a is positive constant learning rate; t is time. 2) The rank order method (RO) The rank order (RO) [10] learning rule allows for new connection weights are formed, depending on the order of the incoming spikes, and the number of spikes that follow the first spike, as in (4). 𝜔𝑖,𝑗 = 𝛼. 𝑚𝑜𝑑 𝑜𝑟𝑑𝑒𝑟(𝑗,𝑖) ± 𝑑 (4) where, 𝜔𝑖,𝑗 is the synaptic weight between a pre-synaptic neuron j and the post-synaptic neuron i; α is a learning parameter; mod is a modulation factor that update weight on first spike occurrence; 𝑜𝑟𝑑𝑒𝑟(𝑗, 𝑖) represents the order of the first spike at synapse j and it is zero for the first spike to neuron i then increases according to the input spike order at other synapses; 𝑑 is the drift parameter that updates the connection weight on subsequent spikes. C. The NeuCube SNN Architecture The NeuCube architectural paradigm is inspired by the human brain and its ability to form AM [11]. It consists of: spike encoding module; 3D SNNcube module; SNN classifier/regressor; parameter optimization module. NeuCube learns from data and connects clusters of neurons to capture the > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 3 Fig. 1. The proposed PAMeT-SNN framework, consisting of: spike encoding module; 3D SNNcube; regression module, inspired by NeuCube [11] and tailored for the first case study trading data. The upper scheme is the training part of a full model and the lower is its recalling on smaller data set . complex temporal relationships between input features of the time series data and to provide a more interpretable representation of the underlying dynamics of the time series in order to make a prediction when new input data is presented. The NeuCube functioning includes the following algorithms: 1) Spike encoding of input signals In the threshold-based encoding algorithm, excitatory (positive) and inhibitory (negative) spikes are generated based on substantial changes in the signal intensity over a certain threshold that is calculated. Encoded spikes are formulated only if variation in consecutive input values is different from the threshold (TH), which are then used as inputs to the spatially located neurons from the SNNcube module [13]. 2) Model Initialization The generated spikes are mapped to a predefined number of neurons located in the 3D-SNNcube. The connections between the neurons are initialized using the small-world connectivity (SWC) approach. For time series where there is no spatial information available for the input variables, a vector quantization principle is employed, in which similar temporal input variables are mapped to closer spiking neurons in the SNNCube [9]. 3) Unsupervised Learning Phase [13] Using the STDP learning rule, a SNNcube is trained on the encoded samples. Some hyper-parameters should be defined, such as the neuron’s potential leak rate, the neuron’s refractory time after emitting the spike, the STDP learning rate for updating the connection’s weights, the firing threshold for the neuron to generate a spike, the number of training iterations, and the long-distance probability. 4) Supervised Learning Phase The computationally efficient dynamic evolving Spike Neural network (deSNN) is trained as an output classifier/regressor using labeled information associated with input samples [11]. When a new input vector is encoded as input spikes during the recall phase, the spiking pattern is transmitted to all newly generated neurons during the learning phase in the SNNcube. If an output neuron's membrane potential exceeds its threshold, neuron i generates an output spike at time t. 5) NeuCube parameter optimization For better performance during the learning process of a NeuCube model, several parameters need to be optimized using the prediction accuracy as an objective function [13]. D. Spatio-temporal associative memories (STAM) The idea of using NeuCube for STAM was first introduced in [11]. In [7] some principles, definitions, and evaluation criteria of STAM are presented. STAM is a system that is trained for classification or prediction on all available spatio-temporal data and their variables and recalled only on part of these variables and/or part of their temporal length. Two validation criteria are introduced for STAM in [7]: - association accuracy, measuring how well a STAM system trained on all spatio-temporal data can be recalled of smaller part of the same data; - generalization accuracy, where the above test is applied on a new spatio-temporal data. In the next section we present the proposed PAMeT-SNN method and framework. III. THE PROPOSED PAMET-SNN METHOD The functional framework of PAMeT-SNN is presented in Fig.1. In this section we present the method and also illustrate it on the first case study data, which is fully investigated in section III. The first case study data [23] is a time series trading data set that represents India's major trading partners, including China, Australia, Germany, Iraq, Saudi Arabia, United Arab Emirates, United States, and other countries as Rest of World > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < (ROW). In this model, the share of monthly imports as a proportion of total imports is used instead of the values of imports. The output variable in this case represents the next month share of Chinese imports to India. As explained in section II, learning in the SNNcube is spatio-temporal, i.e., time series data is first spatially mapped into the structure of the SNNcube and then the local learning rule STDP is applied to the encoded into spike sequences data which changes the connectivity between neurons inside the SNNcube in space and time. To map time series data spatially into a 3D SNN model, first existing time series are analyzed in terms of their similarity, before and after spike encoding. Using the graph matching algorithm proposed in [9], similar variables are mapped to closer spiking neurons in a 3D SNN model (Fig. 2a). Based on the spatial mapping of temporal variables and the unique feature of NeuCube for spatio-temporal learning, the main hypotheses addressed in the paper is: Once a PAMeTSNN model is trained on a full scale of time series variables for prediction, it can generalize to partial time series variables, both in terms of number of variables used and their time length. The rationale behind the hypothesis is the following. The SNNcube accommodates the time series similarity information first as spatial coordinates of the input vbariables. Then the SNNcube is trained on the time series data encoded into spikes using spike-time dependent plasticity (STDP) learning, which guarantees that the connection weights between the neurons capture temporal associations between the input variables (Fig. 2a). Once a SNNcube is trained on K time series variables, each measured in a time length of T, the model can be recalled on a smaller number of variables K1<K and time of measurement T1<T, as the already created connections during learning can be indirectly activated even when some input variables are missing in the recall procedure following the principles polychronization [15] and synfire [16]. And these connections can be activated even for a shorter time of the temporal variables, rather than for the full time used in the training of the full model. The research question now is: How to find the important variables K1 and their time T1 of measurement for recall of a fully trained model on K variables measured over time T, so that the prediction accuracy on the recalled time series data would not be affected? Can these variables be used as predictive markers? The complexity of the problem comes from the fact that here we deal with temporal variables, rather than with static vector-based data, and also from the fact that only few variables can be used for a recall of a model, trained on a full set. In this way some limitations of the traditional deep neural networks can be overcome [17]. The proposed PAMeT-SNN methodology consists of the following procedures: 1. Defining the set of K time series variables for the 4 prediction of a targeted output time series variable or a future event. 2. Encoding the K time series variables into spike sequences. 3. Spatially mapping the K input variables into the SNNcube using the graph mapping algorithm [9] (see Fig. 2a). 4. Training the 3D SNNcube with the STDP unsupervised learning and the deSNN output regressor with supervised learning. 5. Analyze the impact of the input variables in the SNNcube model through calculating a feature interaction network (FIN) as a graph of the number of spike exchange (information exchange) between the clusters of the input variables during learning [13]. The more spikes are exchanged between 2 variables, the more important they are for the predictive model. 6. Rank the importance of variables based on the information from step 5. 7. Remove lower ranked variables and recall the NeuCube model only on the left K1 variables, check the association and generalization accuracy. 8. Recall the model on the K1 variables, measured over time T1 < T and calculate the association and generalization accuracy. 9. Repeat step (8) for smaller time intervals until satisfactory accuracy. 10. Repeat steps (7), (8) and (9) to estimate the final number of important variables K1 and their time of measurement T1 that can be used to recall a NeuCube model without sacrificing the regression accuracy. 11. Evaluate the K1 variables as potential predicting markers according to the prediction problem in hand. When smaller number of variables K1 is entered into a fully trained PAMeT-SNN model for a recall (step 8), the other variables are zeroed, so that they are represented by no spikes, meaning no changes. This is a significant difference between the PAMEeT-SNN and traditional neural networks, that makes the proposed model flexible in both space (variables) and time (their temporal length). When a PAMeT-SNN model, trained with full set of variables, is recalled on a smaller subset of them, the already created connections in the SNNcube are activated even with smaller number of inputs, if these connections represent strong association between the input variables (see Fig.8). In this respect, by zeroing an input variable during a recall process does not mean “ignoring” its impact on the output as this variable may have a strong connection with the output variable, learned in the full model. This is a biologically plausible principle [40]. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 5 (2a) (2b) Fig. 2a. Spatial mapping of 8 time series variables (country import to India from the case study data [23] ) into a SNNcube of 1,000 neurons with the use a graph matching algorithm [7] and connectivity between neurons after unsupervised training of the SNNcube, where connections represent spatio-temporal interaction between input variables. Fig. 2b. represents the (x, z) projection of the SNNcube connectivity. Blue lines are positive connections (excitatory), while red lines are the evolved negative connections (inhibitory). IV. PAMET-SNN MODEL ON CASE STUDY 1 TRADE DATA A. Case Study on Time Series Trade Imports Data Dataset Description The trading case study data was briefly introduced in section III. In 2022, India has emerged as one of the world’s fastest-growing economies and has become a major player in the global economy with a high growth rate [18] [19] [20]. India's imports in 2020 accounted for 2.4 percent of global imports, with the top imports being crude petroleum, electronics, and machinery, imported mostly from China, the United Arab Emirates, and the United States [21] [22]. A PAMeT-SNN model is developed here on this case study, involving predictive modeling of trade data, particularly when limited features are presented on recall. From January 2011 to July 2022, 139 observations of monthly imports of goods were obtained from the United Nations Comtrade [23]. B. Dataset Pre-processing Using a sliding window of 12 months, one data point at a time, 127 overlapping samples have been generated from the original dataset. The 8 input features/variables are the monthly import to India from each of the 8 countries, while the target prediting feature is the next month import from China as the biggest importer to India. C. Dataset Modeling Conforming to the PAMeT-SNN method from section III, the developed here model encodes spikes using the thresholdbased representation (TBR) encoding algorithm. The TBR generates positive and negative spike sequences in response to an increase or decrease in the real value of two consecutive import values from the input time series (Fig.3). The advantage of using the PAMeT-SNN model with the TBR encoding algorithm is that it can deal with noisy data, ignoring small perturbations in it. The algorithm is suitable for trade data analysis as it focuses solely on identifying differences between successive temporal characteristics [25]. In this model, the spike threshold is set at 0.5, and the generated spikes are initially mapped to the randomly connected LIF-based neurons of the NeuCube using a small world (SW) connectivity rule with a radius of 2.5. STDP learning is used in an unsupervised learning stage to capture spatio-temporal associations from encoded inputs. Here the used STDP learning rate is 0.01 and the firing threshold of the LIF neurons is set to 0.5. When this threshold is reached, the neuron emits a spike, and its membrane potential is reset to zero for the duration of its refractory time, which is set at 6 units. The membrane potential leaks between spikes at a potential leak rate of 0.002. The dynamic evolving deSNN regressor is utilized in the supervised mode, with each output neuron associated with a single training sample that is connected to every other neuron in the SNNcube. The connections between connected neurons [i,j] are initialised using the Rank-Order (RO) method of the first incoming spike with the modulation factor (Mod) set to 0.8. The spike-driven synaptic plasticity (SDSP) learning rule adjusts these connections with a drift value δ = 0.005, increasing their weight when spikes arrive at next time points and decreasing it if no spikes occur [11]. D. Experimental Results, Analysis, and Visualization The set of samples was split in half so that 50% of the data were used for training, and the other 50% were used for incremental learning and validating the model. After the unsupervised learning phase is completed, the connections between the input neurons in the trained SNNcube are displayed in Fig. 2a. Positive (excitatory) connections between neurons are shown as blue lines, and negative (inhibitory) connections are represented in red. Fig. 2b. represents a crosssection projection of the 3D SNNcube offering a clear view of the inner connections. A spike raster plot in Fig. 3a is an example representing spiking activity of all input neurons over for one sample of 12 months, where each dot corresponds to the occurrence of a spike from a particular input neuron. Fig. 3b shows the > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 6 amount of positive and negative spikes being emitted by each neuron during the learning stage. (3a) (3b) Fig. 3a. A spike raster plot provides a visual representation of the spiking activity across all 1000 neurons during learning of one sample of 12 time points, where each dot corresponds to the occurrence of a spike at a particular neuron. The horizontal axis represents time, while the vertical axis represents different neurons. Fig. 3b. shows the positive and negative spike amount emitted by each neuron. The learned interactions between the input neurons that represent the trading partners' imports are illustrated by the feature interaction network (FIN) graph in Fig. 4a. The influence of activity between trading partners is represented by connected straight lines. The denser the connection, the greater the effect between the partners. Fig.4b shows the average spike interaction between the clusters around the input variables. In this graphic representation, China-ROW corridor shows a strong influence between each other. During unsupervised STDP learning in the SNNcube, spikes are transmitted across synapses between neurons, resulting in changes to the connection weights. Fig. 5a shows the clusters created around each input neurons and Fig.5b shows the size of these clusters. The larger the size, the more impact this input variable has on the model. Once a full PAMeT-SNN model is created with the use of all K temporal variables and time series data available of length T, the model can be recalled using a smaller number of K1 < K variables, measured at a shorter times T1<T, which is demonstrated in the next sub-section. E. Using the PAMeT-SNN to predict future imports with a small number of input trade time series variables at shorter times The PAMeT-SNN model here aims to predict India's imports from China for the forthcoming time period using an incomplete dataset of independent features after a PAMeTSNN model is created using eight time series variables representing imports from Australia, China, Germany, Iraq, Saudi Arabia, the United Arab Emirates, the United States, and other countries. Using a graph mapping algorithm, the K-selected time series variables are encoded as spike sequences and mapped into the 3D SNNcube (Fig.2a). Following unsupervised learning with STDP and supervised learning to train the deSNN, the proposed algorithm learned and memorized all time series samples for training and relationships between the 8 variables in the model. The experiments involved an iterative process to recall the model using reduced sets of input variables in order to predict the target variable for the next time period. The recall process is repeated for smaller data sets (K1 < K) until a satisfactory level of accuracy is attained. For simplicity, the time interval T1 is set to 1 (T1 = 1) out of the total time-frame of T = 12. Through this iterative process, the final number of important variables (K1) is estimated for a given measurement time (T1=1). Table 1 shows the actual Chinese imports (when K=8) as well as a number of predicted values based on the results of several recall procedures based on various feature availability scenarios. It demonstrates that K1 = 3 is the best scenario for predicting India's imports from China when only the U.A.E., U.S.A., and historical trends of China are provided. Consequently, it can be considered a potential predictive marker with a minimum error. The PAMeT-SNN model is able to activate connections based on input variables in order to retrieve stored patterns or memories and then recall related patterns even when only a subset of input variables is provided. The PAMeT-SNN model effectively preserves associations as spatio-temporal connections between neurons and retains its initial connections that were established during the learning phase when recalling the model on a subset of input variables. The model adeptly retrieves patterns associated with partially connected input features and demonstrates a remarkable capacity for precise predictions. Notably, the dissimilarity between the predicted import patterns of China, obtained using the complete input dataset (Fig. 6a) and the reduced input dataset (Fig. 6b), remains minimal. Fig.7 shows the predicted and the actual values of the import from China when only the input variables are used for a reduced time T=1 with an overall RMSE of 0.02. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < (4a) (4b) Fig. 4a. Total information exchanged between the neuronal clusters of input variables; the thicker the lines, the more spikes are exchanged between the clusters, meaning a temporal association/correlation of changes in the trade. Fig 4b. Average information exchange. (5a) (5b) Fig.5a. Neuronal clusters created through unsupervised learning in the SNNcube. Fig. 5b. Cluster proportion: larger clusters represent more impact of the input vbariable on the the model. . No. of Available Input Variables K=8 K1 = 7 K1 = 6 K1 = 6 K1 = 4 K1 = 3 K1 = 2 K1 = 1 Imports from China [$B] 0.154 0.148 0.147 0.134 0.147 0.153 0.172 0.162 TABLE I THE PREDICTED PATTERN OF INDIA’S IMPORTS FROM CHINA Prediction Error: Type Names of the input variables K1 for time T1=1 to recall the model Value[K] - value[K1] Australia, China, Germany, Iraq, Saudi Arabia, UAE, USA, ROW PAMeT predicted 0.006 Australia, China, Germany, Iraq, Saudi Arabia, UAE, USA PAMeT predicted 0.007 Australia, China, Iraq, Saudi Arabia, UAE, USA PAMeT predicted 0.020 China, Iraq, Saudi Arabia, UAE, USA PAMeT predicted 0.007 Australia, China, Iraq, Saudi Arabia, USA, PAMeT predicted 0.001 China, UAE, USA PAMeT predicted 0.018 Australia, China PAMeT predicted 0.008 China (a) (b) Fig. 6a. Activated connections when all 8 variables are used for a recall of a fully trained model on the K=8 variables for T=12 data points. Fig.6b. Activated connections when only UAE, USA, and China’s (K1=3) time series are used to recall the model for T1=1. 7 > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 8 C. Dataset Pre-processing Utilizing a sliding window approach with a step size of 1, a total of 149 overlapping samples have been generated from the original dataset, each of them represented as 8 time series of length T=12 time points. FIG.7. PREDICTED VS ACTUAL VALUES OF THE IMPORT FROM CHINA TO INDIA WHEN THE FULL PAMET-SNN MODEL IS RECALLED WITH THE USE OF K1=3 INPUT VARIABLES FOR A RECALL TIME OF T1=1. V. PAMET-SNN ON A CASE STUDY 2 DATA A. Case Study on Time Series Commodity Price Expanding the scope of this research and aligning it with diverse domains in trade, economics, and finance, a second case study on commodity market prices was conducted to demonstrate the adaptability of the proposed PAMeT-SNN method. B. Dataset Description Complex interaction of economic forces contributes to the dynamic nature of commodity markets. According to a recent assessment by the World Bank [25], the energy sector encountered a notable increase, with energy prices escalating by 6% in July 2023. For this case study, market prices of eight commodities have been downloaded from the International Monetary Fund (IMF) Commodity Portal [27][26]. The data spans from January 2010 to May 2023, with a total of 161 observations of the monthly market prices of the eight selected commodities (Table II). No. 1 2 3 4 5 6 7 8 TABLE II THE SELECTED COMMODITIES AS SPATIAL FEATURES Spatial feature Description PBANSOP Bananas, US$ per kiloton PGOLD Gold, US$ per thousand troy ounces PSAWMAL Sawn wood, US$ per thousand cubic meters PPLAT Platinum, US$ per thousand troy ounces PROIL Rapeseed oil, US$ per kiloton PSOIL Soybean Oil, US$ per kiloton PSUNO Sunflower oil, US$ per kiloton PWOOLC Wool coarse, US cents per kiloton D. Dataset Modeling The generated samples were split in half, where 50% of the samples was used for training, and the remaining half for incremental learning and validating the model. The proposed PAMeT-SNN model encodes spikes using the thresholdbased representation (TBR) algorithm with a spike threshold of 0.5 [11]. Initially, the generated spikes are spatially mapped into the NeuCube using the graph matching algorithm [9] with a small world (SW) connectivity radius of 2.5. During unsupervised learning, the following NeuCube hyperparameters were set: STDP learning rate was set to 0.01, potential leak rate was set to 0.002, firing threshold was set to 0.5, and refractory time was set to 6 units. The modulation factor was set to 0.95 with a drift parameter of 0.03 during the supervised mode. E. Experimental Results, Analysis, and Visualization After the spatial mapping of temporal variables according to the similarity of the time series as spatial coordinates, the spike-time dependent plasticity (STDP) learning is performed on the encoded spikes. Consequently, the connection weights between neurons reflect the temporal relationships between the input commodities (Fig. 8a). The similarity between the Platinum and Gold, as precious metals, are spatially mapped next to each other in the SNNcube. The feature interaction network (FIN) graph in Fig. 8b shows the learned interactions between the input neurons that represent the commodity market prices. According to the FIN graph, the analyzed feature, ‘Platinum’, has the highest activity with other features. There is a strong direct influence with PGOLD, PSUNO, PSAWMAL, and a modest relationship with PWOOLC, PROIL, and PSOIL. PGOLD, the other precious metal, is strongly connected with PPLAT and PROIL. Custers in the SNNcube around the input variables are depicted in Fig. 9a, where Platinum’s cluster accounts for 13% of all spike activity. The accuracy achieved by the PAMeT-SNN model was 84%. Subsequently, the model was evaluated by performing 2-fold, 3-fold, and 4-fold crossvalidation. The latter provided the highest level of accuracy, 93% (Fig. 9b). > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 9 (9a) (9b) Fig. 8a. The connectivity in SNNcube during unsupervised learning on market prices K= 8 commodities. Fig. 8b. Total information exchanged between the neuronal clusters of commodities, weighted by the thickness of the connections. (10a) (10b) Fig.9a. Neuronal clusters created through unsupervised clustering in a SNNcube, where the Platinum’s cluster accounts for 13% of total spike activity. Fig. 9b. Regression results with a prediction accuracy is 93% after 4-fold cross validation with all 8 variables used.. F. Using the PAMeT-SNN to predict future price with a smaller number of input commodity time series variables The efficiency of the PAMeT-SNN model in capturing and forecasting complex dynamics in the commodities market is validated by its application for analyzing and predicting commodity market prices. This case study applies many experiments on the PAMeT-SNN method with the objective to estimate the future market price of the Platinum commodity utilizing an incomplete dataset of independent features. In accordance with the methodology outlined in section III, the PAMeT-SNN model was originally employed for the whole input stream (K = 8). Subsequently, the experiments involved a repeating procedure whereby the model was recalled by using reduced datasets of input variables. The recall procedure is iterated for subsets of data that are smaller in size (K1 < K) until a desirable degree of precision is achieved. According to the feature interaction network (FIN) graph (Fig. 8b), it can be observed that variables exhibiting a higher number of spikes exchanged, are indicated with thick connections, and deemed to possess more significance in terms of their predictive capabilities throughout the whole sample. The time period T1 is set to 1 out of whole timeframe T = 12. Table III demonstrates that K1 = 4 is the best option for predicting the market price of platinum when only the PBANSOP, PSAWMAL, PSOIL, and PPLAT historical values are available. Consequently, it can be regarded a potential predictive marker with a minimum error rate of 4.2%. In contrast, the traditional linear regression analysis of univariate variables predicts a 32% error in the market price of platinum based solely on their historical trends, as shown in Table III's final row. The PAMeT-SNN model successfully maintains associations by preserving the spatio-temporal connections between neurons. Furthermore, when recalling the model on a subset of input variables, it retains the initial connections established during the learning phase (Fig. 10a, b). When all analyzed variables (K=8) are present in the input stream, the model effectively captures the complexities of commodity price dynamics and depicts the model's full connectivity (Fig. 10a). The displayed connections have a threshold weight of 0.1. During recall, where only a subset of input variables is available (K1 = 4), certain connections are not present (Fig. 10b), but that does not prevent an accurate prediction. > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 10 (11a) (11b) Fig. 10a. The connectivity evolution in NeuCube during global training on prices data recorded from the 8 commodities. Fig. 10b. The connectivity during local recall on prices recorded from only 4 commodities (PBANSOP, PSAWMAL, PPLAT, PSOIL). No. of Available Input Variables K=8 K1 = 7 K1 = 6 K1 = 5 K1 = 4 K1 = 3 K1 = 2 K1 = 1 K1 = 8 TABLE III THE PREDICTED MARKET PRICE OF PLATINUM COMMODITY WITH REDUCED INPUT DATASETS Imports from Prediction Error: Type Names of the available input variables to recall the model China [$B] Value[K1] - value[K] PBANSOP,PGOLD,PSAWMAL,PPLAT,PROIL,PSOIL,PSUNO,PWOOLC 1.0590 Real value PBANSOP,PGOLD,PSAWMAL,PPLAT,PROIL,PSOIL,PWOOLC 1.5148 PAMeT predicted +0.456 PBANSOP,PSAWMAL,PPLAT,PROIL,PSOIL,PWOOLC 1.3744 PAMeT predicted +0.315 PBANSOP,PSAWMAL,PPLAT,PSOIL,PWOOLC 0.8996 PAMeT predicted -0.160 PBANSOP,PSAWMAL,PPLAT,PSOIL 1.1015 PAMeT predicted +0.042 PSAWMAL,PPLAT,PSOIL 1.3043 PAMeT predicted +0.245 PPLAT,PROIL 1.2064 PAMeT predicted +0.147 PPLAT 0.9827 PAMeT predicted -0.076 PBANSOP,PGOLD,PSAWMAL,PPLAT,PROIL,PSOIL,PSUNO,PWOOLC 0.7415 Linear Reg. predicted -0.318 This distinctive behavior of the PAMeT-SNN model to make good predictions with incomplete input allows it to recall relevant information accurately even when operating with a reduced set of variables without compromising the model’s accuracy. Thus, PAMeT-SNN represents a paradigm shift in the field of artificial intelligence. VI. DISCUSSION AND CONCLUSION PAMeT-SNN model offers an effective approach for time series analysis and prediction. Its ability to learn and recall complicated temporal patterns qualifies it for a wide range of applications in various domains such as finance, engineering, and healthcare. The proposed here method can be used to extract temporal fuzzy rules, to be used for a comparison between fully trained and partially recalled PAMeT, adding to the explainability of the model [27]. In [28] [29], time series environmental data collected over 40 days, such as pollution, temperature, wind, solar eruption and others, are used to predict individual stroke occurrence. The proposed PAMeT-SNN can be applied to make on-line predictions of individual stroke prediction based only on few of these variables and a shorter time, saving human lives. In [30] an earthquake prediction system is described that uses 100 days seismic spatio-temporal data to predict an earthquake event several hours ahead. Here a PAMeT-SNN model can be applied for an early event prediction based on shorter time series data and smaller number of seismic input variables. In [31], London city pollution prediction system is described and in [32] a similar pollution prediction system is presented for the areas of Beijing and Shanghai. They use a large number of spatially located sensors to predict pollution several hours ahead using their time series measurements. PAMeT-SNN model can be applied for an early prediction based on shorter time series data and smaller number of pollution sensors. In [33], different methods were discussed for predicting and modeling long time series data on Bulgarian petroleum oil imports from 19 key trading partners in order to identify abnormal patterns over time. The suggested PAMeT-SNN can be used to forecast a crucial commodity based on fewer factors, which is especially useful when data is scarce during crises and pandemics. In [34], financial time series data and online news were integrated to improve forecast accuracy and data interpretation. The proposed PAMeT-SNN will be applied to forecast a stock index using news data and reduced time series. PAMeT-SNN is suitable for wider range of applications where other methods were previously applied [35-38]. In summary, the paper offers a new trend in AI, where instead of creating and recalling many models for the same problem, each using same set of temporal variables for training and recall, but different variables across the models, as it is the case with the deep neural networks, here one > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < PAMeT-SNN is created globally on a large number of time series variables and recalled and updated many times locally on smaller subsets of variables. This introduces a new direction for AI of a global model training and a local model recall. ACKNOWLEDGEMENTS The work of NK and IA is sponsored by the Knowledge Engineering Cons. Ltd (https://knowledgeengineering.ai) and their respective organisations. NeuCube software can be obtained from: www.kedri. aut.ac.nz/ neucube; https://github.com/KEDRI-AUT/ Neu Cube-Matlab; and https://github.com/KEDRI-AUT/NeuCube-Py. BSB acknowledges the support of BITS Pilani K. K. Birla Goa Campus in conducting the Course CS F382 (Brain-inspired Deep Learning) that resulted in her contribution to this project. TB and RG acknowledge the course CS F382 at BITS Pilani K. K. Birla Goa Campus that they undertook as a part of their undergraduate degree and that provided them the opportunity to contribute to this project. REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] J. Y. Lee, et al., “Dopamine facilitates associative memory encoding in the entorhinal cortex,” Nature, doi/10.1038/ s41586-021-03948-8, 2021. H. Sompolinsky, O. L. White, In Les Houches book series, 2005. W. Gerstner, J. L. Van Hemmen, “Associative memory in a network of 'spiking' neurons.” Network 3 139, doi/10.1088/ 0954-898X/3/2/004, 1992. C. H. Ang, C. Jin, P. H. W. Leong, A. Van Schaik, “Spiking neural network-based auto-associative memory using FPGA interconnect delays”. International Conference on Field-Programmable Technology, New Delhi, India, 2011, pp. 1-4, doi/10.1109/FPT. 2011.6132701. J. Hu, H. Tang, K. C. Tan, S. B. Gee, “A Spiking Neural Network Model for Associative Memory Using Temporal Codes.” In: Proceedings of the 18th Asia Pacific Symposium on Intelligent and Evolutionary Systems, Vol.1, Springer, Cham, doi/10.1007/978-3-31913359-1_43, 2015. E. Agnes, R. Erichsen, L. Brunnet, “Model architecture for associative memory in a neural network of spiking neurons.” Physica A: Statistical Mechanics and its Applications, Vol.391, Issue 3, pp 843-848, ISSN 0378-4371, doi/10.1016/j.physa.2011.08.036, 2012. N. K. Kasabov, “Spatio-Temporal Associative Memories in Braininspired Spiking Neural Networks: Concepts and Perspectives.” TechRxiv. Preprint. https://doi.org/10.36227/techrxiv.23723208.v1, 2023. S. Song, K. Miller, L. Abbott, et al., “Competitive Hebbian learning through spike-timing-dependent synaptic plasticity.” Nat. Neurosci. 3, 919–926, 2000. E. Tu, N. Kasabov, J. Yang, “Mapping Temporal Variables Into the NeuCube for Improved Pattern Recognition, Predictive Modeling, and Understanding of Stream Data.” in IEEE Transactions on Neural Networks and Learning Systems, vol. 28, no. 6, pp. 1305-1317, doi/10.1109/TNNLS.2016.2536742, 2017. S. Thorpe, J. Gautrais, “Rank order coding. In: Computational neuroscience.” Springer, pp 113–118, 1998. N. K. Kasabov, “NeuCube: A spiking neural network architecture for mapping, learning and understanding of spatio-temporal brain data.” Neural Networks, Vol.52, pp. 62-76, doi/10.1016/j.neunet. 2014.01.006, 2014. C. Tan, M. Šarlija, N. Kasabov, “Spiking Neural Networks: Background, Recent Development and the NeuCube Architecture.” Neural Processing Letters 52:1675–1701, Springer, Science and Business Media, LLC, doi/10.1007/s11063-020-10322-8, 2020. N. Kasabov, “Time-Space, Spiking Neural Networks and BrainInspired Artificial Intelligence.” Springer Nature, p 750, https://www.springer.com/gp/book/ 9783662577134, 2019. 11 [14] https://kedri.aut.ac.nz/neucube [15] E. M. Izhikevich, “Polychronization: computation with spikes.” Neural Comput. 18(2), 245–282, 2006. [16] Y. Ikegaya, G. Aaron, R. Cossart, I. Aronov, D. Lampl, D. Ferster et al., “Synfire chains and cortical songs: temporal modules of cortical activity.” Science 304, 559–564, N. 2004. [17] G. Indiveri, F. Corradi, N. Qiao, “Neuromorphic architectures for spiking deep neural networks.” IEEE Int. Electron Devices Meeting (IEDM), 2015. [18] https://www.worldbank.org/en/country/india [19] https://www.oecd.org/economy/india-economic-snapshot/ [20] https://economictimes.indiatimes.com/news/economy/indicators/verystrong-economy-imf-division-chief-daniel-leigh-on-indias-growthtrajectory/articleshow/99420331.cms/ [21] https://www.theglobaleconomy.com/India/share_world_ imports/ [22] https://www.oec.world/ [23] https://comtradeplus.un.org/TradeFlow [24] B. Petro, N. Kasabov, R.M. Kiss, “Selection and Optimization of Temporal Spike Encoding Methods for Spiking Neural Networks.” IEEE Transactions on Neural Networks and Learning Systems, vol. 31, no. 2, pp. 358-370, Feb. 2020, doi: 10.1109/TNNLS.2019.2906158, 2020. [25] World Bank Commodities Price Data (The Pink Sheet); https://www. worldbank.org/en/research/commodity-markets, August 2023. [26] Https://www.imf.org/en/Research/commodity-prices. [27] N. Kasabov, Y. Tan, M. Doborjeh, et al. “Transfer Learning of Fuzzy Spatio-Temporal Rules in the NeuCube Brain-Inspired Spiking Neural Network: A Case Study on EEG Spatio-temporal Data.” IEEE Transactions on Fuzzy Systems, 1-11, 1941-0034, DOI :https://doi.org/10.1109/TFUZZ.2023.3292802, July 2023. [28] N. Kasabov, et al. “Evolving spiking neural networks for personalised modelling, classification and prediction of spatio-temporal patterns with a case study on stroke.” Neurocomputing, 134, 269-279, 2014. [29] M. Doborjeh, Z. Doborjeh, A. Merkin, R. Krishnamurthi, R. Enayatollahi, V. Feigin & N. Kasabov, “Personalized spiking neural network models of clinical and environmental factors to predict stroke.” Cognitive Computation, 14(6), 2187-2202, 2022. [30] N. Kasabov et al., “Evolving spatio-temporal data machines based on the NeuCube neuromorphic framework: Design methodology and selected applications.” Neural Networks 78: 1-14, 2016. [31] P. S. Maciąg, N. Kasabov, M. Kryszkiewicz, R. Bembenik, “Air pollution prediction with clustering-based ensemble of evolving spiking neural networks and a case study for London area.” Environmental Modelling & Software, 118, 262-280, 2019. [32] H. Liu, G. Lu, Y. Wang, N. Kasabov, “Evolving spiking neural network model for PM2. 5 hourly concentration prediction based on seasonal differences: A case study on data from Beijing and Shanghai.”, 2021. [33] I. Abouhassan, N. Kasabov, G. Popov, R. Trifonov, “Why Use Evolving Neuro-Fuzzy and Spiking Neural Networks for incremental and explainable learning of time series? A case study on predictive modelling of trade imports and outlier detection.” IEEE 11th International Conference on Intelligent Systems (IS), Warsaw, Poland, 2022, pp. 1-7, doi: 10.1109/IS57118.2022.10019673, 2022. [34] N. Kasabov, I. AbouHassan, V. Jagtap, P. Kulkarni, “Spiking Neural Networks for Predictive and Explainable Modelling of Multimodal Streaming Data with a Case Study on Financial Time-series and Online News.” Preprint, Research Square, doi:10.21203/rs.3.rs-2262084/v1, 2022. [35] Z. Zografski, T. Durrani, Learning associative memory predictors from time series, Signal Processing, Vol. 59, Issue 2, 1997, 243-249, ISSN 0165-1684, https://doi.org/10.1016/S0165-1684(97)00050-9. [36] I. López-Yáñez, L. Sheremetov, C. Yáñez-Márquez, A novel associative model for time series data mining, Pattern Recognition Letters, Vol. 41, 2014, 23-33, ISSN 0167-8655, https://doi.org/10.1016/j.patrec.2013.11.008. [37] Lee, J., Kim, W., Cha, J., Kim, G., Choi, H. (2004). A New Recurrent Fuzzy Associative Memory for Recognizing Time-Series Patterns Contained Ambiguity. In: Laganá, A., Gavrilova, M.L., Kumar, V., Mun, Y., Tan, C.J.K., Gervasi, O. (eds) Computational Science and Its Applications – ICCSA 2004. ICCSA 2004. Lecture Notes in Computer Science, vol 3043. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-24707-4_3 [38] S. J. Silver-Warner, R. J. Glover and T. J. Stonham, "Associative memory neural networks for time series prediction," Proceedings of > REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < [39] [40] [41] [42] Third International Conference on Electronics, Circuits, and Systems, Rhodes, Greece, 1996, pp. 651-654 vol.2, doi: 10.1109/ICECS.1996.584446. Tam, D. Temporal associative memory (TAM) by spike-timing dependent plasticity. BMC Neurosci 11 (Suppl 1), P105 (2010). https://doi.org/10.1186/1471-2202-11-S1-P105. G. Palm, Neural associative memories and sparse coding, Neural Networks, vol.37, 2013, 165-171, ISSN 0893-6080, https://doi.org/10.1016/j.neunet.2012.08.013. D.J. Amit, A. Treves, Associative memory neural network with low temporal spiking rates, Proc. Nati. Acad. Sci. USA Vol. 86, pp. 78717875, October 1989 Biophysics. B.Kosko, (1988). "Bidirectional Associative Memories" . IEEE Transactions on Systems, Man, and Cybernetics. 18 (1), 49– 60. doi:10.1109/21.87054. Iman AbouHassan (IEEE M’16) is a Ph.D. candidate at the Technical University of Sofia / Faculty of Computer Systems and Technology, and a Researcher at knowledgeengineering.ai. Her research focuses on the practical applications and advancements of artificial intelligence and neural network technologies in the analysis of economic and financial data. Apart from her academic pursuits, she holds a master's degree in computer engineering from TU-Sofia. She is also the Head of the External Sector Division at Banque du Liban. Her expertise is acknowledged by the International Monetary Fund (IMF) and the World Bank, particularly in the domain of balance of payments and related subjects. Nikola K. Kasabov (IEEE M’93, SM’98, F’10, LF’22) received his MSc and Ph.D from the Technical University of Sofia, Bulgaria. He is the Founding Director of KEDRI and Professor of Knowledge Engineering at Auckland University of Technology, New Zealand. He holds also professorial positions at the University of Ulster UK, IICT Bulgarian Academy of Sciences and Dalian University, China. He is Director of the knowledgeengineering.ai. His research areas are computational intelligence, neuroinformatics, knowledge discovery, spiking neural networks, with more than 320 journal publications. Tanmay T. Bankar is a final year undergraduate student at BITS Pilani, Goa, India. He is majoring in Computer Science and Engineering (BE. Honors) and is also a visiting researcher at the KEDRI Lab at Auckland University of Technology. His research areas are machine learning, spiking neural networks, and multi-modal predictive time-series modeling. 12 Rishabh Garg received his Bachelor of Engineering degree with a Minor in Finance from Birla Institute of Technology and Science, Pilani, India (2023). He is currently working as a Software Engineer with ServiceNow, having over two years of experience in the IT industry, and previously worked as a Site Reliability Engineering intern at Flipkart. He has also authored the book Blockchain for Real World Applications with the publisher John Wiley and Sons, US. He is a recipient of the National Award for Exceptional Achievement in Innovation from the President of India and National CSIR Innovation Award from the Prime Minister of India. Basabdatta Sen Bhattacharya (IEEE M 2009, SM 2015) received her PhD from the School of Computer Science, University of Manchester, UK (2008), Masters in Engineering from Jadavpur University, Kolkata, India (2002), and Bachelors of Engineering from National Institute of Technology (NIT), Silchar, Assam, India (1992). She was born (1969) and raised in Silchar. She has worked as a Researcher with the University of Manchester (UK), University of Bordeaux (France) and the University of Ulster (Northern Ireland, UK). She has worked as an Academic at the University of Lincoln (UK) and as a Lecturer at the NIT, Rourkela, Odisha, India. She has worked as a Manager (Technical Category) with the Steel Authority of India Ltd. (SAIL), posted at the Electronics Engineering Department, Rourkela Steel Plant, Rourkela, Odisha. Her primary research area is brain-inspired neural networks with applications to computational neuroscience and Artificial Intelligence.
https://openalex.org/W2734956157	http://www.journalijar.com/uploads/319_IJAR-18331.pdf	English	null	ARE DRIVERS OF PSYCHOLOGICAL WELL-BEING SIMILAR TO THE DRIVERS OF ILL-BEING?	International journal of advanced research	2,017	cc-by	2,295	ISSN: 2320-5407 ISSN: 2320-5407 Int. J. Adv. Res. 5(6), 2355-2358 Tools And Techniques Used:- Tools And Techniques Used:- 1. Secondary tools were used. 2. Use of Journals, Books, Magazines, Internet was made. 3. Some articles were taken from sites like Sciencedirect.com and Google scholar as well 3. Some articles were taken from sites like Sciencedirect.com and Google scholar as we Personality Factors:- y One of the strongest predictors of our usual emotional style is Personality.Two dimensions of Extraversion and Neuroticism were taken. Costa & Mc. Crae(1980) found the following things in the 10 year longitudinal study. y One of the strongest predictors of our usual emotional style is Personality.Two dimensions of Extraversion and Neuroticism were taken. Costa & Mc. Crae(1980) found the following things in the 10 year longitudinal study.  Extraversion is strongly associated with positive emotional style.  Extraversion is strongly associated with positive emotional style.  Neuroticism is strongly associated with negative emotional style (Diener et al., 1999).  Neuroticism is strongly associated with negative emotional style (Diener et al., 1999).  Gardner et. al., 2006, Mental ill-being is caused due to link between childhood neuroticism or adolescent neuroticism and psychological distress later in life.  Neuroticism is strongly associated with negative emotional style (Diener et al., 1999).  Gardner et. al., 2006, Mental ill-being is caused due to link between childhood neu g y g y  Gardner et. al., 2006, Mental ill-being is caused due to link between childhood neuroticism or adolescent neuroticism and psychological distress later in life.  Gardner et. al., 2006, Mental ill-being is caused due to link between childhood neurot  Gardner et. al., 2006, Mental ill-being is caused due to link between childhood neuroticism or adolescent neuroticism and psychological distress later in life. neuroticism and psychological distress later in life.  Neuroticism drives negative mood and common mental disorder.  Extraversion drives positive emotional characteristics.  Ryff's Scales of Psychological Well-Being (Ryff, 1989) cover the dimensions of Autonomy, Environmental Mastery, Personal growth, Personal relationships with others, purpose in life and self-acceptance. A recent longitudinal study using the Ryff scale, shows much larger effect of Extraversion than of Neuroticism for psychological well-being (Abbottet. al., 2008). Description of the Variables;- Personality Factors:- Hypothesis Of The Study:- Hypothesis Of The Study:- Hypothesis Of The Study:- s hypothesized that „ Personality' is a common indicator for well-being and ill-being respectively. s hypothesized that „Demographic factors' are similar for the well-being and ill-being respectively. yp g p g g p y 3. It is hypothesized that 'Socio-economic factors' are responsible for both well-being and ill-being respectively. yp g p g g p y s hypothesized that 'Socio-economic factors' are responsible for both well-being and ill-being respe 4. It is hypothesized that „Intentional activities' are responsible for both psychological well-being and ill-being respectively. 4. It is hypothesized that „Intentional activities' are responsible for both psychological well-being and ill-being respectively. Need For The Study:- Need For The Study:- y 1. Recognition that since well-being is more than the absence of ill-being, it needs to be studied in its own right. y 1. Recognition that since well-being is more than the absence of ill-being, it needs to be studied in its own right. g pp p g p y g g 3. Evidence that many of the drivers of well-being are not the same as the drivers of ill-being. g pp p g p y g g 3. Evidence that many of the drivers of well-being are not the same as the drivers of ill-being. Obiective Of The Study:- Obiective Of The Study:- y The indicators of psychological well-being are similar to the indicators of psychological ill-being. The indicators of psychological well-being are similar to the indicators of psychological ill-being. ARE DRIVERS OF PSYCHOLOGICAL WELL-BEING SIMILAR TO THE DRIVERS OF ILL-BEING? Nitika. Economics, Khalsa College For Women, Civil Lines, Ldh India ( Punjab). Nitika. Economics, Khalsa College For Women, Civil Lines, Ldh India ( Punjab). ……………………………………………………………… Psychological well being is about lives going well. It is a combination of being good and functioning effectively. However, it does not mean feeling good all the time, the experience of painful emotions like disappointment, failure, grief is a normal part of life and being able to manage these negative emotions is essential for long-term well being. So the need for the study of this topic is to explain that many of the drivers of well being are not the same as the drivers of ill-being. The study is hypothesized that, personality, demographic factors, socio- economic factors and international activities are responsible for both psychological well-being and ill-being respectively. To study the objectives, secondary tools were used. With study carried out, with the review of Literature, it suggests that personality factors, demographic factors and socio-economic factors are equally responsible for psychological well-being and ill-being respectively but international activities are only responsible for psychological well-being and not for ill-being. Received: 20 April 2017 Final Accepted: 23 May 2017 Published: June 2017 Copy Right, IJAR, 2017,. All rights reserved. Copy Right, IJAR, 2017,. All rights reserved. Introduction To The Concept:- Psychological well being is about lives going well. It is a combination of being good and functioning effectively. However, it doesn't mean feeling good all the time, the experience of painful emotions like disappointment, failure, grief is a normal part of life and being able to manage these negative or painful emotions is essential for long-term well being. Psychological well-being not only corporate the positive emotions of happiness and contentment but also such emotions like interest, engagement, confidence and affection. If we talk in psychological sense, it includes. 1. Development of one's potential. 2. A control over one's life. 3. Having a sense of purpose. 4. Experiencing positive relationships. According to WHO (2001), "Mental health as a state of well-being in which an individual realizes his/her own abilities, can cope with normal stresses of life, can work productively and fruitfully, and is able to make a contribution to his/her community.” 2355 ISSN: 2320-5407 Int. J. Adv. Res. 5(6), 2355-2358 Demographicfactors:- Demographic characteristics also show differential aspects for well-being and ill-being. Demographic characteristics also show differential aspects for well-being and ill  Large surveys showed little evidence of gender differences (Helliwell, 2003).  Large surveys using single-item measures of well-being (eg. Overall rating of life satisfaction) find a U shaped relationship with age.  According to Oswald, 2008, Younger and older people tend to have higher well-being scores than middle-age, although there may be decline in well-being among the very old.  Data from British Health and Lifestyle Survey showed that compared to middle-aged and younger men, older men have the lowest number of symptoms of psychological distress but also the lowest scores on a measure of positive psychological well-being.  However, Zimmerman & Easterlin (2006) found that “Marriage” is also an important factor for psychological well-being, they found that while getting married is good for one's psychological well-being, being married may not be. 2356 ISSN: 2320-5407 Int. J. Adv. Res. 5(6), 2355-2358 Int. J. Adv. Res. 5(6), 2355-2358 Conclusion:- Science of well-being which focuses on what makes people flourish on human asset rather than deficits is a promising new area of research. However, Review of Literature and various case-studies suggests that Personality factors, Demographic factors and Socio-economic factors are equally responsible for psychological well-being and illbeing respectively but the aspects may vary. However, Intentional activities are only responsible for psychological well-being and not for ill-being. Intentional Activities;- Intentional Activities;- ; Lyubomirsky (2006) found that intentional activities are divided into three broad groups 1. Behaviors-such as taking regular exercise, or being kind to others. 2. Cognitions-such as interpreting events in a positive light or feeling gratitude. 3. Motivations-such as striving towards goal which reflect deeply held values rather than being driven by external rewards. 3. Motivations-such as striving towards goal which reflect deeply held values rather than being driven by external rewards. Lyubomirsky found that evidence of such activities increases happiness and psychological well-being but psychological ill-being in either caused or not that is not quantified. References:- 1. Abbott. R.A., Ploubidis, G.B., Croudace, J.J., Kih, D., Wadsworth, M.E.J., & Huppert, F.A. (2008). The relationship between early personality and mildlife psychological wellbeing: Evidence from a U.Kbirth cohort study, “Social Psychiatry and Psychiatric Epidemiology',43(9), 679-687. 1. Abbott. R.A., Ploubidis, G.B., Croudace, J.J., Kih, D., Wadsworth, M.E.J., & Huppert, F.A. (2008). The relationship between early personality and mildlife psychological wellbeing: Evidence from a U.Kbirth cohort study, “Social Psychiatry and Psychiatric Epidemiology',43(9), 679-687. y y y y p gy ( ) 2. Chevalier, A., & Feinstein, L. (2006). Sheepskin or Prozac: The causal effect of education o Report published by the Centre for the Economics of 3. Education, London School of Economics, London. 3. Education, London School of Economics, London. 4. Clark, A. (2003). Unemployment as a Social Norm: Psychological evidence from panel 5 d t “J l f L b E i " 21 323 351 4. Clark, A. (2003). Unemployment as a Social Norm: Psycholog 5 d “J l f L b E i " 21 323 351 4. Clark, A. (2003). Unemployment as a Social Norm: Psychological evidence from panel 5. data. “Journal of Labour Economics", 21,323-351. . C a , . ( 003). U e p oy e t as a Soc a No : syc o og ca ev de ce o pa e 5. data. “Journal of Labour Economics", 21,323-351. 6. Costa, P.T., & Mc. Crae, R.R. (1980). Influence of extraversion and neuroticism on subjective well-being: Happy and unhappy people. "Journal of Personality and Social 7 Psychology ”38 668 678 6. Costa, P.T., & Mc. Crae, R.R. (1980). Influence of extraversion and neuroticism on subjective well-being: Happy and unhappy people. "Journal of Personality and Social 7. Psychology,”38,668-678. 8. Diener, E., Suh E.M., Lucas, R.E., & Smith, H.L. (1999). Subjective Well-Being: Three decades of progress- “Psychological Bulletien", 125,276-302. 8. Diener, E., Suh E.M., Lucas, R.E., & Smith, H.L. (1999). Subjective Well-Being: Three decades of progress- “Psychological Bulletien", 125,276-302. 9. Dolon, P., Peasgood, T., & White, M. (2008). Do we really know what makes us happy? A review of the economic literature on the factors associated with subjective well-being. 9. Dolon, P., Peasgood, T., & White, M. (2008). Do we really know what makes us happy? A review of the economic literature on the factors associated with subjective well-being. "Journal of Economic Psychology', 29, 94-122. y gy 11. Helliwell, J.F.(2003). How's life? Socio-Economicfactors:-  According to Dolon et al., 2008, Higher levels of income and socio-economic status are associated with higher levels of well-being and lower rates of disorder, although this effect diminishes at progressively higher levels of income.  According to Dolon et al., 2008, Higher levels of income and socio-economic status are associated with higher levels of well-being and lower rates of disorder, although this effect diminishes at progressively higher levels of income.  While most studies suggests that higher educational qualifications are protective factors against poor health, a few have revere gradient. Chevalier & Feinstein (2006) found that men with a higher level of education were more likely to be depressed than those with less education. They suggest that the increase in depression associated with the highest level of education may be an indication of job-related stress of occupations requiring a degree. o Wilkinson et al., (2006), Higher national income inequality is linked to higher prevalence of s and lower scores on well-being measures. g  According to Wilkinson et al., (2006), Higher national income inequality is linked to h mental illness and lower scores on well-being measures. g  According to Wilkinson et al., (2006), Higher national income inequality is linked to higher prevalence of mental illness and lower scores on well-being measures.  Unemployment has also been associated with the presence of mental health problems and lower levels of lif satisfaction. According to Clarket al., (2004), Longitudinal studies demonstrate that people who started out relatively happy become unhappy after they were unemployed.  Unemployment has also been associated with the presence of mental health problems and lower levels of lif satisfaction. According to Clarket al., (2004), Longitudinal studies demonstrate that people who started out relatively happy become unhappy after they were unemployed. 15. World Health Organization (2001). The World health report on Mental Health: New understanding, new hope. Geneva: World Health Organization. 16. Zimmerman, A.C., & Eastlin, R.E. (2006). Happily ever after? Cohabitation, marriage, divorce & happiness in Germany, "Population and Development Review'', 32,511-528. 15. World Health Organization (2001). The World health report on Mental Health: New understanding, new hope. Geneva: World Health Organization. 16. Zimmerman, A.C., & Eastlin, R.E. (2006). Happily ever after? Cohabitation, marriage, divorce & happiness in References:- Combining individual and national variations to explain 11. Helliwell, J.F.(2003). How's life? Combining individual and national variations to explain 11. Helliwell, J.F.(2003). How's life? Combining individual and national va Helliwell, J.F.(2003). How's life? Combining indi 12. subjective well-being. “Economics Modelling", 20,331-360. j g g 13. Lyubomirsky, S., King, L., & Diever, E. (2005). The benefits of frequent positive effect: Does happiness leads to success? "Psychological Bulletien?", 131,803-855. j g g 13. Lyubomirsky, S., King, L., & Diever, E. (2005). The benefits of frequent positive effect: Does happiness leads to success? "Psychological Bulletien?", 131,803-855. 14. Ryff, C.D., (1989). Happiness is everything, or is it? Exploration on the meaning of psychological well-being. "Journal of Personality and Social Psychology', 57, 1069-1081. 14. Ryff, C.D., (1989). Happiness is everything, or is it? Exploration on the meaning of psychological well-being. "Journal of Personality and Social Psychology', 57, 1069-1081. 2357 Int. J. Adv. Res. 5(6), 2355-2358 Int. J. Adv. Res. 5(6), 2355-2358 g 16. Zimmerman, A.C., & Eastlin, R.E. (2006). Happily ever after? Cohabitation, marriage, divorce & happiness in Germany, "Population and Development Review'', 32,511-528. ISSN: 2320-5407 ISSN: 2320-5407 2358
https://openalex.org/W4293567428	https://digital.csic.es/bitstream/10261/295507/1/CCP4%20Cloud%20for%20structure%20determination_Krissinel_PV_Art2022.pdf	English	null	<i>CCP</i>4 Cloud for structure determination and project management in macromolecular crystallography	Acta crystallographica. Section D, Structural biology	2,022	cc-by	10,104	ISSN 2059-7983 Eugene Krissinel,a* Andrey A. Lebedev,a Ville Uski,a Charles B. Ballard,a Ronan M. Keegan,a Oleg Kovalevskiy,a Robert A. Nicholls,b Navraj S. Pannu,c Pavol Skuba´k,c John Berrisford,d Maria Fando,e,f,g Bernhard Lohkamp,h Marcin Wojdyr,i Adam J. Simpkin,j Jens M. H. Thomas,j Christopher Oliver,k Clemens Vonrhein,i Grzegorz Chojnowski,l Arnaud Basle,m Andrew Purkiss,n Michail N. Isupov,o Stuart McNicholas,p Edward Lowe,q Josep Trivin˜o,r Kevin Cowtan,p Jon Agirre,p Daniel J. Rigden,j Isabel Uson,r,s Victor Lamzin,l Ivo Tews,g Gerard Bricogne,i Andrew G. W. Leslieb and David G. Brownt Received 18 May 2022 Accepted 8 August 2022 Received 18 May 2022 Accepted 8 August 2022 Received 18 May 2022 Accepted 8 August 2022 Edited by S. Antonyuk, Institute of Integrative Biology, University of Liverpool, United Kingdom Edited by S. Antonyuk, Institute of Integrative Biology, University of Liverpool, United Kingdom aScientific Computing Department, Science and Technology Facilities Council UK, Didcot OX11 0FA, United Kingdom, bStructural Studies Division, MRC Laboratory for Structural Biology, Cambridge CB2 0QH, United Kingdom, cLeiden University Medical Center, 2333 ZA Leiden, The Netherlands, dEuropean Bioinformatics Institute, Hinxton CB9 1SD, United Kingdom, eInstitute of Protein Research, Pushchino 142290, Russian Federation, fTranslational and Clinical Research Institute, Newcastle University, Framlington Place, Newcastle upon Tyne NE2 4HH, United Kingdom, gBiological Sciences, Institute for Life Sciences, University of Southampton, Southampton SO17 1BJ, United Kingdom, hDepartment of Medical Biochemistry and Biophysics, Karolinska Institutet, SE-171 77 Stockholm, Sweden, iGlobal Phasing Limited, Sheraton House, Castle Park, Cambridge CB3 0AX, United Kingdom, jInstitute of Systems, Molecular and Integrative Biology, University of Liverpool, Liverpool L69 7ZB, United Kingdom, kSchool of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, United Kingdom, lEuropean Molecular Biology Laboratory, Hamburg Unit, Notkestrasse 85, 22607 Hamburg, Germany, mNewcastle University Biosciences Institute Medical School, Newcastle upon Tyne NE2 4AX, United Kingdom, nStructural Biology Science Technology Platform, The Francis Crick Institute, 1 Midland Road, London NW1 1AT, United Kingdom, oBiosciences, University of Exeter, Stocker Road, Exeter EX4 4QD, United Kingdom, pYork Structural Biology Laboratory, Department of Chemistry, University of York, York YO10 5DD, United Kingdom, qDepartment of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, United Kingdom, rCrystallographic Methods, Institute of Molecular Biology of Barcelona (IBMB–CSIC), Barcelona Science Park, Helix Building, Baldiri Reixac 15, 08028 Barcelona, Spain, sICREA: Institucio´ Catalana de Recerca i Estudis Avanc¸ats, Pg. Lluis Companys 23, 08010 Barcelona, Spain, and tServier Research Institute in Croissy-sur-Seine, 125 Chemin de Ronde, 78290 Croissy, France. *Correspondence e-mail: eugene.krissinel@stfc.ac.uk Keywords: CCP4; CCP4 Cloud; macromolecular crystallography; distributed computing; data management; project management. Nowadays, progress in the determination of three-dimensional macromolecular structures from diffraction images is achieved partly at the cost of increasing data volumes. This is due to the deployment of modern high-speed, high- resolution detectors, the increased complexity and variety of crystallographic software, the use of extensive databases and high-performance computing. This limits what can be accomplished with personal, ofﬂine, computing equipment in terms of both productivity and maintainability. There is also an issue of long- term data maintenance and availability of structure-solution projects as the links between experimental observations and the ﬁnal results deposited in the PDB. In this article, CCP4 Cloud, a new front-end of the CCP4 software suite, is presented which mitigates these effects by providing an online, cloud-based environment for crystallographic computation. CCP4 Cloud was developed for the efﬁcient delivery of computing power, database services and seamless integration with web resources. It provides a rich graphical user interface that allows project sharing and long-term storage for structure-solution projects, and can be linked to data-producing facilities. The system is distributed with the CCP4 software suite version 7.1 and higher, and an online publicly available instance of CCP4 Cloud is provided by CCP4. Acta Cryst. (2022). D78, 1079–1089 Published under a CC BY 4.0 licence research papers CCP4 Cloud for structure determination and project management in macromolecular crystallography 1.2. Computational resources MX structure solution is a complex process, and consider- able effort has been invested in its automation. Today, most structures can be solved automatically, yet the corresponding algorithms can take hours and even days to be executed on ordinary PCs. The limitations are particularly apparent if large volumes of data, as are typical for modern beamlines, need to be processed. Cloud-based computations can be parallelized on many computational nodes and thereby deliver the power of centralized high-performance computing (HPC) facilities to end users. Distributed computing is already being used within the ﬁeld of MX and cryo-EM. Some examples are as follows. (i) EMBL Hamburg provides the ARP/wARP server (Langer et al., 2008) for the automatic building of protein and nucleic acid structures, and the Auto-Rickshaw server (Panjikar et al., 2005), which offers an extensive collection of phasing methods. (i) EMBL Hamburg provides the ARP/wARP server (Langer et al., 2008) for the automatic building of protein and nucleic acid structures, and the Auto-Rickshaw server (Panjikar et al., 2005), which offers an extensive collection of phasing methods. (ii) Global Phasing Ltd runs online services for their GRADE (geometrical restraints generation; Smart et al., 2011), STARANISO (anisotropy of the diffraction limit and Bayesian estimation of structural amplitudes; Tickle et al., 2018) and PDBpeep (analysis of PDB data sets; Tickle et al., 2018) applications. research papers recent COVID-19 pandemic, when existing trends towards remote working received a truly dramatic acceleration. The main advantages offered by cloud-based setups include (i) zero maintenance burden for end users, (ii) access to signiﬁ- cant computational resources, (iii) efﬁcient data logistics and (iv) a long, potentially indeﬁnite, lifetime of the obtained results with highly efﬁcient access, sharing and dissemination possibilities. All of these are attractive for applications in macromolecular crystallography (MX). recent COVID-19 pandemic, when existing trends towards remote working received a truly dramatic acceleration. The main advantages offered by cloud-based setups include (i) zero maintenance burden for end users, (ii) access to signiﬁ- cant computational resources, (iii) efﬁcient data logistics and (iv) a long, potentially indeﬁnite, lifetime of the obtained results with highly efﬁcient access, sharing and dissemination possibilities. All of these are attractive for applications in macromolecular crystallography (MX). cised in practice. This makes it particularly important to present the overall workﬂow, including abstracted project structures and all derived data, in a clear and concise way. The problem becomes more obvious if more than one computer is used during a structure-solution project or if several researchers work on the same project. In such cases, a cloud- based setup is an elegant solution due to all projects and associated data being kept in a single instance accessible from any convenient device and geographic location. 1.4. Data security, safety, access and dissemination As a method, MX has seen steady and active development over the past several decades and has produced a considerable variety of computational approaches and exemplary pieces of software, allowing one to cope with all computational aspects of experimental MX studies. For example, the CCP4 software suite (Winn et al., 2011) version 8.0 presents a user with more than 500 executable modules and requires about 6.3 GB of disk space in a minimalistic setup. Full installation requires a set of additional databases and third-party software compo- nents, which require additional maintenance and increase the disk-space demand to over 50 GB. In a distributed framework, all software and databases are stored and maintained remotely (i.e. in the cloud) and thus are always kept in a ready and duly updated state. While data security in the cloud represents a topic of continual concern and research, signiﬁcant progress has been achieved in this area during the last decade (Ramachandra et al., 2017). Modern centralized setups are protected from attacks on a considerably higher level than personal computing items, which in fact represent the weak link in the security system. Data in the cloud, running over, for example, the company’s internal network, are as safe as anywhere else on the company’s premises. The lifetime of data stored in the cloud is considerably longer (effectively inﬁnite) than that usually achieved with locally maintained hardware. Equally importantly, a cloud-based setup provides efﬁcient access to its resources, which can be conveniently exploited for archiving purposes. In addition to the ﬁnal results, note that whole structure-solution projects can be archived to be provided as complementary information for scientiﬁc publishing, which is potentially useful for revising and revisiting projects, or subsequently repurposed, for example as educational mate- rial. 1. Introduction Cloud-based computation has gained increasing popularity and has rapidly become the modus operandi almost every- where. The signiﬁcance and advantages of distributed, server- based computing have become particularly clear owing to the https://doi.org/10.1107/S2059798322007987 1079 Acta Cryst. (2022). D78, 1079–1089 Acta Cryst. (2022). D78, 1079–1089 2.1. Front-End Server (iii) Bias-free MR with fragments: ARCIMBOLDO (Milla´n et al., 2015). The Front-End Server (FE) is the central element of the system that is responsible for the overall data logistics (but not computations) in the system. The FE keeps user accounts, structure-solution projects and data, maintains connection with UIs on user machines, provides access to external data resources and web services, prepares data and metadata for running jobs on NCs, dispatches jobs to NCs and balances their load, and relays communication between the UI and any running jobs. Typically, a CCP4 Cloud setup has a single FE, but the system may be conﬁgured to use several FEs if necessary for coping with extra workload or providing groups of users with their own gateways. (iv) MR with ab initio modelling: AMPLE (Bibby et al., 2012). (v) MR with database searches: SIMBAD (Simpkin et al., 2018). (vi) Validation and analysis: PISA (Krissinel, 2015) and Zanuda (Lebedev & Isupov, 2014). This compilation of online resources is far from complete, yet the number and variety of existing developments does suggest that cloud-based computing represents an attractive solution that is already appreciated by the MX research community. All MX-related web services that we are aware of are designed to perform speciﬁc tasks. Even taken together, they do not form a complete system for solving MX structures, and there is no single facility for developing full MX projects online. Individual web services were not generally developed with a common framework for data exchange in mind, which is not convenient for their integration. Nevertheless, a few web services were integrated as part of the WestLife initiative (Morris et al., 2019), and the BALBES and MoRDa servers at CCP4 can send phased structures to the ARP/wARP web server at EMBL Hamburg for subsequent model building, using ad hoc communication protocols. 2.2. Number Crunchers Number Crunchers (NCs) are servers whose sole task is to run computational jobs. They receive input data and execution instructions from the FE(s), perform calculations on the attached back-end and push results back to the FE that sent the task. NCs can be conﬁgured to work with virtually any computational back-end, for example, a cluster queue (SGE, SLURM etc.), another machine with SSH access or just a shell on the NC’s host machine. There must be at least one NC in a CCP4 Cloud setup. However, the system is designed to have multiple computa- tional nodes, and can be heterogeneous in respect to the hardware platforms, operating systems and back-ends used. NCs can accept jobs from multiple FEs, making sure that the results are pushed back to the sender. In this publication, we present CCP4 Cloud, the new online solution for MX computation developed by the Collaborative Computational Project Number 4 in protein crystallography (CCP4 UK). This development was aimed at providing a generic, complete, distributable, highly versatile and scalable framework for performing MX computations on a set of remote servers, which would allow the creation, development and storage of full MX structure-solution projects online. Further ambitions that have been realized include a rich, ergonomic graphical user interface with the possibility of sharing projects for simultaneous teamwork, seamless access to external web resources, a framework for accommodating software not licenced by CCP4, integrated documentation and educational materials, and the abilities to link to data- producing facilities and to run locally as a conventional desktop GUI. research papers reﬁnement from low-resolution data sets (O’Donovan et al., 2012). reﬁnement from low-resolution data sets (O’Donovan et al., 2012). data via the http(s) protocol. The components are combined into a working system by specifying the URL addresses of interacting partners in their conﬁguration ﬁles. data via the http(s) protocol. The components are combined into a working system by specifying the URL addresses of interacting partners in their conﬁguration ﬁles. (vii) The FragFit server for the prediction, visualization and selection of missing segments in protein cryo-EM density maps (Tiemann et al., 2018). (vii) The FragFit server for the prediction, visualization and selection of missing segments in protein cryo-EM density maps (Tiemann et al., 2018). There are four types of logical components in CCP4 Cloud: (i) Front End Server (FE), (ii) Number Cruncher (NC), (iii) CCP4 Cloud Client (CC) and (iv) User Interface (UI). The CCP4 Cloud conﬁguration is deﬁned by the number of logical components of each type in the system and their distribution over hardware hosts. Below, we brieﬂy consider the role of each component and a few representative conﬁguration templates. In addition, CCP4’s online services (Krissinel et al., 2018) comprise a collection of automated tools for the following. (i) Molecular replacement (MR): MrBUMP (Keegan & Winn, 2008), BALBES (Long et al., 2008) and MoRDA (Vagin & Lebedev, 2015). (ii) Experimental phasing: Crank2 (Skuba´k & Pannu, 2013) and SHELX (Sheldrick et al., 2012; Uso´n & Sheldrick, 2018). 1.3. Data logistics There are three main aspects related to data maintenance in MX: (i) the manipulation of the large volumes of data and numbers of ﬁles typical of crystallographic computing, (ii) maintaining the structure-solution workﬂow and presenting it in a form convenient for analysis and development, and (iii) supporting teamwork and coworking from different physical and virtual locations. Modern X-ray detectors produce signiﬁcant amounts of data, reaching terabyte scales per crystal. Ideally, such volumes should either be processed where they are produced or archived, without moving the data to local devices. In the cloud, this problem can be conveniently solved by linking the data-producing facility and the computing centre. The volumes of derived and intermediate data in crystallographic projects are usually not so critically large compared with the raw data, but they may be difﬁcult to maintain due to the large number of ﬁles and the nature of the highly iterative, trial-and-error, approach that is often exer- ) (iii) The TLS server at the University of Washington (Painter & Merritt, 2006), which identiﬁes rigid bodies in crystal structures and parametrizes their ‘motions’. (iv) The PISA server at the EBI (Krissinel & Henrick, 2007), which analyses crystal packing, macromolecular inter- faces and protein oligomeric states. (v) The PDB-REDO server at the Netherlands Cancer Institute (Joosten et al., 2014), which optimizes crystallo- graphic structures with automated algorithms for reﬁnement, rebuilding and validation. (vi) The SBGrid web portal, which provides access to the Wide Search Molecular Replacement (Stokes-Rees & Sliz, 2010) and Deformable Elastic Network tools for structure 1080 Eugene Krissinel et al. CCP4 Cloud Acta Cryst. (2022). D78, 1079–1089 research papers 2.3. CCP4 Cloud Client CCP4 Cloud Client (CC) is a special type of NC that runs on the user’s device and enables the use of CCP4 graphical applications with CCP4 Cloud: Coot (Emsley et al., 2010), CCP4mg (McNicholas et al., 2011), DUI (Winter et al., 2018), iMosﬂm (Battye et al., 2011) and ViewHKL (Evans & Kris- sinel, unpublished work). In the current version of CCP4, these applications cannot be executed within a browser and thus must run from a local installation of the CCP4 software suite. The CC provides a seamless integration of the local CCP4 installation with the rest of CCP4 Cloud, making all communication and data exchange completely automatic. 2. CCP4 Cloud architecture and implementation CCP4 Cloud can be viewed as a set of logical components which function independently of each other and exchange Eugene Krissinel et al. CCP4 Cloud 1081 research papers makes uploading them from the user’s machine to remote servers impractical. Therefore, in addition to image-processing GUIs (DUI, iMosﬂm and optionally XDSGUI; Sparta et al., 2016), the CC can also run the automatic image-processing pipeline xia2 (Winter, 2010) locally (note that xia2 can also be run in a remote NC using diffraction images uploaded by other means; for example, directly from a synchrotron beamline via a special link in the background). perspective, acts like any other program with a GUI. This setup is suitable, for example, for users who do not wish to transmit their data to remote servers. The single-host conﬁg- uration is included in CCP4 version 7.1 or higher and works out of the box. Fig. 1(b) presents a schematic of a CCP4 Cloud conﬁgura- tion that might be suitable for medium-sized research laboratories. This conﬁguration extends that in Fig. 1(a) by exposing the FE to the internet, which allows multiple users to connect to it from their own machines. Both the FE and NC run on the same machine, typically a powerful workstation or a cluster’s head node. The CC launcher is distributed as part of the CCP4 software suite. Besides starting the CC, it also starts a chosen web browser and connects to CCP4 Cloud automatically. This means that starting a full-featured CCP4 Cloud session is as easy as starting a desktop GUI. An example of the CCP4 Cloud setup maintained at the CCP4 site in Rutherford Appleton Laboratory (RAL) Harwell is shown in Fig. 1(c). In this type of setup the FE and NCs are placed on dedicated hardware hosts. This allows a high level of responsiveness for the users, whilst achieving a high number of running and queuing jobs. The FE has access to a dedicated ﬁle system (‘Data’ in Fig. 1c) that is visible to users as a read-only ‘Cloud Storage’. This storage is different from the disk space allocated for user data and projects. The main purpose of Cloud Storage is to keep large volumes of data collected at the synchrotron and make them available for the data-processing pipeline xia2 in user projects. Each user can see only the part of Cloud Storage that was made acces- sible to them by the CCP4 Cloud administrator. In addition, Cloud Storage contains public areas that are usually used for keeping demo projects, tutorials and similar resources. 2.5. CCP4 Cloud configuration and typical deployment scenarios 2.5. CCP4 Cloud configuration and typical deployment scenarios Server nodes (FEs, NCs and CCs) are conﬁgured indivi- dually. They are plugged into the CCP4 Cloud installation by specifying the URL addresses of the communicating partners in conﬁguration ﬁles. For example, an FE conﬁguration ﬁle contains the URLs of the NCs that can accept jobs from that FE. This allows the CCP4 Cloud conﬁguration to be modiﬁed very easily and to scale dynamically according to the current workload. NCs may be switched on/off without stopping the whole system, with the FE reallocating new jobs automatically. This mechanism is used for working with CCs, which are added automatically when the user connects to CCP4 Cloud from a CC-enabled device and are automatically removed from the system when the user ends their session. research papers CCP4 Cloud at RAL Harwell is publicly available and can be conveniently accessed via the pre-conﬁgured CC that is included as part of the CCP4 software suite. Although very practical, CCs are optional elements of CCP4 Cloud. They require a local installation of the CCP4 software suite and so cannot be used when working from a device that cannot run CCP4 software (for example a tablet or smartphone). A session without a CC can be started by navigating to the CCP4 Cloud URL from within a web browser (including from a device without a CCP4 installation, such as a tablet or smartphone). This can be useful for reviewing a project, checking on the progress of a running job or launching non-GUI tasks when away from the desk. The FE server automatically detects whether a CC is running on the user’s device and enables the corresponding tasks accordingly. 2.4. User Interface CCP4 Cloud is operated via a graphical User Interface (UI) in a common web browser running on the user’s device. The interface is presented in more detail below. The CCP4 Cloud conﬁguration in Fig. 1(d) provides indi- vidual FEs for each user or groups of users but utilizes a common line of NCs for running jobs. This scheme is suitable in special situations, for example when user data must be kept strictly conﬁdential, in predeﬁned locations (such as user home directories) or protected by ﬁle-system permissions. This is achieved by running individually conﬁgured FEs under the corresponding user permissions. 2. CCP4 Cloud architecture and implementation 2. CCP4 Cloud architecture and implementation CCs are important for image-processing tasks. Typically, X-ray diffraction images have a considerable size, which CCP4 Cloud can be viewed as a set of logical components which function independently of each other and exchange Eugene Krissinel et al. CCP4 Cloud 1081 1081 1081 Acta Cryst. (2022). D78, 1079–1089 2.6. Testing Principles of the CCP4 Cloud interface and project development for this: ﬁrstly, the CCP4 Cloud ﬁle system is not accessible to users, and secondly a suitable choice of data-object types simpliﬁes the overall data logistics. As a result, speciﬁcation of job input becomes very minimal and in many cases automatic. A data object may correspond to a ﬁle, but it can also repre- sent a selected part of a ﬁle or be a composition of several ﬁles. At a low level, data objects are represented by JSON- formatted ﬁles containing references to ﬁles with all necessary metadata such as selection ranges. Once produced, the ﬁles never change and remain in the original job directories; instead, it is data objects that are communicated within the system. Interconversion between objects and ﬁles is performed automatically as an internal part of computational In CCP4 Cloud, work is organized into Projects. Within a Project there is a single tree of Jobs (the Project Tree), which provides a logical, rather than chronological, representation of the structure-solution pathway. This organizational concept is reinforced by using the concept of a Structure Revision, which will be described later. 2.6. Testing Testing is made an integral part of CCP4 Cloud and is performed automatically on a nightly basis. This was found to be necessary due to the rapid development of both CCP4 Cloud and the underlying CCP4 software suite. The test system is based on the Selenium framework (Islam & Quadri, 2020), which is suitable for automatically imitating a user’s actions in a web browser, such as logging on to CCP4 Cloud, creating projects, setting up and starting tasks. Each test represents the development of a project with a few jobs, and subsequent checks assert that all jobs complete and produce results within set tolerances. Test projects are added to the system with every new task, and most programs are tested in multiple projects featuring different contexts or exploiting different modes and combinations of parameters. CCP4 Cloud can be conﬁgured in a number of different ways and tailored to suit a particular situation, determined mostly by the number and type of available computational back-ends, disk storage for user projects and data, and the mechanism for accessing X-ray diffraction images. Fig. 1 presents a few typical CCP4 Cloud conﬁgurations. In the simplest case (Fig. 1a) all CCP4 Cloud components are placed on a single hardware host, which can be a laptop or a work- station. In this conﬁguration, CCP4 Cloud runs all computa- tional jobs on the host machine and, from the user’s 1082 Eugene Krissinel et al. CCP4 Cloud Acta Cryst. (2022). D78, 1079–1089 Acta Cryst. (2022). D78, 1079–1089 research papers for this: ﬁrstly, the CCP4 Cloud ﬁle system is not accessible to users, and secondly a suitable choice of data-object types simpliﬁes the overall data logistics. As a result, speciﬁcation of job input becomes very minimal and in many cases automatic. A data object may correspond to a ﬁle, but it can also repre- sent a selected part of a ﬁle or be a composition of several ﬁles. At a low level, data objects are represented by JSON- formatted ﬁles containing references to ﬁles with all necessary metadata such as selection ranges. Once produced, the ﬁles never change and remain in the original job directories; instead, it is data objects that are communicated within the system. Interconversion between objects and ﬁles is performed automatically as an internal part of computational 3. Principles of the CCP4 Cloud interface and project development 3. research papers research papers See the discussion in Section 3. Data object ID† File equivalent Unmerged reﬂection data UNMERGED MTZ or HKL ﬁle with unmerged reﬂection data Merged reﬂection data HKL Relevant columns of MTZ ﬁle with merged reﬂection data (e.g. h, k, l, F, sigF) or SCA ﬁle Macromolecular sequence SEQUENCE Sequence ﬁles in FASTA or PIR format Macromolecular coordinates XYZ PDB or mmCIF coordinate ﬁle Phases PHASES Relevant columns of MTZ ﬁle (e.g. FWT, PHWT) Ligand description LIGAND CIF ﬁle with ligand description Ligand library LIB CIF or LIB ﬁle with multiple ligand descriptions MR model MODEL Combination of sequence and single-model coordinate ﬁles MR ensemble ENSEMBLE Combination of sequence and multi-model coordinate ﬁles Structure STRUCTURE Combination of related XYZ and PHASES objects Structure revision REVISION Combination of HKL, SEQUENCE and STRUCTURE objects † Used for automatic naming of data objects in CCP4 Cloud. † Used for automatic naming of data objects in CCP4 Cloud. Jobs; this is seamlessly abstracted from the user. Whenever a data object is created on output of a Job, download links are provided for all ﬁles that it refers to. Table 1 lists the data objects currently used in CCP4 Cloud and their ﬁle equiva- lents. All data objects are named automatically and uniquely using templates shown in Fig. 2. default values, including input data objects. In most cases, input data are taken from the Structure Revision that was generated by the previous Job, making the data choice auto- matic and unambiguous. Default values for other data objects are chosen with some heuristics and should always be checked before running a task. The job output (Fig. 3b) is placed in a dedicated tab and shows the progress of the job in real time. Real-time moni- toring is useful for early termination of a job if the output suggests that an acceptable solution will not be achieved. Job reports provide a graphical representation of results, a description of any data objects created by the Job, as well as links to the in-browser graphical viewer UglyMol (Wojdyr, 2017), which allows quick inspection of structural models and electron-density maps. Where appropriate, Job reports include Verdict and Quality Assessment sections. Verdicts present achieved scores, such as R factors, completeness, number of clashes, Ramachandran outliers etc., along with suggestions for the modiﬁcation of job parameters and follow-up steps. 3.1. Data objects Unlike most desktop GUIs, the CCP4 Cloud UI operates with data objects, rather than ﬁles. There are two main reasons Figure 1 Examples of CCP4 Cloud conﬁgurations. FE, Front End; NC, Number Cruncher; UI, User Interface with optional CC (client-side NC); Data, access t collected X-ray diffraction data sets and other data used for MX structure determination. Arrows denote http(s) connections. (a) Single-hos conﬁguration, suitable for an individual working without an internet connection. (b) A multi-user setup using a central host machine, suitable for small t medium-sized laboratories. (c) Fully distributed, multi-component setup with single point of access, suitable for large facilities and research centres; i can optionally allow data acquisition from external sources. (d) Fully distributed setup with multiple access points. Conﬁgurations (a) and (c) work out o the box using the CCP4 software suite installed on the user’s machine [using the CCP4 Cloud setup at RAL Harwell in case (c)]. All conﬁgurations ma be tailored to speciﬁc conditions and requirements and can be built on top of the CCP4 Cloud components that are included in the CCP4 software suite See the discussion in the text. Figure 1 Figure 1 Examples of CCP4 Cloud conﬁgurations. FE, Front End; NC, Number Cruncher; UI, User Interface with optional CC (client-side NC); Data, access to collected X-ray diffraction data sets and other data used for MX structure determination. Arrows denote http(s) connections. (a) Single-host conﬁguration, suitable for an individual working without an internet connection. (b) A multi-user setup using a central host machine, suitable for small to medium-sized laboratories. (c) Fully distributed, multi-component setup with single point of access, suitable for large facilities and research centres; it can optionally allow data acquisition from external sources. (d) Fully distributed setup with multiple access points. Conﬁgurations (a) and (c) work out of the box using the CCP4 software suite installed on the user’s machine [using the CCP4 Cloud setup at RAL Harwell in case (c)]. All conﬁgurations may be tailored to speciﬁc conditions and requirements and can be built on top of the CCP4 Cloud components that are included in the CCP4 software suite. See the discussion in the text. Eugene Krissinel et al. CCP4 Cloud 1083 Acta Cryst. (2022). D78, 1079–1089 research papers Jobs can see data objects produced by all their ancestors in the Project Tree, except for Structure Revisions, which can be taken only from ﬁrst ancestors (i.e. the immediately previous Job higher up in the Project Tree, which is always unique). A Structure Revision is a special data object which provides the latest representation of the crystal (reﬂection and sequence data, atomic coordinates and phases) obtained at a particular point of the Project Tree. The limited scope of the visibility of Structure Revisions is necessary for keeping the Project Tree logically branched. Thus, if calculations need to be repeated with the same input data but modiﬁed parameters or using a different program, it can be performed simply by creating a new job from the point of decision (and only from that point), resulting in a new branch being created within the Project Tree. Figure 2 Naming conventions for data objects (a) and structure revisions (b) in CCP4 Cloud. All ﬁelds are generated automatically, and only ‘name’ can be overridden by the user. Sufﬁxation is used when a job produces multiple outputs, for example for the original and inverted hands in experimental phasing. research papers Whenever a structural model is modiﬁed, a Quality Assess- ment is performed that utilizes a standard set of programs, including MolProbity (Chen et al., 2010), EDStats (Tickle et al., 1998) and B-factor analysis (Winn et al., 2011), and presents their results in a summary table. 3.3.3. Definition of the asymmetric unit. This task sets the expected content of the asymmetric unit and performs Matthews analysis (Matthews, 1968). As a result, the initial Structure Revision is created, containing only the reﬂection data set and asymmetric unit composition. Having an initial- ized Structure Revision is necessary to proceed to the following stages. 3.3.3. Definition of the asymmetric unit. This task sets the expected content of the asymmetric unit and performs Matthews analysis (Matthews, 1968). As a result, the initial Structure Revision is created, containing only the reﬂection data set and asymmetric unit composition. Having an initial- ized Structure Revision is necessary to proceed to the following stages. 3.2. Jobs CCP4 Cloud Jobs are usually centred around primary components, such as Phaser (McCoy et al., 2007), REFMAC (Murshudov et al., 2011), Coot etc., and use several more programs behind the scenes to perform ﬁle-format inter- conversions, prepare intermediate data ﬁles and analyse results. This organiza- tion results in the user being provided with a reﬁned set of crystallographic tasks, which is more digestible than an exhaustive and overwhelming list of software tools. Task interfaces follow a common pattern (Fig. 3a): all input data objects are placed in the upper part of the interface, with task parameters below them. All input ﬁelds are given g Naming conventions for data objects (a) and structure revisions (b) in CCP4 Cloud. All ﬁelds are generated automatically, and only ‘name’ can be overridden by the user. Sufﬁxation is used when a job produces multiple outputs, for example for the original and inverted hands in experimental phasing. 1084 Eugene Krissinel et al. CCP4 Cloud Acta Cryst. (2022). D78, 1079–1089 3.3. Project development g g 3.3.4. Solving the phase problem. Initial phases may be found with either molecular replacement or experimental phasing, or their combination. CCP4 Cloud includes both automatic structure solvers (MrBUMP, MoRDa, BALBES, ARCIMBOLDO and Crank2) and fundamental tasks (SHELX, Phaser and MOLREP; Vagin & Teplyakov, 2010). Obtained approximations for phases and atomic coordinates are added to the Structure Revision to be improved in subsequent stages. CCP4 Cloud projects (exempliﬁed in Fig. 4) are developed by adding Jobs to the nodes of the Project Tree. This should be performed while being mindful of the visibility of produced data objects for subsequent Jobs, as described above. If data objects required for a Job are not found in the current branch of the tree, then such a Job cannot be added. Typically, CCP4 Cloud projects contain a few stages. 3.3.1. Data import. Data ﬁles may be uploaded to CCP4 Cloud from a local machine or imported from Cloud Storage. The main function of these tasks is to check the validity of ﬁles and create associated data objects. Diffraction images are not uploaded to CCP4 Cloud due to their considerable size. Instead, they are processed locally on the CC with xia2, DUI, iMosﬂm or XDSGUI, and the resulting reﬂection data sets are imported automatically at the end. 3.3.1. Data import. Data ﬁles may be uploaded to CCP4 Cloud from a local machine or imported from Cloud Storage. The main function of these tasks is to check the validity of ﬁles and create associated data objects. Diffraction images are not uploaded to CCP4 Cloud due to their considerable size. Instead, they are processed locally on the CC with xia2, DUI, iMosﬂm or XDSGUI, and the resulting reﬂection data sets are imported automatically at the end. 3.3.5. Density modification. CCP4 Cloud includes several tasks for density improvement: Parrot (Cowtan, 2010), SHELXE (Uso´n & Sheldrick, 2018) and ACORN (Yao et al., 2005). These tasks result in a new version of the Structure Revision in which the density map(s) are replaced accordingly. 3.3.6. Model building and refinement. Usually, the struc- ture is built automatically in the ﬁrst attempt using one of a few available tools: Buccaneer (Cowtan, 2006), ARP/wARP, CCP4Build (Krissinel & Lebedev, unpublished work) and ModelCraft (Bond & Cowtan, 2022). However, ultimately the structural model will require multiple manipulations with REFMAC and Coot. 3.3. Project development The process is highly iterative, and after running each model-building or reﬁnement task a new 3.3.6. Model building and refinement. Usually, the struc- ture is built automatically in the ﬁrst attempt using one of a few available tools: Buccaneer (Cowtan, 2006), ARP/wARP, CCP4Build (Krissinel & Lebedev, unpublished work) and ModelCraft (Bond & Cowtan, 2022). However, ultimately the structural model will require multiple manipulations with REFMAC and Coot. The process is highly iterative, and after running each model-building or reﬁnement task a new 3.3.2. Data preparation. This stage includes tasks such as scaling and merging reﬂections imported from unmerged MTZ ﬁles with AIMLESS (Evans & Murshudov, 2013) and the preparation of models for molecular replacement using MrParse (Simpkin et al., 2022), MrBUMP, CCP4mg and similar. Acta Cryst. (2022). D78, 1079–1089 Eugene Krissinel et al. CCP4 Cloud 1085 Figure 3 Schematic of CCP4 Cloud task interfaces. (a) input panel, (b) output panel (Job report). Eugene Krissinel et al. CCP4 Cloud 1085 Acta Cryst. (2022). D78, 1079–1089 4. CCP4 Cloud as a community resource 4. CCP4 Cloud as a community resource CCP4 has a mission to facilitate research in MX primarily by maintaining, developing and distributing the integrated CCP4 software suite, but also by educating and training scientists in experimental structural biology and via the wide dissemina- tion of new ideas, techniques and practices. CCP4 Cloud ﬁts the CCP4 mission very neatly. research papers by a CCP4 Cloud administrator per user request. There is no expiry date for Projects in CCP4 Cloud. Structure Revision is created that contains the improved atomic coordinates and electron-density maps. 3.3.7. Validation, analysis and deposition. At the end of the structure-solution process, atomic coordinates from the last Structure Revision, along with reﬂection data, sequences and ligand descriptions, should be deposited in the Protein Data Bank (PDB; Berman et al., 2000). Deposition ﬁles, in mmCIF format, are prepared with a dedicated CCP4 Cloud task. However, before deposition, and iteratively during the model- building and reﬁnement procedure, the structure should be validated using the tools available in Coot and by inspection of the Quality Assessment section in the Job report for the REFMAC task. Furthermore, the space-group hypothesis may be checked using Zanuda, carbohydrate structures are vali- dated with Privateer (Agirre et al., 2015) and additional analysis and validation may be performed with PISA. 1086 Eugene Krissinel et al. CCP4 Cloud 4.3. Documentation and educational capabilities CCP4 Cloud documentation articles are written in RST format (https://docutils.sourceforge.io/rst.html) and placed in a dedicated GitLab repository. On modiﬁcation, documenta- tion is automatically compiled and moved into appropriate parts of the system, becoming linked with graphical interfaces. The documentation includes four main parts. CCP4 Cloud documentation articles are written in RST format (https://docutils.sourceforge.io/rst.html) and placed in a dedicated GitLab repository. On modiﬁcation, documenta- tion is automatically compiled and moved into appropriate parts of the system, becoming linked with graphical interfaces. The documentation includes four main parts. Designing a modern, data-driven interface was one of the main ambitions of the CCP4 Cloud project. CCP4 Cloud’s GUI eliminates the burden of managing many ﬁles, as are typically produced in MX projects, by abstracting them with composite data types and managing all data ﬂows within CCP4 Cloud. MX projects are presented in the form of branched Project Trees, using a logically clear system for naming the data objects produced during the course of a project. Jobs are highlighted with summary statistics where appropriate to qualify or quantify Job success. Altogether, this visual repre- sentation makes projects quickly and easily navigable. Project development is made simple by way of growing the Project Tree via adding and removing Jobs with primitive graphical actions. Project branching is made unambiguous by introdu- cing the concept of Structure Revisions, which can be considered as check points corresponding to speciﬁc states of the structure-solution workﬂow. (i) Task reference contains articles related to speciﬁc tasks, focusing on their position in the overall structure-solution workﬂow. This includes method descriptions, which should be sufﬁcient for the general understanding of the main para- meters. Where possible, links to original publications and the author’s manuals are provided for in-depth investigation. (ii) User guide aims at explaining the main principles of the CCP4 Cloud graphical interface. (iii) Developer reference is available only to users tagged as ‘developer’ and contains setup instructions and a description of the framework for writing task interfaces. (iv) Tutorials include examples of structure-solution projects demonstrating the use of structure-solution tech- niques in CCP4 Cloud, such as data processing, automatic and fundamental MR and experimental phasing, MR-SAD, reﬁnement and model building etc. Tutorials are presented as seed projects with pre-loaded data, which users can import into their accounts. The user would then develop the project, following the instructions in the accompanying document. 5. Discussion and conclusions 5. Discussion and conclusions Although CCP4 Cloud does not currently expose all compo- nents of the CCP4 software suite through its graphical inter- face, it represents a sufﬁciently complete system for structure solution and includes all principal stages of MX computing from image processing to model deposition. A few alternatives exist for each stage, leading to functional redundancy but a higher probability of solving structures in difﬁcult cases for the novice or experienced MX user. We therefore argue that CCP4 Cloud is functionally complete for use in the majority of practical cases. CCP4 Cloud is distributed as part of the CCP4 software suite (version 7.1 and higher) and may be used out of the box for working with the public instance of CCP4 Cloud maintained at RAL Harwell (https://cloud.ccp4.ac.uk). In addition, any CCP4 7.1+ setup may be conﬁgured as a server featuring a local instance of CCP4 Cloud, allowing users to beneﬁt from in-house computational resources and avoid transmitting sensitive data to public servers. CCP4 Cloud can be installed on a range of hardware plat- forms and operating systems (Linux, Mac OS X and MS Windows) and allows mixed setups including individual workstations, in-house clusters, centralized HPC facilities and generic clouds. CCP4 Cloud setups can be adjusted to a wide range of available resources; in the simplest conﬁguration, CCP4 Cloud acts as a mere GUI on a single laptop or PC. 4.1. Integration of third-party resources and software CCP4 Cloud is designed as a highly synergetic platform for the integration of software components, databases and web services from CCP4 and other projects. The system includes generic frameworks for importing data from external sources [for example, the wwPDB (Berman et al., 2003) and the AlphaFold database (Varadi et al., 2022)] and running programs that are not included in the CCP4 software suite, such as XDS (Kabsch, 2010) and XDSGUI, OpenFold (Ahdritz et al., 2021) or AlphaFold2 (Jumper et al., 2021) for structure prediction, and BUSTER from Global Phasing Ltd (Bricogne et al., 2017) for model reﬁnement. The frameworks include mechanisms for requesting or validating user licences online (for example, authorization from Global Phasing Ltd is required before the BUSTER task is made available to a user). By assembling complex software in an integrated and harmonized system with 24/7 availability, CCP4 Cloud facil- itates the delivery of novel methods and developments to a wide structural biology community. CCP4 Cloud Projects are objects themselves that can be exported as single ﬁles (ZIP archives), encapsulating all Jobs, data and metadata. The exported Projects are cross-platform compatible with all Microsoft Windows, Linuxes and Mac OS X. They can be re-imported in any instance of CCP4 Cloud; for example, that running on the user’s own machine. Note that importing/exporting Projects is different from project sharing, which is described below. The main purpose of this facility is to mitigate limitations of the disk allocation quota given to users in centralized installations of CCP4 Cloud and also to serve as an additional backup. For example, in the CCP4 Cloud instance maintained at RAL Harwell, disk space is initially capped at 15 GB (conﬁgurable). If all of the space is consumed, no new Jobs can be created. By exporting completed Projects and deleting them in CCP4 Cloud, their disk space is released back to the user. If a Project size grows beyond the initial allocation, the disk quota can be increased Figure 4 Screenshot of a CCP4 Cloud project. An indication of the principal structure solution stages is displayed on the left. See the discussion in the text. 1086 Eugene Krissinel et al. CCP4 Cloud Acta Cryst. (2022). D78, 1079–1089 research papers Eugene Krissinel et al. CCP4 Cloud 4.2. Teamwork CCP4 Cloud documentation to facilitate education in MX either on a local (for example a university) or a global level. 4.2. Teamwork CCP4 Cloud provides a convenient environment for orga- nizing and storing structure-solution projects conducted by a research group or institution. Once in CCP4 Cloud, the projects are automatically backed up, and their lifetime is only limited by local data policies and IT support. All metadata in CCP4 Cloud are versioned and designed to be backward compatible: completed projects can therefore be revisited and revised or extended at any point in the future. CCP4 Cloud projects can be shared between two or more CCP4 Cloud users. In this way, several researchers can work on a project simultaneously and see each other’s actions in real time. This unique feature is particularly convenient for distributed teamwork, exchange between collaborating organizations, training young scientists and reporting software bugs. Subject to user consent, failed jobs with the corresponding input data can be automatically retained in a dedicated area and reported to the CCP4 Cloud maintainer. This mechanism, in combi- nation with project sharing, was found to be extremely useful for software debugging and obtaining feedback from the research community, which is essential for further develop- ment of CCP4 software. Acta Cryst. (2022). D78, 1079–1089 research papers will greatly depend on beamline automation and the perfor- mance of automatic structure-solution pipelines. Automatic methods have improved quite noticeably during the last decade, and it is expected that further advances will be asso- ciated with the availability of AlphaFold-predicted models for use in molecular replacement (Milla´n et al., 2021; Pereira et al., 2021; McCoy et al., 2022; Medina et al., 2022). In addition, the logically strict concepts behind project structure in CCP4 Cloud provide good potential for automating project devel- opment and adjusting to the variety of data properties and structural features. Progress towards the above targets represents yet another direction of future work. Docs or GitHub. This feature is also important for education and training, as CCP4 Cloud can store virtually indeﬁnite volumes of relevant documents and data integrated with demo projects. Thirdly, CCP4 Cloud provides a framework for the long-term storage of structure-solution projects and results, which is maintained to industry standards. This feature is important for research projects that span many years and helps data management in research laboratories and compa- nies. The signiﬁcance of retaining as much data as possible cannot be overestimated. Quite often, structural experiments cannot be repeated under identical conditions, and misplaced data may not be recovered. Historically, the PDB aimed to keep only the ﬁnal atomic coordinates, which is sufﬁcient for most downstream studies. From 1999, merged reﬂection data could be deposited in addition to atomic coordinates; this became mandatory in 2008 (Burley et al., 2017). This move was useful for the development of crystallographic software, but also allowed researchers to investigate structural data on a deeper level, permitting alternative interpretation of electron- density maps and further structure improvement in some cases (see, for example, PDB-REDO; Joosten et al., 2009). The deposition of unmerged reﬂection data, which started in 2020, is another important step in this direction. Thus, the PDB is reaching a state where it captures experimental data in the form of indexed reﬂections and ﬁnal interpretation in the form of an atomic model and electron-density maps. However, the PDB does not capture the way in which this interpretation was performed: information about the workﬂow and modelling decisions made during a structure-solution project is lost. We would like to advocate the idea that retaining structure- solution projects is as important as retaining experimental data. 6. Data access Source code for CCP4 Cloud is included in the CCP4 software suite version 7.1 and higher, as obtainable from https:// www.ccp4.ac.uk. It can also be found in GitLab repositories at https://gitlab.com/CCP4/jsCoFE. CCP4 Cloud documentation is maintained in GitLab at https://gitlab.com/CCP4/jscofe-doc. CCP4 Cloud and CCP4 Cloud documentation are provided free to academic users under the terms of the general CCP4 Licence (https://www.ccp4.ac.uk/licensing/academic_ software_licence.pdf). research papers This can be useful for the subsequent understanding of ambiguous structural features, revisiting and revising struc- tural studies, education, facilitating the peer-review publica- tion process, training and software development. We further suggest that CCP4 Cloud provides a convenient framework for archiving structure-solution projects, complementing data repositories at the wwPDB, and we consider the imple- mentation of such archival as one possible future direction. 4.3. Documentation and educational capabilities While CCP4 Cloud might be perceived as a mere web-based HPC-powered GUI for the CCP4 software suite, there are signiﬁcant differences upon closer inspection. Firstly, the centralized online setup provides a natural basis for combining CCP4, web resources and third-party software. Examples include the PDB, the AlphaFold database at the EBI and software from Global Phasing and AlphaFold2. Secondly, CCP4 Cloud provides an environment for teamwork, where team members can be located anywhere but can work on the same projects, similar to other online systems such as Google (v) CCP4 Cloud documentation is set as a parallel project with moderated access for contributors, which makes it easy to adjust content for rapidly developing tasks, add new articles and make changes instantly available to end users. The tutorial framework is widely used at CCP4 MX schools and in some university courses. Any relevant tutorials with demo projects (including data) can be placed by participating contributors in (v) CCP4 Cloud documentation is set as a parallel project with moderated access for contributors, which makes it easy to adjust content for rapidly developing tasks, add new articles and make changes instantly available to end users. The tutorial framework is widely used at CCP4 MX schools and in some university courses. Any relevant tutorials with demo projects (including data) can be placed by participating contributors in Eugene Krissinel et al. CCP4 Cloud 1087 Acta Cryst. (2022). D78, 1079–1089 Acknowledgements The authors are grateful to all CCP4-associated developers and to CCP4 Working Group 2 for their suggestions and useful discussions. Many feedback and bug reports from users have helped to bring CCP4 Cloud to a stable production state. Special thanks to Global Phasing Ltd for providing access to the BUSTER reﬁnement software in kind for use by CCP4 Cloud’s academic users. The third-party inclusion framework was developed in cooperation with the ARP/wARP team at EMBL Hamburg. The CCP4 Cloud setup at RAL Harwell is hosted in the Research Complex at Harwell, with the help of RCaH IT support. Computational facilities are provided by CCP4, Scientiﬁc Computing Department of the Science and Facilities Research Council UK (STFC), STFC Cloud and IRIS Community Project. The full power of CCP4 Cloud is unlocked when it is conﬁgured to receive data directly from synchrotron beam- lines. The combination of MX facilities equipped with remote access, CCP4 Cloud and the wwPDB creates a closed online environment for macromolecular crystallography, which streamlines data logistics and maintenance. In its current implementation, CCP4 Cloud includes a facility to push data, such as reﬂections, sequences and models, to the user’s account and to then automatically create new structure- solution projects and start automatic structure-solution pipelines (see the details at https://cloud.ccp4.ac.uk/manuals/ html-userguide/jscofe_cloudrun.html). In an ideal hypo- thetical situation, a user could send samples to a CCP4 Cloud- connected synchrotron and then simply check the results of automatic data processing and structure solution in their CCP4 Cloud account (for example on their smartphone or tablet). The success and practical signiﬁcance of such a scheme Agirre, J., Iglesias-Ferna´ndez, J., Rovira, C., Davies, G. J., Wilson, K. S. & Cowtan, K. D. (2015). Nat. Struct. Mol. Biol. 22, 833–834. References Agirre, J., Iglesias-Ferna´ndez, J., Rovira, C., Davies, G. J., Wilson, K. S. & Cowtan, K. D. (2015). Nat. Struct. Mol. Biol. 22, 833–834. research papers g Burley, S. K., Berman, H. M., Kleywegt, G. J., Markley, J. L., Nakamura, H. & Velankar, S. (2017). Methods Mol. Biol. 1607, 627– 641. Sheldrick, G. M., Gilmore, C. J., Hauptman, H. A., Weeks, C. M., Miller, M. & Uso´n, I. (2012). International Tables for Crystallo- graphy, Vol. F, edited by E. Arnold, D. M. Himmel & M. G. Rossmann, pp. 413–429. Dordrecht: Kluwer Academic Publishers. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2010). Acta Cryst. D66, 12–21. Simpkin, A. J., Simkovic, F., Thomas, J. M. H., Savko, M., Lebedev, A., Uski, V., Ballard, C., Wojdyr, M., Wu, R., Sanishvili, R., Xu, Y., Lisa, M.-N., Buschiazzo, A., Shepard, W., Rigden, D. J. & Keegan, R. M. (2018). Acta Cryst. D74, 595–605. Cowtan, K. (2006). Acta Cryst. D62, 1002–1011. Cowtan, K. (2010). Acta Cryst. D66, 470–478. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. (2010). Acta Cryst. D66, 486–501. Simpkin, A. J., Thomas, J. M. H., Keegan, R. M. & Rigden, D. J. (2022). Acta Cryst. D78, 553–559. Evans, P. R. & Murshudov, G. N. (2013). Acta Cryst. D69, 1204–1214. Skuba´k, P. & Pannu, N. S. (2013). Nat. Commun. 4, 2777. Islam, M. N. & Quadri, S. M. (2020). Adv. Sci. Technol. 5, 226–232. Joosten, R. P., Long, F., Murshudov, G. N. & Perrakis, A. (2014). IUCrJ, 1, 213–220. Smart, O. S., Womack, T. O., Sharff, A., Flensburg, C., Keller, P., Paciorek, W., Vonrhein, C. & Bricogne, G. (2011). Grade, version Smart, O. S., Womack, T. O., Sharff, A., Flensburg, C., Keller, P., Paciorek, W., Vonrhein, C. & Bricogne, G. (2011). Grade, version 1.2.20. Global Phasing Ltd, Cambridge, United Kingdom. Paciorek, W., Vonrhein, C. & Bricogne, G. (2011). Grade, v 1.2.20. Global Phasing Ltd, Cambridge, United Kingdom. Joosten, R. P., Womack, T., Vriend, G. & Bricogne, G. (2009). Acta Cryst. D65, 176–185. Sparta, K. M., Krug, M., Heinemann, U., Mueller, U. & Weiss, M. S. (2016). J. Appl. Cryst. 49, 1085–1092. y Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Zˇ ı´dek, A., Potapenko, A., Bridgland, A., Meyer, C., Kohl, S. A. A., Ballard, A. research papers Ahdritz, G., Bouatta, N., Kadyan, S., Xia, Q., Gerecke, W. & AlQuraishi, M. (2021). aqlaboratory/openfold: Openfold v1.0.0. https://doi.org/10.5281/zenodo.5709540. Morris, C., Andreetto, P., Banci, L., Bonvin, A. M. J. J., Chojnowski, G., Cano, L. D., Carazo, M. J., Conesa, P., Daenke, S., Damaskos, G., Giachetti, A., Haley, N. E. C., Hekkelman, M. L., Heuser, P., Joosten, R. P., Kourˇil, D., Krˇenek, A., Kulha´nek, T., Lamzin, V. S., Nadzirin, N., Perrakis, A., Rosato, A., Sanderson, F., Segura, J., Schaarschmidt, J., Sobolev, E., Traldi, S., Trellet, M. E., Velankar, S., Verlato, M. & Winn, M. (2019). J. Struct. Biol. X, 1, 100006. Battye, T. G. G., Kontogiannis, L., Johnson, O., Powell, H. R. & Leslie, A. G. W. (2011). Acta Cryst. D67, 271–281. Berman, H. M., Henrick, K. & Nakamura, H. (2003). Nat. Struct. Mol. Biol. 10, 980. ( ) Murshudov, G. N., Skuba´k, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N. & Bourne, P. E. (2000). Nucleic Acids Res. 28, 235–242. Murshudov, G. N., Skuba´k, P., Lebedev, A. A., Pannu, N. S., Steiner, , , , , , , , , , R. A., Nicholls, R. A., Winn, M. D., Long, F. & Vagin, A. A. (2011). Acta Cryst. D67, 355–367. , Bibby, J., Keegan, R. M., Mayans, O., Winn, M. D. & Rigden, D. J. (2012). Acta Cryst. D68, 1622–1631. y O’Donovan, D. J., Stokes-Rees, I., Nam, Y., Blacklow, S. C., Schro¨der, G. F., Brunger, A. T. & Sliz, P. (2012). Acta Cryst. D68, 261–267. g ( ) y Painter, J. & Merritt, E. A. (2006). J. Appl. Cryst. 39, 109–111. ( ) Bond, P. S. & Cowtan, K. D. (2022). Acta Cryst. D78, https://doi.org/ 10.1107/S2059798322007732. Panjikar, S., Parthasarathy, V., Lamzin, V. S., Weiss, M. S. & Tucker, P. A. (2005). Acta Cryst. D61, 449–457. Bricogne, G., Blanc, E., Brandl, M., Flensburg, C., Keller, P., Paciorek, W., Roversi, P., Sharff, A., Smart, O. S., Vonrhein, C. & Womack, T. O. (2017). BUSTER. Global Phasing Ltd, Cambridge, United Kingdom. Pereira, J., Simpkin, A. J., Hartmann, M. D., Rigden, D. J., Keegan, R. M. & Lupas, A. N. (2021). Proteins, 89, 1687–1699. Ramachandra, G., Iftikhar, M. & Khan, F. A. (2017). Procedia Comput. Sci. 110, 465–472. research papers J., Cowie, A., Romera-Paredes, B., Nikolov, S., Jain, R., Adler, J., Back, T., Petersen, S., Reiman, D., Clancy, E., Zielinski, M., Steinegger, M., Pacholska, M., Berghammer, T., Bodenstein, S., Silver, D., Vinyals, O., Senior, A. W., Kavukcuoglu, K., Kohli, P. & Hassabis, D. (2021). Nature, 596, 583–589. y Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Zˇ ı´dek, A., ( ) pp y Stokes-Rees, I. & Sliz, P. (2010). Proc. Natl Acad. Sci. USA, 107, 21476–21481. Potapenko, A., Bridgland, A., Meyer, C., Kohl, S. A. A., Ballard, Tickle, I. J., Flensburg, C., Keller, P., Paciorek, W., Sharff, A., Vonrhein, C. & Bricogne, G. (2018). STARANISO. Global Phasing Ltd, Cambridge, United Kingdom. g g Tickle, I. J., Laskowski, R. A. & Moss, D. S. (1998). Acta Cryst. D54, 243–252. Silver, D., Vinyals, O., Senior, A. W., Kavukcuoglu, K., Kohli, P. & Hassabis, D. (2021). Nature, 596, 583–589. ( ) Kabsch, W. (2010). Acta Cryst. D66, 125–132. Tiemann, J. K. S., Rose, A. S., Ismer, J., Darvish, M. D., Hilal, T., Spahn, C. M. T. & Hildebrand, P. W. (2018). Nucleic Acids Res. 46, W310–W314. ( ) n, R. M. & Winn, M. D. (2008). Acta Cryst. D64, 119–1 ( ) Keegan, R. M. & Winn, M. D. (2008). Acta Crys ( ) Krissinel, E. (2015). Nucleic Acids Res. 43, W314–W319. Uso´n, I. & Sheldrick, G. M. (2018). Acta Cryst. D74, 106–116. Krissinel, E. & Henrick, K. (2007). J. Mol. Biol. 372, 774–797. Vagin, A. & Lebedev, A. (2015). Acta Cryst. A71, s19. Krissinel, E., Uski, V., Lebedev, A., Winn, M. & Ballard, C. (2018). Acta Cryst. D74, 143–151. agin, A. & Teplyakov, A. (2010). Acta Cryst. D66, 22–2 Langer, G., Cohen, S., Lamzin, V. S. & Perrakis, A. (2008). Nat. Protoc. 3, 1171–1179. Varadi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., Zˇ ı´dek, A., Green, T., Tunyasuvunakool, K., Petersen, S., Jumper, J., Clancy, E., Green, R., Vora, A., Lutﬁ, M., Figurnov, M., Cowie, A., Hobbs, N., Kohli, P., Kleywegt, G., Birney, E., Hassabis, D. & Velankar, S. (2022). Nucleic Acids Res. 50, D439–D444. Lebedev, A. A. & Isupov, M. N. (2014). Acta Cryst. D70, 2430–2443. Long, F., Vagin, A. A., Young, P. & Murshudov, G. N. (2008). Acta Cryst. D64, 125–132. Matthews, B. W. (1968). J. Mol. Biol. 33, 491–497. research papers McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658–674. Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans, P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G. W., McCoy, A., McNicholas, S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A., Powell, H. R., Read, R. J., Vagin, A. & Wilson, K. S. (2011). Acta Cryst. D67, 235–242. McCoy, A. J., Sammito, M. D. & Read, R. J. (2022). Acta Cryst. D78, 1–13. McNicholas, S., Potterton, E., Wilson, K. S. & Noble, M. E. M. (2011). Acta Cryst. D67, 386–394. Winter, G. (2010). J. Appl. Cryst. 43, 186–190. Winter, G. (2010). J. Appl. Cryst. 43, 186–190. Medina, A., Jime´nez, E., Caballero, I., Castellvı´, A., Trivin˜o Valls, J., Alcorio, A., Molina, R., Hermoso, J. A., Sammito, M. D., Borges, R. & Uso´n, I. (2022). Submitted. Winter, G., Waterman, D. G., Parkhurst, J. M., Brewster, A. S., Gildea, R. J., Gerstel, M., Fuentes-Montero, L., Vollmar, M., Michels- Clark, T., Young, I. D., Sauter, N. K. & Evans, G. (2018). Acta Cryst. D74, 85–97. , ( ) Milla´n, C., Keegan, R. M., Pereira, J., Sammito, M. D., Simpkin, A. J., McCoy, A. J., Lupas, A. N., Hartmann, M. D., Rigden, D. J. & Read, R. J. (2021). Proteins, 89, 1752–1769. Milla´n, C., Sammito, M. & Uso´n, I. (2015). IUCrJ, 2, 95–105. ( ) Milla´n, C., Keegan, R. M., Pereira, J., Sammito, M. D., Simpkin, A. J., M C A J L A N H t M D Ri d D J & R d ( ) Milla´n, C., Keegan, R. M., Pereira, J., Sammito, M. D., Simpkin, A. J., McCoy, A. J., Lupas, A. N., Hartmann, M. D., Rigden, D. J. & Read, , Wojdyr, M. (2017). J. Open Source Softw. 2, 350. g p McCoy, A. J., Lupas, A. N., Hartmann, M. D., Rigden, D. J. & Read, R. J. (2021). Proteins, 89, 1752–1769. j y , ( ) p f , Yao, J.-X.., Woolfson, M. M., Wilson, K. S. & Dodson, E. J. (2005). Acta Cryst. D61, 1465–1475. R. J. (2021). Proteins, 89, 1752–1769. Milla´n, C., Sammito, M. & Uso´n, I. (2015). IUCrJ, 2, 95–105. ( ) , , Milla´n, C., Sammito, M. & Uso´n, I. (2015). IUCrJ, 2, 95–105. Eugene Krissinel et al. Funding information The following funding is acknowledged: Biotechnology and Biological Sciences Research Council (grant No. BB/L007037/ 1; grant No. BB/S007040/1; grant No. BB/S007083/1; grant No. BB/S005099/1; grant No. BB/S007105/1; award No. BBF020384/1); Medical Research Council (grant No. MC_UP_A025_1012; grant No. MC_U105184325); Ro¨ntgen- A˚ ngstro¨m Cluster (grant No. 349-2013-597); Nederlandse Wetenschappelijke Organisatie (grant No. TKI 16219). Agirre, J., Iglesias-Ferna´ndez, J., Rovira, C., Davies, G. J., Wilson, K. S. & Cowtan, K. D. (2015). Nat. Struct. Mol. Biol. 22, 833–834. 1088 Eugene Krissinel et al. CCP4 Cloud Acta Cryst. (2022). D78, 1079–1089 research papers research papers CCP4 Cloud 1089 Acta Cryst. (2022). D78, 1079–1089
https://openalex.org/W2990750953	https://www.nature.com/articles/s41598-019-54608-x.pdf	English	null	Nucleic acids enrichment of fungal pathogens to study host-pathogen interactions	Scientific reports	2,019	cc-by	9,130	Antonio Rodríguez1, Brecht Guillemyn2, Paul Coucke2 & Mario Vaneechoutte 1 Fungal infections, ranging from superficial to life-threatening infections, represent a major public health problem that affects 25% of the worldwide population. In this context, the study of host- pathogen interactions within the host is crucial to advance antifungal therapy. However, since fungal cells are usually outnumbered by host cells, the fungal transcriptome frequently remains uncovered. We compared three different methods to selectively lyse human cells from in vitro mixes, composed of Candida cells and peripheral blood mononuclear cells. In order to prevent transcriptional modification, the mixes were stored in RNAlater. We evaluated the enrichment of fungal cells through cell counting using microscopy and aimed to further enrich fungal nucleic acids by centrifugation and by reducing contaminant nucleic acids from the host. We verified the enrichment of fungal DNA and RNA through qPCR and RT-qPCR respectively and confirmed that the resulting RNA has high integrity scores, suitable for downstream applications. The enrichment method provided here, i.e., lysis with Buffer RLT followed by centrifugation, may contribute to increase the proportion of nucleic acids from fungi in clinical samples, thus promoting more comprehensive analysis of fungal transcriptional profiles. Although we focused on C. albicans, the enrichment may be applicable to other fungal pathogens. Superficial fungal infections represent a major public health problem worldwide, affecting 25% of the population. These infections are caused mainly by several species of dermatophytes of which Trichophyton rubrum is the most common1. Though superficial fungal infections are rarely life-threatening, they generally require very long and tedious antifungal treatments2. Mucosal infections such as vulvovaginal candidiasis are also very common, affecting millions of women every year. In fact, it is the second most common type of vaginal infection after bac- terial vaginosis3. This infection develops into recurrent vulvovaginal candidiasis in many patients (5–8%) which impacts directly on their social lives and can be fatal in immunocompromised individuals4,5. Systemic, severe and life-threatening fungal infections, have also a significant impact on public health with cryptococcosis, aspergil- losis and candidiasis as main diseases, with candidiasis as the most common of them6. They have a high rate of mortality, reaching 46% in candidaemia cases in Europe7. Candidiasis is mainly caused by C. albicans, although other Candida species have emerged during the last years, such as C. glabrata, C. parapsilosis and C. tropicalis, and very recently the multidrug-resistant pathogen C. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports OPEN Antonio Rodríguez1, Brecht Guillemyn2, Paul Coucke2 & Mario Vaneechoutte 1 Results B i f Brief outline of the study. Figure 1 summarizes the plan of the study. Three different methods to enrich fungal cells and further enrich nucleic acids from pathogenic fungi were compared. First, we prepared aliquots of mixes composed of 106 yeast cells/ml and 106 PBMCs/ml, and stored these mixtures in aliquots of 1 ml in RNAlater at −80 °C.f For the three different lysis treatments, we determined the number of yeast cells and PBMCs that survived the treatment by means of microscopy. Subsequently, we also investigated whether fungal nucleic acids could be further enriched by centrifugation to precipitate intact yeast cells and to reduce human DNA and RNA from the lysed PBMCs, and we quantified DNA and RNA by qPCR and RT-qPCR respectively, and analysed RNA integrity with a Fragment Analyzer.fi y q q p y y g y g y Finally, we determined the efficacy of these enrichment procedures for mixtures containing yeast cells reduced 10- and 100-fold relative to the human counterpart. Enrichment of fungal cells through lysis of human cells. Unlike fungal cells, human cells do not have a protective cell wall to deal with adverse conditions. This distinction can be used to enrich fungal cells through the differential lysis of human cells, which can potentially constitute the first step to enrich nucleic acids from the pathogenic fungus in a clinical sample. Here, we tested the effect of three different methods, all at room temper- ature, to lyse human cells: i) chemically with Triton X-100 for 20 min and ii) with Buffer RLT supplemented with 1% β-mercaptoethanol (i.e., 143 mM) for at least 1 min (in house protocol), and iii) mechanically by bead beating with 0.5 mm zirconium beads in saline for 10 min. Results are depicted in Fig. 2. Using microscopy, we observed that the number of yeast cells remained invariable among all three methods considered in the study.f y g y For PBMCs, we observed that Buffer RLT completely lysed human cells (0 cells/ml) and that bead beat- ing lysed most of them (5.3 × 104 residual cells/ml). This reduction was significant as compared to the control (9.4 × 105 cells/ml) (p < 0.05) and in addition, the differences between Buffer RLT and bead beating were sig- nificant as well. Triton X-100 did not effectively lyse human cells (4.9 × 105 residual cells/ml). Antonio Rodríguez1, Brecht Guillemyn2, Paul Coucke2 & Mario Vaneechoutte 1 We evaluated the efficiency of the enrichment through cell counting, quantitative PCR (qPCR) and quantitative reverse transcription PCR (RT-qPCR). Finally, we checked RNA quality of fungal RNA obtained after the enrichment Nucleic acids extraction RiboPure Yeast Kit Degradation of human cells (PBMCs) Centrifugation to pellet fungal cells 20 000 g for 8 min Mixtures of 104, 105 or 106 yeast cells and 106 PBMCs per ml Storage in 1 ml aliquots of RNAlater at - 80 °C Buffer RLT for 1 min Triton X-100 for 20 min Bead beating for 10 min qPCR (human vs fungal) CXCL-1 vs ITS-1 RT-qPCR (human vs fungal) CXCL-1 vs ITS-1 RNA quality (Fragment Analyzer) Cell counting by microscopy + DNase treatment No treatment + DNase and RNase treatment + DNase treatment Figure 1. A schematic overview of the study set up. Nucleic acids extraction RiboPure Yeast Kit Degradation of human cells (PBMCs) Centrifugation to pellet fungal cells 20 000 g for 8 min Mixtures of 104, 105 or 106 yeast cells and 106 PBMCs per ml Storage in 1 ml aliquots of RNAlater at - 80 °C Buffer RLT for 1 min Triton X-100 for 20 min Bead beating for 10 min qPCR (human vs fungal) CXCL-1 vs ITS-1 RT-qPCR (human vs fungal) CXCL-1 vs ITS-1 RNA quality (Fragment Analyzer) Cell counting by microscopy + DNase treatment No treatment + DNase and RNase treatment + DNase treatment Figure 1. A schematic overview of the study set up. Mixtures of 104, 105 or 106 yeast cells and 106 PBMCs per ml RT-qPCR (human vs fungal) CXCL-1 vs ITS-1 Figure 1. A schematic overview of the study set up. profiles from the pathogen side. To our knowledge, there are only three studies that have attempted to solve this question and developed different enrichment methods. Amorim-Vaz et al.21 used a large set of Candida-specific probes to selectively enrich fungal mRNAs, in combination with subsequent RNA-seq, Andes et al.22 developed a simple approach to enrich fungal RNA through the lysis of human cells with Triton X-100 and Hebecker et al.23 used buffer RLT (Qiagen) with a homogenizer to lyse murine cells, although none of these studies started from RNAlater.f profiles from the pathogen side. To our knowledge, there are only three studies that have attempted to solve this question and developed different enrichment methods. Antonio Rodríguez1, Brecht Guillemyn2, Paul Coucke2 & Mario Vaneechoutte 1 Amorim-Vaz et al.21 used a large set of Candida-specific probes to selectively enrich fungal mRNAs, in combination with subsequent RNA-seq, Andes et al.22 developed a simple approach to enrich fungal RNA through the lysis of human cells with Triton X-100 and Hebecker et al.23 used buffer RLT (Qiagen) with a homogenizer to lyse murine cells, although none of these studies started from RNAlater.f In this paper, we compared three different methods to enrich fungal nucleic acids from in vitro mixes com- posed of different amounts of Candida cells and peripheral blood mononuclear cells (PBMCs) that had been stored in RNAlater in order to halt transcription. We evaluated the efficiency of the enrichment through cell counting, quantitative PCR (qPCR) and quantitative reverse transcription PCR (RT-qPCR). Finally, we checked RNA quality of fungal RNA obtained after the enrichment. Antonio Rodríguez1, Brecht Guillemyn2, Paul Coucke2 & Mario Vaneechoutte 1 auris has been involved in several outbreaks8,9.f Given the clinical relevance of fungal infections, there have been efforts to elucidate host-pathogen interac- tions in recent years, involving candidiasis10–12, aspergillosis13–16 and cryptococcosis17,18. However, most of these studies were focused on the transcriptome of the host. Transcriptomic studies which identify comprehensively differently expressed genes from the fungus are predominantly limited to in vitro experiments in which cells are cultured19 or animal models infected with a high inoculum of the pathogen20, which is usually not representative for the low amount of fungal cells encountered in the host. Although in vitro experiments contribute to improve our knowledge regarding the mechanisms of fungal pathogenicity and the immune response of the host, they reflect within-host interactions only to a limited extent, because these are further influenced by signalling path- way networks involving different metabolites, cell types and other microorganisms that are present in the ecosys- tem of the host. The use of a high burden of the pathogen in animal studies counterbalances the proportion of the fungus that is present in clinical samples. In fact, RNA content from the fungal pathogen was recently found to constitute only about 0.1% of total RNA21. Therefore, for true host-pathogen interaction studies, carried out on clinical samples, there is a need to enrich fungal RNA to obtain a more comprehensive analysis of transcriptional 1Laboratory Bacteriology Research, Department of Diagnostic Sciences, Faculty of Medicine and Health Sciences, Ghent University, Ghent, 9000, Belgium. 2Center for Medical Genetics Ghent, Ghent University Hospital, Department of Biomolecular Medicine, Ghent, 9000, Belgium. email: antoniojrodri86@gmail.com Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ profiles from the pathogen side. To our knowledge, there are only three studies that have attempted to solve this question and developed different enrichment methods. Amorim-Vaz et al.21 used a large set of Candida-specific probes to selectively enrich fungal mRNAs, in combination with subsequent RNA-seq, Andes et al.22 developed a simple approach to enrich fungal RNA through the lysis of human cells with Triton X-100 and Hebecker et al.23 used buffer RLT (Qiagen) with a homogenizer to lyse murine cells, although none of these studies started from RNAlater. In this paper, we compared three different methods to enrich fungal nucleic acids from in vitro mixes com- posed of different amounts of Candida cells and peripheral blood mononuclear cells (PBMCs) that had been stored in RNAlater in order to halt transcription. Results B i f cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Table 1. Effect of different enrichment methods on PCR-based amplificationa of human DNA, fungal DNA, human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates composed of mixes of 106 PBMCs and 106 Candida cells. Statistically significant differences are marked with one (p < 0.005) or two asterisks (p < 0.001) (Linear mixed model). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Treatment with Buffer RLT is the most efficient method to enrich fungal DNA. As we observed that Buffer RLT and bead beating in saline did reduce the number of human cells, while not affecting the number of yeast cells, we next questioned whether DNA from the fungal pathogen would be enriched as well. For this purpose, the mixes were treated with the three methods of cell lysis mentioned above, but followed by a centrifu- gation step (20 000 g, 8 min, room temperature), aiming at concentrating fungal cells and reducing nucleic acids and cell debris from human cells. qPCR results of nucleic acids extracted with the RiboPure Yeast Kit are shown in Table 1 (human DNA and fungal DNA), targeting CXCL1 (human gene) and the internal transcribed spacer region 1 (ITS-1) DNA (yeast gene), respectively. Treatment with Buffer RLT followed by centrifugation induced a significant reduction of human DNA (p < 0.05), resulting in a Cq value of 35.9 which corresponds to only 0.09% human DNA as compared to the control (Cq value of 25.3) (Table 1). Enrichment with bead beating also reduced significantly human DNA (Cq value of 27.8) but a paired samples t-test showed that Buffer RLT is significantly more efficient (p < 0.001). Triton X-100 reduced human DNA to 63.8% (Cq value of 26.0), but this was not statistically sig- nificant. Additional incorporation of a DNase treatment to remove contaminating human DNA from samples treated with Buffer RLT did not further improve the enrichment, as Cq values of 36.21 and 36.50 were obtained before and after DNase treatment (see Supplementary Table S1). Results B i f Thus, these results indicate that Buffer RLT and bead beating selectively lyse human cells and therefore, these methods can be used to enrich fungal cells after RNAlater storage. Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ Figure 2. Number of yeast cells and PBMCs per ml determined by a haemocytometer, at magnification of 400x, after different cell lysis methods. Error bars represent the standard deviations of results from six biological replicates. Statistically significant differences are marked with one (p < 0.05) or two (p < 0.005) asterisks (Friedman test) and with one bracketed asterisk (p < 0.05) (Wilcoxon test). Figure 2. Number of yeast cells and PBMCs per ml determined by a haemocytometer, at magnification of 400x, after different cell lysis methods. Error bars represent the standard deviations of results from six biological replicates. Statistically significant differences are marked with one (p < 0.05) or two (p < 0.005) asterisks (Friedman test) and with one bracketed asterisk (p < 0.05) (Wilcoxon test). Human DNA Cq valueb SD Log cells/mlc CV DNA %c CV No enrichment 25.31 0.46 5.99 0.02 100.00 0.26 Buffer RLT 35.89** 0.59 2.96 0.06 0.09 0.32 Triton X-100 26.00 0.28 5.79 0.01 63.82 0.17 Bead beating 27.84* 1.62 5.27 0.09 18.93 0.66 Fungal DNA Cq value SD Log cells/mlc CV DNA %c CV No enrichment 23.20 0.22 6.05 0.01 100.00 0.13 Buffer RLT 23.50 0.47 5.97 0.02 83.08 0.25 Triton X-100 23.87 0.66 5.87 0.03 66.03 0.34 Bead beating 23.60 1.62 5.94 0.07 77.91 0.63 Human RNA Cq valueb SD Log cells/mlc CV RNA %c CV No enrichment 30.17 0.37 6.13 0.02 100.00 0.20 Buffer RLT 35.00 0.61 4.85 0.03 5.20 0.31 Triton X-100 34.53 0.60 4.98 0.03 6.93 0.31 Bead beating 34.25 1.05 5.05 0.06 8.20 0.48 Fungal RNA Cq valueb SD Log cells/mlc CV RNA %c CV No enrichment 20.25 0.48 5.93 0.03 100.00 0.29 Buffer RLT 20.97 0.22 5.71 0.01 59.94 0.14 Triton X-100 37.70 0.77 0.52 0.46 0.00 0.42 Bead beating 34.02 0.82 1.66 0.15 0.01 0.44 Table 1. Effect of different enrichment methods on PCR-based amplificationa of human DNA, fungal DNA, human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates composed of mixes of 106 PBMCs and 106 Candida cells. Results B i f Statistically significant differences are marked with one (p < 0.005) or two asterisks (p < 0.001) (Linear mixed model). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Human DNA Cq valueb SD Log cells/mlc CV DNA %c CV No enrichment 25.31 0.46 5.99 0.02 100.00 0.26 Buffer RLT 35.89 0.59 2.96 0.06 0.09 0.32 Triton X-100 26.00 0.28 5.79 0.01 63.82 0.17 Bead beating 27.84* 1.62 5.27 0.09 18.93 0.66 Fungal DNA Cq value SD Log cells/mlc CV DNA %c CV No enrichment 23.20 0.22 6.05 0.01 100.00 0.13 Buffer RLT 23.50 0.47 5.97 0.02 83.08 0.25 Triton X-100 23.87 0.66 5.87 0.03 66.03 0.34 Bead beating 23.60 1.62 5.94 0.07 77.91 0.63 Human RNA Cq valueb SD Log cells/mlc CV RNA %c CV No enrichment 30.17 0.37 6.13 0.02 100.00 0.20 Buffer RLT 35.00 0.61 4.85 0.03 5.20 0.31 Triton X-100 34.53 0.60 4.98 0.03 6.93 0.31 Bead beating 34.25 1.05 5.05 0.06 8.20 0.48 Fungal RNA Cq valueb SD Log cells/mlc CV RNA %c CV No enrichment 20.25 0.48 5.93 0.03 100.00 0.29 Buffer RLT 20.97 0.22 5.71 0.01 59.94 0.14 Triton X-100 37.70 0.77 0.52 0.46 0.00 0.42 Bead beating 34.02** 0.82 1.66 0.15 0.01 0.44 Table 1. Effect of different enrichment methods on PCR-based amplificationa of human DNA, fungal DNA, human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates composed of mixes of 106 PBMCs and 106 Candida cells. Statistically significant differences are marked with one (p < 0.005) or two asterisks (p < 0.001) (Linear mixed model). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Table 1. Effect of different enrichment methods on PCR-based amplificationa of human DNA, fungal DNA, human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates composed of mixes of 106 PBMCs and 106 Candida cells. Statistically significant differences are marked with one (p < 0.005) or two asterisks (p < 0.001) (Linear mixed model). Results B i f aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates. Statistically significant differences are marked with one asterisk (p < 0.05) (paired- samples Wilcoxon test). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Table 2. Effect of Buffer RLT treatment on DNA and RNA enrichment from yeast cells for mixtures composed of different ratios of PBMCs and Candida cells, as determined by qPCR-based amplificationa of human DNA and fungal DNA, and by RT-qPCR based amplificationa of human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates. Statistically significant differences are marked with one asterisk (p < 0.05) (paired- samples Wilcoxon test). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Table 2. Effect of Buffer RLT treatment on DNA and RNA enrichment from yeast cells for mixtures composed of different ratios of PBMCs and Candida cells, as determined by qPCR-based amplificationa of human DNA and fungal DNA, and by RT-qPCR based amplificationa of human RNA and fungal RNA. aHuman nucleic acids amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. bCq values are means of six biological replicates. Statistically significant differences are marked with one asterisk (p < 0.05) (paired- samples Wilcoxon test). cLog cells/ml, DNA percentage and RNA percentage were calculated by extrapolation of Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. treatment with Buffer RLT enriched fungal cells as well as fungal DNA, by reducing the human DNA content of the mixes, with Buffer RLT treatment being the most efficient method. Treatment with Buffer RLT additionally enriches fungal RNA. After testing the enrichment of cells and of DNA from the pathogenic fungus with three different methods, and with the addition of a centrifugation step, we evaluated whether RNA was also enriched together with DNA. Therefore, the nucleic acids obtained from the differently treated mixes after extraction by means of the RiboPure Yeast Kit, were further treated with DNase to obtain pure RNA, and quantitative reverse transcription PCR (RT-qPCR) was carried out. Results B i f In addition, Table 1 shows that fungal DNA was not affected by any of the enrichment methods considered in this study. We concluded that bead beating and Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ Human DNA Cq valueb SD Log cells/mlc CV DNA %c CV 106 PBMCs + 105 Candida cells No enrichment 24.39 0.67 6.08 0.03 100.00 0.34 Buffer RLT 32.52* 0.55 3.89 0.04 0.65 0.29 106 PBMCs + 104 Candida cells No enrichment 24.87 0.13 5.95 0.01 100.00 0.08 Buffer RLT 32.95* 0.65 3.77 0.05 0.67 0.33 Fungal DNA Cq valueb SD Log cells/mlc CV DNA %c CV 106 PBMCs + 105 Candida cells No enrichment 23.74 0.59 5.02 0.03 100.00 0.32 Buffer RLT 24.31 0.58 4.85 0.03 68.56 0.32 106 PBMCs + 104 Candida cells No enrichment 27.77 0.10 3.85 0.01 100.00 0.07 Buffer RLT 28.72 0.25 3.57 0.07 53.05 0.45 Human RNA Cq valueb SD Log cells/mlc CV RNA %c CV 106 PBMCs + 105 Candida cells No enrichment 30.65 0.88 6.09 0.04 100.00 0.45 Buffer RLT 35.15* 0.98 4.78 0.06 4.91 0.48 106 PBMCs + 104 Candida cells No enrichment 30.74 1.65 4.37 0.10 100.00 0.64 Buffer RLT 34.57* 0.71 3.34 0.06 9.30 0.36 Fungal RNA Cq valueb SD Log cells/mlc CV RNA %c CV 106 PBMCs + 105 Candida cells No enrichment 22.51 0.52 5.06 0.03 100.00 0.28 Buffer RLT 23.26 0.46 4.84 0.03 61.47 0.26 106 PBMCs + 104 Candida cells No enrichment 25.47 0.54 4.22 0.04 100.00 0.30 Buffer RLT 27.10 0.82 3.76 0.06 34.69 0.41 Table 2. Effect of Buffer RLT treatment on DNA and RNA enrichment from yeast cells of different ratios of PBMCs and Candida cells, as determined by qPCR-based amplifica and fungal DNA, and by RT-qPCR based amplificationa of human RNA and fungal RN amplified with CXCL-1 primers and fungal nucleic acids amplified with ITS-1 primers. of six biological replicates. Statistically significant differences are marked with one aster samples Wilcoxon test). cLog cells/ml, DNA percentage and RNA percentage were calcu Cq values with a standard curve. SD: Standard deviation. CV: Coefficient of variation. Table 2. Effect of Buffer RLT treatment on DNA and RNA enrichment from yeast cells for mixtures composed of different ratios of PBMCs and Candida cells, as determined by qPCR-based amplificationa of human DNA and fungal DNA, and by RT-qPCR based amplificationa of human RNA and fungal RNA. Results B i f We observed that treatment with Buffer RLT, Triton X-100 and bead beating resulted in very low amounts of human RNA, i.e. 5.2, 6.9 and 8.2% respectively (Table 1). Double enzymatic pretreatment prior to the lysis of fungal cells, first with DNase and subsequently with RNase, to remove remaining human nucleic acids, did not further improve the RNA enrichment (Supplementary Table S1). Furthermore, Table 1 shows no significant differences between treat- ment with Buffer RLT and the untreated control for fungal ITS-1 RNA, whereas fungal RNA drastically decreased with both bead beating and Triton X-100 treatment, i.e. no residual RNA and 0.01% residual RNA, respectively, indicating that not only human but also fungal RNA was lost. These results suggest that only treatment with Buffer RLT can be used to enrich fungal RNA. Enrichment with Buffer RLT in 1:10 and 1:100 ratios. After having shown that Buffer RLT enrichment can be used to enrich nucleic acids from the fungal pathogen in samples with equal numbers of fungal and human cells, through the lysis of human cells, the enrichment procedure was tested in Candida/PBMC ratios of 1:10 and 1:100. To this purpose, we prepared mixtures of 105 Candida cells/106 PBMCs per ml and 104 Candida cells/106 PMBCs per ml. Cq values after qPCR of nucleic acids extracts for human and fungal DNA and after RT-qPCR of human and fungal RNA are shown in Table 2. DNA from human cells was significantly reduced to 0.65% and 0.67% of the initial number of cells (p < 0.05) for the 105 and 104 yeast cells/ml mixtures, respectively, while DNA from yeasts was (not significantly) reduced to 68.6% and 53.1%. Accordingly, human RNA significantly dropped to 4.9% and 9.3% (p < 0.05), while fungal cDNA was not significantly reduced to 61.5% and 34.7% for the 105 and Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ Figure 3. RNA integrity of some RNA samples after enrichment with buffer RLT, Triton X-100 and bead beating. Electropherogram profiles were determined with a Fragment Analyzer. Quality for Fragment Analyzer is shown as the RQN value. Reported on a scale of 1 to 10, with higher values indicating a better quality of total RNA. Values above 7 are considered to represent high quality and non-degraded RNA. Figure 3. RNA integrity of some RNA samples after enrichment with buffer RLT, Triton X-100 and bead beating. Electropherogram profiles were determined with a Fragment Analyzer. Results B i f Quality for Fragment Analyzer is shown as the RQN value. Reported on a scale of 1 to 10, with higher values indicating a better quality of total RNA. Values above 7 are considered to represent high quality and non-degraded RNA. 104 yeast cells/ml mixtures. Together, these results demonstrate that enrichment with Buffer RLT enriches fungal nucleic acids from mixtures with human cells, also in cases when low proportions of fungal cells are present. RNA quality is suitable for transcriptomic analysis. Obtaining high quality RNA is essential to per- form comprehensive and reliable transcriptomic analyses. After having shown that enrichment with Buffer RLT is an efficient method to enrich both DNA and RNA from the pathogen fungus from mixes with 10- to 100-fold more human cells. We analysed RNA integrity to check to what extent it can be used for high-quality RNA-demanding experiments such as RNA-seq. Electropherograms of some RNA samples enriched with differ- ent methods are shown in Fig. 3. RQN values were around 10 for Buffer RLT enrichment and no degradation was detected. Contrarily, Triton X-100 and bead beating enrichments resulted in poor electropherogram profiles with evident RNA degradation. We concluded that enrichment with Buffer RLT confers high-quality RNA which can be used in transcriptomic analysis. www.nature.com/scientificreports/ Our contradictory results can be explained by the influence of RNAlater which can impair cell lysis and enzymatic treatments.f yl y y Using Buffer RLT treatment, followed by centrifugation, fungal DNA and RNA were reduced to - statistically non-significant - 83.1% and 59.9% of the original amount present, as assessed by qPCR and RT-qPCR. In sum- mary, the Buffer RLT enrichment in combination with centrifugation was most efficient to reduce human RNA and can be performed in less than 10 min, further reducing the risk of unwanted transcriptomic profile changes. i Furthermore, we simulated the low proportion of fungal cells that is encountered in the host during colonisation or infection by reducing the number of fungal cells 10- and 100-fold relative to the number of human cells. This strat- egy has been used in other in vitro experiments38–40, and gives a better approximation of host-pathogen ratios present in vivo. It should be noted that many in vitro studies used a number of pathogen cells higher than that of the host cells, most likely, in order to, for example, increase the signal of luminescence during ROS production assays41, to increase the signal obtained during chromatographic analyses to study metabolite profiles42 or to stimulate a stronger immune response43,44. Most importantly, we obtained similar results to our experiments when using a 1:1 ratio.f p p y p g Although we aimed to mimic the in vivo situation by mixing PBMCs and yeast cells in different ratios, a limi- tation of our study may be that these mixtures did not reflect possible interactions which may occur in vivo, such as PBMC stimulation, lysis, invasion of human cells by yeast cells or adherence between human and yeast cells.tf y y y y Finally, we obtained high-quality RNA after Buffer RLT enrichment, as assessed with Fragment Analyzer, demonstrating that RNA integrity was not compromised during the enrichment process. This facilitates RNA-seq experiments and any other downstream application.f p y pp Apart from RNA, we also enriched DNA. Although we did not analyse DNA quality, we assume that Buffer RLT enrichment may be used for genomic studies as well, since DNA is more stable than RNA. Indeed, also genomic studies have to cope with the overwhelming amount of host DNA contamination in clinical samples. www.nature.com/scientificreports/ temperatures for RNA extraction purposes. This is especially relevant in clinical settings in which samples cannot be frozen immediately in liquid nitrogen.hff temperatures for RNA extraction purposes. This is especially relevant in clinical settings in which samples cannot be frozen immediately in liquid nitrogen.hff y q g The methods compared for differential cell lysis were Buffer RLT treatment, Triton X-100 incubation and bead eating in saline.f g Triton X-100 is a non-ionic surfactant widely used to lyse different types of human cells such as red blood cells25, monocytes26 and dendritic cells27.f y Bead beating has also been used extensively to lyse different microorganisms including bacteria28,29 and fungi30,31 and less often human cells32. Although bead beating is usually applied in combination with a lysis buffer, we used saline instead to avoid lysis of yeast cells.fff y y Although Buffer RLT is a commercial lysis buffer which has been used in different protocols and kits for the lysis of cells and tissues prior to RNA extraction33–35, its effectiveness as a differential lysis buffer for human and fungal cells has not been studied yet.i g y Cell lysis was previously determined indirectly by quantification of released haemoglobin25, of released LDH26 and of metabolic activity and efficiency of protein extraction27. Here, we used microscopy and cell counting to directly quantify the number and type of cells that were lysed. We found that yeast cells were resistant to all three lysis methods, as we had expected, since i) the fungal cell wall confers protection against different stress condi- tions36, ii) the cell mixtures had been stored in RNAlater, which makes cells harder to lyse37, and iii) a combina- tion of both bead beating and an efficient lysis buffer is needed to effectively lyse Candida cells37. We found that Buffer RLT was the most efficient method to lyse human cells. www.nature.com/scientificreports/ In addition, Buffer RLT was also the fastest method, since pipetting up and down several times was sufficient to lyse 100% of human cells, while Triton X-100 and bead beating required protocols lasting for 20 and 10 min respectively.hf g q p g p y The reduced effectiveness in lysis of the PBMCs (<100%) that we obtained with Triton X-100 and bead beat- ing in saline is probably due to the pre-storage in RNAlater, since Triton X-100 is a well-known detergent widely used in cell lysis but its use for samples stored in RNAlater has not been evaluated previously. In fact, there are no studies analysing the efficiency of cell lysis in samples stored in RNAlater apart from our recent study with C. albicans37. Buffer RLT is a more powerful chemical agent (chaotropic) that is not affected by RNAlater.tf f gf y In addition, using qPCR and RT-qPCR, we demonstrated that after cell lysis with Buffer RLT, both fungal DNA and RNA can be further enriched by means of a centrifugation step that effectively reduced human nucleic acids. However, we observed that pellets of fungal cells were not strongly attached at the bottom of the tube, which may be due to the presence of large amounts of (human) DNA in the supernatant, which may increase viscosity, and interfere with the centrifugation process. Therefore, care should be taken at removal of the supernatants, to avoid removal of the pellet. For this reason, we also recommend to not include more than three samples per run of centrifugation. Using the enrichment with Buffer RLT and centrifugation, only 0.1% human RNA and 5.2% human DNA were left, which represents an important reduction of (human) nucleic acids. In a further effort to completely remove contaminating human nucleic acids, we also added DNase and RNase digestion after Buffer RLT treatment and centrifugation prior to RNA extraction. Andes et al.23, who also used Triton X-100 to enrich fungal cells, had shown that RNase digestion not only decreased human GADPH mRNA, but surprisingly also increased fungal actin mRNA, according to RT-qPCR. However, we did not succeed to reduce human RNA with Triton X-100, and enzymatic digestion did not further increase fungal RNA. Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x Discussion Host-pathogen interactions in vivo are frequently hampered by the low proportion of microbial cells as com- pared to host cells, resulting in poor transcriptional profiles from the pathogen side. In this study, we approached this problem by comparing different methods to selectively lyse human cells in order to increase the proportion of fungal cells. This enrichment of fungal cells through differential cell lysis was further complemented with centrifugation to decrease contaminant cell debris and nucleic acids of the human host cells. In our study, we started from mixtures containing 106 Candida cells and 106 PBMCs, that had been stored in RNAlater, a solution that has been shown to be as efficient as snap-frozen methods to keep transcriptomic profiles unmodified24. Its use is increasing during recent years because it allows short-term storage of clinical samples at above-freezing Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ Methods C l f Culture of Candida albicans. Candida albicans reference strain ATCC 90028 was grown on Sabouraud Glucose Agar plates with Chloramphenicol (50 μg/ml) and transferred to Yeast extract - Peptone Dextrose broth. Cell suspensions were subcultured at 32 °C in static conditions overnight until the logarithmic growth phase was reached. Cells were counted with a microscope by using a haemocytometer (Bürker chamber). Cell suspensions were adjusted with the same broth to 2 × 106, 2 × 105 and 2 × 104 cells/ml and centrifuged at 8000 g for 10 min, whereafter pellets were resuspended in 1-ml aliquots in RNAlater (Invitrogen, Carlsbad, CA) and placed at 4 °C following immediate preparation of yeast cells/PBMCs mixtures. PBMC isolation. PBMCs were isolated from a buffycoat as described previously48 after informed consent from all participants. Cell pellets were resuspended in RNAlater to a concentration of 2 × 106 cells/ml and divided into 1-ml aliquots and placed at 4 °C following immediate preparation of yeast cells/PBMCs mixtures. We confirm that all methods were carried out in accordance to relevant guidelines and regulations and that all experimental protocols were approved by the ethical committee of the University of Ghent (EC/2016/0192). Preparation of yeast cells/PBMCs mixtures and RNAlater treatment. Volumes of 500 µl of 1 ml-aliquots containing 2 × 106, 2 × 105 and 2 × 104 yeast cells/ml and 500 µl of 1 ml-aliquots with 2 × 106 PBMCs/ ml were mixed together to obtain 1-ml aliquots of 106:106, 105:106 and 104:106 yeasts cells:PBMCs mixtures. These mixtures were stored overnight at 4 °C to enable the RNAlater to penetrate into the cells, and subsequently stored at −80 °C. Cell lysis of human cells. Mixtures of yeast cells and PBMCs that had been stored at −80 °C, were thawed, centrifuged for 10 min at 20 000 g and resuspended in 1-ml aliquots of i) Triton X-100 (final concentration, 1%), ii) Buffer RLT (Qiagen, Hilden, Germany) supplemented with 1% β-mercaptoethanol and iii) saline. Buffer RLT and Triton X-100 samples were incubated for at least 1 min and 20 min respectively, while saline samples were transferred to prefilled tubes with 0.5 mm zirconium beads and further bead beaten for 10 min in a hands-free vortex genie-2 (MO BIO Laboratories, Carlsbad, CA), fitted with a vortex adapter that allows bead beating in a horizontal position. Microscopic visualization and cell viability determination. www.nature.com/scientificreports/ gene and splicing variant of the fungal transcriptome22. As such, these authors carried out an enrichment after and not prior to RNA extraction. This gives an important advantage since transcriptomic profiles may change during the enrichment process. However, disadvantages of this method include certain cross-hybridization that is inherent to the technique and the cost of the probes. Using RNA-seq in two animal models, they obtained from 0.03% to 0.1% of reads aligned to C. albicans before the enrichment, and fungal reads were enriched to 58–69% with the probes. Since they do not provide how many of the 31–42% of unmapped reads were mapped as human, the provenance of these reads is unknown. p In summary, the clinical relevance of fungal infections has boosted the interest in host-pathogen interactions. Expanding our knowledge in this field will contribute to develop new diagnostic tools, identify potential vaccine candidates and improve the antifungal treatments we have used for more than 30 years and that are becoming increasingly obsolete47. The enrichment described here may enable more comprehensive analyses of transcrip- tional profiles from fungal pathogens within the host. Furthermore, we should evaluate whether this approach is also applicable for dermatophytes, whose cell walls are even tougher than Candida. Methods C l f After the different cell lysis treatments, cell suspensions were centrifuged for 10 min at 8000 g to collect intact yeast cells and PBMCs. They were then resus- pended in 1 ml of saline prior to cell counting with a microscope. Ten µl of this saline cell suspension was loaded into a haemocytometer. Yeast cells and PBMCs could be observed with light microscopy and were counted in 25 small squares with 0.2 mm sides (together representing a 0.1 µl volume) at a magnification of 400x. The resulting number of cells/ml was calculated as follows: cells/ml = cell number (in 25 small squares) × 1 (dilution) × 104. Centrifugation conditions for enrichment. Centrifugation at 20 000 g for 8 min at room temperature was used to pellet fungal cells. Supernatants containing cell debris from human cells that had been lysed with the three different methods were discarded very carefully, since fungal pellets were very loose and they can be detached very quickly. No more than three samples were centrifuged per run of centrifugation to minimize the risk of fungal pellets being removed together with cell debris from human cells. DNase and RNase treatment. Pellets of intact yeast cells were treated to digest contaminating human DNA and RNA. First, pellets were resuspended in 86 µl nuclease-free water, and DNase treatment was performed by addition of 10 µl of 10x DNase I buffer and 4 µl of DNase I (2 U/µl) and by incubation for 30 min at 37 °C. Second, the yeast pellet was processed for RNase treatment by addition of 98 µl of nuclease-free water and 2 µl of RNase A Solution (4 mg/ml). The mixture was then incubated for 1 h at 37 °C. Nucleic acids extraction. Nucleic acids were isolated using the RiboPure Yeast Kit (Invitrogen) and fol- lowing the manufacturer’s instruction of the kit. Half of the resulting nucleic acids was used directly for DNA quantification by means of qPCR, and the other half was treated with DNase I to obtain pure RNA. RNA was then analyzed for quality with a NanoDrop spectrophotometer and a Fragment Analyzer (DNF-472 High-Sensitivity Total RNA, Applied Biosystems) and subsequently quantified with RT-qPCR. qPCR and RT-qPCR analysis. www.nature.com/scientificreports/ For example, a recent study reported about the need of DNA enrichment in metagenomic studies of sputum, and described a microbe enrichment microfluidic device coupled with DNA amplification45. However, although this enrichment is suitable for different bacteria, fungi are not so well recovered because of their larger size.h f g g The enrichment method described here is based on the selective lysis of human cells, taking advantage of the protective properties that the cell wall confers to fungal cells. A comparable protocol has been published using RNAPure instead of Buffer RLT, but without explicitly showing the efficiency46. Andes et al.23 also enriched fungal RNA through the lysis of human cells. However, their protocol requires two incubations during 20 min at 37 °C and as such gene expression prior to RNA extraction may be changed, which would no longer match the expres- sion profile encountered during the infection. Another study used 55.000 Candida-specific probes to match every Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ Methods C l f Reverse transcription was performed for 5 min at 25 °C followed by 60 min at 42 °C with random hexamer primers and with 5 µl of total RNA according to the instructions of the manufacturer Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ of the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA). qPCR from both cDNA and DNA was performed in duplicate with a LightCycler 480 (Roche). We used CXCL1_F forward (5′-GGA AAG AGA GAC ACA GCT GCA-3′) and CXCL1_R reverse (5′-AGA AGA CTT CTC CTA AGC GAT GC-3′) primers, previously described (Van Belleghem et al.48), targeting the human gene encoding the cytokine CXCL1, and CA_rRNA F (5′-TTT GCT TGA AAG ACG GTA-3′) and CA_rRNA R (5′-TTG AAG ATA TAC GTG GTG G-3′) targeting the ITS-1 gene of C. albicans (adapted from49). PCR master mixes to quantify both human cDNA/DNA and yeast cDNA/DNA were prepared similarly, i.e. Roche LC480 high resolution melting (HRM) mix (Roche, Basel, Switzerland), 0.5 µM primers and 2 µl sample (cDNA or DNA) in a final volume of 10 µl, with exception of MgCl2 which was used at a concentration of 2 mM and 3 mM to amplify human cDNA/ DNA and yeast cDNA/DNA respectively. The thermal cycling program consisted of a pre-incubation step for 10 min at 95 °C, amplification for 45 cycles of 30 s at 95 °C, 30 s at 59 °C and 30 s at 72 °C. Results were analyzed with the LightCycler 480 software 1.5 (Roche). Statistical analysis. For statistical comparisons of cell lysis and PCR data (cq values), independent exper- iments performed with six replicates were considered. Data were analyzed using the linear mixed model for repeated measures followed by Bonferroni’s multiple testing correction, and with Friedman’s test and Wilcoxon’s test where the assumption of normality was not acceptable, using the IBM SPSS Statistics software v 25.0 (IBM, Armonk, NY, USA). Received: 30 August 2019; Accepted: 15 November 2019; Received: 30 August 2019; Accepted: 15 November 2019; Received: 30 August 2019; Accepted: 15 November 2019; Published: xx xx xxxx Published: xx xx xxxx References erences avlickova, B., Czaika, V. A. & Friedrich, M. Epidemiological trends in skin mycoses worldwide. Mycoses. 4, 2–15 (2008).i 1. Havlickova, B., Czaika, V. A. & Friedrich, M. Epidemiological trends in skin mycoses worldwide. Mycoses. 4, 2–15 (2008). 2 Pires C A et al Clinical epidemiological and therapeutic profile of dermatophytosis An Bras Dermatol 89 259 264 (2014) 1. Havlickova, B., Czaika, V. A. & Friedrich, M. Epidemiological trends in skin mycoses worldwide. Mycoses. 4, 2–15 (2008). 2. Pires, C. A. et al. Clinical, epidemiological, and therapeutic profile of dermatophytosis. An Bras Dermatol. 89, 259–264 (2014 3. Anderson, M. R., Klink, K. & Cohrssen, A. Evaluation of vaginal complaints. JAMA. 291, 1368–1379 (2004). g 4. Sobel, J. D. Vulvovaginal candidosis. Lancet. 369, 1961–1971 (2007). g p 4. Sobel, J. D. Vulvovaginal candidosis. Lancet. 369, 1961–1971 (2007). g 5. Aballéa, S. et al. Subjective health status and health-related quality of life among women with recurrent vulvovaginal candidosi (RVVC) in Europe and the USA. Health Qual Life Outcomes. 11, 169 (2013). 6. Pfaller, M. A. & Diekema, D. J. Epidemiology of invasive candidiasis: a persistent public health problem. Clin Microbiol Rev. 20, 133–163 (2007). 7. Koehler, P. et al. Morbidity and mortality of candidaemia in Europe: an epidemiologic meta-analysis. Clin Microbiol Infect. 25 1200–1212 (2019). 8. Schelenz, S. et al. First hospital outbreak of the globally emerging Candida auris in a European hospital. Antimicrob Resist Infec Control. 5, 35 (2016).i , ( ) 9. Lockhart, S. L. et al. Simultaneous emergence of multidrug-resistant Candida auris on 3 continents confirmed by whole-genome sequencing and epidemiological analyses. Clin Infect Dis. 64, 134–140 (2017).il q g p g y f 0. Bruno, V. M. et al. Transcriptomic analysis of vulvovaginal candidiasis identifies a role for the NLRP3 inflammasome. MBio. 6 e00182–15 (2015). 11. Liu, Y. et al. New signaling pathways govern the host response to C. albicans infection in various niches. Genome Res. 25, 679–689 (2015).l 2. Kirchner, F. R. et al. Persistence of Candida albicans in the oral mucosa induces a curbed inflammatory host response that is independent of immunosuppression. Front Immunol. 10, 330 (2019).i 13. Chen, F. et al. Transcriptome profiles of human lung epithelial cells A549 interacting with Aspergillus fumigatus by RNA-Seq. PLoS One. 10, e0135720 (2015). 14. Kale, S. D. et al. Modulation of immune signaling and metabolism highlights host and fungal transcriptional responses in mouse models of invasive pulmonary aspergillosis. References Sci Rep. 7, 17096 (2017).ihh p y p g p 5. Shankar, J., Cerqueira, G. C., Wortman, J. R., Clemons, K. V. & Stevens, D. A. RNA-Seq profile reveals Th-1 and Th-17-type o immune responses in mice infected systemically with Aspergillus fumigatus. Mycopathologia. 183, 645–658 (2018). 16. Chung, M. et al. Targeted enrichment outperforms other enrichment techniques and enables more multi-species RNA-Seq analyses. Sci Rep. 8, 13377 (2018). p 17. Liu, T. B. et al. Cryptococcus inositol utilization modulates the host protective immune response during brain infection. Cell Commun Signal. 12, 51 (2014). g ( ) 8. Chen, S. et al. Cryptococcus neoformans infection and immune cell regulation in human monocytes. Cell Physiol Biochem. 37 537–547 (2015). 19. Linde, J. et al. Defining the transcriptomic landscape of Candida glabrata by RNA-Seq. Nucleic Acids Res. 43, 1392–1406 (2015).i nde, J. et al. Defining the transcriptomic landscape of Candida glabr Ji g p p g y q ( ) 0. Jiang, C. et al. Significance of hyphae formation in virulence of Candida tropicalis and transcriptomic analysis of hyphal cells Microbiol Res. 192, 65–72 (2016). 1. Amorim-Vaz, S. et al. RNA enrichment method for quantitative transcriptional analysis of pathogens in vivo applied to the fungu Candida albicans. MBio. 6, e00942–15 (2015). 22. Andes, D., Lepak, A., Pitula, A., Marchillo, K. & Clark, J. A simple approach for estimating gene expression in Candida albicans directly from a systemic infection site. J Infect Dis. 192, 893–900 (2005). y y f 3. Hebecker, B. et al. Dual-species transcriptional profiling during systemic candidiasis reveals organ-specific host-pathogen interactions. Sci Rep. 6, 36055 (2016). interactions. Sci Rep. 6, 36055 (2016). 24. Chowdary, D. et al. Prognostic gene expression signatures can be measured in tissues collected in RNAlater preservative. J ( ) p 4. Chowdary, D. et al. Prognostic gene expression signatures can be measured in tissues collected in RNAlater preservative. J Mo Diagn. 8, 31–39 (2006). g ( ) 25. Preté, P. S. et al. Multiple stages of detergent-erythrocyte membrane interaction–a spin label study. Biochim Biophys Acta. 1 164–170 (2011). 25. Preté, P. S. et al. Multiple stages of detergent-erythrocyte membrane interaction–a spin label study. Biochim Biophys Acta. 1808, 164–170 (2011). 26 B h t M t l M ll lid i d id ll d th i h i t d ll Bi M d Ch 22 ( ) 6. Bohnert, M. et al. Data availability y he datasets used and analysed during the current study are available from the corresponding author on easonable request. Received: 30 August 2019; Accepted: 15 November 2019; References 4, a019703 (2014).l p 8. Van Belleghem, J. D., Clement, F., Merabishvili, M., Lavigne, R. & Vaneechoutte, M. Pro- and anti-inflammatory responses o peripheral blood mononuclear cells induced by Staphylococcus aureus and Pseudomonas aeruginosa phages. Sci Rep. 7, 8004 (2017) peripheral blood mononuclear cells induced by Staphylococcus aureus and Pseudomonas aeruginosa phages. Sci Rep. 7, 8004 (2017). 49. Guiver, M., Levi, K. & Oppenheim, B. A. Rapid identification of Candida species by TaqMan PCR. J Clin Pathol. 54, 362–366 (2001). p p y p y g p g p , ( ) 9. Guiver, M., Levi, K. & Oppenheim, B. A. Rapid identification of Candida species by TaqMan PCR. J Clin Pathol. 54, 362–366 (2001) 49. Guiver, M., Levi, K. & Oppenheim, B. A. Rapid identification of Candida species by TaqMan PCR. J Clin Pathol. 54, 362–366 (2001). Acknowledgementsh This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No. 642095. This work has been supported by the Ghent University Hospital. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement No. 642095. This work has been supported by the Ghent University Hospital. References 17, 475–480 (2013). ( ) 36. Beauvais, A. & Latgé, J. P. Special issue: Fungal cell wall. J Fungi. 4, 91 (2018).fiff is, A. & Latgé, J. P. Special issue: Fungal cell wall. J Fungi. 4, 91 (201 36. Beauvais, A. & Latgé, J. P. Special issue: Fungal cell wall. J Fun g p g g 37. Rodríguez, A. & Vaneechoutte, M. Comparison of the efficiency of different cell lysis methods and different commercial met for RNA extraction from Candida albicans stored in RNAlater. BMC Microbiol. 19, 94 (2019). 38. Moyes, D. L. et al. Candida albicans yeast and hyphae are discriminated by MAPK signaling in vaginal epithelial cells. PLoS One. 6, e26580 (2011).f 39. Gao, Y. et al. Different host immunological response to C. albicans by human oral and vaginal epithelial cells. Mycopathologia. 184, 1–12 (2019). 40. Ho, J. et al. Candidalysin activates innate epithelial immune responses via epidermal growth factor receptor. Nat Commun. 10, 2297 (2 41. Salvatori, O., Pathirana, R. U., Kay, J. G. & Edgerton, M. Candida albicans Ras1 inactivation increases resistance to phagosomal killing by human neutrophils. Infect Immun. 86, 1–14 (2018). g y p f 2. Liew, K. L. et al. In vitro analysis of metabolites secreted during infection of lung epithelial cells by Cryptococcus neoformans. PLoS One. 7, e0153356 (2016). 43. Hernández-Chávez, M. J. et al. Role of protein phosphomannosylation in the Candida tropicalis-macrophage interaction. FEMS Yeast Res. 18, 1–14 (2018). ( ) 44. Zhang, X., He, D., Gao, S., Wei, Y. & Wang, L. Aspergillus fumigatus enhances human NK cell activity by regulating M1 macrophage polarization. Mol Med Rep. 20, 1241–1249 (2019). 44. Zhang, X., He, D., Gao, S., Wei, Y. & Wang, L. Aspergillus fumigatus enhances human NK cell activity by regulating M1 macrophage polarization. Mol Med Rep. 20, 1241–1249 (2019). p p 5. Shi, X. et al. Microfluidics-based enrichment and whole-genome amplification enable strain-level resolution for airway metagenomics. mSystems. 4, 1–14 (2019).i g y 6. Lüttich, A., Brunke, S. & Hube, B. Isolation and amplification of fungal RNA for microarray analysis from host samples. Methods Mol Biol. 845, 411–421 (2012). g y 46. Lüttich, A., Brunke, S. & Hube, B. Isolation and amplification of fungal RNA for microarray analysis from host samples. Methods Mol Biol. 845, 411–421 (2012). 7. Roemer, T. & Krysan, D. J. Antifungal drug development: challenges, unmet clinical needs, and new approaches. Cold Spring Harb Perspect Med. References Melleolides induce rapid cell death in human primary monocytes and cancer cells. Bioorg Med Chem. 22 3856–3861 (2014). 3856 3861 (2014). 27. Zhang, Y. et al. Optimization of human dendritic cell sample preparation for mass spectrometry-based proteomic studies. Anal Biochem. 484, 40–50 (2015). Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x www.nature.com/scientificreports/ 28. De Bruin, O. M. & Birnboim, H. C. A method for assessing efficiency of bacterial cell disruption and DNA release. BMC Micro 16, 197 (2016). 28. De Bruin, O. M. & Birnboim, H. C. A method for assessing efficiency of bacterial cell disruption and DNA release. BMC Microbiol. 16, 197 (2016). 29 Gill C van de Wijgert J H Blow F & Darby A C Evaluation of lysis methods for the extraction of bacterial DNA for analysis of 16, 197 (2016). 29. Gill, C., van de Wijgert, J. H., Blow, F. & Darby, A. C. Evaluation of lysis methods for the extraction of bacterial DNA for analysis of th i l i bi t PL S O 11 0163148 (2016) 9. Gill, C., van de Wijgert, J. H., Blow, F. & Darby, A. C. Evaluation of lysis methods for the extraction of bacterial DNA for analysis o the vaginal microbiota. PLoS One. 11, e0163148 (2016). g 30. Sasidharan, K., Amariei, C., Tomita, M. & Murray, D. B. Rapid DNA, RNA and protein extraction protocols optimized for slow continuously growing yeast cultures. Yeast. 29, 311–322 (2012). y g g y 1. Sasidharan, K., Soga, T., Tomita, M. & Murray, D. B. A yeast metabolite extraction protocol optimised for time-series analyses. PLoS One. 7, e44283 (2012).f ( ) 32. Myers, J. N., Rekhadevi, P. V. & Ramesh, A. Comparative evaluation of different cell lysis and extraction methods for studying benzo(a)pyrene metabolism in HT-29 colon cancer cell cultures. Cell Physiol Biochem. 28, 209–218 (2011). 3. Kasahara, T. et al. Evaluation of methods for duration of preservation of RNA quality in rat liver used for transcriptome analysis. Toxicol Sci. 31, 509–519 (2006).l 4. Mack, E., Neubauer, A. & Brendel, C. Comparison of RNA yield from small cell populations sorted by flow cytometry applying different isolation procedures. Cytometry A. 71, 404–409 (2007). f p y y 35. Van der Poel-van de Luytgaarde, S. C. et al. Addition of β-mercaptoethanol is a prerequisite for high-quality RNA isolation using QIAsymphony technology as demonstrated by detection of molecular aberrations in hematologic malignancies. Genet Test Mol Biomarkers. Competing interestsh p g The authors declare no competing interests. p g The authors declare no competing interests. Author contributions A.R. and M.V. conceived and designed the experiments, analysed the data and wrote the paper. A.R. performed the experiments. B.G. and P.C. performed RNA quality analysis. All authors have read and approved the manuscript. Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41598-019-54608-x. Correspondence and requests for materials should be addressed to A.R. Correspondence and requests for materials should be addressed to A.R. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2019 Scientific Reports \| (2019) 9:18037 \| https://doi.org/10.1038/s41598-019-54608-x
https://openalex.org/W2954382400	http://wrap.warwick.ac.uk/132543/7/WRAP-characterisation-damage-mechanisms-oxide-ceramics-quasi-Dancer-2019.pdf	English	null	Characterisation of damage mechanisms in oxide ceramics indented at dynamic and quasi-static strain rates	Journal of the European Ceramic Society	2,019	cc-by	8,861	A R T I C L E I N F O Keywords: Armour Al2O3 ZrO2 Mechanical properties Residual stress Ceramic materials are known to display rate dependent behaviour under impact. Tests to establish the strain-rate dependent variations in damage mechanisms have been carried out on debased alumina, an alumina-zirconia composite, and 3Y-TZP. Materials were indented dynamically and quasi-statically using identical sharp hardened steel projectiles while recording the load proﬁle. Characteristics typical of both sharp and blunt indentation types were observed using scanning electron microscopy and piezospectroscopic mapping. At dynamic strain rates both the depth of the indentation and the residual stress in the material were lower than for quasi-static tests. This was attributed to temperature-induced softening of the projectile. Unusual behaviour was observed in the 3Y-TZP samples due to the reversible transformation from tetragonal to monoclinic crystal structures during mechanical loading. These eﬀects and the observed superior mechanical strength against impact suggest that zirconia or zirconia-composite materials may have advantages over debased alumina for application as ceramic armour materials. Keywords: Armour Al2O3 ZrO2 Mechanical properties Residual stress in a high degree of destruction of the ceramic material. This severely limits the ability of post-test characterisation to establish the micro- mechanisms occurring which lead to failure. It is therefore diﬃcult to develop physically based models for impact and to understand how the microstructure of the material should be modiﬁed to manufacture a material with greater resistance to damage [16,17]. Due to the limited microstructural information which can be obtained from ballistic testing, in our previous study [12] a test was developed which was carried out at sub-ballistic speeds on alumina samples using smaller projectiles designed to resemble real armour-piercing bullet cores. The sub-ballistic impact velocities allow a greater number of samples to be recovered in an essentially intact state, and therefore post-testing measurement of residual stress and some quantiﬁcation of the degree of deformation of the impacted surface could be carried out and compared to analytical models [11,12]. At the same time, the projectile is sharp, so the strain-rate at the point of impact is locally very high and so gives information relevant in more energetic impacts. Characterisation of damage mechanisms in oxide ceramics indented at dynamic and quasi-static strain rates C.E.J. Dancera,b,⁎, J.N.F. Spawtona, S. Falcoc, N. Petrinicc, R.I. Todda a Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK b International Institute for Nanocomposite Manufacturing, Warwick Manufacturing Group, University of Warwick, Gibbet Hill, Coventry, CV4 7AL, UK c Department of Engineering Science, University of Oxford, Parks Road, Oxford, OX1 3PJ, UK https://doi.org/10.1016/j.jeurceramsoc.2019.06.054 Received 26 March 2019; Received in revised form 27 June 2019; Accepted 28 June 2019 ⁎ Corresponding author at: International Institute for Nanocomposite Manufacturing, Warwick Manufacturing Group, University of Warwick, Gibbet Hill, Coventry, CV4 7AL,UK E-mail address: c.dancer@warwick.ac.uk (C.E.J. Dancer). Available online 02 July 2019 0955-2219/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecom m ons.org/licenses/BY/4.0/). Journal of the European Ceram ic Society 39 (2019) 4936–4945 Journal of the European Ceram ic Society 39 (2019) 4936–4945 Contents lists available at ScienceDirect Corresponding author at: International Institute for Nanocomposite Manufacturing, Warwick Manufacturing Group, University of Warwick, Gibbet Hill, Coventry, 1. Introduction By dividing typical values for the impact velocity and by those for the indentation diameter in our test we calculate a maximum strain rate of 1 × 105 s−1; in comparison the dynamic hardness tester achieves a strain rate of 2.2 × 103 s−1 [18]. This means the strain rate in our dynamic test lies comfortably in the dynamic regime rather than in the region of the static-dynamic transition mentioned above. This will re- sult in our setup allowing a more reliable comparison to ballistic tests and a greater contrast when compared with quasi-static indentations. were obtained in the form of large discs of 100–120 mm diameter and 8 mm thickness. The materials were debased alumina (95 wt% alumina with a glassy phase formed from multiple oxide additions), a composite zirconia-alumina material (90 wt% alumina, 10 wt% stabilised zir- conia), and tetragonal zirconia polycrystals (3Y-TZP) (94 wt% zirconia, 5.4 wt% yttria, 0.25 wt% alumina). To obtain suitable materials for impact tests, discs of approximately 25 mm diameter were drilled from the as-supplied material using a diamond-coated core drill (DK Holdings, UK). Samples were ﬂat-bed ground (Jones & Shipman, UK) on both sides to ensure parallelism, and to obtain a suitable surface for subsequent microscopic examination the surface to be impacted or in- dented was polished to a 3 μm ﬁnish using diamond pastes. Samples of suitable dimensions for other characterisation methods were cut from the as-supplied discs using a diamond blade and polished to a 1 μm ﬁnish except where stated otherwise. The bulk density was measured using Archimedes’ method with water as the immersion medium. Hardness indentations were made using a Vickers hardness indenter (Vickers-Armstrong, UK) with a 10 kg load, and the dimen- sions of the indentations were determined using an optical confocal microscope (InﬁniteFocus Optical Microscope, Alicona, Austria) with a 20x objective lens. The fracture strength was determined by 4-point bending tests using eight samples per material, using a jig which con- formed to British Standard EN 843-1:2006 [25]. The Young’s modulus, shear modulus, and Poisson’s ratio were calculated from measured re- sonance frequencies determined using as-supplied discs by the impulse resonance technique (Grindosonic, J.W. Lemmens N.V, Leuven, Bel- gium) following the methods given in British Standard EN 843-2:2006 [26]. 1. Introduction Ceramic materials are both strong under compressive loads and extremely hard [1], and therefore have excellent resistance to in- dentation. This, combined with other properties such as their low density and resistance to wear, means they are attractive for use in armour systems [2–4]. Ceramics exhibit strain rate-dependent me- chanical behaviour above some critical strain rate, which has been determined as ca. 102-103 s−1 for both polycrystalline alumina and silicon carbide [5,6]. This strain rate-dependent behaviour is one of the reasons why the performance of ceramic materials impacted at high strain rates (e.g. ballistic velocities) is not related in a straightforward way to their mechanical properties as measured at quasi-static low strain rates [7,8]. This is because the sequence of fracture mechanisms occurring in ceramic materials during dynamic testing is known to diﬀer from that in quasi-static testing [9–13]. Ballistic testing involves high velocity (> 700-800 m/s) impact of a sphere, rod or sharpened bullet, sometimes encased in a metal sheath, on relatively large ceramic discs or tiles [14,15]. Such testing, while eﬀective at establishing a qualitative relative performance of material compositions against the speciﬁc threat tested [8], will typically result A dynamic indentation hardness tester using a diamond Vickers hardness indenter mounted in a split Hopkinson pressure bar has pre- viously been developed and used to test a number of materials by Subhash and coworkers [18–20]. However our setup has a number of Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. advantages as an analogue for a ballistic test. First, by using a projectile rather than a diamond-tipped hardness indenter we can study the eﬀect of impact on the projectile as well as on the target, rather than in- denting with such a rigid hard material that the indenter is unaﬀected. Secondly, by using a gas-gun to propel the projectile rather than the split Hopkinson pressure bar, higher strain rates are attained during impact. Koeppel and Subhash [18] deﬁned the strain rate in their dy- namic hardness tests as the ratio of the indenter velocity to the size of the indentation. The maximum strain rate in our test occurs on initial impact, when the velocity is highest and the projectile is still sharp. 1. Introduction The grain sizes of the materials were determined by the linear intercept method described by Mendelson [27] using SEM micrographs of specimens which were polished to 0.25μm ﬁnish, thermally etched in air, and coated with 2 nm of platinum. A multiplication factor of 1.56 was applied to convert the 2D projections to the expected 3D sizes [27]. Silicon carbide, alumina and boron carbide are currently the most widely used armour ceramics [2,4,15]. While boron carbide and silicon carbide have superior performance against ballistic threats compared to alumina, their high temperature processing requirements mean they are signiﬁcantly more expensive to produce [1,16]. Compared to alumina, tetragonal zirconia polycrystals (TZP) have a higher toughness which is known to be an important property for armour applications [2]. The enhanced toughness is due to the fact that TZP undergoes a phase transition from tetragonal to monoclinic when under an applied tensile stress [21–23]. The volume expansion (ca. 4%) which occurs when this transition occurs due to the tensile stress at a crack tip results in a compressive stress, which limits crack growth [22,24]. The toughness of TZP is therefore higher than that of other ceramics such as alumina, where this transition does not occur. However the typical density of 6.10 g/cm3 for yttria-stabilised TZP is signiﬁcantly higher compared with alumina (3.96 g/cm3) and with carbide materials such as boron carbide (2.52 g/cm3) and silicon carbide (3.21 g/cm3), which makes it less attractive for armour applications where weight is generally an important factor [1]. However if superior performance were obtained for the TZP, it may prove possible to use comparatively thinner plates of TZP which would save reduce the weight and mitigate the disadvantage of the higher density. In addition, it is reasonable to suggest that a composite material containing TZP may oﬀer a good compromise be- tween the superior toughness of TZP and the lower density and higher hardness of alumina, or one of the carbide materials. The projectiles used in all tests were manufactured in house from hardened tool steel. This material was chosen for its ability to inﬂict measurable damage on the ceramic materials without shattering during testing. 3 mm diameter centreless-ground silver steel rods (composition BS-1407) were heated to 760 °C at a rate of 15 °C/min, held for 20 min and quenched in a water bath. 1. Introduction The resulting material had a Vickers hardness of 8.6 GPa measured on a cross-sectional slice through one of the rods. To make each projectile, a hardened steel rod was sharpened using a cooled diamond wheel and then sectioned at a suitable length using a cubic boron nitride blade. The cut surface of the projectile was ground using a ﬂatbed grinder to ensure ﬂat, perpendicular bases. The projectiles were approximately 0.2 g in mass and 5 mm in length with a 45° tip angle. The tip points were suﬃciently sharp that they were not resolvable using optical microscopy, i.e. the tip diameter was less than ∼5 μm. The tip diameters of all projectiles were checked using a con- focal optical microscope (InﬁniteFocus Optical Microscope, Alicona, Austria) prior to testing and any with a tip diameter greater than this value were rejected. In this work, we extend our previously-described dynamic test [11,12] developed for alumina to study impact and indentation of both a composite zirconia-alumina material and 3Y-TZP, and compare their behaviour to that of debased polycrystalline alumina tested under identical conditions. In addition we determine the eﬀect on the degree of damage sustained by reducing the impact speed to quasi-static rates for these ceramic materials. In order to compare the dynamic and quasi-static behaviour of the materials complementary tests were carried out at diﬀerent strain rates. In dynamic tests a gas gun apparatus was used (Fig. 1) similar to that described in earlier work [11,12]. The sharpened projectiles described 2. Experimental methods Three oxide ceramic materials sintered from mixed oxide powders Fig. 1. Schematic diagram showing the dynamic impact test. Fig. 1. Schematic diagram showing the dynamic impact test. 4937 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Fig. 2. Schematic diagram of the quasi-static test. collecting spectra at points on lines passing through the impact site but including the undamaged region for comparison. By ﬁtting the peaks in each spectrum using a pseudo-Voigt proﬁle, the changes in position of the Cr3+ R1 ﬂuorescence peak in alumina with position across the impact region were determined. All materials contained suﬃcient alu- mina that a Cr3+ ﬂuorescence signal could be obtained, including the 3Y-TZP in which alumina was present as an impurity. For the hydro- static stress state experienced by these samples, the residual stress is related to the change in wavenumber ΔυR1 of the R1 Cr3+ ﬂuorescence peak from the zero stress condition by the following relationship: = ν Π σ Δ R H H 1 (1) (1) = ν Π σ Δ R H H 1 where σH is the hydrostatic stress component and ΠH is the hydrostatic piezospectroscopic coeﬃcient, taken as 7.59 cm−1 GPa−1 [28]. The position of the zero stress R1 peak was determined by averaging the results of 4 spectra taken near the edges of each specimen, far from the impacted or indented region. Fig. 2. Schematic diagram of the quasi-static test. In addition to the Cr3+ ﬂuorescence spectroscopy using the 633 nm red laser, for zirconia-containing materials Raman spectra were also acquired using the 422 nm He-Cd laser to determine the relative quantities of tetragonal and monoclinic zirconia present inside and outside the impacted region. Line scans of spectra across the impact site were acquired using the Renishaw inVia Reﬂex Raman microscope as above. The relative proportion of monoclinic and tetragonal phases present was determined by using pseudo-Voigt peak ﬁtting to obtain the relative areas of the monoclinic doublet at 181 and 190 cm−1 and the tetragonal peaks at 147 and 264 cm−1. These peak areas were used to calculate the relative volume fraction of monoclinic zirconia present using the following equation [29,30]: above were mounted to nylon sabots using double-sided tape and ﬁred at ceramic targets of dimensions 25 mm diameter by 6 mm thickness mounted against a backing bar. The projectile and sabot were separated by means of a splitter. 2. Experimental methods The load history was recorded using calibrated strain gauges mounted on the backing bar. The tests were ﬁlmed using a Phantom 7.1 High Speed Video Camera (Vision Research, USA) to conﬁrm that the test was carried out successfully. The projectile speeds were measured at the light gates and from the images taken using the camera. Typically the velocity of the projectile measured at the light gates was around 250 m/s. Tests were carried out on two identically prepared discs of each material. = + + + + V I I k I I I I ( ) ( ) m m m t t m m 181 190 147 267 181 190 (2) In quasi-static tests a projectile mounted in an aluminium holder was pushed against a ceramic disc using a 5582 Universal Testing System (Instron, USA) ﬁtted with a 50 kN load cell operating in com- pression mode (Fig. 2). The displacement rate was 10 μm/s and the motion was reversed when the load on the sample matched the max- imum load experienced during the dynamic test (5.5 kN). The load and displacement were logged every 0.1 s using Bluehill Control Software (Instron, USA). Tests were carried out on three identically prepared discs of each material. (2) where k is a constant. This correcting factor has been determined by comparison against known standards to be k = 0.33 ± 0.03 [29,30]. 3.2. Behaviour during dynamic impact and quasi-static indentation tests 3.2. Behaviour during dynamic impact and quasi-static indentation tests Typical dynamic test force-time proﬁles recorded from the strain- gauged backing bar are shown in Fig. 4. Both the debased alumina and the composite material force-time proﬁles consist of a single compres- sive peak with a magnitude of approximately 5.5 kN. However the plots for the TZP samples have a double compressive peak, with a peak of around 3 kN occurring approximately 10 μs after the ﬁrst peak of around 4 kN. The quasi-static tests all had very similar force-time proﬁles as they primarily reﬂect the deformation of the projectile against the hard and stiﬀceramics plates. One example is given in Fig. 4. The plot is non- linear owing to plasticity in the projectile tip, and has a sharp peak at the maximum load, at which point the testing machine was reversed. Projectiles were recovered from quasi-static tests but could not usually be recovered from dynamic tests due to the set-up required to enable ﬁlming of the test, with one exception. The tip shape of the projectile which was retrieved was examined using confocal optical microscopy and compared to examinations of typical projectiles prior to testing and after being used for a quasi-static test (Fig. 5). In compar- ison to the initially sharp tip following testing the projectiles are sig- niﬁcantly blunter after testing. The projectile from the quasi-static test has a blunted tip diameter around half that of the projectile recovered from the dynamic test. However as other dynamic test projectiles could not be recovered it is not possible to draw any general conclusions about the eﬀect of diﬀerent materials on the projectile blunting. All dynamic tests were ﬁlmed using high-speed video to verify successful impact had occurred. The two samples of the same material generally showed similar behaviour and impact regions. In one test carried out on a TZP sample, a bright ﬂash of light is observed just after impact of the projectile (Fig. 6). This phenomenon was not observed for other tests on TZP or other materials. Possible causes for this observa- tion are discussed in Section 4.3 below. Fig. 6. High speed video images showing the observed ﬂash of light during impact on a TZP sample. Table 1 Measured material properties of as-supplied samples. ng’s Modulus (GPa) Shear Modulus (GPa) Poisson’s Ratio Fracture Strength (MPa) ± 30 80.4 ± 3 0.33 ± 0.06 950 ± 100 ± 50 142 ± 6 0.26 ± 0.04 360 ± 40 ± 40 132 ± 6 0.25 ± 0.04 360 ± 30 Fig. 5. Comparison of confocal optical microscope line proﬁles across the as- manufactured sharp projectile tip and the blunted tips of projectiles after use in the dynamic and quasi-static tests. Material Density (g/cm3) Hardness (GPa) Grain Size (μm) Young’s Modulus (GPa) Shear Modulus (GPa) Poisson’s Ratio Fracture Strength (MPa) 3Y-TZP 6.06 ± 0.01 12.8 ± 0.2 0.69 ± 0.07 213 ± 30 80.4 ± 3 0.33 ± 0.06 950 ± 100 Zirconia-Alumina Composite 4.06 ± 0.01 18.3 ± 0.5 1.9 ± 0.3 357 ± 50 142 ± 6 0.26 ± 0.04 360 ± 40 Debased Alumina 3.74 ± 0.01 12.5 ± 0.3 5.8 ± 0.9 330 ± 40 132 ± 6 0.25 ± 0.04 360 ± 30 Fig. 4. Typical force-time curves for dynamic tests carried out on each material. A typical plot for a quasi-static test is also shown; note the diﬀerent time axis Fig. 5. Comparison of confocal optical microscope line proﬁles across the as- manufactured sharp projectile tip and the blunted tips of projectiles after use in the dynamic and quasi-static tests. 3) Hardness (GPa) Grain Size (μm) Young’s Modulus (GPa) Shear Modulus (GPa) Poisson’s Ratio Fracture Strength (MPa) Zirconia-Alumina Composite 4.06 ± 0.01 18.3 ± 0.5 1.9 ± 0.3 3 Debased Alumina 3.74 ± 0.01 12.5 ± 0.3 5.8 ± 0.9 3 Fig. 4. Typical force-time curves for dynamic tests carried out on each material. A typical plot for a quasi-static test is also shown; note the diﬀerent time axis scale for the dynamic and quasi-static tests. Fig. 5. Comparison of confocal optical microscope line proﬁles across the as- manufactured sharp projectile tip and the blunted tips of projectiles after use in the dynamic and quasi-static tests. Fig. 4. Typical force-time curves for dynamic tests carried out on each material. A typical plot for a quasi-static test is also shown; note the diﬀerent time axis scale for the dynamic and quasi-static tests. Fig. 6. High speed video images showing the observed ﬂash of light during impact on a TZP sample. 3.1. Properties of ceramic materials Following dynamic and static testing, discs were cleaned using acetone, and any residual steel from the projectile was removed by soaking in 16 M HCl solution for 5 min. The samples were examined using confocal optical microscopy (InﬁniteFocus Optical Microscope, Alicona, Austria) to determine the depth and extent of the impact craters. The cracks and other damage in and around the impact craters were also examined using scanning electron microscopy (JEOL 840 F) on gold-coated samples. Samples of the materials supplied were examined by a number of characterisation methods to aid interpretation of dynamic and quasi- static impact and indentation test results. Their microstructures are shown in Fig. 3 and a summary of measured properties is given in Table 1. The 3Y-TZP sample had a sub-micron grain size with a few larger grains (> 3 μm), usually interpreted as being cubic [22]. The debased alumina material contains large, blocky grains interspersed with a glassy matrix. The alumina-zirconia composite consisted of ∼2 μm alumina grains with smaller zirconia grains both in between and inside the alumina grains. The elastic properties reﬂected the major con- stituent of each material. The composite was the hardest material and the 3Y-TZP the strongest, by a signiﬁcant margin. Optical luminescence microscopy was carried out using a LabRAM ARAMIS Raman microscope (Jorbin Horiba, Japan) with a 633 nm red He-Ne laser and a Renishaw inVia Reﬂex Raman Microscope with a 633 nm red He-Ne laser and a 422 nm blue He-Cd laser. A 50x optical lens was used on both instruments. Line scans were obtained by Fig. 3. SEM micrographs showing the microstructure of the materials. a) TZP, b) debased alumina, c) alumina-zirconia composite. Samples have been polished and thermally etched. Fig. 3. SEM micrographs showing the microstructure of the materials. a) TZP, b) debased alumina, c) alumina-zirconia composite. Samples have been polished and thermally etched. 4938 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Table 1 Measured material properties of as-supplied samples. 3.3. Post-test characterisation Proﬁlometry line scans across the impacted regions measured using the confocal optical microscope show signiﬁcant variation in the da- mage inﬂicted on each material by the quasi-static and dynamic tests 4939 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Fig. 7. Confocal optical proﬁlometry of damaged regions in tested samples. Top row: Quasi-statically tested samples; bottom row: dynamically tested samples. (a,b) 3Y-TZP, (c,d) debased alumina, (e,f) alumina-zirconia composite. Sharp peaks in the spectra are due to debris on the surface which remained after cleaning with HCl. C.E.J. Dancer, et al. Journal of the European Ceram ic Society 39 (2019) 4936–4945 Fig. 7. Confocal optical proﬁlometry of damaged regions in tested samples. Top row: Quasi-statically tested samples; bottom row: dynamically tested samples. (a,b) 3Y-TZP, (c,d) debased alumina, (e,f) alumina-zirconia composite. Sharp peaks in the spectra are due to debris on the surface which remained after cleaning with HCl. Fig. 7. Confocal optical proﬁlometry of damaged regions in tested samples. Top row: Quasi-statically tested samples; bott 3Y-TZP, (c,d) debased alumina, (e,f) alumina-zirconia composite. Sharp peaks in the spectra are due to debris on the surfac (Fig. 7). In general for each material the impact craters were deeper for quasi-static tests than for dynamic tests. Repeated tests at each strain rate generated qualitatively similar damage for each set of identical samples. The extent of damage, assessed from the depth and width of the crater, is larger for the debased alumina samples than for the zir- conia-containing materials. samples the cracks were signiﬁcantly longer and resulted in greater loss of material (Fig. 9f). The ring cracks had smaller diameters for the quasi-statically tested samples (Fig. 9c) than for the dynamically tested samples (Fig. 9d). The ring crack diameters were similar to the pro- jectile blunted tip diameter for the corresponding projectiles (see Sec- tion 3.2), being larger for the dynamically tested samples than for the quasi-static test. Damaged regions in each material after quasi-static and dynamic tests were also examined in detail using low magniﬁcation optical mi- croscopy and SEM. Some residual steel is observed in the sample sur- faces (black regions in Fig. 8a–b, Fig. 9a–b, and Fig. 10a–b) despite cleaning with HCl. Damage to the zirconia-containing samples was less extensive and therefore more diﬃcult to observe, particularly using optical microscopy where little cracking was observed (Figs. 8 and 10a- b). 3.3. Post-test characterisation In the 3Y-TZP sample the impact region was only distinguishable from the rest of the sample by slight roughness and residual staining from the projectile but no micro- or macro-cracking could be observed by optical microscopy (Fig. 8) or in SEM (not shown). In the composite samples (Fig. 10), a single ring crack was observed using SEM on a dynamic tested sample and some cracking was observed inside this ring crack (Fig. 10d), however overall it should be noted that the levels of damage were signiﬁcantly lower than for the debased alumina. Cr3+ ﬂuorescence measurements of R1 peak shift were converted to hydrostatic residual stresses using Eq. (1) and are shown in Fig. 11. The most striking diﬀerence between the materials is that the residual stress ﬁelds at the centre of the indentations in alumina-zirconia composite and quasi-statically tested 3Y-TZP are tensile (positive shift) (Fig. 11a–b) while that for the debased alumina samples is compressive (Fig. 11c–d). The residual stress in the impacted region has a larger magnitude (whether compressive or tensile) for the quasi-statically tested samples than for the dynamically tested samples. While damaged regions in the debased alumina and alumina-zirconia composite mate- rials can be fairly easily observed from these plots, the data for the dynamically tested 3Y-TZP is more dispersed, with both tensile and compressive residual stresses measured (Fig. 11a–b). This is consistent with the visual observation that limited levels of damage were The damage to the debased alumina samples (Fig. 9) was extensive and easily observed even using low magniﬁcation optical microscopy (Fig. 9c-d). In addition to multiple concentric ring cracks, around 10 radial surface cracks extending approximately 1 mm into the sur- rounding material were observed for the quasi-statically tested sample (Fig. 9c). Crack deﬂections were observed in these radial cracks (Fig. 9e) following the glass-crystal boundary before returning to the original direction. Either side of this deﬂection intragranular fracture is visible. While fewer radial cracks were observed for the dynamic tested Fig. 8. Selected optical images of damaged regions in selected 3Y-TZP samples after (a) quasi-static testing and (b) dynamic testing. ected optical images of damaged regions in selected 3Y-TZP samples after (a) quasi-static testing and (b) dynamic testing. 4940 C.E.J. Dancer, et al. Journal of the European Ceram ic Society 39 (2019) 4936–4945 Fig. 9. 3.3. Post-test characterisation Optical (a, b) and scanning electron microscopy (c–f) images of damaged regions in selected debased alumina samples after quasi-static testing (a, c, e) and dynamic testing (b, d, f). Note the diﬀerent magniﬁcations due to the varying scale of the defects observed. Fig. 9. Optical (a, b) and scanning electron microscopy (c–f) images of damaged regions in selected debased alumina samples after quasi-static testing (a, c, e) and dynamic testing (b, d, f). Note the diﬀerent magniﬁcations due to the varying scale of the defects observed. Fig. 9. Optical (a, b) and scanning electron microscopy (c–f) images of damaged regions in selected debased alumina sam dynamic testing (b, d, f). Note the diﬀerent magniﬁcations due to the varying scale of the defects observed. sustained by this sample (Fig. 8). sustained by this sample (Fig. 8). static) has occurred in the damaged regions of the samples. In the quasi- static sample the damaged region has a higher average monoclinic ZrO2 content than for the dynamically tested sample but across a narrower region. The width of the aﬀected region is of the same order as that of the region of tensile residual stress observed in the Cr3+ ﬂuorescence spectra (Fig. 11a–b). The monoclinic ZrO2 phase contents in the impacted or indented regions in the 3Y-TZP and alumina-zirconia composite samples were calculated using Eq. (2) using the measured tetragonal and monoclinic peak areas from Raman spectra acquired for Zr-containing samples. Fig. 12 shows the results of these calculations plotted as a function of position across the damaged site for the zirconia-alumina composite samples. It is clear that in the damaged regions of both quasi-statically and dynamically tested samples signiﬁcantly elevated levels of mono- clinic ZrO2 are observed. Away from the impact site the monoclinic zirconia content of these samples was around 17 vol.% (dynamic) and 29 vol.% (quasi-static), which indicates that a signiﬁcant increase in the monoclinic content to up to 93 vol.% (dynamic) and 98 vol.% (quasi- The results for the Raman spectra taken from the 3Y-TZP samples are shown in Fig. 13. Unlike the zirconia-alumina composite, few spectra from the damaged region contained peaks due to monoclinic ZrO2, leading to the limited experimental data obtained. Monoclinic zirconia could not be detected in the 3Y-TZP samples before impact. As the detection limit for monoclinic zirconia using this method is lower than 10 vol.% [29] so it is clear that signiﬁcantly less transformation as Fig. 10. Selected optical (a, b) and scanning electron microscopy (c, d) images of damaged regions in selected alumina-zirconia composite samples after quasi-static testing (a, c) and dynamic testing (b, d). Note the diﬀerent magniﬁcations due to the varying scale of the defects observed. Fig. 10. Selected optical (a, b) and scanning electron microscopy (c, d) images of damaged regions in selected alumina-zirconia composite samples after quasi-static testing (a, c) and dynamic testing (b, d). Note the diﬀerent magniﬁcations due to the varying scale of the defects observed. 4941 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Fig. 11. Cr3+ ﬂuorescence data linescans across the indented/impacted regions of the ceramic discs showing the distribution of residual stresses. 4.1. Comparison of damage in the diﬀerent ceramics The debased alumina showed the most visible damage after testing (Figs. 7 and 9). Multiple ring cracks (associated with blunt indentation) were observed, along with radial cracks (associated with sharp in- dentation) originating from the ring-crack region [11]. The quasi-sta- tically tested samples had shorter radial cracks which appeared to follow the grain boundaries quite closely. The radial cracks indicate some permanent deformation, probably associated with the glassy phase present in this material. The dynamically-tested samples by contrast had a smaller number of longer cracks which appeared wider as grains had been expelled from the cracked region. This diﬀerence probably stems from the greater blunting and hence longer range stresses of the projectile in the dynamic tests. sustained by this sample (Fig. 8). Top row - quasi- statically tested samples; bottom row - dynamically tested samples. (a,b) 3Y-TZP, (c,d) debased alumina, (e,f) alumina-zirconia composite. Dashed vertical lines indicate the approximate region of damage on each sample. Fig. 11. Cr3+ ﬂuorescence data linescans across the indented/impacted regions of the ceramic discs showing the distribution of residual stresses. Top row - quasi- statically tested samples; bottom row - dynamically tested samples. (a,b) 3Y-TZP, (c,d) debased alumina, (e,f) alumina-zirconia composite. Dashed vertical lines indicate the approximate region of damage on each sample. a proportion of the zirconia present has occurred in the 3Y-TZP sample compared to the zirconia-alumina composites. As for the zirconia-alu- mina composite samples, peaks for monoclinic zirconia are observed over a large region of the surface for the dynamically-tested 3Y-TZP sample compared to the quasi-statically tested sample. plastic deformation and therefore less radial cracking and more pro- jectile blunting and therefore less ring cracking. No microcracking was observed and the levels of damage were fairly similar for the quasi- static and dynamically tested samples. Overall the 3Y-TZP showed the most resistance to damage at both strain rates (Fig. 8), followed by the alumina-zirconia composite, and ﬁnally the debased alumina. The suppression of cracking can be ex- plained in simple terms through the higher strength and toughness of this ceramic. This indicates that incorporating zirconia into alumina materials, or using 3Y-TZP if weight is not an issue, may result in ar- mour with improved properties compared to pure alumina while still retaining the processing advantages of oxide-based materials. 4.3. Physical origin of peaks in the force-time proﬁle The shape of the impact curve is noticeably diﬀerent for the TZP samples compared to the composite and debased alumina samples, consisting of two peaks rather than one. In the case of the composite and debased alumina samples, the compressive peak can be considered as a simple balance between the plastic deformation of the projectile causing an increase in the force as the projectile blunts and hence the rising curve, and the elastic rebound of the projectile, which pushes the projectile away from the surface of the sample and hence accounts for the falling side of the peak. This is consistent with the fact that in the zirconia-alumina com- posite samples, the region in which a tensile residual stress was mea- sured by Cr3+ ﬂuorescence from the alumina also contains the trans- formed monoclinic zirconia. In addition the quasi-statically tested zirconia-alumina composite sample, which had a higher tensile residual stress across a smaller area, showed higher volume fractions of mono- clinic zirconia within that same area compared to the dynamically tested sample, where in both cases peak values were lower but the ef- fect spread over a wider region owing to the greater blunting of the sample. Similar measurements by Huang et al. [35] on the residual stress in Vickers indentations in ZTA samples indicated that residual stress inside the indentation region is compressive. However, the de- grees of plastic deformation and constraint and hence the indentation- induced stress levels are much greater for Vickers indentations than for the present impact sites and therefore masked the eﬀect of the zirconia transformation. In the case of TZP, Fig. 4 shows that the force-time curve has a lower peak value than for the other ceramics but stops the projectile by ap- plying force for a longer duration. The stiﬀness of TZP is of the same order as that of debased alumina and shows little damage after the test, so this surprising behaviour must be attributed to the unique mechan- ical properties of TZP. We suggest that the initial high tensile stresses around the impact site trigger the tetragonal to monoclinic transfor- mation, allowing considerable deformation which absorbs the initial impact. 4.2. Projectile deformation The diameter of the blunted tip of the projectile following the dy- namic test was approximately twice that in the equivalent quasi-static test (see Fig. 5). This can be simply explained in terms of the heating associated with plastic deformation. In quasi-static heating where the test is slow there is time for, the excess thermal energy to be conducted away from the tip region, so the eﬀect of the heating is only observed during dynamic testing which occurs over a much shorter timescale. The blunting can be modelled by considering the deformation of the projectile tip from a sharp point to a blunt cylinder during the test. The derivation of the model is given in the Appendix. This model calculates a temperature change due to the deformation of 444 K. In the alumina-zirconia composite less damage was observed with only a single ring crack evident and no radial cracks (Fig. 10). The reduction in cracking is presumed to be a consequence of the higher hardness compared with the debased alumina, which leads to less Fig. 12. Volume fraction of monoclinic phase (as a proportion of the total zirconia content only) varying with distance across damaged regions for the alumina-zirconia composite samples. (a) Quasi-statically tested sample, (b) dynamically tested sample. Values were cal- culated using Eq. (2) from peak areas measured from Raman spectra. Dashed vertical lines in- dicate the approximate region of damage on each sample. 4942 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Fig. 13. Volume fraction of monoclinic phase varying with position across damaged regions for the 3Y-TZP samples as measured from Raman spectra (a) quasi-statically tested sample, (b) dynamically tested sample. Due to the limited number of spectra containing monoclinic peaks, ﬁve line scans were taken for each sample with 50μm spacings between each line. Spectra from each linescan are re- presented by diﬀerent shaped and coloured points. Dashed vertical lines indicate the ap- proximate region of damage on each sample. Fig. 13. Volume fract varying with position for the 3Y-TZP samp Raman spectra (a) sample, (b) dynamical the limited number monoclinic peaks, ﬁv for each sample with each line. Spectra fro presented by diﬀeren points. Dashed vertica proximate region of da Fig. 13. 4.3. Physical origin of peaks in the force-time proﬁle As the projectile comes to a halt, the force it exerts on the ceramic decreases and constraint from the untransformed material around the impact site in conjunction with the local temperature in- crease forces most of the monoclinic phase to transform back to the tetragonal polytype. The thermodynamic forces involved coupled with the time taken for the reversal to occur produce the second “kickback peak” seen in Fig. 4. Reverse transformation in TZP at room tempera- ture has been seen previously [31] and this explanation is supported by the small amount of monoclinic zirconia seen in Fig. 13, despite the high forces involved. The low volume fraction of zirconia in the com- posite prevents it from producing the same eﬀect. While the proportion of the zirconia remaining in the transformed, monoclinic phase was signiﬁcantly lower in the TZP than in the com- posite, it is important to account for the diﬀerences in zirconia content between the two materials. The 3Y-TZP samples were almost 100% Y2O3-ZrO2 solid solution with a small amount of alumina impurity. The zirconia-alumina composites, by contrast, contained approximately 6.8 vol.% zirconia, with the balance being alumina. Therefore while a signiﬁcantly greater proportion of the zirconia has transformed in the zirconia-alumina composites, the actual volume of material undergoing the transformation in each case may not be signiﬁcantly diﬀerent, perhaps restricted due to the orientation of grains relative to the sur- face, as observed by Deville et al. [36]. This, and the greater stiﬀness of the alumina matrix surrounding the transformed material in the com- posite may account for the higher tensile stresses it contained (Fig. 11). The same super-elastic cycle is assumed to have occurred in the quasi static tests as well but since the displacement in the unloading part of the tests was dominated by the elastic compliance of the testing machine, no eﬀect was observed in the load vs. displacement curves. The transformation may also be associated with the capture in the high-speed video images of one test on this material of a ﬂash of light during the impact of the projectile (Fig. 6). This may simply be sparking of the steel projectile as it impacts the target but could also be me- chanoluminescence, which is well known anecdotally to occur during grinding or sawing of TZP and has recently been investigated in more detail [32]. It is also known to occur in Ti-doped monoclinic zirconia 4.2. Projectile deformation Volume fraction of monoclinic phase varying with position across damaged regions for the 3Y-TZP samples as measured from Raman spectra (a) quasi-statically tested sample, (b) dynamically tested sample. Due to the limited number of spectra containing monoclinic peaks, ﬁve line scans were taken for each sample with 50μm spacings between each line. Spectra from each linescan are re- presented by diﬀerent shaped and coloured points. Dashed vertical lines indicate the ap- proximate region of damage on each sample. [33,34] though the mechanisms involved are yet to be identiﬁed. [33,34] though the mechanisms involved are yet to be identiﬁed. [33,34] though the mechanisms involved are yet to be identiﬁed. This is an approximate value due to the assumptions made in the calculation but it does show that large temperature rises are likely in the projectile tip during impact. This must be considered when com- paring the eﬀects of quasi-static and dynamic testing as the thermal eﬀects will lower the eﬀective yield stress of the projectile material. The ceramic would therefore appear to be stronger during dynamic testing than during quasi-static testing, whereas the correct interpretation is that the projectile becomes weaker, blunting more, and hence causing lower levels of damage. The compressive residual stresses around the impact site in the debased alumina are similar to those observed previously [11,12] and result from limited plastic deformation under the impact site. Although the overall hardness of the debased alumina in the present case is much greater than the yield stress of the projectile, the glassy phase is much softer and allowed suﬃcient deformation to produce the relatively modest stresses measured here. The tensile stresses seen in the zirconia-containing materials (Fig. 11) can be explained when it is considered that the Cr3+ ﬂuor- escence measurements probe only the alumina in the specimens. Both the composite and the TZP contain some transformed monoclinic zir- conia and this would be heavily under compression owing to the ˜ 4% expansion associated with the transformation. It follows that the re- maining material, including the alumina, would be under tension to maintain the balance of forces. 5. Conclusions Tests designed to directly compare the eﬀects of indentation at diﬀerent strain rates have been carried out on three ceramic materials 4943 C.E.J. Dancer, et al. Journal of the European Ceram ic Society 39 (2019) 4936–4945 which combines the lower density of alumina or a carbide material with the superior mechanical properties of zirconia, and this will be the subject of future research. using a sharpened rod of hardened steel as the indenter/projectile. Of the materials tested, 3Y-TZP displayed the best resistance to impact, followed by alumina-zirconia composite and then the debased alumina. This ranking was determined by comparing the results of measurements of residual stress, extent and nature of cracking in the surface, and the size and depth of the damaged zone. Acknowledgements The authors would like to thank Chris Hampson of Morgan Technical Ceramics for the provision of the materials studied in this work. We also thank Dr GeoﬀWest, Prof. Clive Siviour, Dr Euan Wielewski, Dr Robert Paynter, Dr Igor Dyson, Dr Andrew Norton, Richard Duﬃn, Dick Froud, Laurie Walton, and Bob Lloyd for their help and advice. This study was carried out as part of the Understanding and Improving Ceramic Armour (UNICAM) project, jointly funded by the EPSRC (Grant Number EP/G042675/1) and DSTL United Kingdom Defence Science and Technology Laboratory. CEJD would also like to acknowledge EPSRC grant EP/R029873/1 for funding the work carried out at Warwick. The Renishaw inVia Reﬂex Raman Microscope used in this research was obtained through “Birmingham Science City: Innovative Uses for Advanced Materials in the Modern World” with support from Advantage West Midlands (AWM) and part funded by the European Regional Development Fund (ERDF). Materials displayed better resistance to impact when tested dyna- mically than when tested quasi-statically. This can be accounted for by considering the eﬀects of heating of the projectile which results in changes to the projectile during the test, limiting the damage caused. This indicates that potential changes to the projectile must be con- sidered when comparing the apparent strength of ceramic materials measured by similar tests. The superior performance of the 3Y-TZP to damage during these tests was attributed to reversible transformation of tetragonal zirconia to the monoclinic phase during the test. This is assumed to explain the “double peak” shape of the dynamic impact test force-time curve, in which the momentum of the impact was absorbed over a longer period, reducing the maximum force and the extent of the damage. This be- haviour was not observed in the alumina-zirconia composite material, which was attributed to the low tetragonal zirconia content of the material used in this work. However a composite with a higher tetra- gonal zirconia content may be suitable for use as an armour material Data published in this article can be freely downloaded from http:// wrap.warwick.ac.uk/120450. Derivation of the temperature change involved in the deformation of the projectile as stated in Section 4.2. The projectile shape changes during impact as shown in Fig. 5, which is represented mathematically as in the schematic Fig. A1 below. From the assumed geometry shown in Fig. A1 above: Appendix A Derivation of the temperature change involved in the deformation of the projectile as stated in Section 4.2. The projectile shape changes during impact as shown in Fig. 5, which is represented mathematically as in the schematic Fig. A1 below. From the assumed geometry shown in Fig. A1 above: = + πr b πr a b 1 3 ( ) 2 2 2 2 ∴ = a b 2 = + πr b πr a b 1 3 ( ) 2 2 2 2 ∴ = a b 2 where a, r2 and b are the dimensions of the deforming parts of the projectile as shown in Fig. A1. It is also evident that: = + = r a b θ a θ ( ) tan 3 2 tan 2 (A.3) = + = r a b θ a θ ( ) tan 3 2 tan 2 Assuming a mean yield stress σy the plastic work done in blunting the end is: Assuming a mean yield stress σy the plastic work done in blunting the end is: ∫ ∫ = = πr σ da π θ σ r dr π θ σ r 2 3 tan 2 9 tan a y r y y 0 2 2 0 2 2 2 2 3 2 (A.4) Fig. A1. Schematic diagram of the deformations considered in analysing the projectile tip shape changes during quasi-static and dynamic testing. a and r1 are the height and base radius of the cone-shaped tip of the projectile which is ﬂattened during the test. b and r2 are the height and radius of the blunt end of the projectile resulting from this deformation. ∫ ∫ = = πr σ da π θ σ r dr π θ σ r 2 3 tan 2 9 tan a y r y y 0 2 2 0 2 2 2 2 3 2 (A.4) Fig. A1. Schematic diagram of the deformations considered in analysing the projectile tip shape changes during quasi-static and dynamic testing. a and r1 are the height and base radius of the cone-shaped tip of the projectile which is ﬂattened during the test. b and r2 are the height and radius of the blunt end of the projectile resulting from this deformation. Fig. A1. Schematic diagram of the deformations considered in analysing the projectile tip shape changes during quasi-static and dynamic testing. References [21] R.C. Garvie, R.H. Hannink, R.T. Pascoe, Ceramic Steel? Nature 258 (1975) 703–704. [1] S.M. Walley, Historical review of high strain rate and shock properties of ceramics relevant to their application in armour, Adv. Appl. Ceram. 109 (2010) 446–466, https://doi.org/10.1179/174367609X422180. [22] R.H.J. Hannink, P.M. Kelly, B.C. Muddle, Transformation toughening in zirconia- containing ceramics, J. Am. Ceram. Soc. 83 (2000) 461–487. [2] J.C. LaSalvia, J. Campbell, J.J. Swab, J.W. McCauley, Beyond hardness: ceramics and ceramic-based composites for protection, J. Mater. 62 (2010) 16–23. [23] C. Huang, Y. Chen, Eﬀect of varied alumina / zirconia content on ballistic perfor- mance of a functionally graded material, Int. J. Refract. Metals Hard Mater. 67 (2017) 129–140, https://doi.org/10.1016/j.ijrmhm.2017.04.002. p p [3] I.G. Crouch, Body armour - New materials, new systems, Def. Technol. 15 (2019) 241–253, https://doi.org/10.1016/j.dt.2019.02.002. [24] G.R. Anstis, P. Chantikul, B.R. Lawn, D.B. Marshall, A critical evaluation of in- dentation techniques for measuring fracture toughness: I, direct crack measure- ments, J. Am. Ceram. Soc. 64 (1981) 533–538. [4] A. Healey, J. Cotton, S. Maclachlan, P. Smith, J. Yeomans, Understanding the ballistic event: methodology and initial observations, J. Mater. Sci. 52 (2017) 3074–3085, https://doi.org/10.1007/s10853-016-0594-0. [25] Advanced Technical Ceramics — Mechanical Properties of Monolithic Ceramics at Room Temperature — Part I: Determination of Flexural Strength, (2006), p. 3. 3074–3085, https://doi.org/10.1007/s10853-016-0594-0. kf d h bl f d d [5] J. Lankford, Mechanisms responsible for strain-rate-dependent compressive strength in ceramic materials, J. Am. Ceram. Soc. 64 (1981) C33–C34. [26] Advanced technical ceramics — Mechanical Properties of Monolithic Ceramics at Room Temperature — Part 2: Determination of Young’s modulus, Shear Modulus and Poisson’s Ratio, (2006), pp. 1–40 Vol. 3. [6] J. Lankford, High strain rate compression and plastic ﬂow of ceramics, J. Mater. Sci. Lett. 15 (1996) 745–750, https://doi.org/10.1007/BF00274593. [27] M.I. Mendelson, Average grain size in polycrystalline ceramics, J. Am. Ceram. Soc. 52 (1969) 443–446, https://doi.org/10.1111/j.1151-2916.1969.tb11975.x. [7] S.K. Chung, Fracture characterization of armor ceramics, Am. Ceram. Soc. Bull. 69 (1990) 358–366. [8] J. Sternberg, Material properties determining the resistance of ceramics to high velocity penetration, J. Appl. Phys. 65 (1989) 3417, https://doi.org/10.1063/1. 342659. [28] J. He, D.R. Clarke, Determination of the piezospectroscopic coeﬃcients for chro- mium-doped sapphire, J. Am. Ceram. Soc. 78 (1995) 1347–1353. [29] J.A. Muñoz Tabares, M.J. Anglada, Quantitative analysis of monoclinic phase in 3Y- TZP by Raman spectroscopy, J. Am. Ceram. Soc. 93 (2010) 1790–1795, https://doi. org/10.1111/j.1551-2916.2010.03635.x. [9] N.K. Bourne, J.C.F. Millett, M. Chen, J.W. References McCauley, D.P. Dandekar, On the Hugoniot elastic limit in polycrystalline alumina, J. Appl. Phys. 102 (2007) 073514, , https://doi.org/10.1063/1.2787154. [30] C.S. Lim, T.R. Finlayson, F. Ninio, J.R. Griﬃths, In-situ measurement of the stress- induced phase transformations in magnesia-partially-stabilized zirconia using Raman spectroscopy, J. Am. Ceram. Soc. 75 (1992) 1570–1573. [10] D.E. Munson, R.J. Lawrence, Dynamic deformation of polycrystalline alumina, J. Appl. Phys. 50 (1979) 6272–6282. [11] C.E.J. Dancer, H.M. Curtis, S.M. Bennett, N. Petrinic, R.I. Todd, High strain rate indentation-induced deformation in alumina ceramics measured by Cr3+ ﬂuores- cence mapping, J. Eur. Ceram. Soc. 31 (2011) 2177–2187, https://doi.org/10. 1016/j.jeurceramsoc.2011.06.002. [31] M. Saran, Strengthening of Zirconia at High Temperatures, University of Oxford, 2008. [32] K. Wang, L. Ma, X. Xu, S. Wen, Y. Zhang, Enhancement of triboluminescence in the presence of CO < inf > 2 < /inf > by sliding between silica and yttria-stabilized zirconia, Langmuir 31 (2015) 8224–8227, https://doi.org/10.1021/acs.langmuir. 5b01335. [12] C.E.J. Dancer, H.M. Curtis, S.M. Bennett, N. Petrinic, R.I. Todd, Measurement of deformation in alumina samples indented at high strain rates, Ceram. Eng. Sci. Proc. 32 (2011) 227–236. [33] M. Akiyama, C.-N. Xu, K. Nonaka, Intense visible light emission from stress-acti- vated ZrO[sub 2]:Ti, Appl. Phys. Lett. 81 (2002) 457, https://doi.org/10.1063/1. 1494463. [13] B. Lawn, R. Wilshaw, Indentation fracture: principles and applications, J. Mater. Sci. 10 (1975) 1049–1081, https://doi.org/10.1007/BF00823224. [34] B.P. Chandra, Persistent mechanoluminescence induced by elastic deformation of ZrO2:Ti phosphors, J. Lumin. 130 (2010) 2218–2222, https://doi.org/10.1016/j. jlumin.2010.06.023. [14] P.J. Hazell, Ceramic Armour: Design and Defeat Mechanisms, Argos Press, 2006. d d k l f b ll [15] E. Medvedovski, Alumina ceramics for ballistic protection part I, Am. Ceram. Soc. Bull. 81 (2002) 27–32. [16] W.W. Chen, A.M. Rajendran, B. Song, X. Nie, Dynamic fracture of ceramics in armor applications, J. Am. Ceram. Soc. 90 (2007) 1005–1018, https://doi.org/10.1111/j. 1551-2916.2007.01515.x. [35] S. Huang, J.G.P.G.P. Binner, B. Vaidhyanathan, R.I.I. Todd, Quantitative analysis of the residual stress and dislocation density distributions around indentations in alumina and zirconia toughened alumina (ZTA) ceramics, J. Eur. Ceram. Soc. 34 (2014) 753–763, https://doi.org/10.1016/j.jeurceramsoc.2013.09.021. [17] P. Forquin, G. Rossiquet, J. Zinszner, B. Erzar, Microstructure inﬂuence on the fragmentation properties of dense silicon carbides under impact, Mech. Mater. 123 (2018) 59–76, https://doi.org/10.1016/j.mechmat.2018.03.007. [36] S. Deville, G. Guénin, J. Chevalier, Martensitic transformation in zirconia Part I. Nanometer scale prediction and measurement of transformation induced relief, Acta Mater. 52 (2004) 5697–5707, https://doi.org/10.1016/j.actamat.2004.08. 034. [18] B. Koeppel, G. Appendix A a and r1 are the height and base radius of the cone-shaped tip of the projectile which is ﬂattened during the test. b and r2 are the height and radius of the blunt end of the projectile resulting from this deformation. 4944 Journal of the European Ceram ic Society 39 (2019) 4936–4945 C.E.J. Dancer, et al. Assuming uniform and adiabatic heating of the deformed material, and substituting using the geometric relationships in Eqs. (A.2) and (A.3), the temperature rise ΔT is: = = ΔT π θ σ r cρπr b σ cρ 2 9 tan 2 3 y y 2 3 2 2 (A.5) where c is the heat capacity and ρ is the density of the steel projectile material. It can be assumed that 90% of the work done becomes heat (after Rogers [37]) and that all of the heat remains in the deformed region throughout the deformation time and so gives an increase in temperature. For the steel used for the projectile c = 490 J kg−1 K-1 and ρ =8000 kg m-3, and the mean yield stress σy of 2.9 GPa can be assessed using the relationship that Vickers hardness is approximately 3σy for metals (as measured in Section 2 as 8.6 GPa). Therefore from Eq. A.5 ΔT is calculated to be 444 K. [20] G. Subhash, Dynamic indentation testing, ASM Handb. Vol. 8 Mech. Test. ASM, 2000, pp. 519–529. [37] H.C. Rogers, Adiabatic plastic deformation, Annu. Rev. Mater. Sci. 9 (1979) 283–311, https://doi.org/10.1146/annurev.ms.09.080179.001435. References Subhash, Characteristics of residual plastic zone under static and dynamic Vickers indentations, Wear 224 (1999) 56–67, https://doi.org/10.1016/ S0043-1648(98)00328-7. [37] H.C. Rogers, Adiabatic plastic deformation, Annu. Rev. Mater. Sci. 9 (1979) 283–311, https://doi.org/10.1146/annurev.ms.09.080179.001435. [19] R.J. Anton, G. Subhash, Dynamic Vickers indentation of brittle materials, Wear 239 (2000) 27–35. 4945
https://openalex.org/W4378070687	https://periodicos.univali.br/index.php/rtva/article/download/18671/11356	Portuguese	null	Gestão da Experiência do Cliente: Perspectivas Contemporâneas na Hotelaria	Turismo: visão e ação/Turismo : visão e ação	2,023	cc-by	11,951	Thais Bandinelli Vargas Lopes de Oliveira¹, Ana Paula Lisboa Sohn¹ ¹Universidade do Vale do Itajaí, Balneário Camboriú, Santa Catarina, Brasil Data de submissão: 13/06/2022 – Data de aceite: 15/12/2022 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ URISMO ISÃO & ÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 GESTÃO DA EXPERIÊNCIA DO CLIENTE: PERSPECTIVAS CONTEMPORÂNEAS NA HOTELARIA CUSTOMER EXPERIENCE MANAGEMENT: CONTEMPORARY PERSPECTIVES IN HOSPITALITY GESTIÓN DE LA EXPERIENCIA DEL CLIENTE: PERSPECTIVAS CONTEMPORÁNEAS EN HOSTELERÍA Thais Bandinelli Vargas Lopes de Oliveira¹, Ana Paula Lisboa Sohn¹ ¹Universidade do Vale do Itajaí, Balneário Camboriú, Santa Catarina, Brasil Data de submissão: 13/06/2022 – Data de aceite: 15/12/2022 TURISMO VISÃO & AÇÃO Keywords: Abstract: Customer experience management; Guest experience; Hospitality; Experience economy; Systematic review. Customer experience management has been a key strategy for hotel compa- nies considered leaders in their segments. To understand this phenomenon, we propose a systematic review aimed at identifying the practices, benefi ts and trends of customer experience management in the hotel industry. Con- ducted in the Scopus database, the research included a sample of 241 arti- cles (time frame 2011-2021). The twenty most cited articles in the sample were accessed for analysis, and bibliometric fi ndings from the total sample complemented the discussion. The results indicate that the main benefi ts are the maintenance of competitive advantage and business diff erentiation. Best practices are associated with human capital, use of technology, and service design. Trends point to the evolution of the concept of experience, greater attention to the employee experience, the search for mutual benefi ts, and the use of technology to encourage engagement and co-creation. In addition to the research agenda, this study advances by signaling the growing impor- tance of experience management in hospitality, that its success depends on joint eff orts, and that it is necessary to act actively in the face of the constant transformation of services, the economy, and consumer behavior in order to keep the hotel company competitive and aligned with the expectations of its guest. REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Resumo: Gestão da experiência do cliente; Experiência do hóspede; Hotelaria; Economia da experiência; Revisão sistemática. A gestão da experiência do cliente tem sido estratégia-chave de empresas hoteleiras tidas como líderes em seus segmentos. Para entender esse fe nômeno, propõe-se uma revisão sistemática visando a identificar as práti cas, os benefícios e as tendências da gestão da experiência do cliente na hotelaria. Realizada na base de dados Scopus, a pesquisa contemplou uma amostra de 241 artigos (recorte temporal de 2011-2021). Os 20 artigos mais citados da amostra tiveram seu conteúdo acessado para análise e achados bibliométricos da amostra total complementaram a discussão. Os resultados assinalam como principal benefício a manutenção da vantagem competitiva e a diferenciação do negócio. Melhores práticas estão associadas ao capital humano, uso de tecnologia e design de serviço. Tendências apontam a evolu ção do conceito de experiência, maior atenção à experiência do colaborador, busca por benefícios mútuos e o uso da tecnologia para incentivar enga jamento e cocriação. Além da agenda de pesquisa, este estudo avança ao sinalizar a crescente importância da gestão da experiência na hotelaria, que seu sucesso depende de esforços conjuntos, sendo preciso agir ativamente frente à constante transformação dos serviços, da economia e do compor tamento do consumidor, a fim de manter a empresa hoteleira competitiva e alinhada às expectativas do seu hóspede. Thais: Mestranda do Programa de Pós-graduação em Turismo e Hotelaria (PPGTH) da Universidade do Vale do Itajaí (UNIVALI). E-mail: tbvloliveira@gmail.com. Orcid: https://orcid.org/0000-0003-1856-1968. Ana Paula: Doutora em Engenharia de Produção pela Universidade Federal de Santa Catarina, Professora titular do Programa de Pós-graduação em Turismo e Hotelaria da Universidade do Vale do Itajaí (UNIVALI). E-mail: ana sohn@univali.br. Orcid: https://orcid.org/0000-0002-7319-8869. 359 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL Resumen Entretanto, conceituar experiência do cliente é uma tarefa não tão simples: ela é uma combinação de elementos que, juntos, envolvem o consumidor de maneira emocional, psicológica, intelectual e espiritual (Mossberg, 2007). Ao proporcionar experiên cias memoráveis criadas por meio da combinação de estratégias e ativos do meio de hospedagem e envolver os hóspedes emocionalmente, é possível gerar uma vantagem competitiva e despertar um comportamento de recompra (Kariru et al., 2017). Tendo sido reconhecida como um fator importante que afeta comportamentos positivos dos clien tes, como lealdade e recomendação (Cetin & Walls, 2016), a experiência do cliente (EC) tem recebido maior atenção das empresas do setor de serviços, principalmente em função do impacto relevante e percebido das experiências positivas de seus consumidores. Com os avanços tecnológicos e tudo que a era digital tem permitido realizar, o cliente tem esperado por personalização de serviços em cada ponto de contato dele com o negócio. Uma pesquisa da Gartner (2014 apud Kandampully et al., 2018) aponta que a experiência do clien te era vista, por cerca de 89 por cento das empresas participantes do estudo, como sendo a base competitiva dos seus negócios. Destaca-se que a gestão da experiência do cliente (GEC) já tem, há algum tempo, aparecido como estratégia-chave de muitas empresas do ramo da hospitalidade, notoriamente naquelas que se preocupam com o nível de excelência de seus serviços e são tidas como empresas líderes em seu segmento. Marriot, Hilton, Starwood e Disney são exemplos de em presas de área de hospitalidade que se sobressaem na criação de experiências para seus clientes (Kandampully et al., 2018). Também é preciso falar sobre a economia da experiência e como ela tem impactado o setor de serviços. Pine e Gilmore (1998) alertavam desde o século passado que, com o aumento das expec tativas dos consumidores e uma grande oferta de produtos e serviços, empresas só se destacariam se oferecessem experiências memoráveis. Nos últimos anos, com a economia da experiência cada vez mais emergente, é notável o movimento de parte do setor de hospitalidade para atender de forma mais personalizada aos gostos e necessidades de seus hóspedes, adicionando mais valor à sua oferta (Bharwani & Mathews, 2016). Resumen Gestión de la experiencia del cliente; Experiencia del huésped; Hostelería; Economía de la experiencia; Revisión sistemática. La gestión de la experiencia del cliente ha sido una estrategia clave para las empresas hoteleras consideradas líderes en sus segmentos. Para entender este fenómeno, proponemos una revisión sistemática destinada a identifi car las prácticas, benefi cios y tendencias de la gestión de la experiencia del clien- te en la industria hotelera. Realizada en la base de datos Scopus, la investi- gación incluyó una muestra de 241 artículos (marco temporal de 2011-2021). Se accedió al contenido de los veinte artículos más citados de la muestra para su análisis y se complementó la discusión con los hallazgos bibliométri- cos de la muestra total. Los resultados señalan como principal benefi cio el mantenimiento de la ventaja competitiva y la diferenciación empresarial. Las mejores prácticas se asocian al capital humano, el uso de la tecnología y el diseño de servicios. Las tendencias apuntan a la evolución del concepto de experiencia, una mayor atención a la experiencia del empleado, la búsqueda de benefi cios mutuos y el uso de la tecnología para fomentar el compromiso y la cocreación. Además de la agenda de investigación, este estudio avan- za señalando la creciente importancia de la gestión de la experiencia en la hostelería, que su éxito depende de esfuerzos conjuntos, y que es necesa- rio actuar activamente ante la constante transformación de los servicios, la economía y el comportamiento del consumidor, para mantener la empresa hotelera competitiva y alineada con las expectativas de su huésped. 360 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO INTRODUÇÃO Um dos desafios mais críticos da indústria hoteleira é relacionado à gestão das experiências dese jadas pelos hóspedes (Cetin & Walls, 2016). Para que uma empresa prestadora de serviços, como é o caso de um meio de hospedagem, consiga se manter competitiva atualmente, é preciso redefinir suas estratégias, mudando o foco de serviços para experiências – as quais têm sido vistas como pro postas de valor distintas e mais eficazes do que as formas tradicionais de diferenciação no setor de turismo e hospitalidade (Cetin & Walls, 2016). Resumen Entender a EC e garantir que ela será gerida de forma efetiva em cada ponto de interação do hóspe de com o meio de hospedagem tornou-se um objetivo para as empresas que desejam ser líderes de seus mercados (Kandampully et al., 2018). Uma pesquisa da Forrester Research (2014) citada pelo estudo de Kandampully et al. (2018), revelou que a maioria das iniciativas em experiência do cliente estudadas falhou e que poucas companhias atingiam um estágio de maturidade na GEC. Com o advento da economia da experiência e a comoditização dos serviços, a experiência do cliente tem sido cada vez mais vital para a indústria hoteleira e seu estudo se faz necessário para que a adaptação necessária no setor flua (Mody et al., 2019), buscando-se meios de interpretar a deman ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL 361 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 da e adaptar a oferta ao novo estilo de vida dos consumidores (Gilmore & Pine II, 2002), cuja satisfa ção e disposição de pagar está, cada dia mais, relacionada à qualidade da experiência (Pizam, 2010). da e adaptar a oferta ao novo estilo de vida dos consumidores (Gilmore & Pine II, 2002), cuja satisfa ção e disposição de pagar está, cada dia mais, relacionada à qualidade da experiência (Pizam, 2010). Dada a relevância da temática, o presente artigo propõe uma revisão sistemática híbrida visando a identificar as práticas e os benefícios obtidos com a gestão da experiência do cliente na hotelaria, bem como tendências associadas à temática. Com um protocolo e fluxo de revisão previamente estabelecidos, a estratégia de busca foi aplicada na base de dados Scopus, com recorte temporal de 2011-2021, resultando em uma amostra de 241 artigos, que foram analisados com ferramentas bibliométricas. Para dar suporte às discussões dos achados, os 20 artigos mais citados da amostra tiveram seu conteúdo analisado e suas contribuições foram trazidas para enriquecer os resultados. REVISÃO TEÓRICA 362 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Indo para o contexto hoteleiro, um setor marcado por atividades com um elevado grau de envol vimento, pode-se conceituar operacionalmente a experiência do hóspede como um conjunto de eventos ou impressões memoráveis que engajam clientes, de forma emocional e pessoal durante a sua hospedagem, influenciando futuras intenções de compra e de compartilhamento da experi ência – fidelidade e marketing boca-a-boca (recomendação) (Cetin & Dincer, 2014). Fato é que os hóspedes não procuram mais apenas uma boa cama, um bom café da manhã, um bom chuveiro – procuram experiências que elevem essas características funcionais do hotel. Por isso, experiências são a razão de alguns hóspedes pagarem um valor maior por essas características. Assim, o estudo dos fatores que influenciam experiência do hóspede é de interesse tanto acadê mico quanto mercadológico, assim como fidelidade e recomendação, dois outputs de experiências que são constantemente mensurados em pesquisas acadêmicas, assim como a satisfação. Sobre a última, Cetin e Dincer (2014) afirmam que ela não é mais suficiente para criar fidelidade – e é nesta lacuna que a experiência vem ganhando terreno. A hospitalidade está deixando de ser um ramo focado no produto, com forte uso de ativos físicos, para ser um ramo focado no consumidor e na sua experiência, canalizando seus esforços para pro porcionar uma oferta inovadora e holística, lapidada pela subjetividade de cada hóspede, criando uma conexão pessoal e emocional com cada um deles, na intenção de criar a experiência memorá vel (Bharwani & Mathews, 2016), tal como conceito apresentado por Pine e Gilmore (1999). Esse olhar inovador e holístico se deve à adaptação necessária ao mutável comportamento do con sumidor – o hóspede contemporâneo anseia por experiências diferenciadas, de qualidade superior, multiculturais, tendo suas percepções básicas de serviço já consideravelmente alteradas (Bharwani & Mathews, 2016). REVISÃO TEÓRICA A experiência do cliente pode ser entendida como a resposta subjetiva do cliente ao encontro holís tico e indireto com a empresa (Lemke et al., 2011). Experiências e serviços são ofertas econômicas distintas, como argumentam Pine e Gilmore (1998, p. 3): “uma experiência ocorre quando uma empresa usa intencionalmente os serviços como palco e bens como adereços, para engajar clientes individuais de uma forma que crie um evento memorável”. Pine e Gilmore (1999) examinaram as experiências do cliente de acordo com quatro categorias: es tética, educação, entretenimento e escapista. Essas categorias estão dispostas em duas dimensões de entendimento da experiência: participação (ativa e passiva) e conexão (absorção e imersão). Para criar experiências autênticas seria preciso mesclar as quatro categorias na composição da oferta (Silva & Barreto, 2019), engajando o consumidor em situações e de formas diferentes, dentro da experiência como um todo. Assim, dentro do conceito de Economia da Experiência, a empresa passa a ofertar, além de seus produtos e serviços-base, sensações personalizadas. O maior diferencial da Economia da Experiên cia é que, diferentemente dos commodities, bens e serviços, que são exteriores ao consumidor, as experiências são intrínsecas, subjetivas, próprias de cada consumidor – não haverá uma mesma ex periência para duas pessoas, pois cada uma delas será afetada de maneiras diferentes ao interagir com a oferta (Martinelli, 2013). O resultado esperado da interação de cada indivíduo com a experiência vai além do valor econômi co que ele se dispõe a pagar para usufrui-la: há ganho de valor para a marca, refletido em outputs como fidelidade, marketing boca-a-boca, com clientes se tornando admiradores e defensores da empresa. Esse tipo de condução ao se criar uma experiência – e ao se construir também uma nova relação com os consumidores – é uma forma de se obter e sustentar uma vantagem competitiva dentro de um mercado com concorrência (Silva & Barreto, 2019). Conforme a economia da experiência avança, é fato que muitos negócios acompanharão o ritmo e outros ficarão pelo caminho. Entender o que de fato representa uma experiência e o que seus clien tes desejam encontrar, mantendo-se em constante atualização, acompanhando o fluxo de novida des e encantamento, é o que pode fazer a diferença na personalização e sobrevivência no mercado (Hwang & Seo, 2016; Kandampully et al., 2018; Rahimian et al., 2021). REVISÃO TEÓRICA A gestão da experiência do consumidor tem se tornado um fator-chave na hospitalidade contempo rânea: é o processo de gerenciar estrategicamente toda a experiência do cliente com um produto ou empresa (Hwang & Seo, 2016; Rahimian et al., 2021). Maior atenção à GEC começou a ser dada quando Pine e Gilmore, no seu livro Economia da Experiência, reforçaram o valor econômico de se oferecer uma experiência e a importância de se performar eventos para criar uma experiência positiva (Pine & Gilmore, 1999; Hemmington, 2007; Bharwani & Mathews, 2016). A cocriação merece destaque como um agente potencializador dessas experiências (Chathoth et al., 2016): a partir de sua interação nos pontos de contato pré, durante e pós-serviço, consumidor e empresa podem criar, juntos, uma experiência única. A mudança de orientação – de produto/ser viço para o foco no cliente – é essencial para a jornada de cocriação de valor entre empresa e con sumidores, uma vez que eles já não querem mais só uma boa experiência: eles querem cocriar de forma ativa ao longo do processo, na procura por autenticidade e uma experiência personalizada, diferenciada e que atenda às suas expectativas (Prahalad & Ramaswamy, 2004). A cocriação requer o engajamento do consumidor na criação da sua própria experiência, o que demanda, por parte da empresa, a busca de um relacionamento, envolvimento, interesse e ação para antecipar soluções e criar condições para que haja motivação para esse engajamento (Roy et al., 2020). Esse movimento é uma das temáticas na agenda de pesquisa de Hwang e Seo (2016), juntamente como a questão do marketing experiencial na hotelaria. Os autores frisam a importância de en tender bem o público-alvo antes de se planejar a experiência e os valores que se deseja entregar, procurando disponibilizá-los de forma holística e integrada. Afirmam também que esse processo de GEC requer dos gerentes hoteleiros atenção com relação à compatibilização da experiência com o conceito e a filosofia da empresa, considerando, dentre outros fatores, o fator humano: diversos stakeholders se envolvem na (co)criação da experiência e eles precisam estar conectados e em har monia para que a qualidade da entrega esteja assegurada. Corroborando os argumentos anteriores quanto à importância da EC e de sua gestão, a revisão de Kandampully et al. REVISÃO TEÓRICA Não estar atento a esse movimento é ficar para trás no mercado, que tem pas sado por importantes transformações – como as novas tecnologias – em meio a intensificação da concorrência (Pine II & Gilmore, 2011; Rahimian et al., 2021). Considerando que a experiência do cliente seria formada por todas as interações dele com a em presa ao longo da prestação do serviço, a estratégia do negócio e a sua posição no mercado, bem como o conhecimento acerca do comportamento e anseios dos clientes são fatores influenciadores da entrega; a autenticidade deve ser colocada como prioridade, para que experiências não se tor nem eventos padronizados, impessoais e comoditizados, perdendo sua essência e efeito desejado – uma entrega diferenciada, memorável e com impacto positivo no comportamento do consumidor. Assim, os negócios do ramo de hospitalidade têm se adaptado à economia da experiência na busca de vantagem competitiva por meio da criação de modelos memoráveis de serviços e/ou produtos que despertem não só o interesse dos clientes, mas também a disposição deles para pagar pelo que será ofertado (Silva & Barreto, 2019). Considerando a dinâmica do comportamento do consumidor, as empresas têm buscado oferecer experiências diferenciadas, mutualmente benéficas, nutrindo relacionamentos de longo prazo (Rahimian et al., 2021; Yoon & Lee, 2017). Importante lembrar que o conceito de experiência do cliente é multidisciplinar (Cetin & Dincer, 2014) e suscita outros constructos, como satisfação e encantamento: um, é o resultado esperado de um serviço; o outro, o resultado esperado de uma experiência. O encantamento já é apontado como 363 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259-381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ um indicador mais eficaz de gestão do relacionamento com o hóspede, sendo entendido como uma resposta positiva do consumidor, fruto de expectativas superadas na entrega do serviço: clientes encantados são aqueles que tiveram experiências de serviço extraordinárias (Kao et al., 2016). um indicador mais eficaz de gestão do relacionamento com o hóspede, sendo entendido como uma resposta positiva do consumidor, fruto de expectativas superadas na entrega do serviço: clientes encantados são aqueles que tiveram experiências de serviço extraordinárias (Kao et al., 2016). REVISÃO TEÓRICA (2018) enfatiza que a EC não é o resultado de um só encontro, mas de todas as situações em que o consumidor esteve em contato com a empresa, ressaltando também o aspecto da cocriação gerada dessa interação. Destaca-se também que, na hotelaria, a experiência não é gerada apenas pelos serviços-‘base’ (alimentos e bebidas, entretenimento, acomodações), mas por todos serviços e processos que suportam a operação, como aporte tecnológico e as relações inter pessoais – do hóspede com a equipe, entre ele e outro hóspede ou, ainda, com outros stakeholders participantes do processo. Assim, o estudo da GEC deve ir além da díade cliente-empresa e encontros de serviço. Essa gestão vai além do que as empresas possam, de fato, controlar, pois envolve processos nos quais ela não pode interferir, como é o caso das interações entre consumidores (online ou offline), as avaliações deixadas por hóspedes em plataformas digitais, entre outras situações (Kandampully et al., 2018). Envolve, também, processos que antecedem os encontros de serviço e/ou ocorrem fora da estrutu ra física do meio de hospedagem. 364 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259-381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359-381 Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Nesse sentido, sobre as experiências que antecedem a prestação física do serviço, as que ocorrem por meio das interações digitais do hóspede com o hotel, quando positivas, podem trazer aumen to da confiança na oferta, construção de relacionamento e aumento no valor agregado à marca hoteleira. Investir em um website ou aplicativo com atributos mais atrativos pode colaborar para experiências online mais prazerosas e divertidas, podendo trazer inúmeros benefícios ao meio de hospedagem, como mais reservas diretas, maior engajamento do hóspede e outras respostas rela cionadas ao comportamento de compra e relacionamento com a marca (Bilgihan et al., 2014, 2015). Considerando que a percepção de valor do serviço já começa antes de realmente se usufruir dele, há de se considerar também que a qualidade do serviço impacta diretamente o comportamento do consumidor: afeta, por exemplo, o apego emocional à marca e esse apego influencia na fidelidade/ lealdade do cliente. REVISÃO TEÓRICA Por isso, são necessários ferramentas e treinamentos para que a equipe adote uma estratégia centrada no cliente, de forma a se buscar uma conexão emocional – criar e nutrir relacionamento – com os hóspedes (Oliveira et al., 2022). Hemsley-Brown e Alnawas (2016, p. 2786) afirmam que a lealdade “pode ser alcançada quando os hotéis se posicionam, projetam suas ins talações e decorações e desenvolvem experiências de hóspedes baseadas em valores simbólicos e profundos aspectos emocionais, e não sobre a satisfação do cliente em si”. É um passo adiante na escala de progressão de valor econômico. Ainda que Drucker (1973 apud Kandampully et al., 2018) tenha afirmado que satisfazer o cliente é missão de qualquer negócio, atualmente a satisfação já tem efeitos limitados, sendo superada pela EC positiva e seus outputs. Em meio a essa transição, oferecer um serviço deixou de ser suficiente (Cetin & Dincer, 2014; Gilmore & Pine II, 2002). A experiência tem sido vista como uma estratégia de diferenciação: muitos hotéis se tornam cada vez mais semelhantes entre si, com padrões e proces sos replicados. Gerenciar meios de hospedagem como locais de experiência seria uma alternativa para evitar essa “armadilha de commodities” e sustentar, de fato, uma vantagem competitiva (Bil gihan et al., 2014; Gilmore & Pine II, 2002). Hemmington (2007, p. 749) resume essa a importância de se despertar para esse movimento: “os clientes não compram serviços, compram experiências; eles não compram qualidade, compram memórias”. METODOLOGIA Para alcançar o objetivo do presente artigo, delineou-se uma revisão sistemática de literatura (RSL), uma metodologia que auxilia na revisão de literatura relevante sobre o tema de estudo e que per mite vislumbrar a amplitude e profundidade dos trabalhos existentes, bem como identificar novas possibilidades de pesquisa. O método também proporciona um nível de compreensão mais amplo e preciso sobre um tema do que uma revisão de literatura tradicional (Pati & Lorusso, 2018). Op tou-se por realizar, como etapa adicional e inicial, uma análise bibliométrica da amostra total, no intuito de complementar a avaliação subjetiva da revisão de literatura e de reforçar a identificação de tendências e padrões (Zupic & Čater, 2015). A pesquisa foi realizada na base de dados Scopus, escolhida por sua considerável representação da produção científica, bem como pelas características do seu sistema de busca e suas funcionalidades bibliométricas, que serviram de apoio à RSL. Foram realizadas buscas com diferentes combinações das palavras-chave pertinentes ao tema, com os termos em português e inglês, dentro de um re corte temporal mais amplo (2011 a 2021), contemplando apenas publicações em revistas acadê 365 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 micas. Os termos combinados na estratégia de busca foram: ‘experience economy’, ‘hotel’, ‘guest experience’, ‘customer experience’, ‘consumer experience’, ‘customer experience management’ e ‘hospitality’. micas. Os termos combinados na estratégia de busca foram: ‘experience economy’, ‘hotel’, ‘guest experience’, ‘customer experience’, ‘consumer experience’, ‘customer experience management’ e ‘hospitality’. Não sendo localizados artigos em língua portuguesa na base escolhida e dada a pouca incidência de artigos de outros idiomas, optou-se por considerar apenas os artigos em língua inglesa. A es tratégia de busca e o protocolo adotado na revisão, que têm por base as recomendações da lista de verificação PRISMA 2020 (Page et al., 2021) e o fluxo para mapeamento científico com métodos bibliométricos (Zupic & Čater, 2015), estão detalhados na Figura 1. Após a aplicação da estratégia de busca, observados os critérios de inclusão (recorte temporal, arti gos publicados em revistas acadêmicas, em língua inglesa) e feito o refinamento do banco de dados, obteve-se uma amostra de 241 artigos. METODOLOGIA Assim, iniciou-se o processo de análise por meio da biblio metria dos dados bibliográficos dos artigos dessa amostra, com a realização de um mapeamento da produção científica sobre a temática abordada. Para essa etapa, foram utilizadas duas ferramentas: o VOSviewer, pela sua capacidade de gerar um mapa graficamente representativo, permitindo observar a correlação entre as palavras-chave (Feng et al., 2020); e a ferramenta Bibliometrix e sua interface gráfica, a Biblioshiny, do Pacote-R (RStudio), para apoio aos estágios de visualização dos dados gerais e relevantes da amostra e apresentação do mapa temático (Aria & Cuccurullo, 2017). Ainda considerando os critérios de inclusão para a RSL, os vinte artigos mais citados da amostra foram localizados por meio da ferramenta Citation Overview, disponibilizada pela base de dados Scopus e escolhida pela sua capacidade de exibir a tendência de citação de um conjunto de docu mentos e analisar o impacto geral de publicações em uma área de pesquisa (Scopus, 2016). Con sideradas pesquisas de relevância dentro da temática proposta, a análise desse subconjunto da amostra enseja a discussão da temática, apoiada pelas análises bibliométricas que destacaram as temáticas mais relevantes da amostra total. 366 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO T Vi ã A ã 25 2 259 381 M i /A 2023 \| htt //d d i /10 14210/ t 25 2 359 381 Figura 1. Protocolo da Revisão, Estratégia de Busca e Fluxo para Mapeamento Científico Figura 1. Protocolo da Revisão, Estratégia de Busca e Fluxo para Mapeamento Científico Fonte: elaborado pelos autores (2021). Fonte: elaborado pelos autores (2021). RESULTADOS E DISCUSSÕES Com o auxílio da ferramenta Bibliometrix e sua interface gráfica, a Biblioshiny, do Pacote-R (RStu dio), a análise bibliométrica dos 241 artigos permitiu visualizar os dados gerais da amostra, que apresenta 599 autores, 99 periódicos e 996 palavras-chave. A produção científica na temática demonstrou um crescimento considerável de 2017 em diante, com uma taxa de crescimento de 20,45% ao longo do recorte temporal (2011-2021). Dos periódicos com mais publicações, desta cam-se o International Journal of Contemporary Hospitality Management e o International Journal of Hospitality Management, ambos com 27 publicações e o Journal of Hospitality Marketing and Management, com 17 artigos. 367 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Figura 2. Crescimento anual da produção científica Fonte: elaborado pelos autores com auxílio do Biblioshiny (2021). Figura 2. Crescimento anual da produção científica Figura 2. Crescimento anual da produção científica Fonte: elaborado pelos autores com auxílio do Biblioshiny (2021). Fonte: elaborado pelos autores com auxílio do Biblioshiny (2021). Sobre a produção acadêmica por países, há uma prevalência de estudos norte-americanos (132), seguido por países asiáticos, que concentram os estudos mais recentes. Países europeus têm par ticipação relevante dentro da amostra total e o Brasil tem apenas um artigo dentro dessa amostra. Sobre a produção acadêmica por países, há uma prevalência de estudos norte-americanos (132), seguido por países asiáticos, que concentram os estudos mais recentes. Países europeus têm par ticipação relevante dentro da amostra total e o Brasil tem apenas um artigo dentro dessa amostra. Figura 3. Autores-destaque em produção ao longo do tempo Fonte: elaborado pelas autoras com auxílio do Biblioshiny (2021). Figura 3. Autores-destaque em produção ao longo do tempo Fonte: elaborado pelas autoras com auxílio do Biblioshiny (2021). ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL 368 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. RESULTADOS E DISCUSSÕES 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Com relação aos autores, os que tiveram produção em destaque no recorte temporal estão repre sentados na Figura 3. Alguns deles, como S. Bharwani, D. Mathews, F. Okumus e A. Bilgihan estão presentes na revisão teórica dessa pesquisa. Parte dos demais autores consta também na autoria dos 20 artigos mais citados da amostra e inclusos na discussão, apresentados no Quadro 1. Esses artigos tiveram seu conteúdo analisado para conduzir a discussão das temáticas suscitadas pela análise da estrutura conceitual da amostra total. Quadro 1. Vinte artigos mais citados da amostra Autores Artigo Citações 1. Xiang et al. (2015) What can big data and text analytics tell us about hotel guest experience and satisfaction? 510 2. Chathoth et al. (2016) Co-creation and higher order customer engagement in hospita lity and tourism services: A critical review 227 3. Lu et al. (2019) Developing and validating a service robot integration willingness scale 220 4. Walls et al. (2011) Understanding the consumer experience: An exploratory study of luxury hotels 201 5. Mody et al. (2017) The accommodation experiencescape: a comparative assess ment of hotels and Airbnb 200 6. Gao & Mattila (2014) Improving consumer satisfaction in green hotels: The roles of perceived warmth, perceived competence, and CSR motive 184 7. Tung & Au (2018) Exploring customer experiences with robotics in hospitality 176 8. Kandampully et al. (2018) Customer experience management in hospitality: A literature synthesis, new understanding and research agenda 168 9. Ren et al. (2016) Exploring customer experience with budget hotels: Dimensiona lity and satisfaction 141 10. Calheiros et al. (2017) Sentiment Classification of Consumer-Generated Online Re views Using Topic Modeling 134 11. Walls (2013) A cross-sectional examination of hotel consumer experience and relative effects on consumer values 118 12. Radojevic et al. (2015) Ensuring positive feedback: Factors that influence customer satisfaction in the contemporary hospitality industry 117 13. Frederik et al. (2014) The limits of attraction 115 14. Hwang & Seo (2016) A critical review of research on customer experience manage ment: Theoretical, methodological and cultural perspectives 106 Quadro 1. RESULTADOS E DISCUSSÕES Vinte artigos mais citados da amostra 369 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 15. Bilgihan et al. (2015) Online experiences: Flow theory, measuring online customer experience in e-commerce and managerial implications for the lodging industry 103 16. Cetin & Dincer (2014) Influence of customer experience on loyalty and word-of-mouth in hospitality operations 103 17. Dedeoglu et al. (2018) The impact of servicescape on hedonic value and behavioral intentions: The importance of previous experience 102 18. Merli et al. (2019) Why should hotels go green? Insights from guests experience in green hotels 102 19. Pencarelli (2020) The digital revolution in the travel and tourism industry 101 20. Birinci et al. (2018) Comparing customer perceptions of hotel and peer-to-peer accommodation advantages and disadvantages 98 Fonte: extraído pelos autores do Citation Overview, Scopus (2021). Iniciando a análise da estrutura conceitual da amostra, apresenta-se a rede de coocorrência de pa lavras-chave, buscando o melhor entendimento dos temas abordados por meio das conexões entre eles dentro do assunto pesquisado e a identificação das questões mais recentes e importantes (Aria & Cuccurullo, 2022; Cobo et al., 2011). Com auxílio do software VOSviewer, criou-se um mapa basea do em dados bibliográficos, considerando a coocorrência de palavras-chave, adotando método full counting com número mínimo de ocorrências de cinco por termo. Das 996 palavras-chave, 40 aten deram ao limite, formando o mapa de coocorrência abaixo, mostrando cinco clusters temáticos. A explanação dos clusters será feita com base nos artigos mais citados da amostra. 15. Bilgihan et al. (2015) Online experiences: Flow theory, measuring online customer experience in e-commerce and managerial implications for the lodging industry 103 16. Cetin & Dincer (2014) Influence of customer experience on loyalty and word-of-mouth in hospitality operations 103 17. Dedeoglu et al. (2018) The impact of servicescape on hedonic value and behavioral intentions: The importance of previous experience 102 18. Merli et al. (2019) Why should hotels go green? Insights from guests experience in green hotels 102 19. Pencarelli (2020) The digital revolution in the travel and tourism industry 101 20. Birinci et al. RESULTADOS E DISCUSSÕES (2018) Comparing customer perceptions of hotel and peer-to-peer accommodation advantages and disadvantages 98 Iniciando a análise da estrutura conceitual da amostra, apresenta-se a rede de coocorrência de pa lavras-chave, buscando o melhor entendimento dos temas abordados por meio das conexões entre eles dentro do assunto pesquisado e a identificação das questões mais recentes e importantes (Aria & Cuccurullo, 2022; Cobo et al., 2011). Com auxílio do software VOSviewer, criou-se um mapa basea do em dados bibliográficos, considerando a coocorrência de palavras-chave, adotando método full counting com número mínimo de ocorrências de cinco por termo. Das 996 palavras-chave, 40 aten deram ao limite, formando o mapa de coocorrência abaixo, mostrando cinco clusters temáticos. A explanação dos clusters será feita com base nos artigos mais citados da amostra. Iniciando a análise da estrutura conceitual da amostra, apresenta-se a rede de coocorrência de pa lavras-chave, buscando o melhor entendimento dos temas abordados por meio das conexões entre eles dentro do assunto pesquisado e a identificação das questões mais recentes e importantes (Aria & Cuccurullo, 2022; Cobo et al., 2011). Com auxílio do software VOSviewer, criou-se um mapa basea do em dados bibliográficos, considerando a coocorrência de palavras-chave, adotando método full counting com número mínimo de ocorrências de cinco por termo. Das 996 palavras-chave, 40 aten deram ao limite, formando o mapa de coocorrência abaixo, mostrando cinco clusters temáticos. A explanação dos clusters será feita com base nos artigos mais citados da amostra. 370 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ Figura 4. Rede de coocorrência de palavras-chave Figura 4. Rede de coocorrência de palavras-chave Fonte: extraído pelos autores, da base de dados Scopus e VOSViewer (2021). Fonte: extraído pelos autores, da base de dados Scopus e VOSViewer (2021). Fonte: extraído pelos autores, da base de dados Scopus e VOSViewer (2021). No cluster vermelho, que tem como destaques ‘tourism’ e ‘hospitality’, há um movimento relaciona do à tecnologia e presença digital em conjunto com o engajamento, cocriação e gestão da experiên cia. RESULTADOS E DISCUSSÕES Estudos da amostra, como os citados a seguir, se dedicam a analisar dados disponíveis online – o conteúdo gerado pelo usuário – para entender melhor os hóspedes e gerir não só a experiência, mas a melhoria da oferta hoteleira como um todo (Radojevic et al., 2015). Estudos como os de Xiang et al. (2015) e Calheiros et al. (2017), que abordam a experiência do hóspede a partir da mineração de big data e análise de sentimento das avaliações de serviço feitas nas plataformas online e redes sociais, a fim de se capturar as percepções subjetivas dos consumidores sobre os serviços usufru ídos. As informações extraídas das avaliações de serviço são de grande valia para o aumento da vantagem competitiva, uso nas estratégias de marketing e relacionamento, como o CRM e criação de um banco de dados de inteligência do cliente, sendo estes últimos fundamentais na GEC e para melhorar a experiência do hóspede (Calheiros et al., 2017). 371 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 O estudo de Bilgihan et al. (2014) traz a temática das experiências online, do engajamento do hós pede com elas, do início da cocriação da experiência já no online. Cocriação e engajamento também são temas do estudo de Chathoth et al., (2016), que traz importantes contribuições acerca da im portância do papel dos colaboradores e sua interação com os hóspedes no processo de cocriação, e das mudanças de cultura e gestão que permitam isso, com o uso de tecnologia para elevar o relacionamento e engajamento na tentativa de se criar valor superior nas transações de serviço, argumento também corroborado por Pencarelli (2020). O segundo cluster, representado pela cor verde e pelo termo ‘customer experience’ têm conexões fortes com todos os clusters e apresenta termos como encantamento, lealdade à marca, inovação e robôs. Estudos como os de Tung e Au (2018) e Lu et al. (2019) explanam sobre como a tecnologia tem se tornado importante na criação de experiência, focando na introdução dos robôs de serviço com inteligência artificial na hotelaria e na aceitação deles por parte dos hóspedes. RESULTADOS E DISCUSSÕES Em ambos os es tudos, o antropomorfismo (dar aparência humana aos robôs) foi identificado como um fator de não aceitação do serviço. Porém, constatou-se que, ao contrário das tecnologias de autoatendimento, os robôs conseguem socializar com o cliente, o que permite que eles apoiem os serviços de hos pitalidade, reforçando o toque humano essencial na experiência do hóspede. O impacto do papel funcional e interacional dos robôs na experiência merece mais estudos e os autores sugerem uma ampla agenda de pesquisa, como aceitação cultural e reflexos na cultura organizacional. O cluster azul, ancorado nas palavras-chave, ‘hotel industry’, ‘experience’, ‘tourist behavior’ e ‘per ception’, traz questões da economia compartilhada, economia da experiência e comportamento do turista. Dois exemplos são artigos de Birinci et al. (2018) e Mody et al. (2017), que estudam o fenôme no da hospedagem em duas modalidades – hotelaria e Airbnb –, ressaltando o comportamento do hóspede quanto às suas preferências, expectativas e percepção desses dois modais de hospedagem. Segundo o estudo de Birinci et al. (2018), autenticidade é percebida como um fator desejado pelos hóspedes e que ambos os tipos de hospedagem devem apostar em formas de trabalhá-la dentro da oferta: um exemplo é engajar os hóspedes com a comunidade e experiências locais. A pesquisa de Mody et al. (2017) segue essa mesma linha e um achado importante é o fato de que os meios de hospedagem precisam trabalhar melhor os fatores socioculturais da experiência do hóspede, visto que esses têm se destacado nas preferências dos viajantes modernos. Já sobre o termo ‘perception’, Walls (2013) investigou o conceito de EC e seu papel influenciando o valor percebido pelos hóspedes. Dentre seus achados, itens referentes ao ambiente físico e às interações humanas foram os que tiveram o impacto mais significativo e positivo na experiência do hóspede, sendo que o primeiro aspecto é onde os hóspedes percebiam grande parte do valor relacionado ao custo-benefício de sua hospedagem. Porém, estudos que vieram após, com o de Cetin & Dincer (2014) e Dedeoglu et al. (2018), já trazem outra leitura da percepção dos hóspedes, demonstrando que os hóspedes esperam equilíbrio entre o arranjo físico e as interações humanas. O estudo de Dedeoglu et al. (2018) detectou que hóspedes frequentes tendem a acumular mais valor emocional quando a equipe o reconhece e lembra de suas preferências, fato importante na GEC e de relacionamento. RESULTADOS E DISCUSSÕES E, quanto ao servicescape, foi reco mendada atenção às tendências de design e preferências do público, especialmente dos hóspedes jovens, para adequação dos espaços a fim de se obter mais efeitos positivos na experiência. 372 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 O artigo de Cetin e Dincer (2014) versa, especialmente, sobre a influência da experiência do cliente na lealdade e marketing boca-a-boca, alertando que até mesmo os clientes satisfeitos mudam de marca em busca de outras experiências. A pesquisa assinala, também, a importância de se compre ender os anseios do público-alvo, uma vez que experiências são suscetíveis a diversos fatores, como a subjetividade do cliente, situacionais, etc. A pesquisa acima mencionada está também ligada ao cluster amarelo, que destaca as palavras ‘guest experience’ e ‘loyalty’. Estudos como os de Walls et al. (2011) e Ren et al. (2016) procuram entender a experiência do hóspede dentro de um determinado nicho. O estudo de Ren et al. (2016) buscou elencar quais itens são importantes na EC em hotéis econômicos e revelou-se que elemen tos básicos das acomodações foram os mais significativos, visto que em hotéis desse porte eles se tornam mais evidentes em função do custo-benefício. Já o estudo de Walls et al. (2011) se dedicou a entender a percepção da EC em hotéis de luxo e ressal tou a importância de habilidades interpessoais da equipe hoteleira para lidar com as características de cada hóspede, entender como ele gostaria de ser atendido e poder, assim, melhorar a sua ex periência. Suporte tecnológico (CRM), alinhamentos e treinamentos foram apontados como meios de se cultivar e disseminar o conhecimento das preferências do consumidor e suas expectativas, de modo que toda equipe possa ter acesso e utilizar essas informações na criação de experiências. Esse mesmo estudo também traz a ideia de que a interação entre hóspedes pode impactar a expe riência individual de cada um. Kandampully et al. (2018) afirma, nesse sentido, que interações entre hóspedes são imperativas na experiência. O quinto e último cluster traz como destaque ‘customer satisfaction’ e reflete questões como a qualidade, esforços de marketing, emoção e sustentabilidade. RESULTADOS E DISCUSSÕES A discussão das temáticas do mapa será também acompanhada pelos artigos mais citados dentro da amostra, indicados próximos aos temas discutidos na sequência. Figura 5. Mapa temático por palavras-chave da amostra Fonte: elaboração própria (2022). Temáticas atreladas à experiência e seus outputs, como o seu design, encantamento e qualidade, assim como as questões relativas à tecnologia no processo da experiência, como o uso de robôs (Lu et al., 2019; Tung & Au, 2018), machine learning e mineração textual, avaliações online e plataformas (Bilgihan et al., 2014; Calheiros et al., 2017; Radojevic et al., 2015; Xiang et al., 2S015), são tidas como significativas e multidisciplinares. O fato de o cluster ‘customer experience’ estar nesse quadrante ressalta o amplo espectro do conceito de experiência e de todas as dinâmicas que orbitam em torno dele como, por exemplo, a de cunho mais gerencial, como a gestão de equipes, que aparece de for ma significativa nos estudos analisados (Hwang & Seo, 2016; Kandampully et al., 2018). Por sua vez, gestão da experiência no turismo e hotelaria, marketing, valor percebido (Walls, 2013), cocriação (Chathoth et al., 2016), lealdade e marketing boca-a-boca (Cetin & Dincer, 2014) são significativas e bem desenvolvidas dentro da amostra. As temáticas de ambos os quadrantes são fortemente pre sentes nos artigos da revisão teórica e base para os mais citados. Figura 5. Mapa temático por palavras-chave da amostra Fonte: elaboração própria (2022). Figura 5. Mapa temático por palavras-chave da amostra Fonte: elaboração própria (2022). Temáticas atreladas à experiência e seus outputs, como o seu design, encantamento e qualidade, assim como as questões relativas à tecnologia no processo da experiência, como o uso de robôs (Lu et al., 2019; Tung & Au, 2018), machine learning e mineração textual, avaliações online e plataformas (Bilgihan et al., 2014; Calheiros et al., 2017; Radojevic et al., 2015; Xiang et al., 2S015), são tidas como significativas e multidisciplinares. O fato de o cluster ‘customer experience’ estar nesse quadrante ressalta o amplo espectro do conceito de experiência e de todas as dinâmicas que orbitam em torno dele como, por exemplo, a de cunho mais gerencial, como a gestão de equipes, que aparece de for ma significativa nos estudos analisados (Hwang & Seo, 2016; Kandampully et al., 2018). RESULTADOS E DISCUSSÕES Dois estudos dos 20 mais citados se dedicam a esse último tema, correlacionando-o às emoções positivas, satisfação, qualidade e marketing. Gao & Mattila (2014) argumentam que hóspedes que compram produtos/serviços sus tentáveis desfrutam de uma sensação altruística e de bem-estar, por estarem contribuindo, de algu ma forma, com o bem coletivo, ao serem socialmente responsáveis em suas decisões de consumo. Ao terem experiências positivas em meios de hospedagem sustentáveis, sua satisfação costuma ser mais elevada do que em meios de hospedagem que não adotem essas práticas. A pesquisa de Merli et al. (2019) complementa essa ideia, ao dizer que práticas sustentáveis potencializam a sa tisfação e intenções comportamentais dos hóspedes, trazendo vantagens competitivas ao meio de hospedagem, desde que ele saiba comunicar efetivamente suas práticas e seu comprometimento com a causa. A próxima análise da estrutura conceitual é uma análise de coocorrência que apresenta clusters te máticos de palavras-chave, cuja densidade e centralidade orientam sua classificação e mapeamento e permite mensurar a relevância das temáticas, de acordo com o quadrante em que se encontram (Aria & Cuccurullo, 2022), conhecida como mapa temático. O mapa temático por palavras-chave da amostra, apresentado na Figura 4, foi extraído da ferra menta Biblioshiny. Das 996 palavras-chaves da amostra, foram consideradas as 500 mais citadas e, desse conjunto, 139 palavras-chave tiveram o número mínimo de ocorrências por termo solicitado (2) e foram organizadas pela ferramenta em 18 clusters. As primeiras palavras de cada círculo re presentam a temática central do cluster e o comportamento desse agrupamento demonstra simila 373 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259-381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ ridade com os clusters da análise da rede de coocorrência apresentada anteriormente. A discussão das temáticas do mapa será também acompanhada pelos artigos mais citados dentro da amostra, indicados próximos aos temas discutidos na sequência. ridade com os clusters da análise da rede de coocorrência apresentada anteriormente. A discussão das temáticas do mapa será também acompanhada pelos artigos mais citados dentro da amostra, indicados próximos aos temas discutidos na sequência. ridade com os clusters da análise da rede de coocorrência apresentada anteriormente. RESULTADOS E DISCUSSÕES Por sua vez, gestão da experiência no turismo e hotelaria, marketing, valor percebido (Walls, 2013), cocriação (Chathoth et al., 2016), lealdade e marketing boca-a-boca (Cetin & Dincer, 2014) são significativas e bem desenvolvidas dentro da amostra. As temáticas de ambos os quadrantes são fortemente pre sentes nos artigos da revisão teórica e base para os mais citados. 374 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259-381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359-381 Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Nota-se o despontar de temáticas atreladas à experiência na economia compartilhada, Airbnb e seus paralelos com a hotelaria (Birinci et al., 2018; Mody et al., 2017), assim como estudos voltados para comportamentos e intenções de compra, associados também ao digital (Dedeoglu et al., 2018; Frederick et al., 2014; Walls et al., 2011; Walls, 2013), assinalados na revisão como uma dinâmica que precisa de atenção e gestão dentro do processo de experiência. Sustentabilidade, responsabilidade social corporativa e brand equity (valor agregado da marca) são temáticas nichadas, desenvolvidas, mas que ainda estão buscando maior relevância e consolidação no domínio de conhecimento, já tendo destaque entre os artigos mais citados (Gao & Mattila, 2014; Merli et al., 2019). Avançando para próxima análise, ao se utilizar a visualização de sobreposição na rede da Figura 4, percebeu-se que os resultados de desenvolvimento das temáticas, quanto ao aspecto temporal e detecção de tendências, foram similares ao mapa temático da Figura 5, principalmente no que diz respeito à busca pelo entendimento dos rastros digitais deixados pelos hóspedes no digital e como isso pode ser usado a favor do processo de criação da experiência. Destaca-se, de forma adicional, estudos mais recentes falando com o termo ‘covid-19’ e entende-se que essa será uma temática constante ao longo dos próximos anos nos estudos acadêmicos, pelos impactos que a pandemia causou no setor hoteleiro, de forma geral, e no comportamento do hóspede. Nota-se também um interesse no estudo das experiências por segmentos hoteleiros. Finalizando a análise, faz-se importante destacar dois dos 20 artigos mais citados: as revisões sobre gestão da experiência do cliente de Hwang e Seo (2016) e de Kandampully et al. (2018). Essas revi sões endossam vários pontos comentados e avançam em algumas reflexões. RESULTADOS E DISCUSSÕES O artigo de Hwang e Seo (2016) traz três conceitos não vistos nos demais artigos mais citados e na análise bibliométrica: inteligência experiencial, experiência total do cliente e experiências transformacionais. Sobre o pri meiro, citando o estudo de Bharwani e Jauhari (2013), afirma-se que colaboradores com alta inteli gência experiencial promovem experiências mais positivas. O conceito envolve quatro dimensões: centralidade na experiência, criatividade e inovação, generosidade e comportamento profissional. A essas dimensões, os autores somam a autenticidade emocional da equipe e clamam por mais estudos acerca do tema. Já com relação à experiência total do cliente, é um conceito que considera os múltiplos estágios da experiência na configuração do negócio, entendendo que ela é parte da vida contínua do cliente e não um incidente isolado, visando a alcançar a fidelidade do cliente a longo prazo. O constructo atenta para a questão da cumulação de experiências (tal como o estudo de Dedeoglu et al., 2018) e como ela interfere no comportamento e satisfação. Alinhado ao pensamento acima, o conceito de experiência transformacional vem para abranger aqueles que já estão saturados de experiências memoráveis, momento previsto por Gilmore e Pine (2013): experiências transformacionais levam os clientes a mudarem alguma dimensão de si mes mos e representam um próximo estágio na progressão de valor econômico, como já é visto em al gumas ações, atualmente como Transformational Travel Council, que conecta pessoas que anseiam em transformar vidas – as suas e outras – por meio de viagens cheias de propósito. Mesmo sendo o próximo passo na progressão de valor econômico, essa temática ainda é pouco estudada – e até mesmo aplicada – no setor hoteleiro. 375 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259 381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359 381 Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 A revisão de Kandampully et al. (2018) traz alguns insights norteadores da GEC na hospitalidade: a hospitalidade envolve um extensivo relacionamento hóspede-colaborador; o serviço é parte inte gral da experiência; colaboradores tem um papel-chave na criação de experiências únicas, memorá veis e positivas. RESULTADOS E DISCUSSÕES A autenticidade emocional da equipe, a questão sociocultural na experiência e nas interações entre os hóspedes, o papel das redes sociais no exame do constructo da EC, a tecnologia como suporte à cocriação e interação são outros insights da revisão. Kandampully et al. (2018) afirmam, em especial, que o capital humano é a força motriz da GEC e que afeta vários resultados da performance do hotel. Por isso, gerenciar seu bem-estar é uma orienta ção que reflete não só na cultura organizacional, mas também na GEC, visto que os colaboradores atuam como embaixadores da marca, entregando valor, garantindo a qualidade do serviço e a su peração das expectativas dos clientes. Isso reforça a importância de se atrair o perfil adequado de colaboradores, mantê-los e incentivá-los a engajar na criação de experiências memoráveis. A expe riência do cliente começa pela experiência do colaborador. CONSIDERAÇÕES FINAIS São, também, o resultado de esforços e práticas voltadas a se entender, conectar e relacionar com um perfil de 376 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259 381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359 381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ hóspede que procura, cada vez mais, experiências autênticas, únicas, transformacionais e não se contenta mais com a oferta hoteleira padronizada (Birinci et al., 2018; Hwang & Seo, 2016; Kariru et al., 2017; Pine & Gilmore, 2013). hóspede que procura, cada vez mais, experiências autênticas, únicas, transformacionais e não se contenta mais com a oferta hoteleira padronizada (Birinci et al., 2018; Hwang & Seo, 2016; Kariru et al., 2017; Pine & Gilmore, 2013). Dentre as práticas extraídas da amostra, indo de encontro aos outputs comentados acima, nota-se a importância de, tanto se gerir e cuidar do capital humano dos meios de hospedagem (Hwang & Seo, 2016; Kandampully et al., 2018), quanto das suas estratégias de marketing e elementos operacionais e de estrutura (Hemsley-Brown & Alnawas, 2016). Nesse sentido, práticas associadas à inovação, responsabilidade social, alta tecnologia e alta interação (Chathoth et al., 2016), construção de relacio namentos de longo prazo e design de serviços físicos e online são algumas das assinaladas na discus são e que tem se mostrado o caminho para entregar experiências memoráveis (Yoon & Lee, 2017). Quanto às tendências, experiências inovativas, transformacionais (Gilmore & Pine II, 2002; Pine & Gilmore, 2013), são vistas como componentes críticos para o sucesso de qualquer meio de hospe dagem em suas entregas de valor. Os meios de hospedagem têm buscado entender melhor emo ções, sensações, comportamento e intenções de compra de seu público, ao passo que tem dado maior atenção ao seu capital humano, por compreender a crescente importância de seu bem-estar no ambiente organizacional e como isso reverbera e potencializa a entrega de experiências memo ráveis (Hwang & Seo, 2016; Kandampully et al., 2018). CONSIDERAÇÕES FINAIS Visando a identificar as práticas, benefícios e tendências relacionadas à Gestão da Experiência do Cliente na hotelaria, o objetivo da pesquisa foi alcançado e conclusões significativas foram assina ladas. Entende-se que equilíbrio entre processos, tecnologia e capital humano é um dos cernes na gestão da experiência do hóspede contemporâneo, com gestores adotando uma visão holística da experiência (Walls et al., 2011). Avança-se no sentido de compreender a importância da multidisciplinariedade da experiência e que o seu sucesso depende dos esforços conjuntos da organização. Depende, também, do en tendimento de que os consumidores não são mais passivos nas relações de consumo e de que os acontecimentos globais têm alterado, cada vez mais rápido, a dinâmica dos serviços, da economia e dos padrões de consumo. Atentar-se e agir ativamente sobre esses fatos é imprescindível para se continuar avançando no estudo do tema, enquanto academia, e na gestão da experiência, enquan to implicação prática. Garantir que a experiência seja gerida de forma efetiva em cada ponto de interação do hóspede com o hotel tornou-se um objetivo prioritário para as empresas que querem liderar seus mercados (Kandampully et al., 2018). O estudo da experiência do cliente aplicada ao setor hoteleiro é essencial na fase de retomada em que a hotelaria mundial se encontra hoje, podendo ser uma importante estratégia para a recuperação do setor, além de colaborar na diferenciação e manutenção da em presa hoteleira, sendo esses dois últimos fatores o principal benefício de se adotar a gestão da experiência, apontado pelos estudos da amostra, ancorados nos outputs da experiência apresen tados ao longo dessa revisão (Hwang & Seo, 2016; Kandampully et al., 2018; Rahimian et al., 2021). Lealdade, encantamento, satisfação, intenção de recompra, apego emocional à marca, etc. con tribuem para a construção de uma vantagem competitiva ancorada em uma oferta diferenciada, afastada da armadilha dos commodities (Bilgihan et al., 2014; Gilmore & Pine II, 2002). CONSIDERAÇÕES FINAIS Em um cenário em que satisfazer o consumidor já não tem mais 377 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur Visão e Ação v25 n2 p259 381 Mai /Ago 2023 \| http://dx doi org/10 14210/rtva v25n2 p359 381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 tanto efeito quanto antigamente, a gestão da experiência do cliente tem despontado como uma abordagem promissora e promover experiências atrativas, autênticas, positivas e superiores é a uma importante chave para garantir a vantagem competitiva (Kandampully et al., 2018). Seja para acompanhar as necessidades dos hóspedes, seja para obter vantagem competitiva e sustentabili dade para o negócio, trabalhar a experiência do cliente dentro da hotelaria comprovadamente traz benefícios ao negócio – e ao hóspede também. tanto efeito quanto antigamente, a gestão da experiência do cliente tem despontado como uma abordagem promissora e promover experiências atrativas, autênticas, positivas e superiores é a uma importante chave para garantir a vantagem competitiva (Kandampully et al., 2018). Seja para acompanhar as necessidades dos hóspedes, seja para obter vantagem competitiva e sustentabili dade para o negócio, trabalhar a experiência do cliente dentro da hotelaria comprovadamente traz benefícios ao negócio – e ao hóspede também. Este artigo encontra como limitação o uso de uma única base de dados para a coleta dos artigos e, por isso, recomenda-se a ampliação da busca para resultados mais abrangentes e culturalmente diversificados sobre o tema. CONSIDERAÇÕES FINAIS Enxergar a experiência como um conjunto de vivências e a extensão do planejamento e relacionamento para longo prazo, visando à fidelidade e benefícios mútuos aos envolvidos no processo também são ações que despontam nos estudos e operação (Dedeoglu et al., 2018; Hwang & Seo, 2016). O impacto da tecnologia e a inserção de robôs nas operações hoteleiras também é movimento que merece maior atenção, principalmente quanto aos seus reflexos na cultura organização e aceitação cultural (Lu et al., 2019; Tung & Au, 2018). O uso da tecnologia com o suporte ao processo de gestão e criação da experiência, incentivando a cocriação e engajamento de todos envolvidos, também é uma ação que merece ser destacada (Calheiros et al., 2017; Pencarelli, 2020; Xiang et al., 2015). Como fruto das análises realizadas e refletindo um pouco as tendências, deixam-se as seguintes temáticas como uma agenda futura de pesquisa: como a cocriação pode ser incentivada à nível organizacional; como a cultura organizacional interfere na experiência do hóspede; como os gesto res podem gerenciar a relação hóspede-hóspede para incrementar experiências; qual o papel dos colaboradores dentro da experiência do cliente, em um estudo por nicho hoteleiro, entendendo a diferença de percepções; como/ se a evolução da economia da experiência para a economia trans formacional afetará a hotelaria; estudar experiência em meios de hospedagem onde necessaria mente os hóspedes não tenham que pagar muito a mais para ter acesso a elas (como hostels, por exemplo); estudar o comportamento do consumidor, quanto às percepções, emoções e anseios na busca de experiências autênticas; uso do marketing sensorial na criação de experiências; como manter a autenticidade e o fator novidade das experiências em caso de recompra; estudar a criação de escalas ou ferramentas de mensuração a partir da diferença entre encantamento e satisfação; pesquisar novas formas e ferramentas para o uso de big data no entendimento da experiência do cliente, principalmente por meio dos comentários; o impacto dos robôs de serviço na experiência do cliente, com viés cultural e demográfico. Fato é que a experiência do cliente será essencial para o sucesso dos hotéis do futuro (Mody et al., 2017; Oskam & Boswijk, 2016). REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Gao, Y. (Lisa), & Mattila, A. S. (2014). Improving consumer satisfaction in green hotels: The roles of perceived war mth, perceived competence, and CSR motive. International Journal of Hospitality Management, 42, 20-31. DOI: https://doi.org/10.1016/j.ijhm.2014.06.003 Gilmore, J. H., & Pine II, B. J. (2002). Differentiating Hospitality Operations via Experience: why selling services is not enough. Cornell Hotel and Restaurant Administration Quarterly, 43(3), 87-96. Hemmington, N. (2007). From service to experience: Understanding and defining the hospitality business. Servi ce Industries Journal, 27(6), 747-755. DOI: https://doi.org/10.1080/02642060701453221 Hemsley-Brown, J., & Alnawas, I. (2016). Service quality and brand loyalty: The mediation effect of brand passion, brand affection and self-brand connection. International Journal of Contemporary Hospitality Management, 28(12), 2771-2794. DOI: https://doi.org/10.1108/IJCHM-09-2015-0466 Hwang, J., & Seo, S. (2016). A critical review of research on customer experience management: Theoretical, me thodological and cultural perspectives. International Journal of Contemporary Hospitality Management, 28(10), 2218-2246. DOI: https://doi.org/10.1108/IJCHM-04-2015-0192 Kandampully, J., Zhang, T., & Jaakkola, E. (2018). Customer experience management in hospitality: A literature synthesis, new understanding and research agenda. International Journal of Contemporary Hospitality Ma nagement, 30(1), 21-56. DOI: https://doi.org/10.1108/IJCHM-10-2015-0549 Kao, C. Y., Tsaur, S. H., & Wu, T. C. (2016). Organizational culture on customer delight in the hospitality industry. In ternational Journal of Hospitality Management, 56, 98-108. DOI: https://doi.org/10.1016/j.ijhm.2016.05.001 Kariru, A. N., Kambona, O. O., & Odhuno, E. (2017). Enhancing Competitiveness Through Guests’ Experiences: A Typology of Customer Experiences in Upscale Hotels. International Journal of Hospitality and Tourism Admi nistration, 18(4), 361-392. DOI: https://doi.org/10.1080/15256480.2017.1289137 Lemke, F., Clark, M., & Wilson, H. (2011). Customer experience quality: An exploration in business and consumer contexts using repertory grid technique. Journal of the Academy of Marketing Science, 39(6), 846-869. DOI: https://doi.org/10.1007/s11747-010-0219-0 Lu, L., Cai, R., & Gursoy, D. (2019). Developing and validating a service robot integration willingness scale. Interna tional Journal of Hospitality Management, 80(January), 36-51. DOI: https://doi.org/10.1016/j.ijhm.2019.01.005 Martinelli, I. F. M. (2013). A ECONOMIA DA EXPERIÊNCIA NO CONTEXTO DOS FESTIVAIS DE MÚSICA [Universidade Federal do Rio de Janeiro]. https://pantheon.ufrj.br/bitstream/11422/645/3/IMartinelli.pdf Merli, R., Preziosi, M., Acampora, A., & Ali, F. (2019). Why should hotels go green? Insights from guests expe rience in green hotels. International Journal of Hospitality Management, 81, 169-179. DOI: https://doi.or g/10.1016/j.ijhm.2019.04.022 Mody, M. A., Suess, C., & Lehto, X. (2017). The accommodation experiencescape: a comparative assessment of hotels and Airbnb. International Journal of Contemporary Hospitality Management, 29(9), 2377-2404. REFERENCES DOI: https://doi.org/10.1509/jmr.12.0061 378 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO REFERENCES Aria, M., & Cuccurullo, C. (2017). bibliometrix: An R-tool for comprehensive science mapping analysis. Journal of Informetrics, 11(4), 959-975. DOI: https://doi.org/10.1016/j.joi.2017.08.007 Aria, M., & Cuccurullo, C. (2022). Science Mapping Analysis with bibliometrix R-package: an example. https://bi bliometrix.org/documents/bibliometrix_Report.html#section-4-the-conceptual-structure---co-word-analy sis Bharwani, S., & Mathews, D. (2016). Customer service innovations in the Indian hospitality industry. Worldwide Hospitality and Tourism Themes, 8(4), 416-431. DOI: https://doi.org/10.1108/WHATT-04-2016-0020 Bilgihan, A., Nusair, K., Okumus, F., & Cobanoglu, C. (2015). Applying flow theory to booking experiences: An integrated model in an online service context. Information and Management, 52(6), 668-678. DOI: https:// doi.org/10.1016/j.im.2015.05.005 Bilgihan, A., Okumus, F., Nusair, K., & Bujisic, M. (2014). Online experiences: Flow theory, measuring online custo mer experience in e-commerce and managerial implications for the lodging industry. Information Techno logy and Tourism, 14(1), 49-71. DOI: https://doi.org/10.1007/s40558-013-0003-3 Birinci, H., Berezina, K., & Cobanoglu, C. (2018). Comparing customer perceptions of hotel and peer-to-peer accommodation advantages and disadvantages. International Journal of Contemporary Hospitality Mana gement, 30(2), 1190-1210. DOI: https://doi.org/10.1108/IJCHM-09-2016-0506 Calheiros, A. C., Moro, S., & Rita, P. (2017). Sentiment Classification of Consumer-Generated Online Reviews Using Topic Modeling. Journal of Hospitality Marketing and Management, 26(7), 675-693. DOI: https://doi.or g/10.1080/19368623.2017.1310075 Cetin, G., & Dincer, F. I. (2014). Influence of customer experience on loyalty and word-of-mouth in hospitality operations. Anatolia, 25(2), 181-194. DOI: https://doi.org/10.1080/13032917.2013.841094 Cetin, G., & Walls, A. (2016). Understanding the Customer Experiences from the Perspective of Guests and Hotel Managers: Empirical Findings from Luxury Hotels in Istanbul, Turkey. Journal of Hospitality Marketing and Management, 25(4), 395-424. DOI: https://doi.org/10.1080/19368623.2015.1034395 Chathoth, P. K., Ungson, G. R., Harrington, R. J., & Chan, E. S. W. (2016). Co-creation and higher order customer engagement in hospitality and tourism services: A critical review. International Journal of Contemporary Hospitality Management, 28(2), 222-245. DOI: https://doi.org/10.1108/IJCHM-10-2014-0526 Cobo, M. J., López-Herrera, A. G., Herrera-Viedma, E., & Herrera, F. (2011). An approach for detecting, quantifying, and visualizing the evolution of a research field: A practical application to the Fuzzy Sets Theory field. Journal of Informetrics, 5(1), 146-166. DOI: https://doi.org/10.1016/j.joi.2010.10.002 Dedeoglu, B. B., Bilgihan, A., Ye, B. H., Buonincontri, P., & Okumus, F. (2018). The impact of servicescape on hedonic value and behavioral intentions: The importance of previous experience. International Journal of Hospitality Management, 72, 10-20. DOI: https://doi.org/10.1016/j.ijhm.2017.12.007 Frederick, S., Lee, L., & Baskin, E. (2014). The limits of attraction. Journal of Marketing Research, 51(4), 487-507. REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 Pati, D., & Lorusso, L. N. (2018). How to Write a Systematic Review of the Literature. Health Environments Rese arch and Design Journal, 11(1), 15-30. DOI: https://doi.org/10.1177/1937586717747384 Pencarelli, T. (2020). The digital revolution in the travel and tourism industry. Information Technology and Tou rism, 22(3), 455-476. DOI: https://doi.org/10.1007/s40558-019-00160-3 Pine, B. J., & Gilmore, J. H. (1998). Welcome to the experience economy. Harvard Business Review, 76(4), 97-105 Pine, B. J., & Gilmore, J. H. (1999). The experience economy. Harvard Business School. Pine, B. Joseph, & Gilmore, J. H. (2013). The experience economy: Past, present and future. Handbook on the Experience Economy, January 2013, 21-44. DOI: https://doi.org/10.4337/9781781004227.00007 Pine II, B. J., & Gilmore, J. H. (2011). The experience economy. Harvard Business Press. Pizam, A. (2010). Creating memorable experiences. International Journal of Hospitality Management, 29(3), 343 DOI: https://doi.org/10.1016/j.ijhm.2010.04.003 Prahalad, C. K., & Ramaswamy, V. (2004). Co-creating unique value with customers. Strategy & Leadership, 32(3) 4-9. DOI: https://doi.org/10.1108/10878570410699249 Radojevic, T., Stanisic, N., & Stanic, N. (2015). Ensuring positive feedback: Factors that influence customer sa tisfaction in the contemporary hospitality industry. Tourism Management, 51, 13-21. DOI: https://doi.or g/10.1016/j.tourman.2015.04.002 Rahimian, S., ShamiZanjani, M., Manian, A., & Esfidani, M. R. (2021). A framework of customer experience mana gement for hotel industry. International Journal of Contemporary Hospitality Management. DOI: https://doi. org/10.1108/IJCHM-06-2020-0522 Ren, L., Qiu, H., Wang, P., & Lin, P. M. C. (2016). Exploring customer experience with budget hotels: Dimensionality and satisfaction. International Journal of Hospitality Management, 52, 13-23. DOI: https://doi.org/10.1016/j. ijhm.2015.09.009 Roy, S. K., Balaji, M. S., Soutar, G., & Jiang, Y. (2020). The Antecedents and Consequences of Value Co-Creation Behaviors in a Hotel Setting: A Two-Country Study. Cornell Hospitality Quarterly, 61(3), 353-368. DOI: https:// doi.org/10.1177/1938965519890572 Scopus. (2016). Five steps to creating a citation overview in Scopus. https://blog.scopus.com/posts/five-steps-to creating-a-citation-overview-in-scopus Silva, F. F., & Barreto, L. M. T. da S. (2019). A Experiência do consumidor como fator influenciador da formação do encantamento do cliente no turismo e na hospitalidade. Revista Iberoamericana de Turismo, 9(1), 78-95. DOI: https://doi.org/10.2436/20.8070.01.130 Tung, V. W. S., & Au, N. (2018). Exploring customer experiences with robotics in hospitality. International Journal of Contemporary Hospitality Management, 30(7), 2680-2697. DOI: https://doi.org/10.1108/IJCHM-06-2017-0322 Walls, A., Okumus, F., Wang, Y., & Kwun, D. J.-W. (2011). Understanding the consumer experience: An exploratory study of luxury hotels. Journal of Hospitality Marketing and Management, 20(2), 166-197. REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO DOI: https://doi.org/10.1108/IJCHM-09-2016-0501 Mody, M., Suess, C., & Lehto, X. (2019). Going back to its roots: Can hospitableness provide hotels competitive advantage over the sharing economy? International Journal of Hospitality Management, 76, 286-298. DOI: https://doi.org/10.1016/j.ijhm.2018.05.017 Mossberg, L. (2007). A Marketing Approach to the Tourist Experience. Scandinavian Journal of Hospitality and Tourism, 7(1), 59-74. DOI: https://doi.org/10.1080/15022250701231915 Oliveira, T. B. V. L., Sohn, A. P. L., & Linderberg, P. F. (2022). Estratégias de Customer-Centricity na Hotelaria. Ateliê Do Turismo, 6(2), 1-18. Oskam, J., & Boswijk, A. (2016). Airbnb: the future of networked hospitality businesses. Journal of Tourism Futu res, 2(1), 22-42. DOI: https://doi.org/10.1108/JTF-11-2015-0048 Page, M. J., Moher, D., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., Mcdonald, S., … Mckenzie, J. E. (2021). PRISMA 2020 explanation and elaboration: Updated guidance and exemplars for reporting systematic reviews. The BMJ, 372. DOI: https://doi.org/10.1136/bmj. n160 379 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO Tur., Visão e Ação, v25, n2, p259-381 Mai./Ago. 2023 \| http://dx.doi.org/10.14210/rtva.v25n2.p359-381 REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO REVISTA CIENTÍFICA DO PROGRAMA DE PÓS-GRADUAÇÃO EM TURISMO E HOTELARIA DA UNIVERSIDADE DO VALE DO ITAJAÍ TURISMO VISÃO & AÇÃO DOI: https://doi.or g/10.1080/19368623.2011.536074 Walls, A.R. (2013). A cross-sectional examination of hotel consumer experience and relative effects on consu mer values. International Journal of Hospitality Management, 32(1), 179-192. DOI: https://doi.org/10.1016/j. ijhm.2012.04.009 Xiang, Z., Schwartz, Z., Gerdes, J. H., & Uysal, M. (2015). What can big data and text analytics tell us about hotel guest experience and satisfaction? International Journal of Hospitality Management, 44, 120-130. DOI: ht tps://doi.org/10.1016/j.ijhm.2014.10.013 Yoon, S.-J., & Lee, H.-J. (2017). Does Customer Experience Management Pay Off? Evidence from Local versus Global Hotel Brands in South Korea. Journal of Hospitality Marketing and Management, 26(6), 585-605. DOI: https://doi.org/10.1080/19368623.2017.1281192 Zupic, I., & Čater, T. (2015). Bibliometric Methods in Management and Organization. Organizational Research Methods, 18(3), 429-472. DOI: https://doi.org/10.1177/1094428114562629 380 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL CONTRIBUIÇÃO DOS AUTORES Thais Bandinelli Vargas Lopes de Oliveira: Concepção da pesquisa, revisão de literatura, análise dos dados, discussão dos resultados. Ana Paula Lisboa Sohn: Orientação da pesquisa, revisão, apoio na discussão dos resultados. 381 ISSN: 1983-7151 \| BALNEÁRIO CAMBORIÚ, SANTA CATARINA, BRASIL
https://openalex.org/W2035383744	https://systematicreviewsjournal.biomedcentral.com/track/pdf/10.1186/2046-4053-2-43	English	null	Health equity: evidence synthesis and knowledge translation methods	Systematic reviews	2,013	cc-by	9,225	Open Access © 2013 Welch et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: At the Rio Summit in 2011 on Social Determinants of Health, the global community recognized a pressing need to take action on reducing health inequities. This requires an improved evidence base on the effects of national and international policies on health inequities. Although systematic reviews are recognized as an important source for evidence-informed policy, they have been criticized for failing to assess effects on health equity. Methods: This article summarizes guidance on both conducting systematic reviews with a focus on health equity and on methods to translate their findings to different audiences. This guidance was developed based on a series of methodology meetings, previous guidance, a recently developed reporting guideline for equity-focused systematic reviews (PRISMA-Equity 2012) and a systematic review of methods to assess health equity in systematic reviews. Results: We make ten recommendations for conducting equity-focused systematic reviews; and five considerations for knowledge translation. Illustrative examples of equity-focused reviews are provided where these methods have been used. Conclusions: Implementation of the recommendations in this article is one step toward monitoring the impact of national and international policies and programs on health equity, as recommended by the 2011 World Conference on Social Determinants of Health. Keywords: Health Equity, Evidence Synthesis, Knowledge Translation, Systematic Reviews Keywords: Health Equity, Evidence Synthesis, Knowledge Translation, Systematic Reviews Health equity: evidence synthesis and knowledge translation methods Vivian A Welch1, Mark Petticrew2, Jennifer O’Neill1, Elizabeth Waters8, Rebecca Armstrong3, Zulfiqar A Bhutta4, Damian Francis5, Tracey Perez Koehlmoos6, Elizabeth Kristjansson1,12, Tomas Pantoja7 and Peter Tugwell9,10,11 Correspondence: Vivian.welch@uottawa.ca 1Institute of Population Health, University of Ottawa, 1 Stewart Street, Ottawa, ON K1N6N5, Canada Full list of author information is available at the end of the article Welch et al. Systematic Reviews 2013, 2:43 http://www systematicreviewsjournal com/con Welch et al. Systematic Reviews 2013, 2:43 http://www systematicreviewsjournal com/con Welch et al. Systematic Reviews 2013, 2:43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Page 2 of 10 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 broader range of electronic and gray literature sources [8]. Methods to assess the influence of context and its relevance for discussion of applicability are needed. characteristics that may indicate a disadvantage, such as age and disability [17,18]. The use of an acronym such as PROGRESS-Plus helps explicitly and systematically con- sider health equity in the design of both primary studies and systematic reviews. Knowledge translation (KT) of the results on system- atic reviews on equity is essential to ensure the results are utilized. KT is defined by the Canadian Institutes of Health Research as a ‘dynamic and iterative process that includes synthesis, dissemination, exchange and ethically- sound application of knowledge to improve health, pro- vide more effective health services and products and strengthen the health care system’ [11]. Comprehensive KT is important to maximize the benefit from funding and conducting knowledge syntheses, both in terms of improving health of disadvantaged populations and for informing research priority setting exercises that consider prioritized effects on health equity. By clearly identifying disadvantaged populations, implementation of new policies can be targeted to those who most need them. Systematic reviews with a major focus on health equity are those designed to: (1)Assess effects of interventions in disadvantaged population(s) (such as school feeding for disadvantaged children) [19]; (2)Assess effects of interventions aimed at reducing social gradients across populations (e.g., interventions to reduce the social gradient in smoking) [20]; and/or smoking) [20]; and/or (3)Assess effects of interventions not aimed at reducing inequity but where it is important to understand the effects of the intervention on equity, either positively or negatively (e.g., an intervention targeted at the whole population but that may have effects on equity, such as the review on obesity prevention in children, which examined the effects of interventions across relevant PROGRESS-Plus factors) [21]. (3)Assess effects of interventions not aimed at reducing inequity but where it is important to understand the effects of the intervention on equity, either positively or negatively (e.g., an intervention targeted at the whole population but that may have effects on equity, such as the review on obesity prevention in children, which examined the effects of interventions across relevant PROGRESS-Plus factors) [21]. Methods d l We have estimated that at least 20% of systematic reviews indexed in MEDLINE meet one or more of the above criteria [15]. We have assembled a selection of exemplar reviews that highlight one or more of the methodological challenges discussed in this article (Table 1). These reviews were identified by participants in the above meetings and by searching for systematic reviews in PubMed, the Cochrane Database of System- atic Reviews and the Campbell Library using the text words “equity” or “inequity”. We developed these recommendations based on meth- odology meetings held between 2005 and 2012 by the Campbell and Cochrane Equity Methods Group, meth- odological recommendations from the Cochrane Public Health Review Group [12], a Cochrane systematic review [13], methods study [14], the WHO Task Force on evidence-informed policies about health systems [2] and a consensus meeting held in Bellagio, Italy, in February 2012 with methodologists, funders, journal editors, clinicians and public health practitioners as part of the development of reporting guidelines for systematic reviews with a focus on health equity to extend the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta- Analyses) statement (PRISMA-E 2012) [15]. Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 The purpose of this article is to provide guidance on how to conduct equity-focused systematic reviews consis- tent with the recommendations of PRISMA-E 2012 to fa- cilitate the use of both guidance documents. This article also discusses challenges related to knowledge translation for equity-focused systematic reviews. Background promote and maintain health [4-6]. In this era of fiscal restraint, there is a critical need for evidence about how to improve health equity in the most efficient way [7]. The recommendations of the World Conference on Social Determinants of Health (Rio de Janeiro, 19–21 October 2011) recognized the pressing need to take action on re- ducing health inequities; one of its key recommendations was to assess the effects of national and international pol- icies on health inequities [1]. Effects of interventions on health equity are also of paramount importance for health systems research and decision-makers [2,3]. The need for considering health equity is recognized for clinical health care and preventive interventions as well as place-based programs in disadvantaged areas or communities, and the social gradient in effects of population-based strategies to Systematic reviews are widely recognized as an effi- cient, reliable and comprehensive source of evidence for decision-making. Few systematic reviews have considered effects on health equity, even though research methods to assess effects on health equity in systematic reviews have been available and recently have been strengthened for use within natural policy experiments and systems approaches [3,8,9]. Several groups have documented methodological chal- lenges when considering effects on equity in systematic reviews. For example, methods are needed to define the underlying theory and the mechanisms by which the intervention is expected to affect health equity [10]. Also, the search strategy may need to encompass a * Correspondence: Vivian.welch@uottawa.ca 1Institute of Population Health, University of Ottawa, 1 Stewart Street, Ottawa, ON K1N6N5, Canada Full list of author information is available at the end of the article How equity was considered Systematic reviews with a major focus on health equity Example exemplar review Systematic reviews with a major focus on health equity Example exemplar review This review included only studies in which the intervention was target at ’predominantly disadvantaged’ children (e.g., living in a rural area or village, or an urban area and described as socioeconomically disadvantaged, from poor areas, if 30% of more of the children in the sample were underweight, or stunted or the average weight, height and body mass index (BMI) were low, or if the studies were implicitly or explicitly aimed at disadvantaged children (and indicators of disadvantage were provided) [19] School feeding for improving the physical and psychosocial health of disadvantaged students [19] (1) Assess effects of interventions in disadvantaged population(s) Female genital mutilation/cutting (FGM/C) is practiced mainly on young girls and has many potential adverse effects. In addition to ethnic, cultural and religious beliefs, there are socioeconomic beliefs that FGM/C is required for marriage or an economic necessity when women are dependent on men. This review examined intervention features and contextual factors that reduce the prevalence of FGM/C [22] This review aimed to determine the impact of microfinance interventions on incomes of the poor, on wider poverty/wealth of the poor and on non-financial outcomes, such as health and food security. The authors found that microfinance had inconclusive effects on savings and income but positive effects on health outcomes [23] (2) Assess effects of interventions aimed at reducing social gradients across populations Population tobacco control interventions and their effects on social inequalities in smoking: systematic review [20] This systematic review applied an “equity lens” to population level interventions to reduce inequalities in smoking rates and extracted outcome, process and implementation data stratified by PROGRESS-Plus. Certain interventions, such as smoking restrictions in schools, restricting sales to minors and increasing the price of tobacco, are more effective in reducing smoking among lower- income adults and those with manual occupations. Other interventions had no effect on reducing social inequalities in smoking [20] The psychosocial work environment has a strong gradient that influences inequalities in health. What is the impact of microfinance on poor people? A systematic review of evidence from Sub-Saharan Africa [23] Working for health? Evidence from systematic reviews on the effects of health and health inequalities of organizational changes on the psychosocial work environment [24] Interventions to reduce the prevalence of female genital mutilation/cutting in African countries [22] Socioeconomic differences in lung cancer incidence: a systematic review and meta- analysis [25] Population tobacco control interventions and their effects on social inequalities in smoking: systematic review [20] Results Recommendations for assessing health equity Health equity can be considered at the following ten steps in the systematic review process. 1. Define conceptual approach to health equity Systematic review authors should consider the relevance of health equity questions at the protocol stage by considering whether social gradients exist in the burden of the disease and whether relative or absolute effects of interventions are likely to differ for disadvantaged populations. When developing the protocol for a systematic review, it is important to define why there is a need to focus on health equity and the method of assessing disadvantage, including whether proxies will be accepted and, if so, which ones are most appropriate. For example, living in a rural village in a low- or middle-income 1. Define conceptual approach to health equity Systematic review authors should consider the relevance of health equity questions at the protocol stage by considering whether social gradients exist in the burden of the disease and whether relative or absolute effects of interventions are likely to differ for disadvantaged populations. When developing the protocol for a systematic review, it is important to define why there is a need to focus on health equity and the method of assessing disadvantage, including whether proxies will be accepted and, if so, which ones are most appropriate. For example, living in a rural village in a low- or middle-income What is an equity-focused systematic review? Health inequities are defined as differences in health outcomes that are avoidable, unfair and unjust [16]. Health inequities persist and are worsening for some conditions across population and individual characteris- tics both within and across countries. The Campbell and Cochrane Equity Methods Group and the Cochrane Public Health Group recommend the PROGRESS-Plus acronym to identify population and individual characteris- tics across which health inequities may exist. PROGRESS- Plus stands for place of residence, race/ethnicity/culture/ language, occupation, gender/sex, religion, socioeconomic status and social capital, and “plus” captures other Page 3 of 10 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Table 1 Exemplar systematic reviews Table 1 Exemplar systematic reviews (Continued) This review extracted data on intervention implementation, cost, equity and outcomes. The authors used PROGRESS-Plus to extract equity- relevant data from the studies and examined equity effects for each age group [21] (3) Assess effects of interventions not aimed at reducing inequity but where it is important to understand the effects of the intervention on equity, either positively or negatively Interventions for preventing obesity in children [21] Interventions for preventing obesity in children [21] Interventions for preventing obesity in children [21] Interventions for preventing obesity in children [21] This review included studies conducted in any country with any population as long as the intervention was delivered by lay health workers and intended to improve maternal and child health. Many of the included studies focused on low income populations and found that lay health workers can improve access to health care for low income groups and, if extrapolated to other settings, may contribute to reducing inequities [26] This review aims to examine the effectiveness of all built environment interventions to increase physical activity. If sufficient data are available, the authors plan to conduct subgroup analyses to explore whether there is likely to be a relationship of effect to disadvantage and whether an equity gradient is present by assessing studies that have included subgroup analyses by ethnicity, occupation, gender, education, socioeconomic status and disability (including individuals with specific morbidities) [27] country was accepted as a proxy of poverty and socioeconomic disadvantage for a review of school feeding [19]. country was accepted as a proxy of poverty and socioeconomic disadvantage for a review of school feeding [19]. country was accepted as a proxy of poverty and socioeconomic disadvantage for a review of school feeding [19]. Outcome and Context or setting (PICO-C) if the review topic focuses on intervention effectiveness- related questions. Conceptualizing the review questions related to health equity requires a description of how the intervention is expected to work and why it may work differently depending on the context for disadvantaged populations or across gradients in socioeconomic status. This requires a consideration of both relative and absolute effects, as well as baseline risk of the health outcome of interest across social gradients. The absolute effect provides the difference in effectiveness between the most and least disadvantaged while the relative effect describes the difference in effectiveness relative to a reference group, such as the whole population [31]. Built environment interventions for increasing physical activity in adults and children (Protocol) [27] Lay health workers in primary and community health care for maternal and child health and the management of infectious diseases [26] How equity was considered This umbrella review examined the impacts of interventions on inequalities in health by socioeconomic status, age, gender and ethnicity and found that some organizational workplace interventions can reduce health inequalities in those who are employed, especially between men and women, and socioeconomic groups [24] The socioeconomic gradient in lung cancer results from differences in exposures and risk factors, such as smoking, occupational/ environmental exposure to inhaled carcinogens and air pollution. This meta-analysis found that lung cancer risk was highest among those in the lowest socioeconomic categories for education, occupation and income [25] Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Page 4 of 10 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Table 1 Exemplar systematic reviews (Continued) Since disadvantaged populations may have worse health status and higher risk of adverse outcomes, interventions may have a greater absolute effect in disadvantaged populations, even if the relative effect is the same. For example, foreign-born Canadians have an incidence rate of tuberculosis that is 20 times higher than for non- Aboriginal Canadians (16 cases per 100,000 versus 0.8 cases per 100,000). Thus, in a Canadian guideline on tuberculosis for immigrants, although the relative effect of isoniazid preventive treatment was assumed to be 0.40 for both immigrants and Canadian-born, the difference in absolute risk means that the expected absolute benefit was 32 2. Develop a theory-based approach, which may include an analytic framework that identifies health equity as an outcome In equity-focused reviews, it is important to define the assumptions and presumed causal pathways that will be assessed by the systematic review and how these are expected to affect health equity. Causal pathway analysis involves an assessment of contextual factors and processes that influence the effect of an intervention on health outcomes. For example, a systematic review of water and sanitation interventions assessed whether the hardware functioned properly to clean or filter the water, whether people used the hardware and finally the effects of diarrhea on health outcomes [28]. A visual representation (analytic framework) of the assumptions, causal pathways and likely effects on health equity may be useful to justify the equity questions, as well as identify important effect modifiers, confounding factors and important contextual factors [29]. An example analytic framework is provided in Figure 1, which shows how deworming of children is expected to improve health equity [30]. 2. Develop a theory-based approach, which may include an analytic framework that identifies health equity as an outcome In equity-focused reviews, it is important to define the assumptions and presumed causal pathways that will be assessed by the systematic review and how these are expected to affect health equity. Causal pathway analysis involves an assessment of contextual factors and processes that influence the effect of an intervention on health outcomes. For example, a systematic review of water and sanitation interventions assessed whether the hardware functioned properly to clean or filter the water, whether people used the hardware and finally the effects of diarrhea on health outcomes [28]. Table 1 Exemplar systematic reviews (Continued) A visual representation (analytic framework) of the assumptions, causal pathways and likely effects on health equity may be useful to justify the equity questions, as well as identify important effect modifiers, confounding factors and important contextual factors [29]. An example analytic framework is provided in Figure 1, which shows how deworming of children is expected to improve health equity [30]. 2. Develop a theory-based approach, which may include an analytic framework that identifies health equity as an outcome 3. Frame the health equity questions Health equity questions must be defined across the Population, Intervention, Comparison, Welch et al. Systematic Reviews 2013, 2:43 Page 5 of 10 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 y , http://www.systematicreviewsjournal.com/content/2/1/43 Figure 1 Example analytic framework. Source: Welch, 2013 [30]. Figure 1 Example analytic framework. Source: Welch, 2013 [30]. Figure 1 Example analytic framework. Source: Welch, 2013 [30]. fewer cases of active tuberculosis per 1,000 people for immigrants compared to only 6 fewer per 1,000 for Canadian-born subjects [32]. assess the health effects of complex housing improvement interventions included non- randomized study designs [35]. fewer cases of active tuberculosis per 1,000 people for immigrants compared to only 6 fewer per 1,000 for Canadian-born subjects [32]. When equity is a main focus, the authors should consider additional study designs. A review of interventions to upgrade slums included both controlled before and after studies and interrupted time series as well as ‘supporting studies’ such as uncontrolled before and after studies and non- randomized, controlled studies with post- intervention outcome data [34]. Nonrandomized study designs provide considerations of the effects of context, setting and underlying mechanisms of action, which are important when evaluating a complex intervention, even if equity is not the main focus of the review. If the authors do not consider nonrandomized study designs, then failure to find assessment of effects on health equity may be due to “the inverse evidence law”, i.e., that there is less evidence available on the interventions that are most likely to influence policy and population health [36]. 4. Include relevant study designs to assess health equity questions Eligible study designs should be included and described according to their “fitness for purpose”, and the rationale should be clearly stated and explained [33]. 5. Identify information sources for health equity questions the intervention. This is most often done using standard systematic review methods, and the role of context can be explored using meta-analysis. Other review methods are also being increasingly applied to the exploration of context and process, such as realist evaluation [41], meta- ethnography [42] and thematic synthesis [43]. Guidelines for how to use these methods as part of a systematic review have been proposed by the Cochrane Qualitative and Implementation Methods Group (http://cqrmg.cochrane.org), the Cochrane Public Health Review Group [12] and others working in the area of theory-based systematic reviews [42]. The choice of method depends on a number of factors such as the types of questions posed, the types of data sources, and the outcomes and processes of interest. However, there is to date no comprehensive comparison of each of these methods, their advantages and disadvantages and how to choose one method over another, though work is underway to compare them [44]. The methods can also be used in tandem. For example, a review of school feeding used two methods to assess the role of process and context. A process evaluation tabulated effect sizes across implementation factors hypothesized to be important such as supervision and caloric content [19]. A realist evaluation was then used to propose policy recommendations about designing successful school feeding programs [45]. the intervention. This is most often done using standard systematic review methods, and the role of context can be explored using meta-analysis. Other review methods are also being increasingly applied to the exploration of context and process, such as realist evaluation [41], meta- ethnography [42] and thematic synthesis [43]. Guidelines for how to use these methods as part of a systematic review have been proposed by the Cochrane Qualitative and Implementation Methods Group (http://cqrmg.cochrane.org), the Cochrane Public Health Review Group [12] and others working in the area of theory-based systematic reviews [42]. The choice of method depends on a number of factors such as the types of questions posed, the types of data sources, and the outcomes and processes of interest. However, there is to date no comprehensive comparison of each of these methods, their advantages and disadvantages and how to choose one method over another, though work is underway to compare them [44]. The methods can also be used in tandem. For example, a review of school feeding used two methods to assess the role of process and context. 5. Identify information sources for health equity questions A process evaluation tabulated effect sizes across implementation factors hypothesized to be important such as supervision and caloric content [19]. A realist evaluation was then used to propose policy recommendations about designing successful school feeding programs [45]. 6. Define search terms for health equity questions It is especially important to consider the risks of missing relevant literature when using filters for any concept, including disadvantage or health equity, as many of the words describing disadvantaged populations or settings are not indexed in the major databases. The use of text words to limit the search to concepts of health equity or disadvantage risks missing relevant studies that have been described using different terms (for example, disparities vs. inequities) [38]. Furthermore, some community-based interventions in low and middle income countries are entirely equity-focused by focusing on improving health outcomes for under-served populations, but are not indexed with any terms to describe health equity, disparities or inequalities. There are no validated health equity search filters, and equity terms are not indexed consistently [39]. Systematic reviewers need to plan for extra time to screen potentially relevant studies for health equity and should avoid using textword limits unless they have been validated, for example, in the Child Health filter [40]. 9. Use synthesis approaches to assess effects on health equity Table 1 Exemplar systematic reviews (Continued) For systematic reviews with a focus on health equity, the type of intervention (e.g., legislation applied at the population level) and the time frame of interest (e.g., long-term outcomes of interest not likely to be assessed in a short-term RCT) may require the inclusion of nonrandomized studies to inform the review. Upstream, policy- level interventions may have been evaluated in nonrandomized evaluations such as natural policy experiments (e.g., effects of privatization of public utilities, interventions to promote cycling and slum upgrading strategies) and thus necessitate the inclusion of a wider array of evidence [9,28,34]. For example, a systematic review that aimed to Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Page 6 of 10 5. Identify information sources for health equity questions Searches related to disadvantage need to draw on social, political, cultural and ethical perspectives. Thus, potentially relevant studies may be found in a wide range of literature sources (such as books, government publications, policy documents and other gray literature), which are difficult to scope in terms of total volume, location and categorization. It is inappropriate and inaccurate to rely solely on conventional databases such as MedLine; topic-specific databases addressing the research topic such as transportation databases for questions about traffic calming could be more relevant than general databases [8,37]. 5. Identify information sources for health equity questions Searches related to disadvantage need to draw on social, political, cultural and ethical perspectives. Thus, potentially relevant studies may be found in a wide range of literature sources (such as books, government publications, policy documents and other gray literature), which are difficult to scope in terms of total volume, location and categorization. It is inappropriate and inaccurate to rely solely on conventional databases such as MedLine; topic-specific databases addressing the research topic such as transportation databases for questions about traffic calming could be more relevant than general databases [8,37]. 9. Use synthesis approaches to assess effects on health equity Questions about effects of interventions on health equity are likely to require additional synthesis approaches. These approaches may include meta- regression, subgroup analyses and sensitivity analyses, which are well described in the Cochrane Handbook [46] and other sources. As with any such analyses, these analyses need to be conducted according to existing quality standards such as a priori specification and use of other evidence to support hypotheses, such as other empiric evidence, within study effects supporting between- study differences and use of interaction tests [47]. Furthermore, these synthesis approaches may be used to test assumptions about the intervention using a causal pathway approach, which may strengthen inferences made based on these analyses. For example, a systematic review of HIV prevention interventions included an assessment of HIV causal pathways. The interventions were matched to HIV prevention goals along the proposed causal pathways to HIV infection [48]. 7. Develop data extraction tools for health equity Data extraction tools should include specific fields for disadvantage and health equity, as well as any within-study assessment of the effect on health equity as an outcome. We recommend defining all factors of interest in a data extraction checklist to reduce the risk of missing important information. This may include proxy indicators for disadvantage, such as nutritional status. Use of the PROGRESS-Plus framework can ensure that this important information is captured. 7. Develop data extraction tools for health equity Data extraction tools should include specific fields for disadvantage and health equity, as well as any within-study assessment of the effect on health equity as an outcome. We recommend defining all factors of interest in a data extraction checklist to reduce the risk of missing important information. This may include proxy indicators for disadvantage, such as nutritional status. Use of the PROGRESS-Plus framework can ensure that this important information is captured. 8. Assess the influence of context and process on health equity outcomes This includes using methods to assess the influence of context and process on the effects of Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Page 7 of 10 Page 7 of 10 10. Collect data related to applicability and equity Judgments about the applicability of findings to different settings and populations must be made by the user of a systematic review. Knowledge translation methods for an equity-focused systematic review (5)With what effect should research knowledge be transferred? There is still controversy about what endpoints should be considered and how they should be measured [55]. Appropriate outcomes for evaluating a specific KT strategy should be selected, and they may vary across different stakeholder groups and occur at individual, organizational and system levels [55]. Disadvantaged groups may differ in the outcomes they value compared to the more affluent. The explicit use of evidence in the policymaking process (recognizing the range of other influential factors to be considered in the process) is a commonly used outcome [55,56]. (5)With what effect should research knowledge be transferred? There is still controversy about what endpoints should be considered and how they should be measured [55]. Appropriate outcomes for evaluating a specific KT strategy should be selected, and they may vary across different stakeholder groups and occur at individual, organizational and system levels [55]. Disadvantaged groups may differ in the outcomes they value compared to the more affluent. The explicit use of evidence in the policymaking process (recognizing the range of other influential factors to be considered in the process) is a commonly used outcome [55,56]. For these equity-focused reviews, the end result will not be an incontrovertible message as the evidence will often be suggestive given that the evidence base is underdevel- oped and may also vary for populations and contexts. Therefore, knowledge translation methods need to as- semble the best available evidence and help end-users to use it to make better decisions about how and where to intervene. A multitude of frameworks for translating systematic review findings to policy and practice have been pub- lished [49-51]. The framework developed by Grimshaw et al. [52], drawing on Lavis et al. (2003) [53], suggests five questions that need to be addressed in developing a knowledge translation strategy. These are particularly pertinent for equity-focused systematic reviews given that they usually have more relevance to minority popu- lations and/or developing country populations where decision-makers may not be aware of systematic reviews. The knowledge translation plan should be specific to the end-users, keeping in mind their awareness of systematic reviews. The five questions above can be used as a general tem- plate by those designing, implementing and evaluating KT interventions. How to report the results of an equity-focused systematic review? The Equity Methods Group has developed reporting guidelines specific to systematic reviews focusing on equity questions: PRISMA-E 2012 [15]. These guidelines were launched at the Second Global Symposium on Health Systems Research in Beijing, China. Additionally, an equity checklist for use in planning systematic re- views is available from the Campbell and Cochrane Equity Methods Group (http://equity.cochrane.org/). (4)How should research knowledge be transferred? There is limited evidence, beyond the clinical context [52], about the effectiveness of knowledge translation strategies in general, let alone in reducing inequities. However, the literature suggests that any strategy is more likely to be successful if an assessment of the likely barriers and facilitators informs the choice of the specific interventions. 9. Use synthesis approaches to assess effects on health equity However, systematic review authors can assist this decision- making by providing details about the settings and populations in the primary studies, as well as exploring the mediating effect of factors identified in the analytic framework. Second, the systematic review can provide an assessment of the applicability to the most likely setting and population, given the body of evidence. This assessment should present both relative effects and absolute effects, which may be crucially important for health equity questions. The transparent reporting of these factors and their mediating role is likely to be helpful for the end-user of the systematic review. (1)What should be transferred? Up-to-date systematic reviews or other syntheses of global evidence are useful for decision-makers who need to consider a range of equity-related issues (i.e., beyond those described in single studies). Products emanating from these reviews may include structured and/or tailored summaries, patient decision aids, clinical practice guidelines and policy briefs. Evidence products should include a consideration beyond “what works” to consider for whom interventions work (or not), why and at what cost. (2)To whom should research knowledge be transferred? Equity-focused systematic reviews could be relevant to many different audiences including national/provincial policymakers in low- and middle- income countries, international aid agencies and practitioners. (3)By whom should research knowledge be transferred? Building credibility as a messenger is an important consideration and requires a tailored approach [54]. Different messengers are needed depending on the nature of the message, especially in a field where the political dimension of the message is an issue to be considered. Competing interests All authors have completed the ICJME unified disclosure form at www.icmje. org/coi_disclosure.pdf (available on request from the corresponding author); they have no financial relationships with any organizations that might have an interest in the submitted work. The corresponding author, MP, PT and JO are members of the Campbell and Cochrane Equity Methods Group, which has an interest in promoting the consideration of health equity in systematic reviews. Summary points Systematic review authors should determine whether equity considerations are relevant for their review at the question formulation stage and then plan their review accordingly. Acknowledgements The authors gratefully acknowledge participants of previous equity meetings: Oslo 2005, Ottawa 2007, Freiburg 2008, Singapore 2009, and Keystone 2010. In addition, the authors acknowledge the members of the PRISMA-Equity Bellagio group: Rifat Atun, Imperial College London, UK; Shally Awasthi, King George's Medical College, India; Virginia Barbour, PLOS Medicine, UK; Zulfiqar Bhutta, Husein Laljee Dewraj Professor and Chairman, Department of Paediatrics and Child Health, Aga Khan University Medical Center, Pakistan; Luis Gabriel Cuervo, Research Promotion & Development, Pan American Health Organization, USA; Jeremy Grimshaw, Ottawa Hospital Research Institute, Canada; Trish Groves, Deputy Editor, BMJ, UK; Tracey Koehlmoos- Perez, Health Systems and Infectious Disease Division, International Centre for Diarrhoeal Disease Research (ICDDR,B), Bangladesh; Elizabeth Kristjansson, School of Psychology, Faculty of Social Sciences, University of Ottawa, Canada; David Moher, Ottawa Hospital Research Institute, Canada; Jennifer O'Neill, Campbell and Cochrane Equity Methods Group, Centre for Global Health, University of Ottawa, Canada; Andy Oxman, Global Health Unit, Norwegian Knowledge Centre for Health Services, Cochrane EPOC Review Group, WHO Advisory Committee on Health Research, Norway; Tomas Pantoja, Universidad Católica de Chile, Santiago, Systematic Review Methods Centre Cochrane EPOC editor, Alliance for Health Systems and Policy Authors’ contributions All authors contributed to the drafting and editing of the manuscript. PT, VW, MP, JO, EW and EK contributed to the chapter in development for the Cochrane Handbook for Systematic Reviews of Interventions on which the article is based. VW will act as guarantor. All authors read and approved the final manuscript. In order to assist these objectives of improving the evi- dence base for health equity-focused policy questions, the Campbell and Cochrane Equity Methods Group has developed reporting guidelines specific for systematic re- views focusing on equity questions: PRISMA-E 2012 [15]. Additionally, guidance on conducting systematic reviews with a focus on health equity is in development to be added to the next major update of the Cochrane Handbook, and a health equity checklist for use in planning of systematic reviews is available online (http://equity.cochrane.org/). Discussion This article proposes ten steps in the systematic review process where reviews can consider effects on health equity including framing the question, choosing methods, collecting data, and assessing the role of context and implementation methods. The 2015 deadline for the Millennium Development Goals (MDGs) is rapidly approaching, yet the 2010 Millennium Development Goals Report revealed that without a major push, many of the MDG targets are likely to be missed [59]. A major obstacle to the progress of the MDGs has been the inability of health systems in many low- and middle-income countries to effectively imple- ment evidence-informed interventions. There are many examples of systematic reviews of high priority topics that can be used to inform policy-making to achieve the Millennium Development Goals. These include the use of zinc for the treatment of childhood diarrhea and of insecticide treated bednets to prevent malaria. However, the differences between the objectives of researchers and policymakers remain difficult to bridge. Equity in- formed reviews and their policy recommendations can help to bridge the knowledge translation gap by provid- ing policymakers with synthesized evidence in a form that identifies effects in disadvantaged groups, thus aiding with the development and implementation of policies and programs aiming to meet priority health objectives [60]. In order to maximize the effects of considering health equity in systematic reviews, knowledge translation steps are recommended that focus on the appropriate end-users and recognize that messages are likely to be suggestive and context-dependent. Knowledge translation methods for an equity-focused systematic review Although the evidence base to guide the choice of KT approaches targeted at policymakers is evolving [55,57], a profusion of innovative approaches exists that warrants further evaluation in the future. Integrated knowledge translation implies that relevant knowledge users (practitioners, policymakers, patients and Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Page 8 of 10 policies and programs intended to take action on social determinants of health and reduce health inequities. Knowledge translation of these equity-focused systematic reviews that takes into account the context-dependent effects on health equity and focuses on appropriate know- ledge users will contribute to increased awareness about the role of systematic reviews for equity-oriented decision- making. public) need to be involved in formulating the systematic review question and methods. For questions relevant to low- and middle-income countries (LMIC), who bear the highest burden of morbidity and mortality for all diseases, there is an urgent need to increase the representation of authors from LMIC in systematic review teams [58] because these authors can assist in actively translating/ transferring and exchanging results with target audi- ences such as policymakers in LMICs. Initiatives by funders such as the WHO Alliance for Health Systems and Policy Research, the International Initiative for Impact Evaluation (3ie), Ausaid, DFID and CIDA are developing the capacity to conduct systematic reviews in low- and middle-income countries. Summary points Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Research, Chile; Mark Petticrew, London School of Hygiene and Tropical Medicine, Campbell and Cochrane Equity methods group, UK; Terri Pigott, Campbell Methods Coordinating Group; Campbell Statistics Group, Loyola University Chicago, USA; Kent Ranson, Alliance for Health Systems and Policy Research, WHO, Switzerland; Tessa TanTorres, World Health Organization, Choosing Interventions that are Cos-Effective (WHO-CHOICE), Switzerland; Prathap Tharyan, Cochrane Evidence Aid, Department of Psychiatry; Cochrane Schizophrenia Group; Coordinator, South Asian Cochrane Network - India, India; David Tovey, The Cochrane Library, UK; Peter Tugwell, University of Ottawa, Cochrane Musculoskeletal Review Group, Campbell and Cochrane Equity methods group, Canada; Jimmy Volmink, Cochrane HIV/ AIDS Review Group; South Africa Cochrane Centre, Medical Research Council of South Africa; Stellenbosch University, South Africa; Liz Wager, Sideview, UK; Elizabeth Waters, Melbourne School of Population Health, The University of Melbourne, Australia; Vivian Welch, Ottawa Hospital Research Institute, Campbell and Cochrane Equity Methods Group, Institute of Population Health, Centre for Global Health, University of Ottawa, Canada; George Wells, Cardiovascular Research Methods Centre, University of Ottawa Heart Institute, Cochrane Non-Randomized Studies Methods Group; Howard White, Wardlaw T: Countdown to 2015 decade report (2000–10): taking stock of maternal, newborn, and child survival. Lancet 2010, 375(9730):2032–2044. Erratum in. Lancet 2010, 376(9743):772. 7. Chopra M, Campbell H, Rudan I: Understanding the determinants of the complex interplay between cost-effectiveness and equitable impact in maternal and child mortality reduction. Global J Health Sci 2012, 2(1):010406. 8. Ogilvie D, Egan M, Hamilton V, Petticrew M: Systematic reviews of health effects of social interventions: 2. Best available evidence: how low should you go? J Epidemiol Community Health 2005, 59(10):886–892. 9. Egan M, Petticrew M, Ogilvie D, Hamilton V, Drever F: “Profits before people”? A systematic review of the health and safety impacts of privatising public utilities and industries in developed countries. [Review] [51 refs]. J Epidemiol Community Health 2007, 61:862–870. p y 10. Snilstveit B: Systematic reviews: from ‘bare bones’ reviews to policy relevance. J Dev Effect 2012, 4(3):388–408. 11. Canadian Institutes of Health Research: About Knowledge Translation [Internet]. 2012. [updated 2012 Mar 9; cited 2012 Aug 7]. Available from: http://www.cihr-irsc.gc.ca/e/29418.html. International Initiative for Impact Evaluation (3ie); Campbell International Development Coordinating Group, India. 12. Armstrong R, Waters E, Jackson N, et al: Guidelines for Systematic Reviews of Health Promotion and Public Health Interventions, Version 2. Melbourne, Australia: Melbourne University; 2007. 13. Author details 1 15. Welch V, Petticrew M, Tugwell P, Moher D, O’Neill J, et al: PRISMA-Equity 2012 Extension: Reporting Guidelines for Systematic Reviews with a Focus on Health Equity. PLoS Med 2012, 9(10):e1001333. doi:10.1371/ journal.pmed.1001333. 1Institute of Population Health, University of Ottawa, 1 Stewart Street, Ottawa, ON K1N6N5, Canada. 2Department of Social and Environmental Health Research, London School of Hygiene and Tropical Medicine, London, UK. 3Victorian Health Promotion Foundation, Carlton, Australia. 4Division of Women and Child Health, Aga Khan University, Karachi, Pakistan. 5Epidemiology Research Unit, University of the West Indies, Mona, Jamaica. 6College of Health and Human Services, George Mason University, Fairfax, Virgina, USA. 7Department of Family Medicine, Pontificia Universidad Catolica de Chile, Santiago, Chile. 8Melbourne School of Population Health, The University of Melbourne, Melbourne, Australia. 9Department of Medicine, University of Ottawa, Ottawa, Canada. 10Epidemiology and Community Medicine, Faculty of Medicine, University of Ottawa, Ottawa, Canada. 11 16. Whitehead M: The concepts and principles of equity and health. Heal Promot Int 1991, 6(3):217–228. 17. Evans T, Brown H: Road traffic crashes: operationalizing equity in the context of health sector reform. Inj Control Saf Promot 2003, 10(1–2):11–12. 18. Oliver S, Kavanagh J, Caird J, Lorenc T, Oliver K, Harden A, et al: Health promotion, inequalities and young people’s health. A systematic review of research. EPPI-Centre. 2008. http://eppi.ioe.ac.uk/cms/Default.aspx? tabid=2410. 19. Kristjansson B, Petticrew M, MacDonald B, Krasevec J, Janzen L, Greenhalgh T, Wells GA, McGowan J, Farmer AP, Shea B, Mayhew A, Tugwell P, Welch V: School feeding for improving the physical and psychosocial health of disadvantaged students. Cochrane Database Syst Rev 2007(1). doi:10.1002/14651858.CD004676.pub2.CD004676. 11Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, Canada. 12School of Psychology, University of Ottawa, Ottawa, Canada. 11Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, Canada. 12School of Psychology, University of Ottawa, Ottawa, Canada. a. 12School of Psychology, University of Ottawa, Ottawa, Canada. Received: 17 April 2013 Accepted: 22 May 2013 Published: 22 June 2013 20. Thomas S, Fayter D, Misso K, Ogilvie D, Petticrew M, Sowden A, Whitehead M, Worth G: Population tobacco control interventions and their effects on social inequalities in smoking: systematic review. Tob Control 2008, 17(4):230–237. Funding Th d The development and evaluation of these recommendations was funded in part by the Canadian Institutes of Health Research who fund the Campbell and Cochrane Equity Methods Group, the Canada Graduate Scholarship for Vivian Welch and the Canada Research Chair Program for Peter Tugwell. The views expressed in this article are those of the authors and not necessarily those of The Cochrane Collaboration or its registered entities, committees or working groups. 14. Welch V, Petticrew M, Ueffing E, Benkhalti Jandu M, Brand K, Dhaliwal B, Kristjansson E, Smylie J, Wells GA, Tugwell P: Does consideration and assessment of effects on health equity affect the conclusions of systematic reviews? PLoS One 2012, 7(3):e31360. doi:10.1371/journal. pone.0031360. Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 Welch V, Tugwell P, Petticrew M, de Montigny J, Ueffing E, Kristjansson B, McGowan J, Benkhalti Jandu M, Wells GA, Brand, Smylie J: How effects on health equity are assessed in systematic reviews of interventions. Cochrane 2010. doi:10.1002/14651858.MR000028.pub2. Conclusions We hope that uptake and implementation of these recom- mendations will contribute to increased production and use of evidence on the effects of national and international Page 9 of 10 Page 9 of 10 25. Sidorchuk A, Agardh EE, Aremu O, Hallqvist J, Allebeck P, Moradi T: Socioeconomic differences in lung cancer incidence: a systematic review and meta-analysis. Canc Causes Contr 2009, 20:459–471. Reference ld 1. World Health Organization: Summary Report: World Conference on Social Determinants of Health. 2012. http://www.who.int/sdhconference/resources/ Conference_Summary_Report.pdf. 21. Waters E, de Silva-Sanigorski A, Hall BJ, Brown T, Campbell KJ, Gao Y, Armstrong R, Prosser L, Summerbell CD: Interventions for preventing obesity in children. Cochrane Database Syst Rev 2011, 12, CD001871. http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD004015.pub3/abstract. y p p 2. Bosch-Capblanch X, Lavis JN, Lewin S, Atun R, Røttingen J-A, et al: Guidance for evidence-informed policies about health systems: rationale for and challenges of guidance development. PLoS Med 2012, 9(3):e1001185. doi:10.1371/journal.pmed.1001185. 3. Milton B, Moonan M, Taylor-Robinson D, Whitehead M: How can the health equity impact assessment of universal policies be evaluated? Insights into approaches and next steps. Europe: Regional Office for World Health Organization; 2010. 4. Tugwell P, de Savigny D, Hawker G, Robinson V: Applying clinical epidemiological methods to health equity: the equity effectiveness loop. Br Med J 2006, 332:358–361. 5. Petticrew M, Tugwell P, Kristjansson E, Oliver S, Ueffing E, Welch V: Damned if you do, damned if you don’t: subgroup analysis and equity. Community Health: J Epidemiol; 2011. 6. Bhutta ZA, Chopra M, Axelson H, Berman P, Boerma T, Bryce J, Bustreo F, Cavagnero E, Cometto G, Daelmans B, de Francisco A, Fogstad H, Gupta N, Laski L, Lawn J, Maliqi B, Mason E, Pitt C, Requejo J, Starrs A, Victora CG, 2. Bosch-Capblanch X, Lavis JN, Lewin S, Atun R, Røttingen J-A, et al: Guidance for evidence-informed policies about health systems: rationale for and challenges of guidance development. PLoS Med 2012, 9(3):e1001185. doi:10.1371/journal.pmed.1001185. 2. Bosch-Capblanch X, Lavis JN, Lewin S, Atun R, Røttingen J-A, et al: Guidance for evidence-informed policies about health systems: rationale for and challenges of guidance development. PLoS Med 2012, 9(3):e1001185. doi:10.1371/journal.pmed.1001185. 22. Berg RC, Denison E: Interventions to reduce the prevalence of female genital mutilation/cutting in African countries. Campbell Systematic Reviews 2012, 9. 3. Milton B, Moonan M, Taylor-Robinson D, Whitehead M: How can the health equity impact assessment of universal policies be evaluated? Insights into approaches and next steps. Europe: Regional Office for World Health Organization; 2010. 23. Stewart R, van Rooyen C, Dickson K, Majoro M, de Wet T: What is the impact of microfinance on poor people? A systematic review of evidence from sub-Saharan Africa. London: EPPI-Centre, Social Science Research Unit, University of London; 2010. 4. Reference ld Lavis JN: Research, public policymaking, and knowledge-translation processes: Canadian efforts to build bridges. J Contin Educ Health Prof 2006, 26(1):37–45. 29. Anderson L, Petticrew M, Rehfuess E, Armstrong R, Ueffing E, Baker P, Francis D, Tugwell P: Using logic models to capture complexity in systematic reviews. Res Synth Meth 2011, 2:33–42. 50. Straus S, Tetroe J, Graham ID: Knowledge translation in health care: moving from evidence to practice. Oxford: Wiley-Blackwell - BMJ Books; 2009. 51. Rycroft-Malone J, Bucknall T: Models and frameworks for implementing evidence-based practice: linking evidence to action. Oxford: Wiley-Blackwell – Sigma Theta Tau International; 2010. 30. Welch VA, Awasthi S, Cumberbatch C, Fletcher R, McGowan J, Krishnaratne S, Sohani S, Tugwell P, Wells GA: Deworming and Adjuvant Interventions for Improving the Developmental Health and Well-being of Children in Low- and Middle-income, A Systematic Review and Network Meta-analysis. Countries: Campbell Library; 2013. 52. Grimshaw JM, Eccles MP, Lavis JN, Hill SJ, Squires JE: Knowledge translation of research findings. Implement Sci 2012, 7(50). 53. Lavis JN, Robertson D, Woodside JM, McLeod CB, Abelson J: How can research organizations more effectively transfer research knowledge to decision makers? Milbank Q 2003, 81(2):221–248. 31. Mackenbach JP, Kunst AE: Measuring the magnitude of socio-economic inequalities in health: An overview of available measures illustrated with two examples from Europe. Soc Sci Med 1997, 44(6):757–551. 54. Shonkoff JP: Science, Policy and Practice: Three cultures in search of a shared mission. Child Dev 2000, 71(1):181–187. 32. Greenaway C, Sandoe A, Vissandjee B, Kitai I, Gruner D, Wobeser W, Pottie K, Ueffing E, Menzies D, Schwartzman K: For the Canadian Collaboration for Immigrant and Refugee Health. Can Med Assoc J 2011, 183(12):e939–e951. 55. Dobbins M, Hanna S, Ciliska D, Manske S, Cameron R, Mercer S, O’Mara L, DeCorby K, Robeson P: A randomized controlled trial evaluating the impact of knowledge translation and exchange strategies. Implement Sci 2009, 4(1):61. 33. Tugwell P, Petticrew M, Kristjansson E, Welch V, Ueffing E, Waters E, Bonnefoy J, Morgan A, Doohan E, Kelly MP: Assessing equity in systematic reviews: realizing the recommendations of the Commission on Social Determinants of Health. BMJ 2010. doi:10.1136/bmj.c4739. 56. Amara N, Ouimet M, Landry R: New evidence on instrumental, conceptual and symbolic utilization of university research in government agencies. Sci Commun 2004, 26(75):75–106. 34. Reference ld Tugwell P, de Savigny D, Hawker G, Robinson V: Applying clinical epidemiological methods to health equity: the equity effectiveness loop. Br Med J 2006, 332:358–361. 24. Bambra C, Gibson M, Sowden AJ, Wright K, Whitehead M, Petticrew M: Working for health? Evidence from systematic reviews on the effects on health and health inequalities of organisational changes to the psychosocial work environment. Prev Med 2009, 48(5):454–461. doi:10.1016/j.ypmed.2008.12.018. 5. Petticrew M, Tugwell P, Kristjansson E, Oliver S, Ueffing E, Welch V: Damned if you do, damned if you don’t: subgroup analysis and equity. Community Health: J Epidemiol; 2011. y 25. Sidorchuk A, Agardh EE, Aremu O, Hallqvist J, Allebeck P, Moradi T: Socioeconomic differences in lung cancer incidence: a systematic review and meta-analysis. Canc Causes Contr 2009, 20:459–471. 6. Bhutta ZA, Chopra M, Axelson H, Berman P, Boerma T, Bryce J, Bustreo F, Cavagnero E, Cometto G, Daelmans B, de Francisco A, Fogstad H, Gupta N, Laski L, Lawn J, Maliqi B, Mason E, Pitt C, Requejo J, Starrs A, Victora CG, Page 10 of 10 Welch et al. Systematic Reviews 2013, 2:43 http://www.systematicreviewsjournal.com/content/2/1/43 26. Lewin S, Munabi-Babigumira S, Glenton C, Daniels K, Bosch-Capblanch X, van Wyk BE, Odgaard-Jensen J, Johansen M, Aja GN, Zwarenstein M, Scheel IB: Lay health workers in primary and community health care for maternal and child health and the management of infectious diseases. Cochrane Database Syst Rev 2010, 3, CD004015. http://onlinelibrary.wiley.com/doi/10.1002/ 14651858.CD004015.pub3/abstract. 45. Greenhalgh T, Kristjansson E, Welch V: Realist review to understand the efficacy of school feeding programmes. BMJ 2007, 335:858–861. 46. Higgins JPT, Green S: Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration 2011. Available from www.cochrane-handbook.org. Collaboration 2011. Available from www.cochrane-handbook.org. 47. Sun X, Briel M, Walter SD, Guyatt GH: Is a subgroup effect believable? Updating criteria to evaluate the credibility of subgroup analyses. BMJ 2010, 340:c117. 27. Tully MA, Kee F, Foster C, Cardwell CR, Weightman AL, Cupples ME: Built environment interventions for increasing physical activity in adults and children (Protocol). Cochrane Database Syst Rev 2013, 1. http://onlinelibrary. wiley.com/doi/10.1002/14651858.CD010330/abstract. 48. Harrison A, Newell M, Imrie J, Hoddinott G: HIV prevention for South African youth: which interventions work? A systematic review of curren evidence. BMC Publ Health 2010, 10:102. 28. Waddington H, Snilstveit B, White H, Fewtrell L: Water, sanitation and hygiene interventions to combat childhood diarrhoea in developing countries. New Delhi, India: 3ie; 2009. 49. Reference ld Turley R, Saith R, Bhan N, Rehfuess E, Carter B: Slum upgrading strategies involving physical environment and infrastructure interventions and their effects on health and socio-economic outcomes. Cochrane Database Syst Rev 2013(1). http://onlinelibrary.wiley.com/doi/10.1002/14651858. CD010067.pub2/abstract. 57. Armstrong R, Waters E, Dobbins M, Lavis JN, Petticrew M, Christensen R: Knowledge translation strategies for facilitating evidence-informed public health decision making among managers and policy-makers (Protocol). Cochrane Database Syst Rev 2011(6). http://onlinelibrary.wiley. com/doi/10.1002/14651858.CD009181/abstract. 35. Thomson H, Thomas S, Sellstrom E, Petticrew M: Housing improvements for health and associated socio-economic outcomes: A systematic review. Campbell Systematic Reviews 2013, 2. www.campbellcollaboration. org/lib/download/2596/. 58. Allen C, Clarke M, Tharyan P: International activity in The Cochrane Collaboration with particular reference to India. Natl Med J India 2007, 20(5):250–255. 36. Nutbeam D: How does evidence influence public health policy? Tackling health inequalities in England. Health Promot J Aust 2003, 14:154–158. 59. Nations U: The Millennium Development Goals Report 2010. New York: United Nations; 2010. 37. Augustincic Polec L, Ueffing E, Welch V, Tanjong Ghogomu E, Pardo Pardo J, Grabowsky M, Attaran A, Tugwell P: Strategies to increase the ownership and use of insecticide treated bednets to prevent malaria (Protocol). Cochrane Database Syst Rev 2011(6). http://onlinelibrary.wiley. com/doi/10.1002/14651858.CD009186/abstract. 60. Panisset U, Koehlmoos TP, Alkhatib AH, Pantoja T, Singh P, Kengeya- Kayondo J, McCutcheon B: Implementation research evidence update and use for policy making. Health Res Pol Syst 2012, 10:20. doi:10.1186/2046-4053-2-43 Cite this article as: Welch et al.: Health equity: evidence synthesis and knowledge translation methods. Systematic Reviews 2013 2:43. 38. Pienaar E, Grobler L, Busgeeth K, Eisinga A, Siegfried N: Developing a geographic search filter to identify randomised controlled trials in Africa: finding the optimal balance between sensitivity and precision. Health Info Libr J 2011, 28:210–215. doi:10.1111/j.1471-1842.2011.00936.x. 39. McGowan J, Robinson V, Kristjansson B, Petticrew M, Tugwell P: Identifying studies that include issues of equity in Cochrane Reviews. Barcelona: XI Cochrane Colloquium; 2003. Poster presentation. 39. McGowan J, Robinson V, Kristjansson B, Petticrew M, Tugwell P: Identifying studies that include issues of equity in Cochrane Reviews. Barcelona: XI Cochrane Colloquium; 2003. Poster presentation. 40. Leclercq E, Leeflang M-MG, van Dalen E-C: [P121] Validation of a PubMed search filter for identifying studies including children. Cochrane Colloquium Abstracts Journal. Keystone, Colorado: 18th Cochrane Colloquium and the 10th Campbell Colloquium. 2010. Available: http://www.ncbi.nlm.nih.gov/ pubmed/23084708. Accessed 08 March 2013. 40. Reference ld Leclercq E, Leeflang M-MG, van Dalen E-C: [P121] Validation of a PubMed search filter for identifying studies including children. Cochrane Colloquium Abstracts Journal. Keystone, Colorado: 18th Cochrane Colloquium and the 10th Campbell Colloquium. 2010. Available: http://www.ncbi.nlm.nih.gov/ pubmed/23084708. Accessed 08 March 2013. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: 41. Pawson R, Greenhalgh T, Harvey G, Walshe K: Realist review–a new method of systematic review designed for complex policy interventions. J Health Serv Res Policy 2005, 10(1):21–34. • Convenient online submission • Thorough peer review 42. Noblit GW, Hare RD: Meta-ethnography: synthesizing qualitative studies. Newbury Park: Sage; 1988. 43. Thomas J, Harden A: Methods for the thematic synthesis of qualitative research in systematic reviews. BMC Med Res Methodol 2008, 8:45. 44. Kastner M, Tricco AC, Soobiah C, Lillie E, Perrier L, Horsley T, Welch V, Cogo E, Antony J, Straus SE: What is the most appropriate knowledge synthesis method to conduct a review? Protocol for a scoping review. BMC Med Res Meth 2012, 12:114.
https://openalex.org/W4220731366	https://zenodo.org/record/6326794/files/%D0%A1%D0%B0%D1%81%D0%BE%D0%B2%2C%20%D0%A1%D1%83%D0%BF%D1%80%D1%83%D0%BD%D0%B5%D0%BD%D0%BA%D0%BE%20-165-175.pdf	Russian	null	КРИМИНОЛОГИЧЕСКАЯ ХАРАКТЕРИСТИКА ЛИЧНОСТИ СЕРИЙНОГО УБИЙЦЫ	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	3,304	УДК 343,9 DOI УДК 343,9 DOI местного самоуправления в современных условиях // Сборник научных работ серии «Право» Вып. 5 / ГОУ ВПО «ДонАУиГС». – Донецк: ГОУ ВПО ДонАУиГС, 2017. – С. 142-152. местного самоуправления в современных условиях // Сборник научных работ серии «Право» Вып. 5 / ГОУ ВПО «ДонАУиГС». – Донецк: ГОУ ВПО ДонАУиГС, 2017. – С. 142-152. местного самоуправления в современных условиях // Сборник научных работ серии «Право» Вып. 5 / ГОУ ВПО «ДонАУиГС». – Донецк: ГОУ ВПО ДонАУиГС, 2017. – С. 142-152. цк: ГОУ ВПО ДонАУиГС, 2017. – С. 142-152. 3. Погодин И.А. Суицидальное поведение : психологические аспекты: учеб. пособие / И.А. Погодин. - 2-е изд., стереотип. - М.: Флинта: МПСИ, 2011. - 336 с. 4. Превенция самоубийств: руководство для лиц, оказывающих первичную медицинскую помощь/ пер. с анг. Ю.О. Донец, ред. В.А. Розанов. - Одесса: ОНУМ, 2005. - 23с. 5. Шнейдман Э.С. Душа самоубийцы / Э.С. Шнейдман; пер. с англ. О.Ю.Донец, ред. А.Н. Моховик. - М.: Смысл, 2001. - 319 с. 5. Шнейдман Э.С. Душа самоубийцы / Э.С. Шнейдман; пер. с англ. О.Ю.Донец, ред. А.Н. Моховик. - М.: Смысл, 2001. - 319 с. САСОВ А.В., канд. юрид. наук, доцент, доцент кафедры административного права факультета юриспруденции и социальных технологий ГОУ ВПО «ДОНАУИГС», г. Донецк, Донецкая Народная Республика САСОВ А.В., канд. юрид. наук, доцент, доцент кафедры административного права факультета юриспруденции и социальных технологий ГОУ ВПО «ДОНАУИГС», г. Донецк, Донецкая Народная Республика СУПРУНЕНКО Е.А., магистрант группы Юр-20 з-2 факультета юриспруденции и социальных технологий ГОУ ВПО «ДОНАУИГС», г. Донецк, Донецкая Народная Республика СУПРУНЕНКО Е.А., магистрант группы Юр-20 з-2 факультета юриспруденции и социальных технологий ГОУ ВПО «ДОНАУИГС», г. Донецк, Донецкая Народная Республика Исследована проблема раскрытия личности умышленного серийного убийцы, которая предполагает изучение криминологических и уголовно- правовых механизмов противоправного поведения, мотивации насильственных преступлений, его индивидуально-психологических, социальных и ситуативных факторов. Ключевые слова: личность преступника; умышленное убийство; серийный убийца; криминологическая и уголовно-правовая структура преступления; ценностные ориентации преступника; мотивация насильственных преступлений 165 CRIMINOLOGICAL CHARACTERIZATION OF THE IDENTITY OF A SERIAL KILLER SASOV A.V., Candidate of Law, Associate Professor of the Department of Administrative Law, Faculty of Jurisprudence and Social Technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic, SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic The problem of revealing the identity of a deliberate serial killer, which involves studying the criminological and criminal legal mechanisms of unlawful behavior, the motivation of violent crimes, its individual psychological, social and situational factors, was investigated. Keywords: identity of the offender; intentional murder; serial killer; criminological and criminal legal structure of the crime; the value orientation of the perpetrator; motivation for violent crime. Постановка задачи. Одной из важнейших составляющих криминологической характеристики преступлений вообще, а в особенности одного из наиболее общественно опасного вида умышленных убийств – серийных, является криминологическая характеристика личности серийного убийцы. Важное значение для практики противодействия данному преступлению является классификация серийных убийств и убийц. Анализ последних исследований и публикаций. В криминологии над проблемой исследования личности убийцы работали В.М. Анисимков, Ю.М. Антонян, А.Г. Бабичев, В.А. Байков, A.A. Бессонов, С.А. Бояров, М.М. Гитинова, М.И. Еникеев, А.С. Капинус, М.С. Корягина, А.П. Левченко, А.В. Старков, A.A. Реймер. Однако отдельные вопросы остаются недостаточно исследованными или дискуссионными, вследствие чего наблюдается неоднозначное толкование тех или иных правовых норм. Таким образом, возникает необходимость углубленного исследования этого проблемного вопроса. Анализ последних исследований и публикаций. В криминологии над проблемой исследования личности убийцы работали В.М. Анисимков, Ю.М. Антонян, А.Г. Бабичев, В.А. Байков, A.A. Бессонов, С.А. Бояров, М.М. Гитинова, М.И. Еникеев, А.С. Капинус, М.С. Корягина, А.П. Левченко, А.В. Старков, A.A. Реймер. Однако отдельные вопросы остаются недостаточно исследованными или дискуссионными, вследствие чего наблюдается неоднозначное толкование тех или иных правовых норм. Таким образом, возникает необходимость углубленного исследования этого проблемного вопроса. 166 Актуальность. Источником социальных потребностей, вызывающих необходимость дальнейшего развития юридической науки в целом и криминологии, в частности является преступность, присущая любому государству и обществу. Убийства, как особо тяжкие насильственные преступления, составляют группу опасных антиобщественных проявлений, борьба с которыми, а также их предупреждение является важной сферой деятельности правоохранительных органов. Личность преступника интересна для научных исследований. Цель предлагаемой статьи состоит в исследовании личности преступника как социально-психологического явления, а также обосновании и разработке теоретических и методических основ изучения классификации серийных убийц. у ф р у Изложение основного материала исследования. В юридических науках уголовного цикла (криминологии, криминалистике, уголовном праве) разработаны многочисленные типологии и классификации убийц. Обобщенную типологию целесообразно формулировать, опираясь на три базовых признака: мотивацию, способ совершения и количество преступных посягательств на жизнь. Также можно выделить отдельные категории убийц, опирающиеся на конкретные признаки (пол, возраст, форма вины и раскаяния и т.п.). Согласно предложенным признакам целесообразно выделить такие основные типы умышленного убийцы, как серийный убийца, ритуальный убийца, террорист-самоубийца, убийца-каннибал, убийца-ксенофоб, убийца по заказу [1, с. 34]. SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic Среди перечисленных типов наиболее сложным в контексте уголовно-психологического изучения и понимания является серийный убийца. 167 Одним из исследователей, впервые попытавшихся основательно исследовать преступные (социально- психологические) типы, был врач-психиатр Ч. Ломброзо, который изучал представителей мест лишения свободы и своеобразие татуировок на их телах. Ч. Ломброзо и его последователь Э. Ферри различали следующие типы преступников: 1) врожденные; 2) «преступники вследствие безумия», психопаты и другие, страдающие психическими аномалиями; 3) преступники из-за страсти; 4) случайные; 5) привычные. Ч. Ломброзо – выдающийся представитель антропологической школы уголовного права, возникшей в 70–80-х гг. XIX ст. Ученый утверждал, что 167 преступный тип личности можно распознать (диагностировать) по специфическим физическим и психическим признакам [5, с. 44]. Понятие «серийный убийца» впервые официально начали применять в США и Западной Европе в конце 70-х гг. ХХ в. Термин «серийный убийца» предложил агент ФБР Роберт Ресслер, используя его для обозначения лиц, совершавших убийства с особой жестокостью и садизмом с целью получения сексуального наслаждения и разрядки. После открытия в 1985 г. при Академии ФБР Национального центра анализа насильственных преступлений и в его рамках отдела по изучению поведения убийц содержание понятия «серийный убийца» приобрело гораздо более широкое значение [2, с. 63-65]. Повышенный интерес американских специалистов, в частности, специалистов по составлению психологического профиля убийцы, можно объяснить тем, что США занимают первое место в мире по количеству серийных убийств и лиц, их совершающих (ежегодно выявляют до 100 человек). К тому же из 200 самых известных в мире серийных убийц более трети являются американцами [5, с. 134]. р [ , В современном понимании серийный убийца – это лицо, которое в течение определенного времени (от нескольких месяцев до года) совершает три и более умышленных убийства отдельных категорий лиц при обстоятельствах, отягчающих наказание. Такими обстоятельствами прежде всего являются: особая жестокость, насильственное удовлетворение половой страсти неестественным способом, надругательство над трупом, несовершеннолетние или малолетние жертвы. Классически, считается, что мотивация серийных убийств включает ведущий и сопутствующий мотивы. К ведущему мотиву можно отнести психологическое и социальное отчуждение в детском и юношеском возрасте, миссионерский мотив (выполнение особой миссии, возложенной на убийцу или произвольно им избранной), месть на сексуальной основе, комплекс неполноценности и т.д. [3, с. 74]. Важной психолого-юридической характеристикой серийного убийцы обычно является тождество или сходство мотивов преступлений и однотипность способов их совершения. Наиболее распространенным способом серийного убийства является удушение, на втором месте – нанесение колото-резаных ран, не совместимых с жизнью, а на третьем – отравление с 168 помощью специальной жидкости, медицинских препаратов, тяжелых металлов (ртуть, таллий и т.п.). SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic Среди менее распространенных способов следует выделить применение огнестрельного оружия. Жертвами серийного убийцы становятся определенные категории лиц, которых преступник предварительно тщательно отбирает. В зависимости от ведущего мотива, предпочтений, извращений и других характеристик серийного убийцы его жертвами становятся женщины определенного возраста, внешности и занятий; несовершеннолетние и малолетние лица (преимущественно женского пола); мужчины определенной профессии, рода занятий и внешности. Обычно серийным убийцам присущи осторожность и минимизация риска, хитрость, способность к ролевому перевоплощению, тщательность планирования преступных действий, эмоциональная асинтонность (тупость) к жертвам. Также у них развиты антиципация, самоконтроль, воображение и фантазия [4, с. 34-40]. Для объяснения схемы поведения серийного убийцы можно использовать разработанную зарубежными специалистами шестиэтапную модель. Первый этап этой модели называется этапом фантазии, отражающей потребности, мотивы, предполагаемые «вознаграждения» от задуманного убийства. Второй этап – сталкинг, заключающийся в выборе и выслеживании жертвы. Характер, длительность и другие особенности поведения серийного убийцы на этом этапе зависят от силы его желания и степени нетерпения, личностных качеств как самого преступника, так и избранной жертвы. Третий этап – похищение или заманивание жертвы, хотя он присутствует не во всех случаях серийных убийств. Четвертый этап – непосредственный акт убийства, предусматривающий два варианта поведения преступников на этом этапе. Одни из них фокусируют внимание на действиях, которые непосредственно приводят к смерти жертвы, а другие – предпочитают сопутствующие действия (пытки, глумление, унижение, сексуальное насилие и др.) в процессе физического взаимодействия с жертвой. Пятый этап – размещение трупа. Некоторые серийные убийцы покидают труп с поличным, а другие обязательно перемещают тело 169 жертвы для того, чтобы его как можно быстрее обнаружили или наоборот. Шестым этапом поведения серийного убийцы считают депрессию или эмоциональное охлаждение. Следует отметить, что именно эмоциональное охлаждение отличает серийного убийцу от других убийц, в том числе множественных. По признанию многих преступников, им часто не удавалось получить от убийства того эффекта, на который они рассчитывали [3, с. 138-143]. По географическому признаку и печальной известности серийных убийц-мужчин условно можно разделить на три группы: всемирно известные зарубежные убийцы, русские и отечественные. Первая группа – это лица европеоидной расы, среднего или низшего уровня образования, в возрасте от 30 лет и старше, невысокой профессиональной квалификации. Более половины таких преступников – военнослужащие (в том числе бывшие), медицинские работники. Некоторые из них в детстве страдали симптомами так называемой «триады Макдональда»: мочевыделение в постели, пытки животных, поджоги. Также характерными признаками являются отсутствие стремления к употреблению алкоголя и наркотиков, социопатичность (замаскированная аномальность характера и поведения при отсутствии ярко выраженных психических расстройств), склонность к сексуальному садизму. SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic Последние три характеристики универсальны и присущи не только указанной, но и другим группам серийных убийц. ру ру р у Наиболее известными серийными убийцами мирового масштаба являются Теодор Банди, Генри Лукас, Джон Гейси, Джеффи Дамер, Томми Селс (США), Гарольд Тимпан, Джон Хейг (Великобритания), Педро Лопес, Даниэль Барбоса (Колумбия), Педро (Бразилия) [3, с. 153]. р ( р ) [ ] В этом списке значительный интерес вызывает личность Джона Гейси, известного как «убийца-клоун». Его отец злоупотреблял алкоголем и прибегал к насилию в семье. В детстве Джона развратила соседская девушка, а приятель отца, который был педофилом, совершил над ним сексуальное насилие. Не окончив школу, Джон переехал в Лас-Вегас, где устроился работать в трупарню. По его словам, там он впервые прибег к некрофилии. С 1975 г. Джон Гейси начал выступать в роли клоуна Пого на детских праздниках, что давало ему возможность присмотреться к будущим 170 жертвам. Он убил 33 человека, предварительно жестоко издеваясь над ними и удовлетворяя собственную половую страсть. 29 тел спрятал в подвале собственного дома, а четыре тайно спустил в реку. Дж. Гейси тщательно обдумывал и занимался подготовкой преступлений. В ходе расследования Джон Гейси утверждал, что у него раздвоение личности и убийства совершал его двойник Джек Хенли [3, с. 154-156]. жертвам. Он убил 33 человека, предварительно жестоко издеваясь над ними и удовлетворяя собственную половую страсть. 29 тел спрятал в подвале собственного дома, а четыре тайно спустил в реку. Дж. Гейси тщательно обдумывал и занимался подготовкой преступлений. В ходе расследования Джон Гейси утверждал, что у него раздвоение личности и убийства совершал его двойник Джек Хенли [3, с. 154-156]. Раздваивание личности – это психическое расстройство, которое негативно влияет на самосознание, память и идентичность личности. Основной причиной его возникновения есть психическое или сексуальное насилие в детском возрасте. Ко второй группе серийных убийц-мужчин принадлежат представители Российской Федерации, занимающей второе место в мире после США по количеству серийных убийц. Начиная с 1991 г., в России было обнаружено более 70 серийных убийц, а город Ростов-на-Дону приобрел печальную славу «Мекки» для этих лиц. у р р у у В советское время первым официально признанным российским серийным убийцей был Василий Комаров, совершавший преступления в Москве и Ростовской области. Больше всего убийств совершил Андрей Чикатило («Ростовский резник»), жертвами которого стали 53 человека. В советское время первым официально признанным российским серийным убийцей был Василий Комаров, совершавший преступления в Москве и Ростовской области. Больше всего убийств совершил Андрей Чикатило («Ростовский резник»), жертвами которого стали 53 человека. SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic После распада СССР и образования Российской Федерации к перечню наиболее известных российских серийных убийц относятся: Александр Пичушкин («Битцевский маньяк», 49 жертв), Михаил Попов («Ангарский маньяк», 22 жертвы), Сергей Ряховский («Балашихский резник», 19 жертв), Владимир Миргород («Душитель», 16 жертв), Абдуфат Заманов («Красноярский Чикатило», 14 жертв), Денис Писчиков («Уральский Раскольников», 13 жертв), Николай Дудин («Обидный маньяк», 13 жертв), Максим Петров («Медбрат», 12 жертв), Евгений Петров («Новоуральский потрошитель», 11 жертв), Сергей Головкин («Удав», 11 жертв), Олег Кузнецов («Лжедмитрий», 10 жертв) и др. [5, с. 153]. Среди вышеназванных преступников внимание привлекает личность Александра Пичушкина. В школьные годы будущий убийца увлекался шахматами, что в дальнейшем своеобразно отразилось в его преступной деятельности. По признанию Пичушкина, к совершению убийств его побудило как собственное 171 желание, так и резонансное дело Андрея Чикатило. Он задумал совершить 64 убийства по количеству клеток на шахматной доске. После каждой удачной попытки он закрывал ячейку. Ведущим мотивом было миссионерство, которое затем соединилось с патологическим наслаждением от акта убийства. Первоначально жертвами Пичушкина были алкоголики, бомжи и другие асоциальные лица, которые, по его мнению, не имели права на жизнь. Затем он начал убивать знакомых и получал от этого особое наслаждение. Показательно то, что к 2005 г. убийца прятал трупы (в канализационных колодцах и других местах), а затем, пытаясь заявить о себе, начал оставлять тела убитых им лиц на открытых местах. А. Пичушкин имел свой преступный почерк: убивал жертву многочисленными ударами молотка по голове, вставляя после этого в открытую рану ветку или бутылку. Наиболее известными и одиозными украинскими серийными убийцами являются Анатолий Оноприенко (52 жертвы), Сергей Ткач (37 жертв), Сергей Довженко (21 жертва), Виктор Саенко (21 жертва), Руслан Хамара (11 жертв) и другие [3, с. 166-167]. Особую категорию серийных убийц образуют женщины. По мировой статистике количественное соотношение серийных убийц мужчин и женщин составляет 9:1. На протяжении ХХ ст. в Европе, США и Канаде было разоблачено более 100 серийных убийц женского пола. Женщинам – серийным убийцам присущи такие характеристики, как хитрость, изящество, жестокость, склонность к перевоплощению, акцентуация по истероидному типу, психические аномалии и расстройства, в частности, синдром Мюнхаузена: убедительная симуляция или сознательный вызов у себя физических заболеваний (инфаркт, лихорадка и т.д.) с целью привлечения внимания к себе, сочувствия, лечения у престижных врачей. Около трети женщин – серийных убийц действуют в составе групп. Чаще это тандем мужчина-женщина. Используя образную терминологию зарубежных специалистов, можно выделить следующие разновидности серийных убийц женского пола: «черная вдова»; «ангел смерти»; «мстительницы»; «детоубийца». SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic «Черная вдова» получила название по аналогии с самкой ядовитого паука черного тарантула, которая убивает партнера-самца после спаривания. Ведущим мотивом этой категории серийных убийц является стремление женщины к увеличению собственного состояния. Жертвами становятся 172 мужчины, любовники, члены семьи и другие лица, с которыми она установила доверительные отношения. Основным способом убийства является применение яда и медицинских препаратов в целях маскировки преступления под естественную смерть [3, с. 189-197]. Одной из известнейших представительниц этой разновидности серийных убийц-женщин является американка норвежского происхождения Белла Соренсон Гинесс. Ее жертвами стали 42 человека, среди которых два ее мужа, собственные дочери, несколько поклонников и любовников. После смерти родных Белла Соренсон Гинесс получала страховку. За деньги, вырученные от страховой компании, она приобрела ферму. Вскоре с помощью любовных писем и объявлений убийца начала привлекать к себе мужчин среднего возраста, после чего с ними случались несчастные случаи [3, с. 199]. Представительницы категории убийц «ангел смерти» обычно работают в больницах, домах престарелых, приютах, хосписах, то есть в местах, где случаи смерти происходят постоянно. Жертвами могут быть как дети, так и старики. Для убийств эти женщины используют медицинские препараты и разное медицинское оборудование. Эта разновидность женщин – серийных убийц ярко представляет Джейн Топпан, работавшая медсестрой. Еще во время учебы она проявляла повышенный интерес к фотографиям по вскрытию тел. Дж. Топпан получала сексуальное удовольствие от введения пациентам наркотиков и доведения их до предела смерти. Когда они находились в бессознательном состоянии, она ощущала сексуальное возбуждение. В конце концов, ее арестовали по обвинению в совершении 11 убийств. В ходе следствия Дж. Топпан призналась, что убила 31 человека. Ее признали невменяемой из-за безумия, остальную жизнь она провела в специальном медицинском учреждении [3, с. 198- 208]. провела в специальном медицинском учреждении [3, с. 198 208]. Ведущим мотивом убийств «мстительниц» является месть за психологическое и социальное отчуждение, оскорбление, издевательство, сексуальные притязания и изнасилования. Жертвами этой группы убийц обычно становятся конкретные лица, совершившие ответные действия или определенные категории лиц по символическому значению. 173 В современной Украине широкий резонанс получила история с «мстительницей» Ниной Л., которая работала уборщицей в столовой одной из столичных школ. Поводом для тревоги послужили случаи групповых жалоб младших школьников на плохое самочувствие после обедов. Многочисленные проверки, анализы и обследования ситуацию не прояснили. После тщательного изучения лиц, причастных к работе в столовой, работники правоохранительных органов обратили внимание на уборщицу. Одинокая женщина на протяжении нескольких лет трижды становилась вдовой, потеряла мать. Все ее родные скончались после тяжелой, но непродолжительной болезни внутренних органов. Список использованных источников 1. Баронин А.С. Психологический профиль убийцы: пособие по криминальной психологии и криминалистике / А.С. Баронин. – К.: Паливода А.В., 2001. – 176 с. , 2. Криминальная психология: учебное пособие / авт.-сост. А.И. Ушатиков, О.Г. Ковалев. – М.: Изд-во Московского психолого-социального института; Воронеж: МОДЭК, 2007. – 496 с. 3. Собольников В.В. Криминальная психология: учебник для вузов / В.В. Собольников. – 2-е изд., перераб. и доп. – М.: Юрайт, 2020. – 379 с. [Электронный ресурс]. – Режим доступа: https://urait.ru/bcode/447910. – (Дата обращения: 16.12.2021). 4. Образцов В.А. Криминалистическая психология: учебное пособие для вузов / В.А. Образцов, С.Н. Богомолова. – М.: Юнити- Дана; Закон и право, 2002. – 447 с. 4. Образцов В.А. Криминалистическая психология: учебное пособие для вузов / В.А. Образцов, С.Н. Богомолова. – М.: Юнити- Дана; Закон и право, 2002. – 447 с. 5. Криминальная психология: хрестоматия / сост.: И.А. Фурманов, И.Е. Метлицкий. – Минск: БГУ, 2018. - 270 с. ПРОБЛЕМЫ РЕАЛИЗАЦИИ НАКАЗАНИЙ В УГОЛОВНО- ПРАВОВЫХ НОРМАХ ДОНЕЦКОЙ НАРОДНОЙ РЕСПУБЛИКИ СТАДНИК Д.В., следователь прокуратуры Куйбышевского района, г. Донецк, Донецкая Народная Республика SUPRUNENKO EA, undergraduate of the group Yur-20 с-2 of the Faculty of Law and social technologies SEE HPE «DAMPA», Donetsk, Donetsk People’s Republic В дальнейшем было установлено, что среди знакомых и подруг Нины Л. также умерли шесть человек. Следствием было установлено, что преступница – акцентируемая личность застревающего типа. После ссор, недоразумений, незначительных оскорблений женщина прибегала к мести весьма своеобразным способом. Она подмешивала в пищу своим жертвам легкорастворимые соли металла таллия, который влияет на человеческий организм подобно ртути. Свои действия в отношении школьников Нина. Л. объяснила их плохим поведением в столовой и недостаточным проявлением уважения к ней самой. [2, с. 189-197]. Вывод. Таким образом, к важным признакам серийных убийств относятся: Вывод. Таким образом, к важным признакам серийных убийств относятся: - периодическая актуализация криминальной агрессии (убийца никогда не останавливается на уже совершенном по собственной инициативе); - избыточность криминальной агрессии (убийство обязательно сопровождается насильственными действиями, непосредственно не связанными с лишением жизни); - высокая мобильность после совершения убийства (убийца быстро исчезает с места преступления как можно дальше): - единичность жертвы («один эпизод» – «одна жертва»), совершение убийства без соучастников (исключениями являются пары, связанные семейными или сексуальными связями), отсутствие очевидцев. 174 Итак, основой уголовно-психологической классификации личности преступников являются доминирующие позиции 174 Итак, основой уголовно-психологической классификации личности преступников являются доминирующие позиции личности, ее побуждения, мотивы, устойчивые цели и способы совершения преступления, мера десоциализированности личности, характер ее антиобщественной направленности. личности, ее побуждения, мотивы, устойчивые цели и способы совершения преступления, мера десоциализированности личности, характер ее антиобщественной направленности. Список использованных источников СТАДНИК Д.В., следователь прокуратуры Куйбышевского района, г. Донецк, Донецкая Народная Республика В статье проанализирована система реализации наказаний в уголовно-правовых нормах и разработаны предложения по повышению ее эффективности в условиях Донецкой Народной Республики. Ключевые слова: уголовное право, законодательная система, Донецкая Народная Республика, наказание, уголовно-правовые нормы. 175
https://openalex.org/W4230166612	http://www.business-inform.net/export_pdf/business-inform-2021-5_0-pages-402_408.pdf	Ukrainian	null	The Methodological Principles of of Logistical Provision Management for the Development of Innovative Cooperation in a Circular Economy	Bìznes ìnform/Bìznes ìnform	2,021	cc-by-sa	5,009	ВИСНОВКИ Кожна компанія, яка хоче сформувати свій бренд, повинна володіти технологією створення, формування бренду. Розробка такої технології базу ється на пізнанні існуючих концепцій, аналізі брен дингу компаній – лідерів світового ринку. Діючими й ефективними концепціями, що пропонують унікальні технології позиціонування, формування особливос тей і конкурентоздатності продуктів на ринку, висту пають концепції «Капітал бренду», «Колесо бренду», «Теорії іміджу», «4-D Брендинг», «Максимізація по тенційних ринків», «Value-Based-Marketing». Аналіз світових брендів компаній-лідерів підтверджує при кладну значущість розроблених концептуальних по ложень і визначених ключових характеристик для конкурентоспроможності їх брендів.  REFERENCES Aaker, D. A., and McLoughlin, D. Strategic Market Manage ment: Global Perspectives. New York: John Wiley and Sons, Ltd, 2010. Aaker, D. Sozdaniye silnykh brendov [Building Strong Brands]. Moscow: Izdatelskiy Dom Grebennikova, 2003. Aleksandrov, Yu. L., Rychkov, V. A., and Tereshchenko, N. N. “Analiz potrebitelskogo rynka khleba i khlebobuloch nykh izdeliy“ [Analysis of the Consumer Market for Bread and Bakery Products]. Marketing v Rossii i za ru bezhom, no. 2 (1998): 83-85. Ambler, T. Prakticheskiy marketing [Practical Marketing]. St. Petersburg: Piter, 1999. Doyl, P., and Shtern, F. Marketing-menedzhment i strategii [Marketing Management and Strategies]. St. Peters burg: Piter, 2007.f МЕНЕДЖМЕНТ І МАРКЕТИНГ КА МЕНЕДЖМЕНТ І МАРКЕТИНГ ЕКОНОМІКА МЕНЕДЖМЕНТ І МАРКЕТИНГ Ellvud, Ya. 100 priyemov effektivnogo brendinga [100 Tricks for Effective Branding]. St. Petersburg: Piter, 2002. Науковий керівник – Отенко І. П., доктор економічних наук, Науковий керівник – Отенко І. П., доктор економічних наук, Науковий керівник – Отенко І. П., доктор економічних наук, професор, завідувачка кафедри міжнародних економічних відносин Харківського національного економічного університету імені Семена Кузнеця професор, завідувачка кафедри міжнародних економічних відносин Харківського національного економічного університету інших країн, наприклад Toyota (Японія), Mercedes- Benz (Німеччина), Samsung (Південна Корея) тощо. Це є підтвердженням того факту, що ці країни є «пе реможцями» в процесі перетворень, які відбуваються внаслідок конвергенції комп’ютерних і телекомуніка ційних технологій і поширюють свій вплив на розви ток таких галузей, як охорона здоров’я, природни чі науки, геноміка. А також ті компанії, що надають можливість безготівкового розрахунку. Тому серед концепцій брендингу, що мають прикладний харак тер і значний вплив на його розвиток, слід виділити такі, що мають глобальне та міжнародне значення в економічному розвитку – «глобальний бренд», «між народний бренд». У такому напрямку будуть реалізо вуватися подальші авторські дослідження. інших країн, наприклад Toyota (Японія), Mercedes- Benz (Німеччина), Samsung (Південна Корея) тощо. Це є підтвердженням того факту, що ці країни є «пе реможцями» в процесі перетворень, які відбуваються внаслідок конвергенції комп’ютерних і телекомуніка ційних технологій і поширюють свій вплив на розви ток таких галузей, як охорона здоров’я, природни чі науки, геноміка. А також ті компанії, що надають можливість безготівкового розрахунку. Тому серед концепцій брендингу, що мають прикладний харак тер і значний вплив на його розвиток, слід виділити такі, що мають глобальне та міжнародне значення в економічному розвитку – «глобальний бренд», «між народний бренд». У такому напрямку будуть реалізо вуватися подальші авторські дослідження. 10. Траут Дж., Райс Эл. Позиционирование. Битва за уз наваемость. СПб. : Питер, 2017. 256 с. 11. Ламбен Ж.-Ж. Стратегический маркетинг. Европей ская перспектива / пер. с франц. Б. И. Лифляндчик, В. Л. Дунаевского, С. А. Бурьяна. СПб. : Наука, 1996. 610 с. 12. Дойль П., Штерн Ф. Маркетинг-менеджмент и стра тегии. 4-е изд. / пер. с англ. С. Жильцова. СПб. : Пи тер, 2007. 544 с. 13. Котлер Ф. Маркетинг в третьем тысячелетии. Как создать, завоевать и удержать рынок. СПб. : АСТ, 2001. 272 с. 14. Aaker D. A., McLoughlin D. Strategic Market Manage� ment: Global Perspectives. New York : John Wiley and Sons, Ltd, 2010. 354 p. ЛІТЕРАТУРА Ged, T. 4D Brending: vzlamyvaya korporativnyy kod setevoy ekonomiki [4D Branding: Breaking the Corporate Code of the Networked Economy]. Moscow: Mann, Ivanov i Ferber, Stokgolmskaya shkola ekonomiki v Sankt-Pe terburge, 2005. 1. Александров Ю. Л., Рычков В. А., Терещенко Н. Н. Анализ потребительского рынка хлеба и хлебобу лочных изделий. Маркетинг в России и за рубежом. 1998. № 2. С. 83–85. 2. Котлер Ф. Основы маркетинга: краткий курс. М. : Диалектика, 2016. 496 с. Kapferer, J.-N., and Bastien, V. The Luxury Strategy: Break the Rules of Marketing to Build Luxury Brands. London ; Philadelphia: Kogan Page Ltd., 2009. 3. Евстратов А. В., Абрамов С. А. Стратегический мар кетинг : учебное пособие. Волгоград : Волгоград ский государственный технический университет, 2010. 85 с. Kotler, F. Marketing v tretem tysyacheletii. Kak sozdat, za voyevat i uderzhat rynok [Marketing in the Third Millen nium. How to Create, Conquer and Hold the Market]. St. Petersburg: AST, 2001. 4. Аакер Д. Создание сильных брендов. М. : Издатель ский Дом Гребенникова, 2003. 440 с. Kotler, F. Osnovy marketinga: kratkiy kurs [Marketing Essen tials]. Moscow: Dialektika, 2016. 5. Огилви Д. О рекламе. М. : Манн, Иванов и Фербер, 2020. 304 с. Lamben, Zh.-Zh. Strategicheskiy marketing. Yevropeyskaya perspektiva [Strategic Marketing. European Perspec tive]. St. Petersburg: Nauka, 1996. ЕКОНОМІКА 6. Kapferer J.-N., Bastien V. The Luxury Strategy: Break the Rules of Marketing to Build Luxury Brands. London ; Philadelphia : Kogan Page Ltd. 2009. 336 p. Ogilvi, D. O reklame [About Advertising]. Moscow: Mann, Ivanov i Ferber, 2020. 7. Гэд Т. 4D Брэндинг: взламывая корпоративный код сетевой экономики. М. : Манн, Иванов и Фер бер, Стокгольмская школа экономики в Санкт- Петербурге, 2005. 232 с. Traut, Dzh., and Rays, El. Pozitsionirovaniye. Bitva za uzna vayemost [Positioning. The Battle for Recognition]. St. Petersburg: Piter, 2017. 8. Эллвуд Я. 100 приемов эффективного брендинга. СПб. : Питер, 2002. 368 с. Yevstratov, A. V., and Abramov, S. A. Strategicheskiy market ing [Strategic Marketing]. Volgograd: Volgogradskiy gosudarstvennyy tekhnicheskiy universitet, 2010. 9. Амблер Т. Практический маркетинг. СПб. : Питер, 1999. 400 с. 401 Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Актуальною проблемою сучасної науки є формування наукової методології управління логістичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки, оскільки прагнення науковців додати власний внесок у науку пояснюється бажанням її розвивати й удосконалювати з метою розширення кола як теоретичних, так і практичних питань, які можуть бути вирішені із застосуванням сучасної тер мінології. Мета статті полягає у формуванні методологічних положень управління логістичним забезпеченням розвитку інноваційного співро бітництва в умовах циркулярної економіки. Проведено дослідження формування сутності методології, проаналізовано етапи розвитку методів і засобів наукового пізнання. Систематизовано та проаналізовано визначення дефініції «методологія», що дозволило визначити основні засади формування методологічних положень управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки, а саме: підходи, принципи, категорії, функції, інструментарій, парадигму. Виокремлено та надано характеристику категоріальному, концептуальному та інструментально-модельному рівням методології управління логістичним забезпеченням розвитку інноваційного співро бітництва в умовах циркулярної економіки. Розглянуто та визначено зв’язок емпіричних і теоретичних методів наукового пізнання. Доведено, що як концептуальний модуль методології управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки варто розглядати метатеоретичну єдність, що базується на онтологічних і гносеологічних ідеалізаціях, – методологічну парадигму. Ключові слова: методологія, методи, управління, логістичне забезпечення, розвиток, циркулярна економіка. DOI: https://doi.org/10.32983/2222-4459-2021-5-402-408 Рис : 4 Табл : 1 Бібл : 17 МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ Шкуренко Ольга Володимирівна – кандидат економічних наук, доцент, доцент кафедри бізнес-логістики та транспортних технологій, Держав ний університет інфраструктури та технологій (вул. Кирилівська, 9, Київ, 04071, Україна) E-mail: dondyy@ukr.net 4 ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖ Shkurenko O. V. The Methodological Principles of Logistical Provision Management for the Development of Innovative Cooperation in a Circular Economy An urgent problem of modern science is the formation of a scientific methodology for managing logistical provision for the development of innovative coopera tion in circular economy, since the desire of scientists to add their own contribution to science is explained by the desire to develop and improve it in order to expand the range of both theoretical and practical issues that can be solved with the use of modern terminology. The article is aimed at forming methodologi cal principles for managing logistical provision for the development of innovative cooperation in a circular economy. The research of formation of the essence of methodology is carried out, the stages of development of methods and means of scientific knowledge are analyzed. БІЗНЕСІНФОРМ № 5 ’2021 www.business-inform.net УДК 167.7:330.34 JEL: B41; M21 УДК 167.7:330.34 JEL: B41; M21 УДК 167.7:330.34 JEL: B41; M21 УДК 167.7:330.34 JEL: B41; M21 МЕТОДОЛОГІЧНІ ПОЛОЖЕННЯ УПРАВЛІННЯ ЛОГІСТИЧНИМ ЗАБЕЗПЕЧЕННЯМ РОЗВИТКУ ІННОВАЦІЙНОГО СПІВРОБІТНИЦТВА В УМОВАХ ЦИРКУЛЯРНОЇ ЕКОНОМІКИ 2021 ШКУРЕНКО О. В. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки The definition of «methodology» is systematized and analyzed, which allows to determine the basics of formation of methodological principles for the management of logistical provision for the development of innovative cooperation in the conditions of circular economy, namely: approaches, principles, categories, functions, instrumentarium, paradigm. The categorical, conceptual and instrumental-model levels of the methodology for managing logistical provision for the development of innova tive cooperation in the conditions of circular economy have been distinguished and characterized. The connection between the empirical and the theoretical methods of scientific cognition is considered and defined. It is proved that as a conceptual module of the methodology for managing logistical support for the development of innovative cooperation in a circular economy should be considered a metatheoretic unity based on the ontological and gnoseological idealiza tions – the methodological paradigm.i Keywords: methodology, methods, management, logistical provision, development, circular economy. ig.: 4. Tabl.: 1. Bibl.: 17. Shkurenko Olga V. – PhD (Economics), Associate Professor, Associate Professor, Department of Business-Logistics and Transport Technologies, State University of Infrastructure and Technologies (9 Kyrylivska Str., Kyiv, 04071, Ukraine) E-mail: dondyy@ukr.net ORCID: http://orcid.org/0000-0002-0460-4800 Researcher ID: https://publons.com/researcher/1742416/olga-shkurenko/ Scopus Author ID: https://www scopus com/authid/detail uri?authorId=57221967879 ЕКОНОМІКА філософсько-світоглядні, методологічний аналіз перетворює їх у засоби усвідомленого вибору та на укового пошуку. Актуальною проблемою сучасної науки є формування наукової методології управління логістичним забезпеченням розвитку інноваційного П П рояснити способи поєднання знання та ді яльності, будову, організацію, способи одер жання та обґрунтування знань можна за до помогою методологічного аналізу. Виявляючи умови та передумови пізнавальної діяльності, у тому числі БІЗНЕСІНФОРМ № 5 ’2021 www.business-inform.net 402 БІЗНЕСІНФОРМ № 5 2021 www.business-inform.net Метою наукового пізнання є отримання об’єк тивного та істинного знання про досліджуваний об’єкт, явище або процес з виключенням упередже ного та тенденційного ставлення до них. Основне завдання наукового пізнання полягає у виявленні об’єктивних законів навколишньої дійсності – при родних, соціальних, економічних тощо, а також за конів пізнання та мислення. Поряд із цим основною функцією наукового пізнання або дослідження є об слуговування потреб практики. Наука як форма піз нання переважно орієнтована на втілення її досяг нень у практику, використання зміни навколишньої дійсності й управління реальними процесами. співробітництва в умовах циркулярної економіки, оскільки прагнення науковців додати власний внесок у науку пояснюється бажанням її розвивати й удоско налювати з метою розширення кола як теоретичних, так і практичних питань, які можуть бути вирішені із застосуванням сучасної термінології. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Методологія має багатоаспектний характер, її складає філософія, яка в поєднанні з діалектикою, гносеологією та логікою виконує в пізнавальному процесі важливе завдання суб’єктивного осмислення об’єктивного розвитку [1]. Науковці [5–13] досліджували різноманітні ас пекти методології, зокрема методологію наукових досліджень, оцінки процесів, явищ і систем. У У гносеологічному плані наукове пізнання або дослідження виступає як складний, супереч ливий процес відтворення знань, що утворю ють струнку систему ідеальних форм і логічних обра зів, закріплених, насамперед, у мові – природній або штучній. Наукове пізнання безперервно відтворює елементи на власній основі, тобто формує відповід но до власних норм і принципів. Такий процес безпе рервного оновлення наукою свого концептуального арсеналу є проявом процесів її розвитку. Однак у наукових працях науковців не приділе но уваги методологічним засадам управління логіс тичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки в орга нізаційній єдності методології наукового пізнання. Метою статті є формування методологічних положень управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки. МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ П р в ц Науковому пізнанню притаманні такі основні характерні ознаки, що наведені на рис. 1. П роцес пізнання, як основа будь-якого науко вого дослідження, є складним і вимагає кон цептуального підходу на основі певної мето дології, застосування певних методів. ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В Е Безперечність наукової та практичної значу щості наукового дослідження забезпечується в тому випадку, коли воно ґрунтується на досконалій мето дології. Наукове пізнання в сучасних умовах є основною формою пізнання реальної дійсності або навколиш нього середовища. Процес наукового пізнання відбу вається не стихійно, а цілеспрямовано і по суті є нау ковим дослідженням, яке має певну природу, струк туру й особливості. Питання, пов’язані з методологією (від грец. methodos – шлях дослідження, спосіб пізнання; logos – слово, поняття, учення) наукового дослідження, по чали вивчати ще за часів античності. Першочерго ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ Д ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ Д О З Н А К ОЗНАКИ Предметність – реалізується через можливість передбачення майбутнього стану предмету тільки за умови глибокого та всебічного вивчення минулого та сьогоденного станів Об’єктивність – прагнення до об'єктивності змушує дослідника критично ставитися до себе і до дійсності, виключати відсталість і догматизм Практичне використання отриманих знань – передбачає розробку рекомендацій за результатами дослідження, що виражаються та закріплюються у практичному досвіді 1 2 3 4 5 Системність – сукупність знань, зведених у певний порядок на основі однорідних принципів, які об'єднують знання в системи Істинність – вища цінність наукового пізнання, яка осягається переважно раціональними методами, хоча і за участю живого споглядання Рис. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки 1. Основні характерні ознаки наукового пізнання Джерело: авторська розробка. ОЗНАКИ Предметність – реалізується через можливість передбачення майбутнього стану предмету тільки за умови глибокого та всебічного вивчення минулого та сьогоденного станів Об’єктивність – прагнення до об'єктивності змушує дослідника критично ставитися до себе і до дійсності, виключати відсталість і догматизм Системність – сукупність знань, зведених у певний порядок на основі однорідних принципів, які об'єднують знання в системи Істинність – вища цінність наукового пізнання, яка осягається переважно раціональними методами, хоча і за участю живого споглядання Практичне використання отриманих знань – передбачає розробку рекомендацій за результатами дослідження, що виражаються та закріплюються у практичному досвіді Рис. 1. Основні характерні ознаки наукового пізнання 403 вого значення стародавні філософи – Платон, Со крат і Арістотель – надавали діалектичній сутності мислення. Арістотель вивчав принципи побудови суджень, питання термінології, правила умовиводів тощо. У XVI–XVII ст. Ф. Бекон та Г. Галілей довели, що наукове пізнання ґрунтується на планомірному експерименті. Раціоналістичний метод дослідження був запропонований Р. Декартом, який закликав «… достовірними вважати тільки ті положення, що чітко й виразно усвідомлені» [2]. вого значення стародавні філософи – Платон, Со крат і Арістотель – надавали діалектичній сутності мислення. Арістотель вивчав принципи побудови суджень, питання термінології, правила умовиводів тощо. У XVI–XVII ст. Ф. Бекон та Г. Галілей довели, що наукове пізнання ґрунтується на планомірному експерименті. Раціоналістичний метод дослідження був запропонований Р. Декартом, який закликав «… достовірними вважати тільки ті положення, що чітко й виразно усвідомлені» [2]. метод і визначається як система принципів і способів організації та створення теоретичної та практичної діяльності людини, а також є вченням про цю систе му [17]. Методологія наукового пізнання представляє собою філософське вчення, що розглядає проблеми наукового пізнання, його логічні принципи, прийоми, методи і форми (рис. 2). Завдяки методології реалізується функція стра тегії пізнання, що дозволяє знайти перспективні на прями пізнання та обґрунтувати їх, при цьому запо бігти використання невірних методів пізнання. Г. Гегель – батько діалектичного матеріалізму – стверджував, що вміння знаходження істини можли ве через раціональну дискусію між співрозмовника ми, що мають різні точки зору [3]. Одним зі шляхів до мети діалектичного матеріалізму є сократів метод, що полягає у вилученні зі списків кандидатів на іс тину (reductio ad absurdum) ту гіпотезу, що веде до протиріччя. Інший діалектичний метод подолання незгоди – протиставити два протилежні тверджен ня, тезу та антитезу, і зняти суперечність через тре тє твердження, що було б синтезом тези та антитези. Г. Гегель утвердив діалектику як загальний метод піз нання та духовної діяльності. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Він дослідив протиріч чя розвитку буття та мислення. М М етодологія наукового пізнання як фунда мент управління логістичним забезпечен ням розвитку інноваційного співробітни цтва в умовах циркулярної економіки є сукупністю наукових теорій, ідей, принципів, методів і моделей пізнання логістичного забезпечення розвитку ін новаційного співробітництва з виділенням катего ріального (визначення структурних компонентів), концептуального (сукупність теоретичних концеп цій), інструментально-модельного (сукупність спо собів, методів і моделей) рівнів. Графічно методоло гія управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки наведена на рис. 3. МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ тє Г. Г на чя сво В скл на кіл бо тув нау гія Дж Отже, кожна конкретна наука використовує свою методологію та спеціальні методи [4]. ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В Е В о т ч В одночас поняття методології не має єдиного трактування в літературних джерелах, у тому числі й у довідниках, що можна пояснити його складністю та багатозначністю (табл. 1). Більше того, навіть у одних і тих самих джерелах часто дається де кілька визначень методології, які не завжди між со бою узгоджені. Неоднозначність у підходах до трак тування методології наукових досліджень ускладнює науковий пошук та оцінювання його результатів. Перший рівень методології управління логіс тичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки перед бачає визначення категорій «управління», «логістич не забезпечення», «розвиток», «інноваційне співро бітництво» та «циркулярна економіка». На другому рівні методології управління логіс тичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки уза гальнюються теоретичні концепції. У Великому економічному словнику методоло гія в перекладі з грецької трактується як вчення про ЕКОНОМІКА МЕТОДОЛОГІЯ МЕТОДОЛОГІЯ НАУКОВОГО ПІЗНАННЯ Теоретична Сукупність теоретичних принципів, логічних прийомів і конкретних способів дослідження Вчення про методи, правила мислення при створенні теорії науки Система теорій, принципів, законів і категорій, що відображають процес пізнання Практична Змістовна Формальна Рис. 2. Сутнісна характеристика методології наукового пізнання Джерело: авторська розробка. МЕТОДОЛОГІЯ НАУКОВОГО ПІЗНАННЯ Сукупність теоретичних принципів логічних прийомів і конкретних способів дослідження ЕКОНОМІКА Формальна Рис. 2. Сутнісна характеристика методології наукового пізнання Джерело: авторська розробка. Рис. 2. Сутнісна характеристика методології наукового пізнання Джерело: авторська розробка. БІЗНЕСІНФОРМ № 5 ’2021 БІЗНЕСІНФОРМ № 5 ’2021 404 www.business-inform.net Таблиця 1 Систематизація визначення дефініції «методологія» ц д ф ц д Автор(-и), джерело Визначення методології Шейко В., Кушнаренко Н. [5] Концептуальний виклад мети, змісту, методів дослідження, які забезпечують отримання максимально об’єктивної, точної, систематизованої інформації про процеси та явища Машков А. [6] Теоретична основа деяких спеціальних, доволі приватних прийомів та засобів Радько Т. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки [7] Спеціальний розділ теорії пізнання, який являє собою логічне обґрунтування історич ного шляху, засобів, прийомів науково-дослідницької діяльності в різних сферах Вишновецька С. [8] З одного боку, – це сукупність методологічних принципів, прийомів, способів, засобів, методів і методик наукового пізнання, а з іншого – наука про методи Комаров С. [9] Теоретичне обґрунтування використовуваних наукою методів пізнання навколишньої дійсності, вчення про наукові методи пізнання. Метод учений розглядає як внутрішньо ор ганізовану систему загальнофілософських, загальнонаукових і приватно-наукових принци пів, прийомів, засобів пізнання державно-правової надбудови. Відповідно, методологія – це застосування обумовленої діалектикою сукупності певних теоретичних принципів, ло гічних прийомів і спеціальних способів дослідження державно-правових явищ Сурмін Ю. [10] Вчення про метод діяльності, який охоплює принципи, методи діяльності та знання, відображає їх і складається з методології пізнання, методології практичної цілісності та методології оцінювання (аксіометодології) Костицький М. [11] Вчення, цілісна теорія з поняттями та категоріями, світогляд, крізь призму яких і здій снюють пізнання за допомогою методів відповідно до технології їх застосування – методики. Тобто є метод, методика та методологія. Вони, з одного боку, знаходяться в ієрархічному зв’язку, з іншого – можуть існувати доволі автономно. Наприклад, метод спостереження застосовують у більшості наук, попри те, що він найбільше пов’язаний з позитивістською методологією, емпіризмом. Тобто, щоб спостерігати за природою, соціумом, процесами та явищами, необов’язково позиціонувати себе як позитивіста чи емпіриста» Поппер К. [12] Виділяє притаманність сучасній науці не лише логічної форми, а й, крім цього, специфіч ної методології Конверський А. [14] Тип раціонально-рефлексивної свідомості, що спрямований на вивчення, удосконален ня та конструювання методів і має два основні значення: по-перше, це система певних правил, принципів і операцій, що застосовують у певній сфері діяльності (науці, політи ці, мистецтві тощо); по-друге, – учення про цю систему, загальна теорія методу Джерело: систематизовано автором. Визначення методології МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ КОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В Джерело: систематизовано автором. На третьому рівні обирається сукупність прин ципів, методів і моделей. тичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки будемо розглядати метатеоретичну єдність, що базується на онтологічних і гносеологічних ідеалізаціях наукового товариства, – методологічну парадигму. Парадиг мальності як ознаці зрілості наукового дослідження притаманні певні функції – забезпечення цілісності, однорідності, функціональності науки, спрямування її руху [13; 16]. Методологічна парадигма, запропоно вана Т. С. Куном, є апогеєм критичного раціоналізму і передбачала трансформацію моделі розвитку науки – від кумулятивістської моделі з поступально-еволю ційною траєкторією змін до революційної моделі [15]. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Методологічна парадигма використовується в багатьох науках і розглядається у вузькому змісті як зразок, а в широкому – як «світоглядний каркас». К. Бейлі розглядає парадигму як «…деякий пер спективний фрейм референції (а perspective frame of reference) для аналізу об’єкта, який складається із су купності концептів і припущень» [16]. тичним забезпеченням розвитку інноваційного спів робітництва в умовах циркулярної економіки будемо розглядати метатеоретичну єдність, що базується на онтологічних і гносеологічних ідеалізаціях наукового товариства, – методологічну парадигму. Парадиг мальності як ознаці зрілості наукового дослідження притаманні певні функції – забезпечення цілісності, однорідності, функціональності науки, спрямування її руху [13; 16]. Методологічна парадигма, запропоно вана Т. С. Куном, є апогеєм критичного раціоналізму і передбачала трансформацію моделі розвитку науки – від кумулятивістської моделі з поступально-еволю ційною траєкторією змін до революційної моделі [15]. Методологічна парадигма використовується в багатьох науках і розглядається у вузькому змісті як зразок, а в широкому – як «світоглядний каркас». К. Бейлі розглядає парадигму як «…деякий пер спективний фрейм референції (а perspective frame of reference) для аналізу об’єкта, який складається із су купності концептів і припущень» [16]. Я Я к було зазначено вище, існують різні рівні ме тодологічного аналізу. Конкретно-наукова ме тодологія зі своїми методиками має справу з технічними прийомами, приписами, нормативами, формулює принципи та методи конкретно-наукової діяльності, описує й обґрунтовує їх. На другому рівні реалізується загальнонаукова методологія як вчення про принципи, методи та форми знання, що функ ціонують у багатьох науках, які відповідають пред мету й об’єкту дослідження. Це методи емпіричного дослідження, а саме: спостереження, вимірювання, експеримент; емпірико-теоретичні методи – аналіз, синтез, індукція, аналогія, дедукція тощо, а також такі форми знання, як поняття і закони, гіпотези і теорії, що наведено на рис. 4. Погоджуємося з думкою Т. Шкуренко О. В. Методологічні положення управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Куна [15] і як кон цептуальний модуль методології управління логіс 405 БІЗНЕСІНФОРМ № 5 ’2021 www.business-inform.net www.business-inform.net www.business-inform.net 406 ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ БІЗНЕСІНФОРМ №5 ’202 Елементи Технології Логічна структура Суб’єкт Об’єкт Предмет МЕТОДОЛОГІЯ УПРАВЛІННЯ ЛОГІСТИЧНИМ ЗАБЕЗПЕЧЕННЯМ РОЗВИТКУ ІННОВАЦІЙНОГО СПІВРОБІТНИЦТВА В УМОВАХ ЦИРКУЛЯРНОЇ ЕКОНОМІКИ П Р И Н Ц И П И ФУНКЦІЇ К А Т Е Г О Р І П І Д Х О Д И ІНСТРУМЕНТАРІЙ Планування Аналізування Розроблення Моніторинг Координація Регулювання Управління Розвиток Інноваційне співробітництво Циркулярна економіка Системний Процесний Ситуаційний Адаптивний Комплексний Предметний Об’єктивність Конкретність Розвиток Системність Детермінізм Методологічний апарат Методи Засоби Форми Соціального партнерства Постіндустріального суспільства Конвергенції Економіки розвитку Стратегічного управління Зовнішні умови та характеристики управління логістичним забезпеченням розвитку інноваційних співробітництва в умовах циркулярної економіки економічні правові політичні соціально-психологічні Єдність теорії та практики Логістичне забезпечення ПАРАДИГМА (філософія науки) Дисциплінарна Мультидисциплінарна Міждисциплінарна Трансдисциплінарна Методи Моделі Концепції Теорії Генерація наукових гіпотез Вирішення наукового протиріччя Протиріччя розв’язання наукової проблематики Рис. 3. Методологія управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Д б ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ Елементи Технології Логічна структура Суб’єкт Об’єкт Предмет МЕТОДОЛОГІЯ УПРАВЛІННЯ ЛОГІСТИЧНИМ ЗАБЕЗПЕЧЕННЯМ РОЗВИТКУ ІННОВАЦІЙНОГО СПІВРОБІТНИЦТВА В УМОВАХ ЦИРКУЛЯРНОЇ ЕКОНОМІКИ П Р И Н Ц И П И ФУНКЦІЇ К А Т Е Г О Р І П І Д Х О Д И ІНСТРУМЕНТАРІЙ Планування Аналізування Розроблення Моніторинг Координація Регулювання Управління Розвиток Інноваційне співробітництво Циркулярна економіка Системний Процесний Ситуаційний Адаптивний Комплексний Предметний Об’єктивність Конкретність Розвиток Системність Детермінізм Методологічний апарат Методи Засоби Форми Соціального партнерства Постіндустріального суспільства Конвергенції Економіки розвитку Стратегічного управління Зовнішні умови та характеристики управління логістичним забезпеченням розвитку інноваційних співробітництва в умовах циркулярної економіки економічні правові політичні соціально-психологічні Єдність теорії та практики Логістичне забезпечення ПАРАДИГМА (філософія науки) Дисциплінарна Мультидисциплінарна Міждисциплінарна Трансдисциплінарна Методи Моделі Концепції Теорії Генерація наукових гіпотез Вирішення наукового протиріччя Протиріччя розв’язання наукової проблематики Рис. 3. Методологія управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Джерело: авторська розробка Рис. 3. Методологія управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки Джерело: авторська розробка Рис. 3. Методологія управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки вторська розробка ЕКОНОМІКА Джерело: авторська розробка. 406 БІЗНЕСІНФОРМ № 5 ’2021 406 www.business-inform.net Метатеоретичні методи Діалектичний метод Метод системного аналізу Теоретичні методи Абстрагування Формалізація Емпіричні методи Спостереження Порівняння Підрахунок Вимірювання Емпірико-теоретичні методи Ідеалізація Узагальнення Експеримент Аналіз і синтез Інформетрія Наукометрія Конкретно-наукова методологія Спеціальні методи Гіпотетичний метод Історичний метод Моделювання Індукція і дедукція Морфологічний аналіз Синектика Метод найменших квадратів Екстра- та інтерполяція МЕТОДИ НАУКОВОГО ПІЗНАННЯ Рис. 4. Методи наукового пізнання Джерело: авторська розробка. Метатеоретичні методи Діалектичний метод Метод системного аналізу Теоретичні методи Абстрагування Формалізація Емпіричні методи Спостереження Порівняння Підрахунок Вимірювання Емпірико-теоретичні методи Ідеалізація Узагальнення Експеримент Аналіз і синтез Інформетрія Наукометрія Конкретно-наукова методологія Спеціальні методи Гіпотетичний метод Історичний метод Моделювання Індукція і дедукція Морфологічний аналіз Синектика Метод найменших квадратів Екстра- та інтерполяція МЕТОДИ НАУКОВОГО ПІЗНАННЯ Рис. 4. Методи наукового пізнання Джерело: авторська розробка. МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ Рис. 4. Методи наукового пізнання Джерело: авторська розробка. 2. Декарт Р. Сочинения : в 2 т. М. : Мысль, 1989. Т. 1. 654 с. Складовими методологічної парадигми, що слу жать теоретичним підґрунтям для управління логіс тичним забезпеченням розвитку інноваційного спів робітництва, є теорії та концепції. На підставі кон цепцій і теорій, враховуючи особливості управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки, здійснюється вибір методів і моделей дослідження, ідентифікація отриманих результатів з розробкою практичних рекомендацій. КОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В 3. Pinto R. C. Argument, Inference and Dialectic: Collect� ed Papers on Informal Logic / an Introduction by Hans V. Hansen. Springer, 2001. 168 p. 4. Ковальчук В. В., Моїсєєв Л. М. Основи наукових до сліджень : навч. посіб. 2-ге вид., перероб. і допов. Київ : Професіонал, 2004. 208 с. 5. Шейко В. М., Кушнаренко Н. М. Організація та ме тодика науково-дослідницької діяльності : підруч ник. 2-ге вид., перероб. і доп. К. : Знання-Прес, 2002. 295 с. 6. Машков А. Проблеми теорії держави і права. Осно ви : курс лекцій. Київ : К. І. С., 2008. 470 с. ВИСНОВКИ У результаті проведеного дослідження обґрун товано формування сутності методології, проана лізовано етапи розвитку методів і засобів науково го пізнання. Визначено основні засади формування методології управління логістичним забезпеченням розвитку інноваційного співробітництва в умовах циркулярної економіки, а саме: підходи; принципи; категорії; функції, інструменти; парадигму (філософію науки). Розглянуто та визначено зв’язок емпіричних і теоретичних методів наукового пізнання. Доведено, що як концептуальний модуль методології управлін ня логістичним забезпеченням розвитку інноваційно го співробітництва в умовах циркулярної економіки варто розглядати метатеоретичну єдність.  7. Радько Т. Н. Теория государства и права : учебник 2-е изд. М. : Проспект, 2009. 752 с. 8. Вишновецька С. В. Методологія науки трудового права : монографія. Київ : Ніка-Центр, 2014. 332 с. 9. Комаров С. А. Общая теория государства и права : учебник 7-е изд. СПб. : Питер, 2008. 512 с. 10. Сурмін Ю. П. Майстерня вченого : підручник для на уковця. Київ : Консорціум з удосконалення менедж мент-освіти в Україні, 2006. 302 с. 11. Костицький М. Деякі питання методології юри дичної науки. Науковий вісник Національної академії внутрішніх справ. 2013. № 1. С. 3–11. URL: http://elar.naiau.kiev.ua/jspui/bitstream/ 123456789/3056/1/2013_1_kostytskyi.pdf ЕКОНОМІКА 12. Поппер К. Логика научного исследования. М. : Рес публика, 2004. 608 с. 1. Тарахонич Т. І. Методологічний інструментарій на укового пізнання. Правова держава. 2017. Вип. 28. С. 118–126. ЛІТЕРАТУРА 13. Савостьянова М. В. Структура та функціональний ді апазон парадигмальної науки. Філософія науки: тра диції та інновації. 2010. № 1 (2). С. 18–26. URL: https:// repository.sspu.edu.ua/bitstream/123456789/757/1/ Struktura%20ta%20funktsionalnyi%20diapazon.pdf 1. Тарахонич Т. І. Методологічний інструментарій на укового пізнання. Правова держава. 2017. Вип. 28. С. 118–126. 407 Mashkov, A. Problemy teorii derzhavy i prava. Osnovy [Prob lems of the Theory of State and Law. Foundations]. Kyiv: K. I. S., 2008. 14. Основи методології та організації наукових дослі джень : навч. посіб. / за ред. А. Є. Конверського. Київ : ЦУЛ, 2010. 352 с.i 15. Kuhn Th. S. The Structure of Scientific Revolutions. 2nd Ed., Enlarged. The University of Chicagoб 1970. 222 р. URL: https://www.lri.fr/~mbl/Stanford/CS477/papers/ Kuhn-SSR-2ndEd.pdf Osnovy metodolohii ta orhanizatsii naukovykh doslidzhen [Fundamentals of Methodology and Organization of Scientific Research]. Kyiv: TsUL, 2010. Scientific Research]. Kyiv: TsUL, 2010. Pinto, R. C. Argument, Inference and Dialectic: Collected Pa pers on Informal Logic. Springer, 2001. 16. Bailey K. D. Methods of Social Research. 4th ed. New York : Free Press, 2007. 592 p. Popper, K. Logika nauchnogo issledovaniya [The Logic of Scientific Research]. Moscow: Respublika, 2004. 17. Большой экономический словарь / под ред. А. Н. Аз рилияна. 7-е изд. М. : Ин-т новой экономики, 2008. 1472 с. Radko, T. N. Teoriya gosudarstva i prava [Theory of State and Law]. Moscow: Prospekt, 2009. REFERENCES Savostianova, M. V. “Struktura ta funktsionalnyi diapazon paradyhmalnoi nauky“ [Structure and Functional Range of Paradigmal Science]. Filosofiia nauky: trady tsii ta innovatsii, no. 1 (2) (2010): 18-26. https://reposi tory.sspu.edu.ua/bitstream/123456789/757/1/Struk tura%20ta%20funktsionalnyi%20diapazon.pdf Bailey, K. D. Methods of Social Research. New York: Free Press, 2007. Bolshoy ekonomicheskiy slovar [Big Dictionary of Econom ics]. Moscow: In-t novoy ekonomiki, 2008. Dekart, R. Sochineniya [Essays], vol. 1. Moscow: Mysl, 1989. Sheiko, V. M., and Kushnarenko, N. M. Orhanizatsiia ta metodyka naukovo-doslidnytskoi diialnosti [Organiza tion and Methods of Research Activities]. Kyiv: Znan nia-Pres, 2002. Komarov, S. A. Obshchaya teoriya gosudarstva i prava [Gene ral Theory of State and Law]. St. Petersburg: Piter, 2008. Kostytskyi, M. “Deiaki pytannia metodolohii yurydychnoi nauky“ [Some Questions of Methodology of Legal Science]. Naukovyi visnyk Natsionalnoi akademii vnu trishnikh sprav, no. 1 (2013): 3-11. http://elar.naiau. kiev.ua/jspui/bitstream/123456789/3056/1/2013_1_ kostytskyi.pdf МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ 408 ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В ЕКОНОМІЦІ Surmin, Yu. P. Maisternia vchenoho [Scientist's Workshop]. Kyiv: Konsortsium z udoskonalennia menedzhment- osvity v Ukraini, 2006. Tarakhonych, T. I. “Metodolohichnyi instrumentarii nauko voho piznannia“ [Methodological Tools of Scientific Cognition]. Pravova derzhava, no. 28 (2017): 118-126. ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛІДЖЕНЬ В Е Kovalchuk, V. V., and Moisieiev, L. M. Osnovy naukovykh doslidzhen [Basics of the Scientific Research]. Kyiv: Pro fesional, 2004.i Vyshnovetska, S. V. Metodolohiia nauky trudovoho prava [Methodology of Labor Law Science]. Kyiv: Nika-Tsentr, 2014. Kuhn, Th. S. “The Structure of Scientific Revolutions“. The University of Chicago. 1970. https://www.lri.fr/~mbl/ Stanford/CS477/papers/Kuhn-SSR-2ndEd.pdf ЕКОНОМІКА МЕТОДОЛОГІЯ НАУКОВИХ ДОСЛ БІЗНЕСІНФОРМ № 5 ’2021 408
https://openalex.org/W2810192505	https://www.scielo.br/j/aabc/a/8wVQQNCHBS9MDzyXVrd4Hnt/?lang=en&format=pdf	English	null	Geographic and Gender Diversity in the Brazilian Academy of Sciences	Anais da Academia Brasileira de Ciências	2,018	cc-by	5,942	ABSTRACT Recent studies have shown that diversity in race and gender is important to improve competitiveness, creativity and innovation. However, many studies have shown that the universe of science is not diverse in none of these criteria. In this work, we quantify the gender and geographic diversity in the Brazilian science. We study in detail the profile of the titular members of the most prestigious academic institution of Brazil, the Brazilian Academy of Sciences (ABC). Our analysis shows that the members of the ABC concentrates 80% in the Southeast region and that the overwhelming majority is composed by men. We show that female and male researchers have roughly the same amount of publication per year but women produce more human resources. Using the Brazilian Researchers’ Fellowship databank we show that the gender and diversity both in gender and in geography decreases along the researchers’ career. Key words: diversity, gender balance, women in science, excellence in science. NATHÁLIA C. FERRARI1, RAQUEL MARTELL1, DANIELA H. OKIDO2, GRASIELE ROMANZINI2, VIVIANE MAGNAN2, MARCIA C. BARBOSA2 and CAROLINA BRITO2 1Escola de Engenharia, Universidade Federal do Rio Grande do Sul, Praça Argentina, 9, 90040-020 Porto Alegre, RS, Brazil 2Instituto de Física, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves, 9500, 91501-970 Porto Alegre, RS, Brazil Manuscript received on March 2, 2017; accepted for publication on July 24, 2017 Manuscript received on March 2, 2017; accepted for publication on July 24, 2017 Correspondence to: Carolina Brito E-mail: carolina.brito@ufrgs.br An Acad Bras Cienc (2018) 90 (2 Suppl. 1) * Contribution to the centenary of the Brazilian Academy of Sciences. INTRODUCTION only reason for the need of women at all career levels. The presence of women and minorities in leadership positions results in more diversity in ideas and strategies to solve different types of problems. By excluding female researchers, we are limiting the available pool of talented people to half of humanity. Science needs women to survive. When a group of people has to solve problems, the individuals bring to the discussion their different information and perspectives. A group of people different in race, gender and other dimensions brings unique information, experiences and perspectives to solve problems. In the particular case of companies, diversity correlates with the increase of the firm value and of the annual profit only reason for the need of women at all career levels. The presence of women and minorities in leadership positions results in more diversity in ideas and strategies to solve different types of problems. By excluding female researchers, we are limiting the available pool of talented people to half of humanity. Science needs women to survive. Women are greatly underrepresented in the senior and leadership levels of the career path both in the private and in the government sectors (Hunt et al. 2015, Joy et al. 2007, Dezso and Ross 2012, NSFoundation 2015, Slavin 2008, Philips 2014, Wallon et al. 2015). But, why should we care? Women who have a passion for any subject have the right to make a living from it and have a career with the same opportunities as the men in any field. Equal opportunities, however, are not the When a group of people has to solve problems, the individuals bring to the discussion their different information and perspectives. A group of people different in race, gender and other dimensions brings unique information, experiences and perspectives to solve problems. In the particular case of companies, diversity correlates with the increase of the firm value and of the annual profit Correspondence to: Carolina Brito E-mail: carolina.brito@ufrgs.br * Contribution to the centenary of the Brazilian Academy of Sciences. An Acad Bras Cienc (2018) 90 (2 Suppl. 1) NATHÁLIA C. FERRARI et al. 2544 (Hunt et al. 2015, Dezso and Ross 2012, Dezso et al. 2017). places, no clear policy exists. In Brazil, affirmative action for women in science is considered unnecessary, because there is a common belief that no procedures for promotion suffer from gender, race or geographic background biases. INTRODUCTION The mechanism behind this effect goes as follows. When a group of people is exposed to a challenge that needs cooperation to be solved, the best performance is obtained when the group has race and gender diversity. Being with similar others leads us to think we all hold the same information and share the same perspective, which makes us less competitive. This perspective, which stopped the all-white groups from effectively processing the information, is what hinders creativity and innovation (Guterl 2014, Philips 2014, Phillips et al. 2006).fi In order to understand if equity is present in the selection process, in this paper we analyze in detail the profile of the titular members of the most prestigious academic institution of Brazil, the Brazilian Academy of Sciences (ABC). Our main goal is to measure the level of diversity existent in the Brazilian Academy of Sciences in terms of geographic region and gender. We first look at the distribution of the titular members of ABC in the regions of Brazil and identify that about 80% of them are in the Southeast region. We compare it with the universe of researchers with a special type of scholarship (it will be explained below), which allows us to infer how this concentration in the Southeast region is created along the career. In the particular case of gender, the affirmative action movement has been working since the 60’s to increase the participation of women in all fields. Unfortunately, the increase of this participation in science is slow. Statistic evidences demonstrate that the high rank position in science belongs to males (NSFoundation 2015, Slavin 2008). Male predominance, however, does not occur at all stages of the career. In Brazil for example, about 54% of the people who complete high school are women, and more than 50% of graduated students are women.1 Women are the majority of graduated persons in most of the countries of the world (Osborn 2000). Unfortunately, they do not keep this proportion along the career. Instead, the proportion of women decays gradually along the career, defining what is called “Scissors effect”. This effect means that the percentage of men increases while the percentage of women decreases when the career progresses (Moss-Racusina et al. 2012, Raymond 2013, Slavin 2008). 1see “Censo da Educação Superior” at http://portal. inep.gov.br/web/guest/censo-da-educacao-superior INTRODUCTION We then concentrate in the gender distribution among different areas of knowledge and identify that the proportion of women is very small in all areas, being at most 25% in the areas where they are more frequent and only 1% in Engineering. We explore some indexes of scientific productivity and identify that, in most fields, women supervise more students than men but in some fields, men on average publish more than women. We discuss the possible causes for the fact that more supervised students do not convert in more publications. We end the paper with comparison of the CNPq research level of the members of the Brazilian Academy of Sciences. We find that there is no woman in the ABC without a CNPq research level but 15% of the men do not have any level. Comments at this result end the paper. Even though the lack of diversity in high positions is a universal phenomenon, the acknowledgment of the problem and taking action toward solving the issue is rather new. In some GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC 2545 • total number of papers published, p • total number of papers published, p science covered by the Academy. Table I specifies the number of male and female titular members of the ABC in all the fields. We split the table in two parts: the left part shows the total number of titular members, and the right part shows a subgroup of the researchers that have less than 35 years after finishing their PhD. The reason to split these numbers is because in Brazil, men retire after 35 years of work and women, after 30 years. We then estimate that until 35 years after obtaining her/his PhD, the researchers are still working, and we use this subgroup to analyze the productivity of the titular members in terms of type of the PQ- scholarship they have. • total number of PhD students supervised, nPhD • total number of masters students supervised, total number of masters students supervised, nm • total number of post-PhD supervised, nPD • total number of undergraduate students supervised under Scientific Initiation Program, nIC • level of the Research Scholarship Program (PQ),fi MATERIALS AND METHODS We selected from the webpage of the ABC the names of the titular members alive of each field of An Acad Bras Cienc (2018) 90 (2 Suppl. 1) GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC An Acad Bras Cienc (2018) 90 (2 Suppl. 1) An Acad Bras Cienc (2018) 90 (2 Suppl. 1) • affiliation address. The Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) is one of the main research funding agencies of Brazil. One of their policies to develop the research in Brazil is to give scholarships for undergraduate students, a program known as “Scientific Initiation” (SI). The scholarships are distributed among professors of the universities and they select the students who will work in their research projects. The students who receive this scholarship are referred to as SI students. In this work, we built Figures 1, 2, 3a-c using the total number of researchers and Figure 3d is built with the subgroup that has less than 35 years of PhD. The Curriculum Vitae Lattes (CV) of each member was downloaded from the CNPq site in July 2016, but we only considered the information up to the end of 2015. The CVs of twenty (20) people were not found, 3 women and 17 men. These members were excluded from the analysis. Another important program to develop the research is the “Research Scholarship” (PQ), which are Monthly scholarships given to some researchers who are recognized in their respective fields. The The information extracted from the CV of each person is the following: • year of the obtainment of the PhD, phd TABLE I Area of knowledge All Titular Members Members < 35 years of PhD Female (Nf T) Male (Nm T) Total Female (Nf T) Male (Nm T) Total Engineering 1 39 40 1 19 20 Agriculture 4 21 25 3 13 16 Biological Sciences 10 25 35 4 10 14 Biomedical Sciences 18 93 111 5 37 42 Health Sciences 6 25 31 5 14 19 Earth Sciences 4 43 47 2 17 19 Physical Sciences 6 77 83 3 16 19 Mathematics 3 61 64 1 25 26 Chemistry 12 47 59 5 16 21 Social Sciences 5 18 23 2 2 4 Total 69 449 518 31 169 200 NATHÁLIA C. FERRARI et al. 2546 Research Scholarship is divided in five levels named as 2, 1D, 1C, 1B and 1A, from lower to upper levels. In addition, the Senior (SR) level was created for retired researchers who have been 1A or 1B levels for more than 15 years. Each researcher is evaluated every three or four years to renew or obtain the Research Scholarship. RESULTS Figure 1a also shows that the concentration of the recognized researchers increases in the Southeast from about 62% of CNPq researchers to 80% of ABC titular members, while it decreases in all other regions of Brazil in a plot that resembles the scissors gender plot (Slavin 2008). In this section, we show the distribution of the members of the ABC in the Brazilian regions and discriminated by gender. We also compare gender and region distribution in the ABC with the universe of PQ-researchers. How can we understand this concentration in the Southeast region of Brazil? One possible reason is the fact that the Southeast is the richest region of Brazil, producing about 60% of the Brazilian GDP.2 • affiliation address. In this work, we use the fact that the Research Scholarship PQ is an indicator of the quality of her/his work. Then we compare the universe of the titular members of the ABC with the universe of the PQ-researchers. This comparison allows us to follow the evolution of the researchers along the career. numbers, the fluctuations in any statistical analysis are huge. In order to gain insight in which stage of the career starts the concentration of recognized researchers in the Southeast region, we compare the percentage of ABC members with the percentage of PQ-researchers in each region. The group of “all PQ-researchers” indicates a select group of researchers, who are distributed in different ages and, in principle, in different localities in Brazil. To capture the most prestigious researchers of CNPq, we look at a subgroup of researchers at the levels 1A and 1B of CNPq. Then, representing the most prestigious researchers of Brazil, we show the titular members of ABC. 2see IBGE: Instituto Brasileiro de Geografia e Estatística at www.ibge.gov.br DISTRIBUTION OF THE PERCENTAGE OF THE ABC MEMBERS AND CNPq PQ-RESEARCHERS BY GEOGRAPHIC REGION Figure 1 - a) Distribution of the total number of CNPq PQ-researchers, 1A and 1B CNPq PQ-researchers and ABC titular members per Brazilian geographic region. Distribution of the percentage of CNPq PQ-researchers, 1A and 1B CNPq PQ-researchers and ABC titular members per Brazilian geographic region. The total number of individuals in each category is 10311 as PQ researchers, where 1818 are in the levels 1A and 1B. The total number of titular members of ABC who have a Brazilian address is 490. b) Map of the quotient of the number of titular female researchers divided by the total number of titular members by state. The color code is shown in the legend. Brazil. The quantity shown is the ratio of the number of female titular members over the total titular members in a given state. The stronger is the orange color, the higher is the percentage of female members in the state. The color code is such that the interval of the gender proportion varies between zero and 0.5, meaning that at best the number of men and women is the same. One can observe that in most states there is no female titular member. Alagoas and Pará are two states with the largest percentages of women, but this is due to the small number of members in these states: in Alagoas, there is only one woman and one man and in Pará there are 4 men and 2 women as titular members of the ABC. the ABC. The process to become a titular member of ABC goes as follows: one titular member of the Academy indicates a candidate and the acceptance of this indication is made by the votes of all the other members of ABC. This implies that the researchers proposed and accepted are most likely close colleagues of the current members. Consequently, the states or regions with more members are the ones with higher probability to have other members of that state. A third reason is that the Southeast region has a more professional structure for science communication and publicity through the media. Therefore, the scientists from that region are more known by the members of ABC. In the case of the CNPq, the committee that selects the PQ-researchers is smaller and selected to guarantee a representation from all states and research areas. DISTRIBUTION OF THE PERCENTAGE OF THE ABC MEMBERS AND CNPq PQ-RESEARCHERS BY GEOGRAPHIC REGION Since most of them do not know the candidates for the PQ fellowship, the selection is based on numeric achievements. At the top level, 1A and 1B, where the numbers are small, regional bias is also observed. Figure 2a shows the percentage of female and male members in ABC for different areas of knowledge. The figure evidences the underrepresentation of women in all fields of the Academia. The Engineering is the worst case, in which only 1% of the members are women while the Biological sciences is the best case in which about 25% of the members are women. DISTRIBUTION OF THE PERCENTAGE OF THE ABC MEMBERS AND CNPq PQ-RESEARCHERS BY GEOGRAPHIC REGION This implies for example that the budget for research only in São Paulo state is roughly the same as for the rest of the country, which is an attractive for good researchers to move to the Southeast region. It is important to note, however, that Brazil does not have a high academic mobility; when a researcher has a position in one university or center of research the tradition is to stay in that location. Therefore, the “bottleneck” that happens in Southeast region along the career cannot be only due to the fact that good researchers migrate to this region during his/her career. Brazil is divided in five geographic regions, which are composed by at least three states each. The regions and their respective percentage of the population are: Southeast (42.3%), Northeast (28%), South (14.3%), North (8.3%), and Central- West (7.1%). Figure 1a shows that the population distribution is not reproduced in the distribution of CNPq PQ-researchers, 1A and 1B CNPq PQ-researchers or the titular members of the ABC. Instead, this figure shows an impressive concentration of the members of ABC in the Southeast region. In the other extreme, there is the North region with a representation so small that the bars cannot be visualized: there are only 180 researchers with PQ scholarships, being 14 at the level B or A. The total number of ABC titular members is 12 in the region. With such small Another possible and complementary explanation for this concentration is the process of selection to the top levels both in the CNPq and in An Acad Bras Cienc (2018) 90 (2 Suppl. 1) GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC 2547 GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC 2547 Figure 1 - a) Distribution of the total number of CNPq PQ-researchers, 1A and 1B CNPq PQ-researchers and ABC titular members per Brazilian geographic region. Distribution of the percentage of CNPq PQ-researchers, 1A and 1B CNPq PQ-researchers and ABC titular members per Brazilian geographic region. The total number of individuals in each category is 10311 as PQ researchers, where 1818 are in the levels 1A and 1B. The total number of titular members of ABC who have a Brazilian address is 490. b) Map of the quotient of the number of titular female researchers divided by the total number of titular members by state. The color code is shown in the legend. An Acad Bras Cienc (2018) 90 (2 Suppl. 1) GENDER DISTRIBUTION Since this division is not quite clear in the CNPq, we adopt the Economy in the CNPq analysis. Figure 2b shows that the percentage of female researchers decreases as the career progresses. The only exception is Social A natural question is why the concentration of men increases when the career advances. To shed light in this point, we analyze the scientific production of the members of the ABC in the next section. GENDER DISTRIBUTION The map in Figure 1b shows the gender distribution of the titular members of the ABC per state in As we have done to understand the origin of the concentration of the members in the Southeast An Acad Bras Cienc (2018) 90 (2 Suppl. 1) 2548 NATHÁLIA C. FERRARI et al. Figure 2 - a) Percentage of male and female members of the ABC in each area; b) percentage of female members of the ABC, of the CNPq PQ-research and of the CNPq 1A and 1B PQ-research members in different areas. Figure 2 - a) Percentage of male and female members of the ABC in each area; b) percentage of female members of the ABC, of the CNPq PQ-research and of the CNPq 1A and 1B PQ-research members in different areas. Sciences, in which the comparison is difficult because at the CNPq we select only Economy while we are not quite sure if the ABC includes other fields of social science in its definition. region, we compare the percentage of female ABC titular members with the percentage of women that have a PQ-scholarship and the women who are at 1A and 1B levels of this scholarship. There are technical points to comment about this comparison. The ABC splits the “Biological Sciences” and “Biomedical” in two areas, while CNPq makes a slightly different division. To be able to compare both, we unified both and as shown in Figure 2 and in Figure 3 indicated as “Biological”. Another point is that in ABC “Social Science” actually means “Applied Social Science”. Since this division is not quite clear in the CNPq, we adopt the Economy in the CNPq analysis. Figure 2b shows that the percentage of female researchers decreases as the career progresses. The only exception is Social region, we compare the percentage of female ABC titular members with the percentage of women that have a PQ-scholarship and the women who are at 1A and 1B levels of this scholarship. There are technical points to comment about this comparison. The ABC splits the “Biological Sciences” and “Biomedical” in two areas, while CNPq makes a slightly different division. To be able to compare both, we unified both and as shown in Figure 2 and in Figure 3 indicated as “Biological”. Another point is that in ABC “Social Science” actually means “Applied Social Science”. QUALITY OF SCIENTIFIC PRODUCTION OF THE ABC MEMBERS In this section, we compare the scientific production of male and female ABC titular members using two variables: the average number of articles and the average number of students supervised per An Acad Bras Cienc (2018) 90 (2 Suppl. 1) GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC 2549 Figure 3 - a) Average number of human resources production and b) average number of articles production in the different areas by the male (x axis) and female (y axis) members of the ABC. c) Average number of students supervised by female and male titular members separated by different types of students: scientific initiation (SI), Master, PhD and post-doc (PD). d) Distribution of researchers in terms of the type of PQ-scholarship. For this figure, we only use the members with less than 35 years of PhD. Figure 3 - a) Average number of human resources production and b) average number of articles production in the different areas Figure 3 - a) Average number of human resources production and b) average number of articles production in the different areas by the male (x axis) and female (y axis) members of the ABC. c) Average number of students supervised by female and male titular members separated by different types of students: scientific initiation (SI), Master, PhD and post-doc (PD). d) Distribution of researchers in terms of the type of PQ-scholarship. For this figure, we only use the members with less than 35 years of PhD. researchers. In most areas of knowledge, women produce more human resources than men. researcher per year, also referred to as human resources. It is quantified as the average number of people that the researcher supervised during his/ her academic life, including all types of students. The precise definition of this quantity is shown in the Appendix. Figure 3b is a scatter plot of the publications per year for female (in x-axis) and male (in y-axis) titular members discriminated by knowledge area. In this scatter plot we observe that in most areas the average number of papers published by men and women lies close to the line y=x, which suggests that they have a similar production. These findings are in agreement with an analysis of the PQ- researchers in physics (Arenzon et al. 2013). There are two areas where the male production is clearly higher than the women production: biomedical and chemistry. CONCLUSIONS Researchers invest time in them because having a mentor correlates with good performance as undergraduate and, therefore, implies forming better future master and PhD students. Since the number of graduate fellowships for each advisor is limited, many of these SI students that were formed and supervised by female researchers end up doing their master and PhD with male scientists. In this paper, we quantify the diversity in the Brazilian Academy of Sciences in terms of gender and geographic location. We show that 80% of the members are concentrated in the Southeast region and that about 86% of the alive titular members are men. We then compare the quality of the research of the female and male members using two indexes, the average number of publications and the average number of students supervised per year. These analyses show that men and women produce roughly the same amount of papers in all areas of knowledge covered by ABC, but women in general are more engaged in supervising students. We note that both genders supervise relatively few post-doc, probably because many Brazilians go abroad after their PhD and few people come to Brazil at this stage of the career. Then we believe that the difference in the number of post-doc does not impact in the papers production. We also evaluate members’ productivity by measuring the percentage of them with each type of PQ-scholarship and show that men have higher proportion of members without any type of PQ- scholarship. All this together allows us to conclude that the Brazilian Academy of Sciences is neither diverse in terms of geographic region nor in terms of gender, even if the male members to do not have higher indexes of scientific productivity compared to female members. We also looked at the academic performance of the titular members in terms of the type of CNPq PQ-scholarship they have. The result discriminated by gender is shown in Figure 3d. We observe that the absolute majority of women has 1A and 1B scholarships. A striking observation is that there are about 15% of men without PQ- scholarship, while only 1% of the women do not have it. It is important to emphasize that we are only considering those members who have less than 35 years after finishing their PhD. If we would include everybody, the percentage of men without PQ-scholarship increases to 40% and to 16% for women. QUALITY OF SCIENTIFIC PRODUCTION OF THE ABC MEMBERS In this last case, the difference is of about 6 papers published by male researchers in average against only 4 by female researchers. Figure 3a is a scatter plot of the Students supervised per year for female (in x-axis) and male (in y-axis) titular members for different knowledge area, as specified in the legend of the figure. If the average production of human resources of men and women members were the same, the data points would lie on the line y=x illustrated as a thin line in the figure. Further up or down is the point from this line, larger is the difference between the human resource production of the female and male An Acad Bras Cienc (2018) 90 (2 Suppl. 1) NATHÁLIA C. FERRARI et al. 2550 have the PQ scholarship when they were selected to become a titular member, but it is possible that they are working in administrative positions for which women are rarely selected. In any case, some women who are titular members of ABC also have administrative jobs, but they manage to keep their PQ-scholarship. We hypothesize that the criteria to be accepted at ABC is more restrictive for women than for men, suggesting a kind of mechanism to explain the “ceiling glass” that is measured through some statistics but is not explicitly written in any official rule. As a whole, Figure 3 indicates that, in addition to the fact that women produce more human resources than men, at best they publish as many papers as their counterparts. In order to understand why more supervisions do not convert in more publications, we look at the type of supervised students for all the members of ABC, shown in Figure 3c. We observe that the total number of master and PhD students supervised by female researchers is roughly the same as the ones supervised by male titular members, but women have almost two times more SI and post-doc students than men. Since SI students are undergraduate, they are not supposed to perform a competitive research and write papers. Researchers invest time in them because having a mentor correlates with good performance as undergraduate and, therefore, implies forming better future master and PhD students. Since the number of graduate fellowships for each advisor is limited, many of these SI students that were formed and supervised by female researchers end up doing their master and PhD with male scientists. QUALITY OF SCIENTIFIC PRODUCTION OF THE ABC MEMBERS We note that both genders supervise relatively few post-doc, probably because many Brazilians go abroad after their PhD and few people come to Brazil at this stage of the career. Then we believe that the difference in the number of post-doc does not impact in the papers production. As a whole, Figure 3 indicates that, in addition to the fact that women produce more human resources than men, at best they publish as many papers as their counterparts. In order to understand why more supervisions do not convert in more publications, we look at the type of supervised students for all the members of ABC, shown in Figure 3c. We observe that the total number of master and PhD students supervised by female researchers is roughly the same as the ones supervised by male titular members, but women have almost two times more SI and post-doc students than men. Since SI students are undergraduate, they are not supposed to perform a competitive research and write papers. An Acad Bras Cienc (2018) 90 (2 Suppl. 1) REFERENCES ARENZON JJ, DUARTE P, CAVALCANTI S AND BARBOSA MC. 2013. Women and physics in Brazil: Publications, citations and H index. AIP Conf Proc 1517: 78-80. CHAMORRO-PREMUZIC T. 2017. The Talent Delusion: Why Data, Not Intuition, Is the Key to Unlocking Human Potential. Piatkus Publishing, UK, 304 p. Potential. Piatkus Publishing, UK, 304 p. DEZSO CL AND ROSS DG. 2012. Does female representation in top management improve firm performance? A panel data investigation. Strategic Manage J 33: 1072-1089. DEZSO CL, ROSS DG AND URIBE J. 2017. Is there an implicit quota on women in top management? A large- sample statistical analysis. Strategic Manage J 37: 98-115. This gradual concentration of male researchers in the richest region of Brazil when the career evolves has some negative consequences. One is the waste of financial resources, because many people are educated and then excluded from the workforce. The second is that, by excluding women and individuals from least developed regions of the country, the pool of talents is decreased, rendering less likely to promote the most competent people of the society (Chamorro-Premuzic 2017). GUTERL F. 2014. Diversity in science: Why it is essential for excellence. Sci Am 311: 38-41. HUNT V, LAYTON D AND PRINCE S. 2015. Diversity Matters: McKinsey and Company Report. http://www. mckinsey.com/business-functions/organization/our- insights/why-diversity-matters. (Accessed on Jan 18, 2017). JOY L, NANCY MC, WAGNER HM AND NARAYANAN S. 2007. The Bottom Line: Corporate Performance and Women’s Representation on Boards. http://www. catalyst.org/system/files/The_Bottom_Line_Corporate_ Performance_and_Womens_Representation_on_Boards. pdf. (Accessed on Jan 18, 2017). Another very important dimension of the social world that we have not investigated here is the race-ethnic identity. It is clear from our daily life in Academia that the percentage of black/indigenous people is very small, even though Brazil has, for example, more than 50% of self-declared black people. Quantify the race and ethnic diversity in academia is still a problem because our main data basis of CV (CNPq Lattes) included recently a possibility to declare the color of the researcher. We hope that in a near future some statistics about the diversity of races in the Brazilian science will be possible. MOSS-RACUSINA CA, DOVIDIO FJ, BRESCOLL VL, GRAHAM MJ AND HANDELSMAN J. 2012. Science faculty’s subtle gender biases favor male students. PNAS 109: 16474-16479. NSFOUNDATION. 2015. Women, minorities, persons with Disabilities in Science and Engineering. https://www. nsf.gov/statistics/2015/nsf15311/digest/. (Accessed on January 18, 2017). OSBORN M. 2000. GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC 2551 the evolution effect in the researchers’ career, we compare the ABC data with the universe of PQ- researchers and observe that diversity decreases along the researchers’ career: the ABC contains more male members and more people from the Southeast region than the universe of prestigious people who have the PQ-scholarship. It shows that not only the ABC has very small diversity but also that this problem only increases when the level of the career gets higher. It is worth noting that among the 15 new titular members announced for 2017, 12 work in the Southeast region, suggesting that the 80% proportion of researchers from the Southeast region will be kept for longer. and more people from other regions of Brazil other than the Southeast region into research. CONCLUSIONS It is very likely that these members used to One could argue that the lack of diversity observed in ABC is just a photography of the science situation in Brazil and a more relevant measure is how this situation evolves. To take into account An Acad Bras Cienc (2018) 90 (2 Suppl. 1) GEOGRAPHIC AND GENDER DIVERSITY IN THE ABC REFERENCES Science policies in the European Union: Promoting excellence through mainstreaming gender equality. http://ufm.dk/en/publications/2000/files-2000/ promoting-excellence-through-mainstreaming-gender- equality.pdf (Accessed on Jan 2, 2017). PHILIPS KW. 2014. How diversity makes us smarter. Sci Am 311: 42-47. PHILLIPS KW, NORTHCRAFT GB AND NEALE MA. 2006. Surface-level diversity and decision-making in groups: When does deep-level similarity help? Group Process Intergr Relat 9: 467-482. Having in mind that diversity is important to solve different kinds of problems and improve efficiency, it is time for Brazilians to think more seriously about policies to include more women RAYMOND J. 2013. Sexist attitudes: Most of us are biased. Nature 495: 33. An Acad Bras Cienc (2018) 90 (2 Suppl. 1) NATHÁLIA C. FERRARI et al. 2552 SLAVIN K. 2008. Gender equality report Sixth Framework Programme. https://ec.europa.eu/research/swafs/pdf/pub_ gender_equality/gender-equality-report-fp6-final_en.pdf (Accessed on Jan 18, 2017). Then Studentsi is averaged over all the titular female members Nf T and male members Nm T of a given area of knowledge, WALLON G, BENDISCIOLI S AND GARNKEL MS. 2015. Exploring Quotas in the Academy. http://www.embo. org/documents/science_policy/exploring_quotas.pdf. (Accessed on Jan 18, 2017). female−Studentsper year = ∑ i Nf T Students i Nf T male−Studentsper year= ∑ i Nm T Students i Nm T An Acad Bras Cienc (2018) 90 (2 Suppl. 1) APPENDIX: MATHEMATICAL DEFINITION OF QUANTITIES SHOWN IN FIGURES 3a AND 3b We also analyze the average number of scientific articles published by each researcher, defined as: In this section, we show the precise definition of the quantities shown in the scatter plot of the Figures 2a and 2b. Publications i= p i (2015−PhD i) The production of human resources is quantified as the average number of people that the titular members supervised during their academic life. This number includes all the students supervised by the researcher divided by the number of years that the person has been working since the obtainment of his/her PhD: where pi is the total number of papers published by the researcher i. To obtain this index by knowledge area and different gender, we average the publications of the members by area: female−Publicationsper year= ∑ i N f T P i Nf T male−Publicationsper year= ∑ i Nm T P i Nm T Students i=(nPhD i +nm i +nSI i +nPD i ) (2015−PhD i) where nPhD i is the total number of PhD students, nm i is the number of master students, nSI i is the number of scientific initiation students, ni PD is the number post-doc supervised by the research i and PhDi is the year that the person i obtained her/his PhD.
https://openalex.org/W3034035436	https://ora.ox.ac.uk/objects/uuid:c9a7c53a-372d-4559-94ed-d44bb92e1bd6/files/srf55z792s	English	null	An improved method to model dislocation self-climb	Modelling and simulation in materials science and engineering	2,020	cc-by	11,731	An improved method to model dislocation self-climb To cite this article: Fengxian Liu et al 2020 Modelling Simul. Mater. Sci. Eng. 28 055012 View the article online for updates and enhancements. This content was downloaded from IP address 146.199.5.55 on 19/06/2020 at 10:41 Modelling and Simulation in Materials Science and Engineering Modelling and Simulation in Materials Science and Engineering This content was downloaded from IP address 146.199.5.55 on 19/06/2020 at 10:41 This content was downloaded from IP address 146.199.5.55 on 19/06/2020 at 10:41 Modelling and Simulation in Materials Science and Engineering Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 (21pp) https://doi.org/10.1088/1361-651X/ab81a8 E-mail: edmund.tarleton@eng.ox.ac.uk E-mail: edmund.tarleton@eng.ox.ac.uk E-mail: edmund.tarleton@eng.ox.ac.uk Received 16 January 2020, revised 18 February 2020 Accepted for publication 20 March 2020 Published 4 June 2020 Abstract 0965-0393/20/055012+21$33.00 © 2020 The Author(s). Published by IOP Publishing Ltd Printed in the UK 4The author to whom any correspondence should be addressed. Content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 0965-0393/20/055012+21$33.00 © 2020 The Author(s). Published by IOP Publishing Ltd Printed in the UK 1 Fengxian Liu1 , Alan C F Cocks2 , Simon P A Gill3 and Edmund Tarleton2,1,4 1 Department of Materials, University of Oxford, Parks Road, Oxford OX1 3PH, United Kingdom 2 Department of Engineering Science, University of Oxford, Parks Road, Oxford OX1 3PJ, United Kingdom 3 School of Engineering, University of Leicester, Leicester, LE1 7RH, United Kingdom 1. Introduction The diffusion of atoms in crystalline materials is a thermally activated process which is dras- tically enhanced by lattice defects [2], such as surfaces, grain boundaries [3], and dislocations [4, 5]. The short-circuit paths provided by these defects are thought to originate from the loose-packed and highly disordered atoms that lower the activation energy for diffusion and enhance the atomic jump frequency. When carried by dislocations, the diffusion mechanism is known as core diffusion or pipe diffusion, as the core of dislocations act as pipes for atoms to diffuse down. However, while surface diffusion and grain boundary diffusion have been studied and discussed in detail [6–12], only limited results are available for pipe diffusion along dislocation networks [13]. It is challenging to extract core diffusivity from experimental observations [14], since core diffusion often occurs simultaneously with lattice diffusion and occupies far fewer atomic sites. Core diffusion is, therefore, usually neglected in most studies [15–18] or treated phenomenologically as an enhancement to bulk diffusion [19]. However, at lower temperatures, core diffusion plays a dominant role in mass diffusion, particularly at high stresses. The recent direct observation of core diffusion effects [5, 20, 21] demon- strated that the core diffusivity in aluminium can be increased by three orders of magnitude near 600 K [5], and the difference can be up to six orders of magnitude in bcc Fe at 700 K [4]. The rapid atomic diffusion along dislocations represents in itself an interesting physical phenomenon and plays a signifcant role in a wide variety of material behaviours, such as creep [19, 22], loop or particle coarsening [4, 23–27], precipitation and phase transforma- tions [5, 28, 29], dynamic strain ageing [30, 31], and solute segregation [32–34]. However, the role of core diffusion on dislocation motion has received little attention and is not well understood. During the core diffusion process, atoms can be absorbed or emitted by the core region, leading to climb out of the original slip plane, known as dislocation self-climb. The self- climb of dislocation loops, as frst proposed by [35], differs from the commonly studied non-conservative climb [18, 36–39] in the sense that atoms can be transferred only along the perimeter of the loop and the total area of the loop projected perpendicular to its Burg- ers vector remains unchanged, namely, the relaxation volume of loops stay constant [40]. Abstract Dislocations can provide short circuit diffusion paths for atoms resulting in a dislocation climb motion referred to as self-climb. A variational principle is presented for the analysis of problems in which fast dislocation core diffu- sion is the dominant mechanism for material redistribution. The linear element based self-climb model, developed in our previous work [1] Liu, Cocks and Tarleton (2020 J. Mech. Phys. Solids 135 103783), is signifcantly accelerated here, by employing a new fnite element discretisation method. The speed-up in computation enables us to use the self-climb model as an effective numerical technique to simulate emergent dislocation behaviour involving both self-climb and glide. The formation of prismatic loops from the break-up of different types of edge dislocation dipoles are investigated based on this new method. We demonstrate that edge dipoles sequentially pinch-off prismatic loops, rather than spontaneously breaking-up into a string of loops, to rapidly decrease the total dislocation energy. Keywords: core diffusion, pipe diffusion, variational principle, self-climb S Supplementary material for this article is available online Content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1 F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 1. Introduction In self-climb, atoms are rapidly rearranged in the dislocation core region to balance the differ- ence of chemical potential along the dislocation line, which is more energetically favorable than mass exchange with the surrounding lattice (bulk diffusion), particularly at low temper- atures where bulk diffusion could be many orders of magnitude slower than core diffusion [4]. Also, the analysis of core diffusion-controlled mechanisms tends to be more straight- forward; for example, the three-dimensional lattice diffusion problem [15, 17, 36, 38, 39, 41] can be reduced to the consideration of one-dimensional diffusion along the dislocation line. Earlier attempts at analysing dislocation self-climb mainly focus on the transport of pris- matic loops, due to the simplicity in dislocation character, and their importance in irradiated materials. Following the pioneering theoretical studies on the dislocation self-climb of pris- matic loops in the 1960s [35, 42, 43], Turnbull [26] constructed a model for loop motion by self-climb, in which loops remain circular throughout the process. This theoretical analysis is further supported by the recent experimental observation [4] that the self-climb velocity of a circular prismatic loop is proportional to the driving force and inverse cube of the loop size; indicating higher mobility for smaller loops as observed in post-irradiation annealing experiments [44]. 2 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al A simple derivation for the transport of a rigid prismatic loop by self-climb was pro- posed in our recent study, based on a variational principle [1]. The theoretical analyses of self-climb mobility, with the rigid loop assumption, provide a description of the coarsen- ing process of prismatic loops, it fails when loops are close or in contact. In real materials, loops change both their positions and shapes to minimise the total dislocation energy dur- ing the coarsening process while conserving the total enclosed area [1, 45]. More impor- tantly, the transport of a prismatic loop on its habit plane is just a special case of self-climb of a pure edge dislocation. For the motion of an arbitrary shaped dislocation line, a more general self-climb model is needed to couple atomic diffusion with dislocation self-climb motion. Molecular dynamic (MD) simulations are well suited to study core diffusion with atomic resolution [46]. 1. Introduction MD has been used to calculate various core diffusion dominated behaviours including: the core diffusivity along different dislocation types [14, 47], the annihilation of prismatic half loops via pipe diffusion during nanoindentation [48], and the conservative climb motion of a cluster of self-interstitial atoms towards an edge dislocation [49]. Collective dislo- cation motion involves both climb and glide, making it far beyond the length and time scales accessible with atomistic simulations. A solution is to develop mesoscopic coarse-grained models that operate at intermediate length and time scales, providing a bridge between atom- istic and macroscopic models. Geslin et al [50] proposed a multiscale approach to model dislocation climb of jogged dislocations with consideration of both bulk and core diffusion between jogs. The analytical climb rate was derived by assuming a periodic distribution of jogs along the dislocation line. The climb rate was then upscaled to a phase feld model to simulating dislocation climb at a larger scale. Later, Niu and co-authors [45, 51] developed a discrete dislocation dynamics (DDD)-based climb model by upscaling from a stochastic jog dynamic model on the atomic scale. Both the pipe diffusion equation with Dirichlet boundary conditions at the jogs and the bulk diffusion equation with Robin boundary conditions near the dislocation were solved in [51], to derive the jog mobility. Note that, among these multi- scale models, dislocation climb motion is represented by the dynamics of the pre-existing jogs (neglecting nucleation), which act as perfect sinks or sources of vacancies. This is a good assumption for bulk diffusion when the jog density is high enough to maintain an equilibrium vacancy concentration along the dislocation core [17]. But for a low jog density, as is likely at lower temperatures and higher stresses, the stresses tend to sweep the jogs away and push them towards the dislocation ends. Consequently in this regime the nucleation of jog-pairs needs to be taken into consideration [52]. In our recent work [1], a self-climb model was proposed, which employs a variational prin- ciple for the evolution of microstructure [53, 54]. In this model, a fnite element based analysis for the one-dimensional core diffusion process was implemented. A dislocation self-climb model was then developed by implementing this core diffusion formulation into the nodal based three-dimensional discrete dislocation dynamics (DDD) framework [55], to extend the traditional DDD method to simulate self-climb and glide of any arbitrarily shaped dislocation network. 1. Introduction However, the DDD method is still limited by its high computational cost [56]. Improv- ing the computational effciency [57, 58] is, therefore, a challenging but necessary task. In a general DDD simulation [59, 60], the most time-consuming parts include: (1) the time integration of the equation of motion, which has been effectively sped up ∼100× by time subcycling [57, 58]; (2) the calculation of seg–seg interactions between different disloca- tion segments, which is an N-body problem and increases dramatically with strain. This problem has been accelerated by a parallelization strategy on the graphical pgrocessing unit (GPU) [58, 61]. In the newly developed DDD-based self-climb model [1], the additional 3 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al time-consuming part is the calculation of nodal climb velocities. In this method, disloca- tions are discretised into a series of straight segments (1D fnite elements). The climb veloc- ity is defned at each node and varies linearly along the segment, and the diffusive fux is defned across the mid-point of each segment. To enforce the fux continuity at nodes where two segments meet, a series of Lagrange multiplier are introduced, one for each node. A set of linear simultaneous equations is then derived as the kinetic equations for self-climb [1], [K] [C]T [C] [0] [Vc] [λ] = [Fc] [0] (1) (1) where [K ] is the (N + m) × (N + m) global viscosity matrix for core diffusion, [C] is an N × (N + m) constraint matrix, [0] is the N × N null matrix, where N is the number of dis- location nodes and m is the number of segment. [Vc] is the (N + m) × 1 vector of unknowns, including the nodal climb velocities and fuxes, [λ] is the N × 1 vector of Lagrange multipliers, [Fc] is the (N + m) × 1 vector of generalised nodal forces and [0] is the (N + m) × 1 zero vec- tor. The advantage of using Lagrange Multipliers to enforce the fux continuity at junctions is that it allows the [K] matrix to be narrowly banded and effectively solved with standard matrix methods. 1. Introduction However, in (1) the leading matrix of the equations is then not positive defnite, due to the introduction of Lagrange multipliers, leading to a signifcant increase in the computa- tional cost which becomes problematic as the number of segments increases. Consequently the purpose of this work is to improve the computational effciency of the self-climb model, to enable the simulation of collective dislocation evolution involving both self-climb and glide motion. The present paper is organised as follows. In section 2, a new discretisation method is devel- oped by the adoption of the paired-linear element based method (PLEBM), to circumvent the usage of Lagrange Multipliers for every node. The PLEBM is validated and compared with the linear element based method (LEBM) to show the improvement in capability and eff- ciency.Finally, in section 3 new physical insight into the break-offof dislocation dipoles, where core diffusion-controlled self-climb is the dominant mechanism, are obtained based on the PLEBM. 2. Methodology The variational principle for core diffusion developed in our previous work [1] is briefy reviewed here. We then implement this formulation with a paired-linear element to derive the kinetic equations for self-climb. 2.1. The variational principle for core diffusion Variational principles have been rigorously formulated for a series of mass transport mecha- nisms, including grain-boundary diffusion [62], surface diffusion [63], grain-boundary migra- tion [64] and coupled grain-boundary and surface diffusion [65]. The essential idea carries over to dislocation core diffusion. That is, among all the virtual velocities of microstructures and virtual diffusive fuxes that satisfy matter conservation, the actual velocity and fux felds minimize the functional of the system Π [65], Π = Ψ + ˙G (2) Π = Ψ + ˙G (2) where Ψ is a rate potential term involving contributions from all possible dissipative processes. ˙G is the energy rate term, which is the origin of the generalised thermodynamic force. In (2) it Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 1. A general situation considered in the present work: an infnite three dimen- sional domain with an arbitrarily shaped dislocation network. (a) A general consecutive dislocation connection, every node has two connections. Segment (13) has end nodes 1 and 3 and a ‘dummy’ central node 2; (b) a triple-junction node with three connections. Figure 1. A general situation considered in the present work: an infnite three dimen- sional domain with an arbitrarily shaped dislocation network. (a) A general consecutive dislocation connection, every node has two connections. Segment (13) has end nodes 1 and 3 and a ‘dummy’ central node 2; (b) a triple-junction node with three connections. is the combination of thermodynamics and kinetics that determines the actual evolution path and the fnal state. Here, the one-dimensional core diffusion-controlled self-climb process is investigated. The general situation analysed in the current paper is schematically shown in fgure 1. An arbitrarily shaped three-dimensional dislocation network, in an infnite domain, is subjected to long-range elastic interactions between dislocation segments. Dislocations can either glide in their original slip plane, or relocate themselves by the diffusional transport of material to move perpendicular to their original slip plane. As stated in the introduction, we con- sider the situation of lower temperatures when lattice diffusion is negligible in comparison with core diffusion; such that mass can only diffuse along the dislocation line in the core area. Here, we state two basics laws before developing the variational functional of the system: Here, we state two basics laws before developing the variational functional of the system: The law for mass diffusion. 2.1. The variational principle for core diffusion We adopt a classical description of the diffusion process, the volumetric fux j(l) driven by the gradient in the chemical potential as described by Fick’s law, j(l) = −acoreDcoreΩ kT ∂µ ∂l = −D∂µ ∂l (3) (3) where l is curvilinear coordinate along the dislocation line, Dcore is the core diffusivity, acore is the cross sectional area of the dislocation core, k is the Boltzmann’s constant, T is the absolute temperature and µ is the excess chemical potential; For core diffusion along dislocation lines, µ in (3) then refers to the change in Gibbs free energy per unit volume for the diffusing material along the dislocation; it is directly proportional to the climb component of the Peach–Koehler 5 elling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 force. The gradient of µ drives the diffusion of material with a unit volume. D = acoreDcoreΩ kT (4) D = acoreDcoreΩ kT (4) is the effective diffusivity. Along the dislocation line, both the local stress and concentration of point defects contribute to the chemical potential µ. is the effective diffusivity. Along the dislocation line, both the local stress and concentration of point defects contribute to the chemical potential µ. The law for dislocation climb. As mass diffuses along the dislocation core, atoms may be deposited or removed from the extra-atomic-plane, leading to climb motion perpendicular to the slip plane. During this process, mass conservation requires the climb velocity is balanced by the variation of the fux, d j dl + vcb sin θ = 0 (5) (5) where dl sin θ is the edge component of the dislocation line increment dl, θ is the angle between the Burgers vector b and dislocation line vector l. As discussed in [1], this ensures only the edge component of a dislocation line can climb and consume the atomic fux. We now develop the variational functional of the system, as derived in [1], where dl sin θ is the edge component of the dislocation line increment dl, θ is the angle between the Burgers vector b and dislocation line vector l. As discussed in [1], this ensures only the edge component of a dislocation line can climb and consume the atomic fux. 2.1. The variational principle for core diffusion We now develop the variational functional of the system, as derived in [1], Π = Ψ + ˙G = Z L j2(l) 2D(l)dl − Z L fc(l)vc(l)dl (6) (6) where fc is the Peach–Koehler force component in the climb direction, vc is the climb veloc- ity and L is the total dislocation line length in the system. The frst term on the right-hand side of (6) is the rate potential due to core diffusion, and the second term is the rate of change of Gibbs free energy during the self-climb motion. Different virtual motions give different values of Π. Of all virtual motions, the actual motion renders Π stationary, that is, where fc is the Peach–Koehler force component in the climb direction, vc is the climb veloc- ity and L is the total dislocation line length in the system. The frst term on the right-hand side of (6) is the rate potential due to core diffusion, and the second term is the rate of change of Gibbs free energy during the self-climb motion. Different virtual motions give different values of Π. Of all virtual motions, the actual motion renders Π stationary, that is, δΠ = 0. (7) (7) δΠ = 0. 2.2. Implementation of the variational principle In the follow- ing discussion, the functional Π is discretised in terms of the degrees of freedom defned above. To avoid using Lagrange multipliers at every node, and avoid the non-positive coeffcient matrix in the kinematic equation (1), a different discretisation method is used here. An addi- tional dummy node is located at the midpoint of each segment, node 2 in fgure 1(a). To standardise the following computation, a local coordinate s is defned. The origin of s is located at the start point of a given segment, and the positive direction points from node 1 →2. s is normalised by the segment length l(12), so that s = 0 at node 1 and s = 1 at node 2. Assuming a linear variation of the climb velocity along both segment (12) and seg- ment (23), so that the climb velocities along segments (12) and (23) are given respectively as, v(12) c (s) = β1 (1 −s) Vc1 + β2sVc2 = h N(12) 1 (s) N(12) 2 (s) i Vc1 Vc2 (8) v(23) c (s) = β2 (1 −s) Vc2 + β3sVc3 = h N(23) 2 (s) N(23) 3 (s) i Vc2 Vc3 (9) (8) (9) where s ∈[0, 1], β1 = N1 · n(12), β2 = N2 · n(12) = N2 · n(23), β3 = N3 · n(23) and N1, N2, N3 are the directions of the climb velocity at nodes 1,2 and 3, respectively. These are assumed to be the weighted average of the segment slip plane normals connected to the node, for example, N2 = l(12)n(12) + l(23)n(23) \|l(12)n(12) + l(23)n(23)\|. (10) (10) Not to be confused with N(12) 1 (s) and N(12) 2 (s), the isoparametric shape functions for segment (12), N(23) 2 (s) and N(23) 3 (s) refer to segment (23), Not to be confused with N(12) 1 (s) and N(12) 2 (s), the isoparametric shape functions for segment (12), N(23) 2 (s) and N(23) 3 (s) refer to segment (23), N(12) 1 (s) = β1 (1 −s) , N(12) 2 (s) = β2s (11) N(23) 2 (s) = β2 (1 −s) , N(23) 3 (s) = β3s. 2.2. Implementation of the variational principle The kinetics, which here is atomic diffusion, proceeds locally. The idea is to transcribe local information onto the total energy landscape, to fnd the global equilibriumstate. The continuum structure is approximated by discrete elements and a viscosity matrix assigned to every point on the total landscape. In using the variational principle, it is impractical to search for the exact motion from all virtual motions. Rather, the solution is selected from a restricted set of virtual motions, and the solution which minimises the variational functional Π approximates the exact motion. We have described how numerical schemes can be developed from (6) and (7) for three- dimensional dislocation networks in our previous work [1]. We limit our attention here to the differences in the paired-linear element discretisation compared to the previous linear ele- ment [1]. As demonstrated in fgure 1, dislocation lines are discretised into a series of straight segments (1D elements). Typical types of dislocation connections are shown in fgures 1(a) and (b). We initially restrict our attention to the majority of situations, demonstrated in fgure 1(a), where each node connects to only two segments. We will see in the following discussion that it proves convenient to isolate a segment, for example, segment (13) in fgure 1(a). Along the segment, the nodal climb velocity Vc and the diffusive fux J are defned at each node. Here, we follow the convention that the nodal values are denoted in uppercase and 6 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al numbered by subscripts, whereas segment values are indicated in lowercase and numbered by superscripts with parentheses. For example, Vc1 is the nodal climb velocity at node 1, and J(12) 1 is the fux at node 1 along segment (12). b(12), l(12) and n(12) denote the Burgers vector, unit line vector and unit normal vector of segment (12), respectively. In the follow- ing discussion, the functional Π is discretised in terms of the degrees of freedom defned above. numbered by subscripts, whereas segment values are indicated in lowercase and numbered by superscripts with parentheses. For example, Vc1 is the nodal climb velocity at node 1, and J(12) 1 is the fux at node 1 along segment (12). b(12), l(12) and n(12) denote the Burgers vector, unit line vector and unit normal vector of segment (12), respectively. 2.2. Implementation of the variational principle (12) (12) Mass conservation requires, Mass conservation requires, Mass conservation requires, d j(l) dl + vc(l)b sinθ = 0 (13) d j(l) dl + vc(l)b sinθ = 0 (13) d j(l) dl + vc(l)b sinθ = 0 (13) and so, and so, d j(s) = −vc(s)bl sin θ ds, (14) (14) and the fux is and the fux is j (12)(s) = J(12) 1 −l(12) e b(12) Z [β1 (1 −s) , β2s]ds Vc1 Vc2 , (15) j (23)(s) = J(23) 2 −l(23) e b(23) Z [β2 (1 −s) , β3s] ds Vc2 Vc3 (16) (15) 7 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al where le = lsinα is the edge component of the segment length l. Note that there are now 7 degrees of freedom for segment (13), where le = lsinα is the edge component of the segment length l. Note that there are now 7 degrees of freedom for segment (13), v(13) = h Vc1, J(12) 1 , Vc2, J(12) 2 , J(23) 2 , Vc3, J(23) 3 i . (17) (17) Flux continuity at node 2 requires that the fux into node 2 must fow out, so that J(12) 2 = J(23) 2 = J2. This removes one of the degrees of freedom. Flux continuity is required at every node. Therefore, for a node connected with two segments, only one degree of free- dom is needed for the fux. One degree of freedom, JI, is therefore used for the fux at each node I. Care is required for triple or quadruple junction nodes with more than 2 seg- ments, such as node 4 in fgure 1(b). In such cases, additional degrees of freedom are needed for the fux along different segments. This is discussed and evaluated further in section 2.4. Mass conservationat two ends of each segmentrequires j (12)(s = 1) = J (12) 2 and j (23)(s = 1) = J (23) 3 . This again, removes two of the degrees of freedom. We choose to remove the degrees of freedom from the dummy node (node 2), Vc2 and J2: Vc2 = 1 β2(l(12) + l(23)) −β1l(12), 2, −β3l(23), −2   Vc1 J1 Vc3 J3   (18) (18) J2 = 1 2(l(12) + l(23)) −β1l(12)l(23), 2l(23), β3l(12)l(23), 2l(12)   Vc1 J1 Vc3 J3  . 2.2. Implementation of the variational principle (19) (19) The dissipation potential for segment (13) is then given by Ψ(13) = 1 2D(12) Z l(12) 0 j (12)(s) 2ds + 1 2D(23) Z l(23) 0 j (23)(s) 2ds (20) = 1 2[v(13) c ]T[k(13)][v(13) c ] (21) (20) (21) where [v(13) c ] = [Vc1, J1, Vc3, J3]T is a 4 × 1 vector of unknowns including nodal velocities and fuxes at the two ends of segment (13). [k(13)] is a symmetric viscosity matrix for core diffusion along segment (13), which is given in A. The total rate potential Ψ is therefore, Ψ = n X i=1 Ψ(i) = n X i=1 1 2[v(i) c ]T[k(i)][v(i) c ] = 1 2[Vc]T[K][Vc] (22) (22) where n is the total segment number. [Vc] is the global column vector of unknowns including climb velocities and fuxes at all nodes. [K ] is the global viscosity matrix assembled from all elementary viscosity matrices [k(i)]. The rate of change of Gibbs free energy during climb motion can also be discretised as, ˙G = − Z L fc(s)vc(s)ds = − N X I=1 VcIFcI = −[Vc]T[Fc] (23) (23) Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al where N is the total number of dislocation nodes, VcI and FcI are the nodal climb velocity and nodal climb force at node I. [Fc] is an N × 1 nodal climb force vector corresponding to the global nodal velocity vector [Vc]. As stated in our previous work [1], the nodal climb force on node I, FcI, is where N is the total number of dislocation nodes, VcI and FcI are the nodal climb velocity and nodal climb force at node I. [Fc] is an N × 1 nodal climb force vector corresponding to the global nodal velocity vector [Vc]. As stated in our previous work [1], the nodal climb force on node I, FcI, is FcI = FI · NI = ˜FI + Fself I + Fcore I + Fapp I · NI (24) (24) where FI is the full nodal force at node I, with normal NI, as defned in (10). ˜FI is the elastic interaction between the (i) segments connected to node I with every other segment (j) ̸= (i). Fself I refers to the elastic self force on node I, due to segments (i). 2.2. Implementation of the variational principle Fcore I is the core force on node I and Fapp I accounts for an applied stress. The variational functional Π can now be expressed in terms of the discretised degrees of freedom by substituting (22) and (23) into (6), Π = 1 2[Vc]T[K][Vc] −[Vc]T[Fc]. (25) (25) Substituting (25) into (6), we obtain δΠ = δ[Vc][K][Vc] −δ[Vc][Fc] = 0, for any small perturbation δ[Vc]. Such that, Substituting (25) into (6), we obtain δΠ = δ[Vc][K][Vc] −δ[Vc][Fc] = 0, for any small perturbation δ[Vc]. Such that, Substituting (25) into (6), we obtain δΠ = δ[Vc][K][Vc] −δ[Vc][Fc] = 0, for any small perturbation δ[Vc]. Such that, [K][Vc] = [Fc] (26) [K][Vc] = [Fc] (26) are the kinematic equations for dislocation self climb motion. A set of linear simulta- neous equations with a narrowly banded matrix [K ], which can be solved with standard matrix methods. Once (26) is solved, the generalized coordinates can be updated for a small time increment. The process is repeated for many increments to evolve the dislocation confguration. are the kinematic equations for dislocation self climb motion. A set of linear simulta- neous equations with a narrowly banded matrix [K ], which can be solved with standard matrix methods. Once (26) is solved, the generalized coordinates can be updated for a small time increment. The process is repeated for many increments to evolve the dislocation confguration. 2.3. Validation of the PLEBM Importantly, (26) is preferable to (1), derived with the linear element based method [1], in terms of improving the coeffcient matrix by eliminating the zero diagonal elements. We now compare the accuracy and computational cost of the paired-linear vs linear element. A benchmark study was performed of the annealing of an isolated elliptical prismatic dislo- cation loop in an infnite domain in α-iron, at T = 750 K. The initial semi-major axis is a0 = 400 nm and the semi-minor axis is b0 = 100 nm (the initial aspect ratio a0/b0 = 4). Other parameters of α-iron are obtained from table 2 in [1]. The loop is discretised into 30 straight segments. The evolution of the loop’s aspect ratio a/b over time is shown in fgure 2(a), the black curve corresponds to the result obtained with the PLEBM, and the red curve with the LEBM. The corresponding loop profles during some points in the evolution are also shown. The loop converges to a circle with R ≃200 nm to conserve the enclosed loop area. The accu- racy of the PLEBM is demonstrated by the excellent agreement between the two methods as shown in fgure 2(a), and the consistency with the previous theoretical [42] and numerical [45] studies. It also worth noting that, although the kinematic equations in the LEBM (1) and the PLEBM (25), give the same results, there is a signifcant reduction in the computational cost achieved with the PLEBM; as shown in fgure 2(b). The computational time required to calculate the climb velocity increases with the segment number with both methods. However, the growth rate of the LEBM is much higher than the PLEBM, which becomes signifcant as the segment number increases. 9 9 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 2. Comparison between the linear element based method (LEBM) and the paired-linear element based method (PLEBM) in terms of (a) accuracy, (b) effciency. Figure 2. Comparison between the linear element based method (LEBM) and the paired-linear element based method (PLEBM) in terms of (a) accuracy, (b) effciency. 2.4. Junction nodes 28 (2020) 055012 F Liu et al F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 summation can be expanded as, summation can be expanded as, λ4 X i J(i) 4 m(i) = λ4 [0 0 0 −1 1 1]   Vx c4 Vy c4 Vz c4 J(34) 4 J(45) 4 J(46) 4   . (29) (29) So that, So that, X I λI X i J(i) 4 m(i) = [λ]T[Cs][Vc] (30) (30) where [λ] is a P × 1 vector of Lagrange multipliers, which are interpreted as the chem- ical potential at junction nodes, with P indicating the total number of triple and quadru- ple junctions and [Cs] is a complimentary matrix. We now reconstruct the functional Π as, Π = 1 2[Vc]T[K][Vc] −[Vc]T[Fc] + [λ]T[Cs][Vc]. (31) (31) Taking the stationary value of Π, for any small perturbations δ[Vc] or δ[λ], Taking the stationary value of Π, for any small perturbations δ[Vc] or δ[λ], δΠ δ[Vc] = 0 (32) δΠ δ[λ] = 0. (33) (32) (33) Thus (26) can be rewritten to derive the general kinematic equations, [K] [Cs]T [Cs] [0] [Vc] [λ] = [Fc] [0] . (34) (34) Although the structure of (34) is similar to (1) from the LEBM, the size of the [0] matrix in the leading coeffcient matrix is now much smaller than in (1). Because multipliers are only introduced for the triple and quadruple junction nodes, which are a small proportion of the total number of nodes. Once (34) is solved, we can then obtain the nodal climb velocity and update the dislocation network for a small time increment, and remesh the topological connections [66] based on the nodal DDD framework. [66] based on the nodal DDD framework. 2.4. Junction nodes In the derivation of the kinematic equations for climb velocity, we have implicitly assumed that: (1) every node only connect with two segments and (2) every segment connected to the same node are on the same habit plane, as in fgure 1(a). However, for triple or quadruple nodes with more than two connections, as in fgure 1(b), (25) and (26) are no longer valid. At a junction node, the direction of the climb velocity N is unknown and (10) is no longer applicable. In this case, a full nodal velocity with 3 components must be used, which is required for nodes connected to segments which are on different habit planes. Moreover, fux continuity at a junction node requires mass fowing into the junction node must fow out. We therefore assign more degrees of freedom for the nodal fux, and further constrain the range of admissible felds by requiring that the fux of material into a triple or quadruple junction node is zero; i.e. at each junction node X i J(i) I m(i) = 0, (27) X i J(i) I m(i) = 0, (27) where the summation is over every segments i connected to node I. m(i) = ±1 so that, m(i) = 1 when J(i) fows away from node I, and m(i) = −1 when J(i) fows into node I. We account for these constraints by employing a series of Lagrange multipliers, one per junction node, so that the functional Π in (25) can be extended to include junction nodes as, where the summation is over every segments i connected to node I. m(i) = ±1 so that, m(i) = 1 when J(i) fows away from node I, and m(i) = −1 when J(i) fows into node I. We account for these constraints by employing a series of Lagrange multipliers, one per junction node, so that the functional Π in (25) can be extended to include junction nodes as, Π = 1 2[Vc]T[K][Vc] −[Vc]T[Fc] + X I λI X i J(i) I m(i) ! (28) (28) where the outer summation in the last term is performed overall triple and higher order junction nodes, I, in the dislocation network. Taking node 4 in fgure 1(b) as an example, the inner 10 10 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Modelling Simul. Mater. Sci. Eng. 3. Applications In this section, we describe how the variational principle presented in the previous section can be used to simulate dislocation evolution in engineering materials. The examples described in the following subsections are chosen to illustrate the proposed method, and to demonstrate that it produces accurate results by comparison with analytical results and available experimental observations reference [13]. 11 F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 Figure 3. An elliptical prismatic dislocation loop in an infnite elastic solid. Figure 3. An elliptical prismatic dislocation loop in an infnite elastic solid. 3.1. Analytical solution for self climb of elliptical prismatic loop 3.1. Analytical solution for self climb of elliptical prismatic loop It is of interest to construct a simple analytical model for the evolution of an elliptical pris- matic loop, rather than the rectangular loop derived previously [1]. Sketched in fgure 3 is an elliptical prismatic loop in an infnite elastic domain. The two semi-axes satisfy a0 ⩾b0. For a dislocation core diffusion problem, the enclosed area of the loop is conserved as the shape of the loop changes. Assume that the loop remains an ellipse during the evolution. Let r0 be the radius of a circle having the same volume as the elliptical loop. This process can be described by, x = a cosθ = r0α cosθ, (35) y = b sinθ = r0α−1 sinθ (36) (35) (35) (36) (36) y = b sinθ = r0α−1 sinθ where α is the shape parameter, with a/r0 = r0/b = α ⩾1, which evolves with time; α = 1 corresponds to a circular loop and α →∞corresponds to an edge dislocation dipole. At a fxed time, (35) and (36) trace the entire ellipse as θ varies in the interval [0, 2π]. Let r be the position vector of a point (x, y) on the loop, which has an outward unit vector normal n. As atoms diffuse along the prismatic loop in the core region, the dislocation segment moves at a velocity where α is the shape parameter, with a/r0 = r0/b = α ⩾1, which evolves with time; α = 1 corresponds to a circular loop and α →∞corresponds to an edge dislocation dipole. At a fxed time, (35) and (36) trace the entire ellipse as θ varies in the interval [0, 2π]. Let r be the position vector of a point (x, y) on the loop, which has an outward unit vector normal n. As atoms diffuse along the prismatic loop in the core region, the dislocation segment moves at a velocity vc = n · ˙r, (37) vc = n · ˙r, (37) where ˙r = dr/dt. With r = (x, y), where ˙r = dr/dt. With r = (x, y), r = (r0α cosθ, r0α−1 sinθ) (38) ˙r = (r0 ˙α cosθ, −r0α−2 ˙α sin θ) (39) n = −dy ds , dx ds = (−r0α−1 cosθ, −r0α sinθ)dθ ds . d js ds + vcbe = 0 (42) 3.1. Analytical solution for self climb of elliptical prismatic loop (40) (38) ˙r = (r0 ˙α cosθ, −r0α−2 ˙α sin θ) (39) n = −dy ds , dx ds = (−r0α−1 cosθ, −r0α sinθ)dθ ds . (40) (40) So that (37) can be rewritten as, So that (37) can be rewritten as, vc(θ) = −r0 2 ˙αα−1 cos 2θdθ ds . (41) (41) Denote js as the volumetric fux of atoms due to core diffusion. Mass conservation requires that, d js ds + vcbe = 0 d js ds + vcbe = 0 (42) (42) 12 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al where be is the edge component of the Burgers vector b, for a prismatic loop, be = b and where be is the edge component of the Burgers vector b, for a prismatic loop, be = b and js(θ) = − Z vcbeds = − Z vc(θ)bdθ ds dθ = r2 0b ˙αα−1 sin 2θ 2 + C, (43) (43) where the integration constant C = 0 due to the symmetry condition, js(θ = 0) = 0. Then the rate potential of the whole loop Ψ becomes, Ψ = Z L 0 j2 s 2Ds ds = 4 Z π 2 0 j2 s(θ) 2Ds dθ ds dθ (44) (44) where L = 4aE(k) is the circumference of the ellipse, with E(k) denoting the second complete elliptic integral with modulus k = √ a2 −b2/a. The complete elliptic integral of the frst and second kind, K and E, with modulus k = √ a2 −b2/a are, where L = 4aE(k) is the circumference of the ellipse, with E(k) denoting the second complete elliptic integral with modulus k = √ a2 −b2/a. The complete elliptic integral of the frst and second kind, K and E, with modulus k = √ a2 −b2/a are, K(k) = Z π 2 0 dθ p 1 −k2 sin2 θ (45) E(k) = Z π 2 0 p 1 −k2 sin2 θdθ. (46) (45) (46) So that, Ψ = r5 0 ˙α2α−2b2 2Ds Z π 2 0 sin2(2θ) p α2 sin2 θ + α−2 cos2 θdθ. 3.1. Analytical solution for self climb of elliptical prismatic loop (47) (47) Defne, Defne, I1 = r5 0α−2b2 2Ds (48) I2 = Z π 2 0 sin2 2θ p α2 sin2 θ + α−2 cos2 θdθ (49) = −4α −2 α4 −1 + α−4 E (k) + (1 + α−4)K (k) 15(α2 −α−2)2 . (50) I1 = r5 0α−2b2 2Ds (48) (48) I2 = Z π 2 0 sin2 2θ p α2 sin2 θ + α−2 cos2 θdθ (49) (49) = −4α −2 α4 −1 + α−4 E (k) + (1 + α−4)K (k) 15(α2 −α−2)2 . (50) (50) So that Ψ can be expressed as, So that Ψ can be expressed as, So that Ψ can be expressed as, So that Ψ can be expressed as, Ψ = I1I2 ˙α2 (51) Ψ = I1I2 ˙α2 (51) while the energy of an elliptical prismatic loop [67, 68] is given as, while the energy of an elliptical prismatic loop [67, 68] is given as, Eel = µb2 L 4π(1 −ν) ln 8r0 rc −1 −α2 + α−2 2α2 K E (52) (52) 13 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 4. Loop profles during the evolution with different initial aspect ratios. Figure 4. Loop profles during the evolution with different initial aspect ratios. where rc is the core radius of the dislocation. The rate of change of Gibbs free energy ˙G is, where rc is the core radius of the dislocation. The rate of change of Gibbs free energy ˙G is, ˙G = dEel dt = dEel dα dα dt = I3 ˙α, (53) (53) where I3 = dEel/dα is given in B. The variational functional now becomes, Π = Ψ + ˙G = I1I2 ˙α2 + I3 ˙α. (54) (54) Taking the stationary value, Taking the stationary value, ∂Π ∂˙α = 2I1I2 ˙α + I3 = 0 (55) ∂Π ∂˙α = 2I1I2 ˙α + I3 = 0 (55) s, ˙α = −I3 2I1I2 . (56) ∂Π ∂˙α = 2I1I2 ˙α + I3 = 0 (55) which gives, which gives, ˙α = −I3 2I1I2 . (56) (56) With the initial value α0 = p a0/b0, the ODE, (56), gives the evolution of the shape parameter α, to describe the loop profle. 3.1. Analytical solution for self climb of elliptical prismatic loop Changes in total dislocation energy tend to reduce the length and curvature of the loop, and mass conservation maintains the area enclosed by the loop, driving the elliptical loop into a circular profle. Next, we compare the results obtained from the analytic solution in (56) with the results from the PLEBM by conducting DDD simulations. Snapshots of an isolated elliptical loop with different initial aspect ratios: α2 0 = a0/b0 = 4 and α2 0 = a0/b0 = 16, evolving from an ellipse into a circle, are shown in fgure 4. It is expected that the loop will evolve into a circular loop with radius √a0b0 if mass is conserved, which is found to be the case with both the analytical and DDD results. With a0/b0 = 4, the analytic solution agrees well with the DDD solution. However, as the initial aspect ratio increases to a0/b0 = 16, with DDD the ellipse initially evolves into a bone- shape to reduce the high curvatureat the tips of the ellipse, and convergesinto a circle gradually, which is much more energetically favorable than the process described by the analytic solu- tion. This is because, in the deviation of the analytic solution, we implicitly assume that the loop remain an ellipse, so that a/r0 = r0/b during the evolution. This is a good approximation 14 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 5. Edge dipoles formed by (a) motion of screw dislocations with large jogs, (b) dislocations gliding on parallel slip planes, [76]. Figure 5. Edge dipoles formed by (a) motion of screw dislocations with large jogs, (b) dislocations gliding on parallel slip planes, [76]. for the loop evolution to some extent, which was also employed by Sun and co-authors [69] and Cocks [70] to describe the growth of an ellipsoidal void controlled by surface diffusion. However, it is too constrained to capture the evolution of an ellipse with high aspect ratio. The additional degrees of freedom in the DDD method allows it to more accurately simulate the physical process. This raises the question: how do loops evolve if the aspect ratio increases further? 3.2. Formation of prismatic loops from dislocation dipoles Dislocation dipoles and multi-poles are ubiquitous features of the dislocation debris in deformed crystals. Edge dipoles are commonly reported [13, 71, 72] as arising from the drag- ging out of jogs on screw dislocations, called jog dragging, as sketched in fgure 5(a), or by trapping edge dislocations of opposite sign moving on parallel basal slip planes, called edge trapping, as shown in fgure 5(b). Theses dipoles are sessile, they remain as prominent debris in deformed crystals, because they cannot self-annihilate by glide. Consequently they play a signifcant role in strain hardening and patterning in stressed metals. At high temperature or during annealing of deformed crystal, these dipoles can annihilate and break up into pris- matic loops by a conservative dislocation climb mechanism, which is primarily controlled by core diffusion along the dislocation line. In certain cases, strings of prismatic loops can form [73]. Attempts to study the annihilation of the dipoles mainly focus on an energetic analysis, in which the dipole fuctuation energetics are evaluated as a function of fuctuation wavelength and amplitude with either a square wave [74] or sinusoidal wave fuctuation [71]. Detailed loop formation mechanisms, by spontaneous break-up or sequential pinch-off from the dipole, still remain controversial [71, 74, 75]. It is therefore of interest to simulate the for- mation of prismatic loops from the break-up of dislocation dipoles or elongated loops with DDD. In the following subsection, we examine two simple representative situations as schemat- ically shown in fgure 5, with the mechanism of dipole annihilation of edge dipoles simulated using the PLEBM. We limit our attention initially to the pure self climb process of jog dragging dipoles shown in fgure 5(a). We then consider the break- up of edge trapping dipoles shown in fgure 5(b), which involves both glide and climb. 3.2.1. Formation of prismatic loops due to pinch-off from the end of a closed-end edge dipole. A jog dragging dipole as shown in fgure 5(a) is introduced as the initial confguration. The typical height of dipoles observed are 10 to 100 nm. The separation distance between two dislocations making up the dipole is therefore set at H = 20nm and the length of the dipole is set at L = 800 nm. The Burgers vector b = 1/2[111]. 3.2. Formation of prismatic loops from dislocation dipoles Snapshots of dislocation profles during the break-up of an open-ended edge dipole. Figure 7. Snapshots of dislocation profles during the break-up of an open-ended edge dipole. and opposite line direction are introduced as the initial confguration. Both of them are pinned at two ends and bow-out when a shear stress is applied. The parallel disloca- tion lines have the same length, L = 480 nm, with a vertical separation H = 10 nm and an in-plane separation D = 100 nm. Evolution of the open-ended dipole under a shear stress τ = Gb/L is simulated, where G is the shear modulus of the simulated material, and Gb/L is the critical shear stress for the activation of a Frank–Read source with length L. It is noteworthy that the evolution of this initial confguration, as shown in fgure 7(a), involves both glide in the slip plane and climb perpendicular to the original slip plane. A linear phonon-drag mobility law for bcc metals is, therefore, employed to calculate the glide velocity. To bridge the large time scale separation between glide and climb motion, as stated in the introduction, an adaptive time scheme [1] is adopted here. Thus, different time scales are used, small time increments ∼ns are used to deal with the glide steps, whereas much larger time increments ∼ms are used for the climb steps. Figure 7 shows snapshots of the dislocation evolution; a movie of the evolution process is available in the supplementary material S2. There is approximatelysix orders of magnitude dif- ference in the time between the different stages, which signifes the dominantmechanism (glide or climb) at different stages. Initially, the dislocations glide towards each other on their original slip planes at a very fast rate, until they are trapped by each other, as shown in fgure 7(b). This mechanism of dipole formation seems to be very frequent at the beginning of plastic deforma- tion when a large number of dislocation loops expand on parallel slip planes. This glide process only takes 9µs and results in a stable edge dipole. Then, the attraction from the anti-parallel edge segments facilitates self-climb normal to the original slip planes, at a much slower rate, as shown in fgure 7(c). This climb motion gives rise to the fuctuation of the dislocation line to conserve mass. 3.2. Formation of prismatic loops from dislocation dipoles The evolution of the total dislocation energy is plotted in fgure 6(a) as a function of time, and the snapshots of loop profles at points 15 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 6. Formation of prismatic loops from pinch-off at the end of a closed-end edge dipole. (a) The energy evolution over time; (b) Dislocation profles during evolution corresponding to the marked points A–D in (a); (c) Experimental observations from Tang et.al [13]. Figure 6. Formation of prismatic loops from pinch-off at the end of a closed-end edge dipole. (a) The energy evolution over time; (b) Dislocation profles during evolution corresponding to the marked points A–D in (a); (c) Experimental observations from Tang et.al [13]. marked A–D are demonstrated in fgure 6(b); available experimental observations from [13] are presented in fgure 6(c); a movie of the evolution process is available in the supplementary material S1. Our simulation results show that loops are pinched-offsequentially from the tip of the dipole rather than the dipole spontaneously breaking-up into a string of loops. This is consistent with the available experimental observations [13, 73]. Initially, the diffusion distance is short, as material fows much faster at the tips than at the middle of the dipole due to the higher curva- ture. The near tip region thus contracts due to mass conservation as shown in illustration B in fgure 6(b). This leads to the frst pinch-off from the two ends. The process repeats itself until the aspect ratio of the remaining loop in is too small to pinch-off, which will converge to a circle as described in fgure 4. The evolution of the dipole is a balance between the dislocation self-energy and the elastic interaction energy. Forming loops would decrease the dislocation line curvature which decreases the elastic interaction energy, but increase the total line length which reduces the dislocation self-energy. It is also worth noting that each pinch-off event is accompanied by a signifcant decrease in energy, indicating that the pinch-off of loops is energetically favorable. 3.2.2. Formation of prismatic loops from break-up of open-ended edge trapping dipoles. In this subsection, we simulate the break-up of an edge trapping dipole. As shown in fgure 5(b), two straight edge dislocations with the same Burgers vector b = 1/2[111] 16 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al Figure 7. 3.2. Formation of prismatic loops from dislocation dipoles In addition, according to (5), a mixed dislocation climbs a larger distance when consuming the same volume of material, compared to a pure edge dislocation. There- fore, the trapped dislocation dipole is broken at the corners where the climb rate is higher than at the middle, turning the edge dipole into an elongated interstitial prismatic loop and two helices, as shown in fgure 7(d). As one may expect, the elongated loop would continue to evolve into small loops by sequential pinching off as discussed in subsection 3.2.1. This allows further glide along the direction parallel to the Burgers vector. This mechanism of dipole 17 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al F Liu et al breakup is believed to play a signifcant role in releasing the strain-hardening in the subsequent annealing of deformed crystals [13]. The remaining helices may shrink or expand perpendic- ular to their Burgers vector by interacting with prismatic loops or by bulk-diffusion controlled climb [38, 77]. breakup is believed to play a signifcant role in releasing the strain-hardening in the subsequent annealing of deformed crystals [13]. The remaining helices may shrink or expand perpendic- ular to their Burgers vector by interacting with prismatic loops or by bulk-diffusion controlled climb [38, 77]. Acknowledgments We thank the Engineering and Physical Sciences Research Council (EPSRC) for funding through project grant EP/R013136/1 and Fellowship grant EP/N007239/1. 4. Conclusions Dislocations can provide short circuit diffusion paths for atoms. This fast atomic trans- port, along the dislocation core region, known as core diffusion, accelerates the diffusion by more than three orders of magnitude compared to lattice diffusion. This allows dislocation motions perpendicular to the original slip system, known as dislocation self-climb (conser- vative climb) and is of particular importance in low-temperature creep and post-irradiation annealing. A variational principle is presented for the analysis of problems in which fast dislocation core diffusion is the dominant mechanism for material redistribution. A new fnite element dis- cretisation method is employed to accelerate the simulation of dislocation self-climb. We have demonstrated here, that the paired-linear element based method has a much higher computa- tional effciency compared with the previous linear element based method [1]. This becomes more evident as the segment number increases. The acceleration in computation enables the simulation of collective dislocation motion, including both glide and self-climb, based on the current method. An analytical solution for the evolution of an isolated elliptical prismatic loop was derived based on the variational principle, and compared with the numerical method. The analytical solution provides a good approximation for predicting the evolution of elliptical loops with small aspect ratios. However, as the aspect ratio increases, the high curvature at the two end of the ellipse cause the loop to evolve into a bone-like shape initially, and gradually converge into a circular loop, which the analytic solution fails to predict. This interesting discrepancy leads to the study of edge dipoles or elongated prismatic loops, for which the aspect ratios are much higher. The evolution of two representative edge dipoles, the close-ended jog drag- ging dipole and the open-ended edge trapping dipole, were examined with the new DDD method. Results show that a string of prismatic loops is formed by the sequential pinching- off of the dipole, rather than spontaneous break-up, which agrees well with experimental observations. Appendix A. Derivation of the viscosity matrix of segment (13) [k(13)] The viscosity matrix of segment (13) obtained from equation (20) is, k11 = l D (β1)2(leb)2 96 , k12 = k21 = −l D 9 160β1leb k11 = l D (β1)2(leb)2 96 , (A.1) k12 = k21 = −l D 9 160β1leb (A.2) (A.1) (A.2) 18 18 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 F Liu et al k13 = k31 = −l D β1β3 (leb)2 160 (A.3) k14 = k41 = −l D 13 480β1leb (A.4) k22 = l D 23 60 (A.5) k23 = k32 = l D 13 480β3leb (A.6) k24 = k42 = l D 7 60 (A.7) k33 = l D 1 96(β3)2(leb)2 (A.8) k34 = k43 = l D 9 160β3leb (A.9) k44 = l D 23 60 (A.10) k13 = k31 = −l D β1β3 (leb)2 160 (A.3) k14 = k41 = −l D 13 480β1leb (A.4) k22 = l D 23 60 (A.5) k23 = k32 = l D 13 480β3leb (A.6) k24 = k42 = l D 7 60 (A.7) k33 = l D 1 96(β3)2(leb)2 (A.8) k34 = k43 = l D 9 160β3leb (A.9) k44 = l D 23 60 (A.10) where D = D(13) is the generalized diffusivity along segment (13). Since the dummy node 2 is chosen to be located at the middle of segment (13), l(12) = l(23) = l(13) 2 . Assuming l(13) = l, b(13) = b, l(13)sin θ(13) = le, the above matrix can be rewritten as, [k(13)] = l D   1 96(β1)2(leb)2 −9 160β1leb −1 160β1β3(leb)2 −13 480β1leb 23 60 13 480β3leb 7 60 1 96(β3)2(leb)2 9 160β3leb sym 23 60   . (A.11) As illustrated in (A.11), the element viscosity matrix [k(13)] depends not only on the quantities of segment (13), such as l(13), θ(13) and D(13), but also on how the segment is connected to other segments, through β1 and β3, leading to connections between elements when assembling the global viscosity matrix. As illustrated in (A.11), the element viscosity matrix [k(13)] depends not only on the quantities of segment (13), such as l(13), θ(13) and D(13), but also on how the segment is connected to other segments, through β1 and β3, leading to connections between elements when assembling the global viscosity matrix. ORCID iDs Fengxian Liu https://orcid.org/0000-0002-0216-3853 Alan C F Cocks https://orcid.org/0000-0002-6245-3406 Simon P A Gill https://orcid.org/0000-0001-9223-5410 Edmund Tarleton https://orcid.org/0000-0001-6725-9373 Fengxian Liu https://orcid.org/0000-0002-0216-3853 Alan C F Cocks https://orcid.org/0000-0002-6245-3406 Simon P A Gill https://orcid.org/0000-0001-9223-5410 Edmund Tarleton https://orcid.org/0000-0001-6725-9373 Appendix B. Expression for I3 in (53) The explicit expression of I3 from equation (55) is, The explicit expression of I3 from equation (55) is, I3 = dEel/dα = µb2 4π(1 −ν) r0 α4(α4 −1) [A1K(k) + A2E(k)] (B.1) (B.1) where 19 19 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 k = p 1 −α−4 (B.2) A1 = −2α8 + 20α4 −8α4ln 8r0 rc −2 (B.3) A2 = α4 −8α4 + 4 α4 + 1 ln 8r0 rc −8 . (B.4) k = p 1 −α−4 (B.2) A1 = −2α8 + 20α4 −8α4ln 8r0 rc −2 (B.3) A2 = α4 −8α4 + 4 α4 + 1 ln 8r0 rc −8 . (B.4) References [1] Liu F, Cocks A C F and Tarleton E 2020 J. Mech. Phys. Solids 135 103783 [2] Shewmon P 2016 Diffusion in Solids (Berlin: Springer) [3] Xie D G, Nie Z Y, Shinzato S, Yang Y Q, Liu F X, Ogata S, Li J, Ma E and Shan Z W 2019 Nat. Commun. 10 1–8 [4] Swinburne T D, Arakawa K, Mori H, Yasuda H, Isshiki M, Mimura K, Uchikoshi M and Dudarev S L 2016 Sci. Rep. 6 30596 [5] Legros M, Dehm G, Arzt E and Balk T J 2008 Science 319 1646–9 [5] Legros M, Dehm G, Arzt E and Balk T J 2008 S [6] Bowden F and Singer K 1969 Nature 222 977 [7] Wang W, Wang M, Ambrosi E, Bricalli A, Laudato M, Sun Z, Chen X and Ielmini D 2019 Nat. Commun. 10 81 [8] Kenrick P 1968 Nature 217 1249 [9] Balluff R 1982 Metall. Trans. B 13 527–53 [10] Kaur I, Gust W and Mishin Y 1995 Fundamentals of Grain and Interphase Boundary Diffusion (New York: Wiley) [10] Kaur I, Gust W and Mishin Y 1995 Fundam 136–53 [18] Keralavarma S and Benzerga A 2015 J. Mech. Phys. Solids 82 1–22 [19] Rajaguru M and Keralavarma S 2018 Modelling Simul. Mater. Sci. Eng. 26 035007 aguru M and Keralavarma S 2018 Modelling Simul. Mater. [20] Garbrecht M, Saha B, Schroeder J L, Hultman L and Sands T D 2017 Sci. Rep. 7 46092 Garbrecht M, Saha B, Schroeder J L, Hultman L and Sands T Arakawa K, Amino T and Mori H 2011 Acta Mater. 59 14 22] Pandey A, Mishra R, Paradkar A and Mahajan Y 1997 [22] Pandey A, Mishra R, Paradkar A and Mahajan Y 1997 Acta Mater. 45 1297–306 [23] Moll S, Jourdan T and Lefaix-Jeuland H 2013 Phys. Rev. Lett. 111 015503 [24] Burton B and Speight M 1986 Phil. Mag. A 53 38 [25] Balluff R 1970 Phys. Status Solidi b 42 11–34 New York: Wiley) y [11] Mishin Y and Herzig C 1999 Mater. Sci. Eng.: A 260 55–71 [12] Kolobov Y, Valiev R and Ivanov M 2007 Grain Boundary Diffusion and Properties of Nanostruc- tured Materials Mater. Sci. Eng.: A 260 55–71 [12] Kolobov Y, Valiev R and Ivanov M 2007 Grain Bou tured Materials Mater. Sci. Eng.: A 260 55–71 [12] Kolobov Y, Valiev R and Ivanov M 2007 Grain Boundary Diffusion and Properties of Nanostruc- tured Materials Mater. Sci. Eng.: A 260 55–71 [12] Kolobov Y, Valiev R and Ivanov M 2007 Grain Boundary Diffusion and Properties of Nanostruc- tured Materials Mater. Sci. Eng.: A 260 55–71 tured Materials Mater. Sci. Eng.: A 260 55–71 g [13] Tang X, Lagerlöf K P D and Heuer A H 2003 J. Am. Ceram. Soc. 86 560–5 [13] Tang X, Lagerlöf K P D and Heuer A H 2003 J. Am. Ceram. Soc. 86 560–5 [13] Tang X, Lagerlöf K P D and Heuer A H 2003 J. Am. g g 14] Pun G P and Mishin Y 2009 Acta Mater. 57 5531–42 [14] Pun G P and Mishin Y 2009 Acta Mater. 57 5531–42 [ ] [15] Bakó B, Clouet E, Dupuy L M and Blétry M 2011 Phil. Mag. 91 3173–91 [ ] [15] Bakó B, Clouet E, Dupuy L M and Blétry M 2011 Phil. Mag. 91 3173–91 5] Bakó B, Clouet E, Dupuy L M and Blétry M 2011 Phil. [16] Mordehai D, Clouet E, Fivel M and Verdier M 2008 Phil. Mag. 88 899–925 [17] Geers M, Cottura M, Appolaire B, Busso E P, Forest S and Villani A 2014 J. Mech. Phys. Solids 70 136–53 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 [31] Schwink C and Nortmann A 1997 Mater. Sci. Eng.: A 234 1–7 32] Blavette D, Cadel E, Fraczkiewicz A and Menand A 1 [33] Mortlock A 1960 Acta Metall. 8 132–4 [34] Hu S and Chen L 2001 Acta Mater. 49 463–72 [35] Johnson C 1960 Phil. Mag. 5 1255–65 g [36] Keralavarma S M, Cagin T, Arsenlis A and Benzerga A A 2012 Phys. Rev. Lett. 109 265504 36] Keralavarma S M, Cagin T, Arsenlis A and Benzerga A g g [37] Mordehai D, Clouet E, Fivel M and Verdier M 2009 Conf. Ser.: Mater. Sci. Eng. 3 012001 rdehai D, Clouet E, Fivel M and Verdier M 2009 Conf. Ser. [38] Liu F, Liu Z l, Lin P and Zhuang Z 2017 Int. J. Plast. 92 2–18 [39] Liu F, Liu Z, Pei X, Hu J and Zhuang Z 2017 Int. J. Plast. 99 102–19 9] Liu F, Liu Z, Pei X, Hu J and Zhuang Z 2017 Int. J. Plas g [40] Dudarev S and Ma P W 2018 Phys. Rev. Mater. 2 033602 [ ] y [41] Gao Y and Cocks A C F 2009 Acta Mech. Solida Sin. 22 426–35 [42] Kroupa F and Price P 1961 Phil. Mag. 6 243–7 [ ] p g [43] Kroupa F, Silcox J and Whelan M 1961 Phil. Mag. 6 971–8 [43] Kroupa F, Silcox J and Whelan M 1961 Phil. Mag p [44] Ferroni F, Yi X, Arakawa K, Fitzgerald S P, Edmondson P D and Roberts S G 2015 Acta Mater. 90 380–93 5] Niu X, Gu Y and Xiang Y 2019 Int. J. Plast. 120 262–77 [46] Huang J, Meyer M and Pontikis V 1989 Phys. Rev. Lett. 63 628 [47] Fang Q and Wang R 2000 Phys. Rev. B 62 9317 [48] Roy S and Mordehai D 2017 Acta Mater. 127 351–8 [ ] y [49] Okita T, Hayakawa S, Itakura M, Aichi M, Fujita S and Suzuki K 2016 Acta Mater. 118 342–9 [49] Okita T, Hayakawa S, Itakura M, Aichi M, Fujita S a [50] Geslin P A, Appolaire B and Finel A 2015 Phys. Rev. Lett. 115 265501 [51] Niu X, Luo T, Lu J and Xiang Y 2017 J. Mech. Phys. (Amsterdam: Pergamon) pp 107–24 g pp [63] Suo Z and Wang W 1994 J. Appl. Phys. 76 3410–21 g pp y [64] Gill S and Cocks A C F 1996 Acta Mater. 44 4777–89 g pp y 64] Gill S and Cocks A C F 1996 Acta Mater. 44 4777–89 [ ] [65] Pan J, Cocks A C F and Kucherenko S 1997 Proc. R. Soc. London, Ser. A 453 2161–84 [66] Bulatov V and Cai W 2006 Computer Simulations of Dislocations (Oxford: Oxford University Press) vol 3 [67] Li J C and Liu G C 1966 Phil. Mag. 14 413–4 [68] Bacon D 1966 Phil. Mag.: A Journal of Theoretical Experimental and Applied Physics 14 715–29 [68] Bacon D 1966 Phil. Mag.: A Journal of Theoretical g f p pp y [69] Sun B, Suo Z and Evans A 1994 J. Mech. Phys. Solids 42 1653–77 g f p [69] Sun B, Suo Z and Evans A 1994 J. Mech. Phys. Solids 42 1653–77 69] Sun B, Suo Z and Evans A 1994 J. Mech. Phys. Solids [70] Cocks A C F 2002 Continuum Damage M [70] Cocks A C F 2002 Continuum Damage Mechanics of Materials and Structures ed O Allix and F Hild (Amsterdam: Elsevier) pp 49–77 [70] Cocks A C F 2002 Continuum Damage Mechanics of Materials and Structures ed O Allix and F Hild (Amsterdam: Elsevier) pp 49–77 Hild (Amsterdam: Elsevier) pp 49–77 6] Turnbull J 1970 Phil. Mag. 21 83–94 [26] Turnbull J 1970 Phil. Mag. 21 83–94 27] Narayan J and Washburn J 1972 Phil. Mag. 26 1179–9 [28] Luo W, Shen C and Wang Y 2007 Acta Mater. 55 2579–86 [28] Luo W, Shen C and Wang Y 2007 Acta Mater. 55 2579–8 [29] Clouet E and Soisson F 2010 C. R. Phys. 11 226–35 [ ] y [30] Curtin W A, Olmsted D L and Hector L G Jr 2006 Nat. Mater. 5 875 [30] Curtin W A, Olmsted D L and Hector L G Jr 2006 Nat. Mater. 5 875 20 F Liu et al F Liu et al Modelling Simul. Mater. Sci. Eng. 28 (2020) 055012 Solids 99 242–58 [52] Caillard D and Martin J L 2003 Thermally Activated Mechanisms in Crystal Plasticity vol 8 (Oxford: Elsevier) [53] Cocks A 1996 J. Mech. Phys. Solids 44 1429–52 [54] Cocks A C, Gill S P and Pan J 1998 Advances in Applied Mechanics vol 36 ed E Van der Giessen and T Y Wu (Amsterdam: Elsevier) pp 81–162 [54] Cocks A C, Gill S P and Pan J 1998 Advances in A and T Y Wu (Amsterdam: Elsevier) pp 81–162 [55] Arsenlis A, Cai W, Tang M, Rhee M, Oppelstrup T, Hommes G, Pierce T G and Bulatov V V 2007 Modelling Simul. Mater. Sci. Eng. 15 553 [56] Sills R B, Kuykendall W P, Aghaei A and Cai W 2016 Multiscale Materials Modeling for Nanomechanics vol 245 ed C Weinberger and G Tucker (Berlin: Springer) pp 53–87 [57] Sills R B, Aghaei A and Cai W 2016 Modelling Simul. Mater. Sci. Eng. 24 045019 ertin N, Aubry S, Arsenlis A and Cai W 2019 Modelling Si [59] Hu J, Zhuang Z, Liu F, Liu X and Liu Z 2019 Comput. Mater. Sci. 159 86–94 [60] Cui Y, Po G and Ghoniem N M 2018 Int. J. Plast. 104 54–67 [61] Ferroni F, Tarleton E and Fitzgerald S 2014 J. Comput. Phys. 272 619–28 [61] Ferroni F, Tarleton E and Fitzgerald S 2014 J. Com [62] Needleman A and Rice J 1983 Perspectives in Creep Fracture ed M Ashby and L Brown (Amsterdam: Pergamon) pp 107–24 Hild (Amsterdam: Elsevier) pp 49–77 [71] Lagerlöf K, Mitchell T and Heuer A 1989 Acta [71] Lagerlöf K, Mitchell T and Heuer A 1989 Acta Metall. 37 3315–25 [72] Basinski Z 1964 Discuss. Faraday Soc. 38 93–102 y [73] Pletka B, Mitchell T and Heuer A 1982 Acta Metall. 30 147–56 [73] Pletka B, Mitchell T and Heuer A 1982 Acta Metall. 30 147–56 73] Pletka B, Mitchell T and Heuer A 1982 Acta Metall. 3 74] Phillips D, Pletka B, Heuer A and Mitchell T 1982 Act [74] Phillips D, Pletka B, Heuer A and Mitchell T 1982 Acta Metall. 30 491–8 [75] Segall R, Partridge P G and Hirsch P 1961 Phil. Mag. 6 1493–513 [76] Kroupa F 1966 Le Journal de Physique Colloques 27 C3–154 76] Kroupa F 1966 Le Journal de Physique Colloques 27 C [77] Haley J, Liu F, Tarleton E, Cocks A C F, Odette G, Lozano-Perez S and Roberts S 2019 Acta Mater. y 181 173–84 21
https://openalex.org/W4384069194	https://egusphere.copernicus.org/preprints/2023/egusphere-2023-303/egusphere-2023-303.pdf	English	null	Coupled hydrological and hydrodynamic modelling application for climate change impact assessment in the Nemunas river watershed–Curonian Lagoon–southeastern Baltic Sea continuum	Ocean science	2,023	cc-by	20,475	ERROR: type should be string, got "https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Coupled hydrological and hydrodynamic modelling application for \nclimate change impact assessment in the Nemunas River watershed–\nCuronian Lagoon–south-eastern Baltic Sea continuum \nRasa Idzelytė \n1, Natalja Čerkasova \n1,2, Jovita Mėžinė \n1, Toma Dabulevičienė \n1, Artūras \nRazinkovas-Baziukas 1, Ali Ertürk 1,3, Georg Umgiesser 1,4 \n5 \n1Marine Research Institute, Klaipėda University, Klaipėda, 92294, Lithuania \n2Texas A&M AgriLife Research, Blackland Research and Extension Center, Temple, TX 76502, USA \n3Department of Inland Water Resources and Management, Istanbul University, Istanbul, 34134, Turkey \n4CNR–National Research Council of Italy, ISMAR–Institute of Marine Sciences, Venice, 30122, Italy Correspondence to: Georg Umgiesser (georg.umgiesser@ismar.cnr.it) Correspondence to: Georg Umgiesser (georg.umgiesser@ismar.cnr.it) \n10 Abstract. We analyse the cumulative impacts of climate change in a complex basin-lagoon-sea system continuum, which \ncovers the Nemunas River basin, Curonian Lagoon, and the south-eastern part of the Baltic Sea. A unique state-of-the-art \ncoupled modelling system, consisting of hydrological and hydrodynamic models, has been developed and used for this \npurpose. Results of four regional downscaled models from the Rossby Centre high-resolution regional atmospheric climate Abstract. We analyse the cumulative impacts of climate change in a complex basin-lagoon-sea system continuum, which \ncovers the Nemunas River basin, Curonian Lagoon, and the south-eastern part of the Baltic Sea. A unique state-of-the-art \ncoupled modelling system, consisting of hydrological and hydrodynamic models, has been developed and used for this \npurpose. Results of four regional downscaled models from the Rossby Centre high-resolution regional atmospheric climate model have been bias-corrected using in situ measurements, and were used as forcing to assess the changes that the continuum \n15 \nwill undergo until the end of this century. Results show that the Curonian Lagoon will be subjected to higher river discharges that in turn increase the outgoing fluxes \ninto the Baltic Sea. Through these higher fluxes, both the water residence time and saltwater intrusion event frequency will \ndecrease. Most of these changes will be more pronounced in the northern part of the lagoon, which is more likely to be 15 influenced by the variations in the Nemunas River discharge. The southern part of the lagoon will experience lesser changes. 20 \nWater temperatures in the entire lagoon and the south-eastern Baltic Sea will steadily increase, and salinity values will \ndecrease. However, the foreseen changes in physical characteristics are not of the scale suggesting significant shifts in the \necosystem functioning, but are expected to manifest in some quantitative alterations in the nutrient retention capacity. However, some ecosystem services such as ice fishing are expected to vanish completely due to the loss of ice cover. Moreover, the Curonian Lagoon and the south-\neastern Baltic Sea are covered by a finite element grid that resolves its hydrodynamic and physical parameters. A seamless \nintegration of two models represents the watershed–lagoon–sea continuum for the simulations of the present state and future \nscenarios based on the ensemble of meteorological datasets produced by climate models. In this article, we present a framework of coupled hydrological and hydrodynamic models that are able to simulate the \n50 \ncontribution of the entire Nemunas River watershed to the Curonian Lagoon. Moreover, the Curonian Lagoon and the south-\neastern Baltic Sea are covered by a finite element grid that resolves its hydrodynamic and physical parameters. A seamless \nintegration of two models represents the watershed–lagoon–sea continuum for the simulations of the present state and future \nscenarios based on the ensemble of meteorological datasets produced by climate models. One of the new key findings of the 2nd climate change assessment \n40 \n(BACC II Author Team, 2015) was that due to the large bias in the water balance projections the projected changes of the \nBaltic Sea salinity remained uncertain, i.e., it is unclear if the Baltic Sea becomes more or less saline (von Storch et al., 2015). Nevertheless, climate change is consequently going to affect marine ecosystem and may reduce its resilience. The magnitude and severity of climate change impacts will vary considerably within the Baltic Sea region (Graham, 2004) salinity values will decrease (BACC Author Team, 2008). One of the new key findings of the 2nd climate change assessment \n40 \n(BACC II Author Team, 2015) was that due to the large bias in the water balance projections the projected changes of the \nBaltic Sea salinity remained uncertain, i.e., it is unclear if the Baltic Sea becomes more or less saline (von Storch et al., 2015). Nevertheless, climate change is consequently going to affect marine ecosystem and may reduce its resilience. The magnitude and severity of climate change impacts will vary considerably within the Baltic Sea region (Graham, 2004) g\ny\ng\np\ny\ny\ng\n(\n,\n)\nmaking basin-specific studies of high importance for understanding climate change induced impacts on a local scale. Likewise, \n45 \nwith increasing computational capacity, the multidisciplinary modelling studies were progressively evolving during the past \nfew decades (Rodrigues et al., 2015). Thus, the need for a combination of different aquatic processes became essential. The \ncoupling of different numerical models allows researchers to replicate the environmental processes for better assessment of \nthe physical parameters’ response to changing climate. making basin-specific studies of high importance for understanding climate change induced impacts on a local scale. Likewise, \n45 \nwith increasing computational capacity, the multidisciplinary modelling studies were progressively evolving during the past \nfew decades (Rodrigues et al., 2015). Thus, the need for a combination of different aquatic processes became essential. The \ncoupling of different numerical models allows researchers to replicate the environmental processes for better assessment of \nthe physical parameters’ response to changing climate. In this article, we present a framework of coupled hydrological and hydrodynamic models that are able to simulate the \n50 \ncontribution of the entire Nemunas River watershed to the Curonian Lagoon. 1 Introduction \n25 Climate change and increasing anthropogenic pressures are expected to cause drastic changes in the global environment in the \nnear future, and thus may affect the discharge and hydrological regime of rivers (Middelkoop et al., 2001), lagoons \n(Jakimavičius and Kriaučiūnienė, 2020; Anthony et al., 2009) and regional seas (Holt et al., 2016). The Baltic Sea, together \nwith its largest lagoon, the Curonian Lagoon, and the Nemunas River watershed are no exceptions to this trend. The impact of \nclimate change is already evident here through the changes in ice conditions in the sea (Merkouriadi and Leppäranta, 2014) \n30 Climate change and increasing anthropogenic pressures are expected to cause drastic changes in the global environment in the \nnear future, and thus may affect the discharge and hydrological regime of rivers (Middelkoop et al., 2001), lagoons \n(Jakimavičius and Kriaučiūnienė, 2020; Anthony et al., 2009) and regional seas (Holt et al., 2016). The Baltic Sea, together \nwith its largest lagoon, the Curonian Lagoon, and the Nemunas River watershed are no exceptions to this trend. The impact of climate change is already evident here through the changes in ice conditions in the sea (Merkouriadi and Leppäranta, 2014) \n30 1 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. and the lagoon (Idzelytė et al., 2019), increase in sea surface temperature (Kniebusch et al., 2019; Belkin, 2009) or \nredistribution of river runoff over the year (Meier et al., 2022a). and the lagoon (Idzelytė et al., 2019), increase in sea surface temperature (Kniebusch et al., 2019; Belkin, 2009) or \nredistribution of river runoff over the year (Meier et al., 2022a). increasingly important for understanding and projecting climate change impacts, also for supporting the formulation of \n35 \nmanagement measures (Vohland et al., 2014). The first climate change scenario simulations for the Baltic Sea were carried out more than 20 years ago (Meier, 2002a, b; \nOmstedt et al., 2000). ln 2008, the first assessment of the climate change in the Baltic Sea region was prepared with the main \nconclusions that by the end of 2100 the annual mean sea surface temperature will be 2–4 °C higher while the ice cover and increasingly important for understanding and projecting climate change impacts, also for supporting the formulation of \n35 \nmanagement measures (Vohland et al., 2014). The first climate change scenario simulations for the Baltic Sea were carried out more than 20 years ago (Meier, 2002a, b; \nOmstedt et al., 2000). ln 2008, the first assessment of the climate change in the Baltic Sea region was prepared with the main \nconclusions that by the end of 2100 the annual mean sea surface temperature will be 2–4 °C higher while the ice cover and increasingly important for understanding and projecting climate change impacts, also for supporting the formulation of \n35 \nmanagement measures (Vohland et al., 2014). The first climate change scenario simulations for the Baltic Sea were carried out more than 20 years ago (Meier, 2002a, b; \nOmstedt et al., 2000). ln 2008, the first assessment of the climate change in the Baltic Sea region was prepared with the main \nconclusions that by the end of 2100 the annual mean sea surface temperature will be 2–4 °C higher while the ice cover and 35 salinity values will decrease (BACC Author Team, 2008). The study site covers part of the south-eastern Baltic Sea region together with the Curonian Lagoon and the Nemunas River \nwatershed (Fig. 1). The drainage basin is shared by five countries: Belarus (48%), Lithuania (46%) and the other 6% is shared \nby Kaliningrad region of Russian Federation, Poland and Latvia (Gailiušis et al., 2001). The river enters the Curonian Lagoon \nat its central-eastern part and divides the system into two sub-basins; its southern part belongs to the Russian Federation \n60 \n(Kaliningrad region, 76%) and its northern part to Lithuania (24%). 2.1 Study area The Curonian Lagoon is a shallow water body with an average depth of 3.8 m (Gasiūnaitė et al., 2008) located on the south- The Nemunas River is the fourth largest river draining into the Baltic Sea and the largest tributary entering the Curonian \n65 \nLagoon, accounting for ~96% of the total freshwater input into it (Jakimavičius and Kovalenkovienė, 2010). The annual \naverage discharge is 16.4 km3 yr−1 (518.3 m3 s−1), calculated for the period 2012–2016 (Vybernaite-Lubiene et al., 2018), and \nthe long-term average of 21.847 km3 yr−1 (692 m3 s−1) for the period 1960–2007 (Jakimavičius and Kovalenkovienė, 2010). The Curonian Lagoon is a shallow water body with an average depth of 3.8 m (Gasiūnaitė et al., 2008) located on the south- eastern coast of the Baltic Sea. With a total area of ~1584 km2 (Žaromskis, 1996), it is considered the largest lagoon in Europe. 70 \nIt is a transnational lagoon whose northern part is influenced by the Nemunas River flow and brackish water intrusions from \nthe Baltic Sea, while the southern part is more stagnant, where hydrodynamics are influenced by the wind (Vybernaite-Lubiene \net al., 2022; Ferrarin et al., 2008). The intrusions of the sea water through the narrow (0.4–1.1 km) Klaipėda Strait occur in the \nnorthern part of the lagoon and reach up to 20 km southward from the strait (Zemlys et al., 2013). eastern coast of the Baltic Sea. With a total area of ~1584 km2 (Žaromskis, 1996), it is considered the largest lagoon in Europe. 70 \nIt is a transnational lagoon whose northern part is influenced by the Nemunas River flow and brackish water intrusions from \nthe Baltic Sea, while the southern part is more stagnant, where hydrodynamics are influenced by the wind (Vybernaite-Lubiene \net al., 2022; Ferrarin et al., 2008). The intrusions of the sea water through the narrow (0.4–1.1 km) Klaipėda Strait occur in the \nnorthern part of the lagoon and reach up to 20 km southward from the strait (Zemlys et al., 2013). The coastal area of the south-eastern Baltic Sea is relatively shallow, up to 20 m depth. The lowest water temperatures here \n75 \nare observed in winter months (in Jan and Feb the average water temperature is ~2 °C), while in summer months, e.g., Jul and \nAug, the average water temperature is ~18–19 °C (Kozlov et al., 2014). 2.1 Study area The study site covers part of the south-eastern Baltic Sea region together with the Curonian Lagoon and the Nemunas River \nwatershed (Fig. 1). The drainage basin is shared by five countries: Belarus (48%), Lithuania (46%) and the other 6% is shared \nby Kaliningrad region of Russian Federation, Poland and Latvia (Gailiušis et al., 2001). The river enters the Curonian Lagoon \nat its central-eastern part and divides the system into two sub-basins; its southern part belongs to the Russian Federation \n60 \n(Kaliningrad region, 76%) and its northern part to Lithuania (24%). The study site covers part of the south-eastern Baltic Sea region together with the Curonian Lagoon and the Nemunas River \nwatershed (Fig. 1). The drainage basin is shared by five countries: Belarus (48%), Lithuania (46%) and the other 6% is shared \nby Kaliningrad region of Russian Federation, Poland and Latvia (Gailiušis et al., 2001). The river enters the Curonian Lagoon \nat its central-eastern part and divides the system into two sub-basins; its southern part belongs to the Russian Federation \n60 \n(Kaliningrad region, 76%) and its northern part to Lithuania (24%). 60 at its central-eastern part and divides the system into two sub-basins; its southern part belongs to the Russian Federation \n60 \n(Kaliningrad region, 76%) and its northern part to Lithuania (24%). 2 2 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Figure 1. The Nemunas River watershed (coloured polygons in the mainland are major sub-basins), Curonian Lagoon, and south-\neastern Baltic Sea with respect to the entire sea area. Basemap source: ESRI. Figure 1. The Nemunas River watershed (coloured polygons in the mainland are major sub-basins), Curonian Lagoon, and south-\neastern Baltic Sea with respect to the entire sea area. Basemap source: ESRI. The Nemunas River is the fourth largest river draining into the Baltic Sea and the largest tributary entering the Curonian \n65 \nLagoon, accounting for ~96% of the total freshwater input into it (Jakimavičius and Kovalenkovienė, 2010). The annual \naverage discharge is 16.4 km3 yr−1 (518.3 m3 s−1), calculated for the period 2012–2016 (Vybernaite-Lubiene et al., 2018), and \nthe long-term average of 21.847 km3 yr−1 (692 m3 s−1) for the period 1960–2007 (Jakimavičius and Kovalenkovienė, 2010). 2.2.1 Climate projection data was constructed according to the methodology presented in the “Renewal of a River Basin Districts Management Plans and \n100 \nProgrammes of Measures” project report (PAIC, 2015) and were split in three groups corresponding to the closest \nmeteorological stations in Lithuania of which measured data were used for bias correction. The bias correction of climate input \ndata for the hydrodynamic model was performed from the averaged data of three meteorological stations closest to the northern \npart of the lagoon that describe the coastal climate conditions. was constructed according to the methodology presented in the “Renewal of a River Basin Districts Management Plans and \n100 \nProgrammes of Measures” project report (PAIC, 2015) and were split in three groups corresponding to the closest \nmeteorological stations in Lithuania of which measured data were used for bias correction. The bias correction of climate input \ndata for the hydrodynamic model was performed from the averaged data of three meteorological stations closest to the northern \npart of the lagoon that describe the coastal climate conditions. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. 2.2 Data \n80 \n2.2.1 Climate projection data \nTo feed our developed modelling system we used meteorological forcing data acquired from CORDEX (Coordinated Regional \nDownscaling Experiment) scenarios for Europe from the Rossby Centre high-resolution regional atmospheric climate model \n(RCA4), which consisted of four sets of simulations (downscaling) driven by four global climate models (specified in Table \n1). The datasets consist of cloud cover, solar radiation, precipitation, surface air pressure, relative humidity, air temperature, \n85 \nand wind speed and direction, and covers a historical period of 1970–2005 and projection period of 2006–2100. The projections \nare derived according to two Representative Concentration Pathway (RCP) scenarios: RCP4.5 and RCP8.5. Abbreviation \nModel \nInstitution \nICHEC \nEC-Earth \nIrish Centre for High-End Computing \nIPSL \nIPSL-CM 5A-MR \nThe Institut Pierre-Simon Laplace \nMOHC \nHadGEM2-ES \nMet Office Hadley Centre \nMPI \nMPI-ESM-LR \nMax Planck Institute for Meteorology \nTable 1. Summary of the global climate model data sources. 2.2 Data \n80 \n2.2.1 Climate projection data \nTo feed our developed modelling system we used meteorological forcing data acquired from CORDEX (Coordinated Regional \nDownscaling Experiment) scenarios for Europe from the Rossby Centre high-resolution regional atmospheric climate model \n(RCA4), which consisted of four sets of simulations (downscaling) driven by four global climate models (specified in Table \n1). The datasets consist of cloud cover, solar radiation, precipitation, surface air pressure, relative humidity, air temperature, \n85 \nand wind speed and direction, and covers a historical period of 1970–2005 and projection period of 2006–2100. The projections \nare derived according to two Representative Concentration Pathway (RCP) scenarios: RCP4.5 and RCP8.5. Abbreviation \nModel \nInstitution \nICHEC \nEC-Earth \nIrish Centre for High-End Computing \nIPSL \nIPSL-CM 5A-MR \nThe Institut Pierre-Simon Laplace \nMOHC \nHadGEM2-ES \nMet Office Hadley Centre \nMPI \nMPI-ESM-LR \nMax Planck Institute for Meteorology \nTable 1. Summary of the global climate model data sources. 2.1 Study area The salinity in the coastal zone is varying between 6–\n7 g kg-1, although in the plume affected waters from the Curonian Lagoon it can drop drastically to nearly fresh water (0–3 g \nkg-1) (Olenin and Daunys, 2004). The coastal area of the south-eastern Baltic Sea is relatively shallow, up to 20 m depth. The lowest water temperatures here \n75 \nare observed in winter months (in Jan and Feb the average water temperature is ~2 °C), while in summer months, e.g., Jul and \nAug, the average water temperature is ~18–19 °C (Kozlov et al., 2014). The salinity in the coastal zone is varying between 6–\n7 g kg-1, although in the plume affected waters from the Curonian Lagoon it can drop drastically to nearly fresh water (0–3 g \nkg-1) (Olenin and Daunys, 2004). 3 3 2.2 Data \n80 2.2 Data \n80 \n2.2.1 Climate projection data \nTo feed our developed modelling system we used meteorological forcing data acquired from CORDEX (Coordinated Regional \nDownscaling Experiment) scenarios for Europe from the Rossby Centre high-resolution regional atmospheric climate model \n(RCA4), which consisted of four sets of simulations (downscaling) driven by four global climate models (specified in Table \n1). The datasets consist of cloud cover, solar radiation, precipitation, surface air pressure, relative humidity, air temperature, \n85 \nand wind speed and direction, and covers a historical period of 1970–2005 and projection period of 2006–2100. The projections \nare derived according to two Representative Concentration Pathway (RCP) scenarios: RCP4.5 and RCP8.5. Abbreviation \nModel \nInstitution \nICHEC \nEC-Earth \nIrish Centre for High-End Computing \nIPSL \nIPSL-CM 5A-MR \nThe Institut Pierre-Simon Laplace \nMOHC \nHadGEM2-ES \nMet Office Hadley Centre \nMPI \nMPI-ESM-LR \nMax Planck Institute for Meteorology \nTable 1. Summary of the global climate model data sources. 2.2.1 Climate projection data The applied modelling system uses standard Gregorian calendar, however, MOHC uses 360-day calendar and IPSL 365-day \ncalendar. Therefore, correction was applied to harmonize the input data. MOHC calendar was adjusted by deleting 30 Feb and \n90 \n29 Feb (if not leap year), and interpolating the 31st of respective months. IPSL was corrected by interpolating the 29 Feb \nduring the leap years. The bias correction of air temperature and precipitation datasets were done separately for the hydrological and hydrodynamic The applied modelling system uses standard Gregorian calendar, however, MOHC uses 360-day calendar and IPSL 365-day \ncalendar. Therefore, correction was applied to harmonize the input data. MOHC calendar was adjusted by deleting 30 Feb and \n90 \n29 Feb (if not leap year), and interpolating the 31st of respective months. IPSL was corrected by interpolating the 29 Feb \nduring the leap years. The bias correction of air temperature and precipitation datasets were done separately for the hydrological and hydrodynamic \nmodels by applying the climate data bias correction tool (Gupta et al., 2019), which uses the quantile mapping approach by \nfitting the daily values to normal distribution function for the air temperature and gamma distribution function for the \n95 \nprecipitation. Due to the availability of measurement data, provided by Lithuanian Hydrometeorological Service, the correction \nperiod was from 1993 to 2005. The data from 18 meteorological stations, which are scattered throughout the Republic of Lithuania, were used for the bias \ncorrection of the climate data and used as inputs for the hydrological model. For the Belarus region the meteorological grid correction of the climate data and used as inputs for the hydrological model. For the Belarus region the meteorological grid \nwas constructed according to the methodology presented in the “Renewal of a River Basin Districts Management Plans and \n100 \nProgrammes of Measures” project report (PAIC, 2015) and were split in three groups corresponding to the closest \nmeteorological stations in Lithuania of which measured data were used for bias correction. The bias correction of climate input \ndata for the hydrodynamic model was performed from the averaged data of three meteorological stations closest to the northern \npart of the lagoon that describe the coastal climate conditions. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. developed by the Rossby Centre and the Oceanographic research group at Swedish Meteorological and Hydrological Institute \n(SMHI). This model was run by the same before-mentioned global climate projections (Table 1). developed by the Rossby Centre and the Oceanographic research group at Swedish Meteorological and Hydrological Institute \n(SMHI). This model was run by the same before-mentioned global climate projections (Table 1). The bias correction for the data of RCA4-NEMO model was done by using Copernicus Marine Environment Monitoring \n110 \nService (CMEMS) Baltic Sea Physical Reanalysis product data for the period of 1993–2005. The correction was done by \nsimply adding the difference between the average values of CMEMS, 𝐶𝑂, and RCA4–NEMO data, 𝐶𝑚, (Lenderink et al., \n2007): The bias correction for the data of RCA4-NEMO model was done by using Copernicus Marine Environment Monitoring \n110 \nService (CMEMS) Baltic Sea Physical Reanalysis product data for the period of 1993–2005. The correction was done by \nsimply adding the difference between the average values of CMEMS, 𝐶𝑂, and RCA4–NEMO data, 𝐶𝑚, (Lenderink et al., \n2007): \n𝐶𝐵𝐶(𝑡) = 𝐶𝑚(𝑡) + (𝐶𝑂−𝐶𝑚) , \n \n \n \n \n \n \n \n \n \n(1) 110 𝐶𝐵𝐶(𝑡) = 𝐶𝑚(𝑡) + (𝐶𝑂−𝐶𝑚) , (1) Bathymetry data of a high-resolution spherical grid topography of the Baltic Sea were used as a bottom boundary (Seifert et \n115 \nal., 2001). Data of ice thickness from ice thermodynamic model ESIM2, run using the meteorological data specified in Table \n1, were used as a top boundary during the ice cover season (Idzelytė and Umgiesser, 2021; Tedesco et al., 2009). Bathymetry data of a high-resolution spherical grid topography of the Baltic Sea were used as a bottom boundary (Seifert et \n115 \nal., 2001). Data of ice thickness from ice thermodynamic model ESIM2, run using the meteorological data specified in Table \n1, were used as a top boundary during the ice cover season (Idzelytė and Umgiesser, 2021; Tedesco et al., 2009). 2.2.3 Watershed-scale data Some data had to be manually digitized, i.e., the stream network of the Nemunas watershed outside \nof the territory of Lithuania. Unfortunately, due to political reasons, we could not acquire any observational data from the \nRepublic of Belarus and the Kaliningrad Region. We resolved the issue by calibrating the model against observed data at the \nnearest border locations within the territory of Lithuania, provided by abovementioned governmental institutions. 130 2.2.2 Boundary data \n105 Water level, temperature, and salinity data for the sea boundary of the hydrodynamic model were acquired from a high \nresolution regional coupled ocean–sea ice–atmosphere model RCA4–NEMO (Gröger et al., 2019; Wang et al., 2015) 4 2.2.3 Watershed-scale data For the development of the hydrological model many basin-scale datasets are required, i.e., the digital elevation model (DEM), For the development of the hydrological model many basin-scale datasets are required, i.e., the For the development of the hydrological model many basin-scale datasets are required, i.e., the digital elevation model (DEM), \nland use and management data, hydrologic grid, soil maps, etc. We obtained the data from several governmental sources in \n120 \ndifferent countries as well as public open access databases. For a full list of datasets and their sources, we refer the reader to \nČerkasova et al. (2021), where the acquired datasets are described in depth. Observed discharge data with varying time step \n(daily, weekly, by-weekly, and monthly) for the majority of the Nemunas River tributaries as well as the main branch were \nobtained from the Lithuanian Hydrometeorological Service. Observed nutrient (TN and TP) concentration values for the \noverlapping periods were provided by the Lithuanian Environmental Protection Agency. 125 land use and management data, hydrologic grid, soil maps, etc. We obtained the data from several governmental sources in \n120 \ndifferent countries as well as public open access databases. For a full list of datasets and their sources, we refer the reader to \nČerkasova et al. (2021), where the acquired datasets are described in depth. Observed discharge data with varying time step \n(daily, weekly, by-weekly, and monthly) for the majority of the Nemunas River tributaries as well as the main branch were \nobtained from the Lithuanian Hydrometeorological Service. Observed nutrient (TN and TP) concentration values for the overlapping periods were provided by the Lithuanian Environmental Protection Agency. 125 \nLarge parts of the watershed are outside of Lithuania, hence open-access data for those regions were identified and used \n(Čerkasova et al., 2021). Some data had to be manually digitized, i.e., the stream network of the Nemunas watershed outside \nof the territory of Lithuania. Unfortunately, due to political reasons, we could not acquire any observational data from the \nRepublic of Belarus and the Kaliningrad Region. We resolved the issue by calibrating the model against observed data at the \nnearest border locations within the territory of Lithuania, provided by abovementioned governmental institutions. 130 pp g p\np\ny\ng\ny\nLarge parts of the watershed are outside of Lithuania, hence open-access data for those regions were identified and used \n(Čerkasova et al., 2021). 2.3.1 Hydrological model High-resolution basin-scale model has been developed for the entire Nemunas watershed (Fig. 1) and implemented using Soil \nand Water Assessment Tool (SWAT) using a set of custom tools and scripts. To set up a comprehensive model, we incorporated \nthe topographic information of the area, land use and soil properties, information on water bodies, land management, water \n135 \nuse, crop growth, livestock production, administrative units, etc. The model was calibrated, validated, and proved to reliably \nrepresent the water balance components as well as nutrient and sediment load estimates, and was used in previous studies \n(Čerkasova et al., 2021, 2019, 2018). High-resolution basin-scale model has been developed for the entire Nemunas watershed (Fig. 1) and implemented using Soil \nand Water Assessment Tool (SWAT) using a set of custom tools and scripts. To set up a comprehensive model, we incorporated the topographic information of the area, land use and soil properties, information on water bodies, land management, water \n135 \nuse, crop growth, livestock production, administrative units, etc. The model was calibrated, validated, and proved to reliably \nrepresent the water balance components as well as nutrient and sediment load estimates, and was used in previous studies \n(Čerkasova et al., 2021, 2019, 2018). the topographic information of the area, land use and soil properties, information on water bodies, land management, water \n135 \nuse, crop growth, livestock production, administrative units, etc. The model was calibrated, validated, and proved to reliably \nrepresent the water balance components as well as nutrient and sediment load estimates, and was used in previous studies \n(Čerkasova et al., 2021, 2019, 2018). 5 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. The basin-scale hydrological model consists of 11 sub-models, each representing one of the tributaries of the Nemunas River, \nconnected to the main branch (Fig. 2). In total, the model consists of 9012 sub-basins, and 148 212 Hydrologic Response Units. 140 \nThese sub-models are linked from upstream to downstream, where the outputs of two downstream models are directly used as \nriver boundaries for the hydrodynamic model. Discharge and temperature outputs were extracted from the list of variables, \nprovided by the Nemunas drainage basin modelling framework. 140 Figure 2. SWAT models for the Nemunas River watershed and their linkage with SHYFEM through river boundaries. 145 Figure 2. SWAT models for the Nemunas River watershed and their linkage with SHYFEM through river boundaries. 145 Figure 2. SWAT models for the Nemunas River watershed and their linkage with SHYFEM th\n145 Figure 2. SWAT models for the Nemunas River watershed and their linkage with SHYFEM through river boundaries. 145 \n2.3.2 Hydrodynamic model 2.3.2 Hydrodynamic model Hydrodynamics of the Curonian Lagoon and the south-eastern Baltic Sea were modelled with an open-source shallow water \nhydrodynamic finite element model SHYFEM (http://www.ismar.cnr.it/shyfem, Umgiesser et al., 2004), which consists of a \nfinite element 3-D hydrodynamic model, a transport and diffusion model, and a radiation transfer model of heat at the water surface. The model resolves the 3-D primitive equations, vertically integrated over each layer, in their formulations with water \n150 \nlevels and transports. Horizontal spatial discretization of the model is based on an unstructured triangular grid and carried out \nusing a finite element method, which makes it suitable for applications to coastal systems with complicated geometry and \nbathymetry. This model has already been successfully applied to many coastal environments (Umgiesser et al., 2014; De Pascalis et al., surface. The model resolves the 3-D primitive equations, vertically integrated over each layer, in their formulations with water \n150 \nlevels and transports. Horizontal spatial discretization of the model is based on an unstructured triangular grid and carried out \nusing a finite element method, which makes it suitable for applications to coastal systems with complicated geometry and \nbathymetry. This model has already been successfully applied to many coastal environments (Umgiesser et al., 2014; De Pascalis et al., 2011; Ferrarin et al., 2013, 2010; Bellafiore and Umgiesser, 2010; Ferrarin and Umgiesser, 2005), as well as validated for the \n155 \nCuronian Lagoon case study in previous works (Mėžinė et al., 2019; Umgiesser et al., 2016; Zemlys et al., 2013; Ferrarin et \nal., 2008). For this study, SHYFEM was applied in its 2-D version, which is sufficient considering the shallow nature of this \nlagoon. The computational grid consists of 3292 triangular elements with 1986 nodes having a much finer resolution in the \nKlaipėda Strait area (Fig. 3). Model produced output data of hydrodynamic properties every 6 hours. 2011; Ferrarin et al., 2013, 2010; Bellafiore and Umgiesser, 2010; Ferrarin and Umgiesser, 2005), as well as validated for the \n155 \nCuronian Lagoon case study in previous works (Mėžinė et al., 2019; Umgiesser et al., 2016; Zemlys et al., 2013; Ferrarin et \nal., 2008). For this study, SHYFEM was applied in its 2-D version, which is sufficient considering the shallow nature of this \nlagoon. The computational grid consists of 3292 triangular elements with 1986 nodes having a much finer resolution in the \nKlaipėda Strait area (Fig. 3). Model produced output data of hydrodynamic properties every 6 hours. 160 Figure 3. Computational grid of the hydrodynamic model SHYFEM. Dark blue arrows denote the location of the river boundaries \n(modelling system connectivity points), where the hydrological model outputs (water temperature and discharge) are used as inputs \nto the hydrodynamic model. 2.3.3 Coupling of models A better representation of hydrodynamic conditions in the Nemunas Delta region is achieved with a hydrodynamic model, thus \n165 \nthe overlapping parts of SHYFEM and SWAT models were modelled entirely by SHYFEM. The daily water temperature and \ndischarge data from the SWAT model were directly used as Minija and Nemunas river boundaries (Fig. 3). However, other \ntwo main rivers discharging in the southern part of the lagoon (Matrosovka, branch of the Nemunas, and Deima) are not part \nof the Nemunas River watershed model, thus the SWAT output data of the Nemunas River were scaled by the ratio between \nMatrosovka-Nemunas and Deima-Šešupė calculated from the data presented in Jakimavičius (2012)\n170 Matrosovka-Nemunas and Deima-Šešupė calculated from the data presented in Jakimavičius (2012). 170 2.3.2 Hydrodynamic model 6 0 \nFigure 3. Computational grid of the hydrodynamic model SHYFEM. Dark blue arrows denote the location of the river boundaries \n(modelling system connectivity points), where the hydrological model outputs (water temperature and discharge) are used as inputs \nto the hydrodynamic model. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. divided into historical (1975–2005), short-term (2020–2050) and long-term (2070–2100), and the results were compared by \naveraging over these timespans. The period of 1970–1975 was discarded from analysis as it was used for the model spin-up. 175 \nThe main air temperature and precipitation patterns were determined. The parameters of the hydrological model (water \ndischarge) and parameters of the hydrodynamic model (water temperature, salinity, level, residence time, and fluxes through \npredefined cross sections), as well as ice thickness, were assessed. The analysis was done using standard statistical methods, \nand computing the percentage of change from the historical period. divided into historical (1975–2005), short-term (2020–2050) and long-term (2070–2100), and the results were compared by \naveraging over these timespans. The period of 1970–1975 was discarded from analysis as it was used for the model spin-up. 175 \nThe main air temperature and precipitation patterns were determined. The parameters of the hydrological model (water \ndischarge) and parameters of the hydrodynamic model (water temperature, salinity, level, residence time, and fluxes through \npredefined cross sections), as well as ice thickness, were assessed. The analysis was done using standard statistical methods, \nand computing the percentage of change from the historical period. 175 3.1 Bias correction Even though the meteorological data from the global climate models were already downscaled by the regional climate model, \nthere was still some bias remaining in comparison with the local measurement data – underestimation of air temperature and \noverestimation of precipitation. Therefore, an additional correction was applied (Fig. 4). The correlation between the modelled \nand measured air temperature datasets was strong (~0.8), however after bias correction there was an evident improvement of \n5 \nthe standard deviation. Similar change is also equivalent to the monthly precipitation, although here we observe that the climate \nmodels have their own internal dynamics, thus the correlation with measurements is weak (~0.25). Even though the meteorological data from the global climate models were already downscaled by the regional climate model, \nthere was still some bias remaining in comparison with the local measurement data – underestimation of air temperature and \noverestimation of precipitation. Therefore, an additional correction was applied (Fig. 4). The correlation between the modelled 185 and measured air temperature datasets was strong (~0.8), however after bias correction there was an evident improvement of \n85 \nthe standard deviation. Similar change is also equivalent to the monthly precipitation, although here we observe that the climate \nmodels have their own internal dynamics, thus the correlation with measurements is weak (~0.25). and measured air temperature datasets was strong (~0.8), however after bias correction there was an evident improvement of \n185 \nthe standard deviation. Similar change is also equivalent to the monthly precipitation, although here we observe that the climate \nmodels have their own internal dynamics, thus the correlation with measurements is weak (~0.25). Figure 4. Taylor diagram showing a statistical comparison of air temperature (daily mean) and precipitation (monthly sum) before \nand after bias correction (BC) with respect to the reference measurement data. Statistics were computed for the period of 1990-\n0 \n2005. Figure 4. Taylor diagram showing a statistical comparison of air temperature (daily mean) and precipitation (monthly sum) before \nand after bias correction (BC) with respect to the reference measurement data. Statistics were computed for the period of 1990-\n \n2005. 190 2.4 Methods of data analysis Model results based on four global climate model forcings (Table 1) were averaged, therefore the final analysis was done for \nthree aggregated outputs referring to the historical and projection (RCP4.5 and RCP8.5) scenarios. The analysis periods were 7 Therefore, annual\n205 Considering the differences between marine and terrestrial parts of the study area, there is a clear pattern of lower air \ntemperatures (on average by ~1–1.5 °C) over the Baltic Sea and the Curonian Lagoon during March through June, while during \nthe rest of the year it is warmer (on average by ~2.2 °C), compared with that of the Nemunas River watershed. During February \nthrough August and December months, less precipitation over the marine part (on average by 10 mm month-1) is observed, \nwhile during the rest of the months it is higher by the same amount compared with the terrestrial area. Therefore, annual \n205 \nprecipitation is similar to both areas. while during the rest of the months it is higher by the same amount compared with the terrestrial area. Therefore, annual \n205 \nprecipitation is similar to both areas. 3.2 Meteorological changes The bias-corrected statistics of both RCP scenarios show a distinct increase in the projected monthly and annual average \nprecipitation and air temperature (Table 2). The differences between the scenarios in the short-term are smaller than in the \nlong-term by an average of 49% under RCP4.5 and ~123% under RCP8.5. Considering precipitation, the monthly \n195 8 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. wetness/dryness pattern is similar between the analysis periods and scenarios. The average air temperature in the long-term \nhas the highest increase, especially in the months from November to April. The highest changes of these parameters are \nprojected under RCP8.5 scenario. Air temperature (oC) \n \n \nJan \nFeb \nMar \nApr \nMay \nJun \nJul \nAug \nSep \nOct \nNov \nDec \nYear \nHistorical (1975-2005) \n-2.6 \n-2.6 \n0.3 \n5.3 \n10.7 \n14.6 \n17.9 \n17.5 \n13.2 \n8.1 \n1.8 \n-2.3 \n6.8 \nShort-term \n(2020-2050) \nRCP4.5 \n-0.5 \n0.4 \n2.4 \n7.4 \n12.2 \n15.9 \n19.2 \n18.7 \n14.6 \n9.7 \n3.8 \n0.0 \n8.6 \nRCP8.5 \n-0.6 \n0.5 \n2.8 \n7.5 \n12.4 \n16.1 \n19.2 \n18.8 \n14.6 \n10.0 \n4.2 \n-0.2 \n8.8 \nLong-term \n(2070-2100) \nRCP4.5 \n0.9 \n1.4 \n3.4 \n8.4 \n13.3 \n16.9 \n20.1 \n19.7 \n15.5 \n10.9 \n4.7 \n1.0 \n9.7 \nRCP8.5 \n2.9 \n3.4 \n5.0 \n10.0 \n14.7 \n18.1 \n21.8 \n21.5 \n17.6 \n12.4 \n6.7 \n3.3 \n11.5 \n \n \nPrecipitation (mm month-1) \n \n \nJan \nFeb \nMar \nApr \nMay \nJun \nJul \nAug \nSep \nOct \nNov \nDec \nYear \nHistorical (1975-2005) \n42.8 \n36.5 \n43.1 \n36.9 \n55.9 \n87.1 \n57.5 \n57.9 \n48.7 \n63.7 \n39.0 \n48.7 \n617.7 \nShort-term \n(2020-2050) \nRCP4.5 \n54.4 \n47.6 \n53.7 \n41.1 \n53.1 \n83.4 \n69.9 \n64.7 \n55.6 \n72.7 \n49.9 \n56.4 \n702.5 \nRCP8.5 \n53.1 \n46.9 \n50.1 \n43.1 \n61.0 \n83.1 \n75.4 \n69.1 \n54.2 \n71.2 \n50.5 \n58.4 \n716.2 \nLong-term \n(2070-2100) \nRCP4.5 \n56.9 \n46.5 \n53.7 \n44.8 \n58.8 \n83.3 \n69.2 \n70.0 \n54.0 \n78.2 \n53.6 \n59.3 \n727.3 \nRCP8.5 \n72.3 \n52.2 \n65.4 \n49.5 \n67.7 \n93.2 \n73.6 \n77.9 \n60.6 \n85.3 \n62.7 \n70.3 \n830.8 \nTable 2. Monthly and yearly mean of air temperature and precipitation over the study region derived by averaging four bias-\ncorrected climate models’ datasets. Table 2. Monthly and yearly mean of air temperature and precipitation over the study region derived by averaging four bias-\ncorrected climate models’ datasets. 200 corrected climate models’ datasets. 200 \nConsidering the differences between marine and terrestrial parts of the study area, there is a clear pattern of lower air \ntemperatures (on average by ~1–1.5 °C) over the Baltic Sea and the Curonian Lagoon during March through June, while during \nthe rest of the year it is warmer (on average by ~2.2 °C), compared with that of the Nemunas River watershed. During February \nthrough August and December months, less precipitation over the marine part (on average by 10 mm month-1) is observed, \nwhile during the rest of the months it is higher by the same amount compared with the terrestrial area. 3.3 Interactions between domains The historical discharges of the Nemunas River watershed to the Curonian Lagoon were compared to the projected averaged \ndischarges under RCP4.5 and RCP8.5 climate change scenarios for all seasons (Fig. 5). The most noticeable change is the \nprojected drastic increase in runoff during winter, which will almost double in the long-term period under both RCPs. This \n210 \nresult is a combination of increased projected average precipitation and temperatures, which will likely result in warmer winters \nwith reduced snow cover throughout the basin. The increased liquid precipitation in winter will likely lead to an increased soil \nerosion and nutrient wash-off, when the soil is bare and exposed to the elements. 9 9 Figure 5. The percentage of change of seasonal average discharge from the Nemunas River for the RCP4.5 and RCP8.5 scenario \n215 \nruns during the short-term (2020–2050) and long-term periods (2070–2100), compared with the historical period (1975–2005). https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Figure 5. The percentage of change of seasonal average discharge from the Nemunas River for the RCP4.5 and RCP8.5 scenario\nhttps://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Figure 5. The percentage of change of seasonal average discharge from the Nemunas River for the RCP4.5 and RCP8.5 scenario \n215 \nruns during the short-term (2020–2050) and long-term periods (2070–2100), compared with the historical period (1975–2005). Figure 5. The percentage of change of seasonal average discharge from the Nemunas River for the RCP4.5 and RCP8.5 scenario \n215 \nruns during the short-term (2020–2050) and long-term periods (2070–2100), compared with the historical period (1975–2005). In contrast, a decrease in average summer outflow is projected under both RCP scenarios, despite an increase in precipitation \nacross the region. The increased average temperatures in summer months will impact the evapotranspiration (ET) over the \nbasin, where an increase in both potential ET and ET is projected, which will lead to lower discharges at the stream gages during this period. Stable and mostly unchanged Nemunas River discharge will continue to supply the lagoon over the rest of \n220 \nthe season, in spring and autumn. Noteworthy are the differences between the RCP scenarios. 3.3 Interactions between domains We see a stronger shift in the increased winter discharges in the \nRCP8.5 scenario in the short-term period, and a stronger shift towards lesser summer flows in the RCP4.5 scenario (Fig. 5). The RCP8.5 in the short-term can be considered as “wetter” scenario, whereas RCP4.5 is “dryer”. In the long-term, however, both RCPs simulate similar outcomes with slightly different magnitudes. 225 \nThe interactions between the study domains can be defined by the water fluxes from the main discharging rivers into the \nCuronian Lagoon, eventually into the Baltic Sea (water outflow), and vice versa (water inflow). The volume of inflowing water \nis much lower than that of the outflowing (Fig. 6), especially in the northern part of the Curonian Lagoon, meaning that the \noutflow is already dominating and this study shows that it will prevail and even increase in the analysed continuum system. 10 230 \nFigure 6. The percentage of the average outflowing and inflowing water volume of the absolute flux of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. The locations of \ncross-section are shown in Fig. 7. The bottom graph is for reference, indicating the average water flux in m3 s-1 during the historical \nperiod (1975–2005). https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Table 3. Seasonal water flux in m3 s-1 averaged over historical (1975–2005), short-term (2020–2050), and long-term (2070–2100) \n255 \nperiods under RCP4.5 and RCP8.5 scenarios. Percentage in the brackets shows the change compared with the historical period. The \nvalues are averaged over four cross-sections in the Curonian Lagoon: Klaipėda Strait, North of Nemunas, Nemunas Delta, and \nLithuanian Russian border (see map insert in Fig. 7). 230 Figure 6. The percentage of the average outflowing and inflowing water volume of the absolute flux of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. The locations of \ncross-section are shown in Fig. 7. The bottom graph is for reference, indicating the average water flux in m3 s-1 during the historical \nperiod (1975–2005). Figure 6. The percentage of the average outflowing and inflowing water volume of the absolute flux of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. The locations of \ncross-section are shown in Fig. 7. The bottom graph is for reference, indicating the average water flux in m3 s-1 during the historical \nperiod (1975–2005). Comparing the simulation results of the scenario runs (short- and long-term) with the historical period, it is evident that the \n235 \nwater inflow in the northern part of the lagoon (cross-sections 1 and 2 in Fig. 7) is projected to decrease. The change of water \ninflow in the centre of the lagoon (cross-sections 3 and 4 in Fig. 7) is very small and can be considered negligible. The \noutflowing water flux is projected to increase and is related to the growing outflow from the Nemunas River and will increase \nalong the pathway towards the Baltic Sea (cross-sections 1, 2, and 3 in Fig. 7). In the southern part of the lagoon, near the Lithuanian–Russian border, the changes in water fluxes are less evident, due to the fact that water circulation in this area is \n240 \nmostly determined by the wind induced currents. 11 11 Figure 7. The percentage of change of projected (under RCP4.5 and RCP8.5 scenarios) inflowing and outflowing water flux from \nthe historical period (1975–2005). The map insert shows the locations of the predefined cross-sections: 1 – Klaipėda Strait, 2 – North \nof Nemunas, 3 – Nemunas Delta, 4 – Lithuanian–Russian border. Outflow is from south to north (sections 1,2,4), and from east to \nwest (section 3), inflow - vice versa. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Fi\n7 Th\nf h\nf\nj\nd (\nd\nRCP4 5\nd RCP8 5\ni\n) i fl\ni\nd\nfl\ni\nfl\nf Figure 7. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. The changes of the outflow during winter, spring, and autumn are expected to increase (Table 3). In the short-term, the changes \nbetween the scenarios are similar, while in the long-term, the change increases nearly twice under RCP8.5, compared with \n260 \nRCP4.5. During summer the outflow changes are negligible, apart from the long-term scenario under RCP8.5 (-8%). 3.4 Salinity During the historical period 48 days year-1 of salt water intrusions were observed, i.e., the average yearly number of days when \nsalinity exceeds the 2 g kg-1 threshold, approximately 20 km below the lagoon sea connection. In the short-term, this number \n265 \nwill decrease by 60% under RCP4.5 scenario (19 days year-1) and 65% under RCP8.5 scenario (17 days year-1) compared with \nthe historical period. While in the long-term, it will decrease by ~67% under RCP4.5 (16 days year-1) and 83% under RCP8.5 \nscenario (8 days year-1). Seasonally, the highest change of saltwater intrusion days is observed during winter and especially spring, when the outflow Seasonally, the highest change of saltwater intrusion days is observed during winter and especially spring, when the outflow \nof the lagoon is dominating and there are no water intrusions from the Baltic Sea (right side of Fig. 8). Saltwater intrusions \n270 \nduring summer season have a larger decrease under RCP8.5 scenario than RCP4.5, which does not have a distinct pattern in \nthe future. In autumn, the changes in the short-term are similar under both scenarios ~27% decrease compared with the \nhistorical period, and it continues to decline in the long-term as well. of the lagoon is dominating and there are no water intrusions from the Baltic Sea (right side of Fig. 8). Saltwater intrusions \n270 \nduring summer season have a larger decrease under RCP8.5 scenario than RCP4.5, which does not have a distinct pattern in \nthe future. In autumn, the changes in the short-term are similar under both scenarios ~27% decrease compared with the \nhistorical period, and it continues to decline in the long-term as well. Figure 8. The percentage of change of seasonal average salinity of saltwater intrusions (solid pattern – left) and number of days \n275 \nwhen salinity exceeded the 2 g kg-1 threshold (hatched pattern – right) in Juodkrantė, approximately 20 km below the strait \nconnecting the Baltic Sea and the Curonian Lagoon. Values denote the percentage of difference of the scenario runs compared with \nthe historical period (1975–2005). Figure 8. The percentage of change of seasonal average salinity of saltwater intrusions (solid pattern – left) and number of days \n275 \nwhen salinity exceeded the 2 g kg-1 threshold (hatched pattern – right) in Juodkrantė, approximately 20 km below the strait \nconnecting the Baltic Sea and the Curonian Lagoon. The percentage of change of projected (under RCP4.5 and RCP8.5 scenarios) inflowing and outflowing water flux from \nthe historical period (1975–2005). The map insert shows the locations of the predefined cross-sections: 1 – Klaipėda Strait, 2 – North \nof Nemunas, 3 – Nemunas Delta, 4 – Lithuanian–Russian border. Outflow is from south to north (sections 1,2,4), and from east to \n \nwest (section 3), inflow - vice versa. 245 The change of water flow through all cross-sections display a similar pattern in all scenarios, the water flux changes from the \nhistorical are very similar in each analysed cross-section and for each flow type. For both the outflow and inflow the highest \nchange from the historical period is observed in the long-term under RCP8.5 scenario. Looking at the water exchange seasonal dynamics, we can see that the highest inflowing water deviations from the historical \n250 \nperiod are predicted during winter and spring, being lower by 10–28% (Table 3). The water exchange is expected to increase \nin summer by 12–22% while in the autumn the changes are very small and similar throughout the analysis periods and \nscenarios. 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \nOUTFLOW \nWinter \n1017 \n1291 (27%) \n1324 (30%) \n1388 (36%) \n1663 (63%) \nSpring \n1046 \n1210 (15%) \n1213 (15%) \n1214 (16%) \n1356 (29%) \nSummer \n858 \n814 (-5%) \n850 (-1%) \n856 (0.3%) \n923 (8%) \nAutumn \n924 \n1007 (9%) \n1030 (12%) \n1076 (17%) \n1190 (29%) \nINFLOW \nWinter \n452 \n387 (-19%) \n383 (-20%) \n380 (-21%) \n357 (-28%) \nSpring \n246 \n225 (-13%) \n220 (-16%) \n233 (-10%) \n225 (-14%) \nSummer \n277 \n309 (13%) \n308 (12%) \n331 (22%) \n315 (14%) \nAutumn \n439 \n449 (2%) \n450 (2%) \n461 (4%) \n456 (2%) Table 3. Seasonal water flux in m3 s-1 averaged over historical (1975–2005), short-term (2020–2050), and long-term (2070–2100) \n255 \nperiods under RCP4.5 and RCP8.5 scenarios. Percentage in the brackets shows the change compared with the historical period. The \nvalues are averaged over four cross-sections in the Curonian Lagoon: Klaipėda Strait, North of Nemunas, Nemunas Delta, and \nLithuanian Russian border (see map insert in Fig. 7). 12 In the coastal waters the highest change, compared with the historical period, is also projected to \n290 \nbe under RCP8.5 scenario. Notably, the highest decrease is projected northward of Klaipėda Strait (lagoon outlet), while \nsouthward the change is lower and salinity values are similar to that of south-eastern Baltic Sea 3.4 Salinity Values denote the percentage of difference of the scenario runs compared with \nthe historical period (1975–2005). Figure 8. The percentage of change of seasonal average salinity of saltwater intrusions (solid pattern – left) and number of days \n275 \nwhen salinity exceeded the 2 g kg-1 threshold (hatched pattern – right) in Juodkrantė, approximately 20 km below the strait \nconnecting the Baltic Sea and the Curonian Lagoon. Values denote the percentage of difference of the scenario runs compared with \nthe historical period (1975–2005). 13 In the Baltic Sea there is no difference of average salinity throughout the season 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \nCoastal waters \nWinter \n6.6 \n5.9 (-11%) \n5.8 (-13%) \n5.4 (-18%) \n4.8 (-27%) \nSpring \n5.4 \n4.9 (-11%) \n4.8 (-12%) \n4.4 (-18%) \n4.0 (-26%) \nSummer \n5.4 \n5.1 (-4%) \n5.0 (-7%) \n4.8 (-12%) \n4.3 (-20%) \nAutumn \n6.3 \n5.9 (-6%) \n5.8 (-8%) \n5.5 (-13%) \n5.0 (-21%) \nBaltic Sea \nWinter \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n6.9 (-11%) \n6.4 (-18%) \nSpring \n7.7 \n7.4 (-5%) \n7.3 (-6%) \n6.9 (-11%) \n6.3 (-18%) \nSummer \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n7.0 (-11%) \n6.4 (-18%) \nAutumn \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n7.0 (-11%) \n6.4 (-18%) \nTable 4. Average seasonal salinity in the Baltic Sea and along the south-eastern Baltic coast. Percentage in the brackets shows the \nh\nf th\ni\nd\nith th hi t\ni\nl\ni d (1975 2005) Table 4. Average seasonal salinity in the Baltic Sea and along the south-eastern Baltic coast. Percentage in the brackets shows the \nchange of the scenario runs compared with the historical period (1975–2005). Salinity in the coastal area is slightly lower compared to that of the Baltic Sea. However, here the difference of seasons is more \napparent, having a higher decease of salinity values during winter and spring when the discharge of freshwater from the \nCuronian Lagoon increases. In the coastal waters the highest change, compared with the historical period, is also projected to \n90 \nbe under RCP8 5 scenario Notably the highest decrease is projected northward of Klaipėda Strait (lagoon outlet) while Salinity in the coastal area is slightly lower compared to that of the Baltic Sea. However, here the difference of seasons is more \napparent, having a higher decease of salinity values during winter and spring when the discharge of freshwater from the Salinity in the coastal area is slightly lower compared to that of the Baltic Sea. However, here the difference of seasons is more \napparent, having a higher decease of salinity values during winter and spring when the discharge of freshwater from the \nCuronian Lagoon increases. In the coastal waters the highest change, compared with the historical period, is also projected to \n290 \nbe under RCP8.5 scenario. Notably, the highest decrease is projected northward of Klaipėda Strait (lagoon outlet), while \nsouthward the change is lower and salinity values are similar to that of south-eastern Baltic Sea Curonian Lagoon increases. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. The severity of the saltwater intrusions is also predicted to decrease (left side of Fig. 8). The pattern is similar to that of the \nnumber of days of saltwater intrusions. Due to the increased outflow from the lagoon during spring, the changes are the highest \n280 \nin the same period. The severity of the saltwater intrusions is also predicted to decrease (left side of Fig. 8). The pattern is similar to that of the \nnumber of days of saltwater intrusions. Due to the increased outflow from the lagoon during spring, the changes are the highest \n280 \nin the same period. Salinity in the south-eastern Baltic Sea show a decreasing pattern throughout the analysis periods. Under RCP4.5 scenario, \nthere is a small salt content decrease projected – 5–6% less in the short-term and 11–18% in the long-term (Table 4), compared \nwith the historical period. Difference between the RCP scenarios are only apparent in the long-term, where the change is \nprojected to be higher under RCP8.5 scenario. In the Baltic Sea there is no difference of average salinity throughout the seasons. 285 \n \n \n1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \nCoastal waters \nWinter \n6.6 \n5.9 (-11%) \n5.8 (-13%) \n5.4 (-18%) \n4.8 (-27%) \nSpring \n5.4 \n4.9 (-11%) \n4.8 (-12%) \n4.4 (-18%) \n4.0 (-26%) \nSummer \n5.4 \n5.1 (-4%) \n5.0 (-7%) \n4.8 (-12%) \n4.3 (-20%) \nAutumn \n6.3 \n5.9 (-6%) \n5.8 (-8%) \n5.5 (-13%) \n5.0 (-21%) \nBaltic Sea \nWinter \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n6.9 (-11%) \n6.4 (-18%) \nSpring \n7.7 \n7.4 (-5%) \n7.3 (-6%) \n6.9 (-11%) \n6.3 (-18%) \nSummer \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n7.0 (-11%) \n6.4 (-18%) \nAutumn \n7.8 \n7.4 (-5%) \n7.3 (-6%) \n7.0 (-11%) \n6.4 (-18%) \nTable 4. Average seasonal salinity in the Baltic Sea and along the south-eastern Baltic coast. Percentage in the brackets shows the \nchange of the scenario runs compared with the historical period (1975–2005). Salinity in the south-eastern Baltic Sea show a decreasing pattern throughout the analysis periods. Under RCP4.5 scenario, \nthere is a small salt content decrease projected – 5–6% less in the short-term and 11–18% in the long-term (Table 4), compared \nwith the historical period. Difference between the RCP scenarios are only apparent in the long-term, where the change is \nprojected to be higher under RCP8.5 scenario. In the Baltic Sea there is no difference of average salinity throughout the seasons. 285 der RCP8.5 scenario. The changes between the analysis periods and RCPs scenarios comparing with the historical \n310 \nperiod are higher than that of the south-eastern Baltic Sea. seasons, on average by 4.9 °C. The changes between the analysis periods and RCPs scenarios comparing with the historical \n310 \nperiod are higher than that of the south-eastern Baltic Sea. 3.5 Water temperature The water temperature in the Curonian Lagoon is relatively homogeneous throughout its area with slightly warmer water (by \non average 0.4 °C) in the northern part of the lagoon. Considering the seasonal dynamics, water temperature is projected to \n295 \nincrease the most during winter, by ~60% in the short-term and more than twice in the long-term (Table 5). During spring and \nautumn, the increase is lower – ~17% in the short-term and 19–27% in the long-term. The lowest increase is projected during \nsummer season – 5–6% in the short-term and 9–13% in the long-term. The more apparent difference between the RCP \nscenarios is observed during the long-term, where RCP8.5 displays a higher temperature increase. 14 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \nLagoon \nWinter \n1.5 \n2.5 (60%) \n2.5 (61%) \n3.1 (102%) \n3.9 (151%) \nSpring \n8.3 \n9.6 (16%) \n9.8 (18%) \n9.9 (19%) \n10.5 (27%) \nSummer \n19.8 \n20.8 (5%) \n20.9 (6%) \n21.6 (9%) \n22.3 (13%) \nAutumn \n10.5 \n11.7 (11%) \n12.0 (14%) \n13.1 (24%) \n13.8 (32%) \nCoastal waters \nWinter \n2.6 \n3.7 (42%) \n3.7 (42%) \n4.4 (72%) \n5.9 (129%) \nSpring \n5.4 \n6.7 (24%) \n6.8 (26%) \n7.0 (30%) \n8.4 (56%) \nSummer \n16.0 \n17.0 (7%) \n17.2 (8%) \n17.7 (11%) \n19.2 (20%) \nAutumn \n12.0 \n13.1 (9%) \n13.4 (11%) \n14.4 (20%) \n15.8 (31%) \nBaltic Sea \nWinter \n4.2 \n5.2 (24%) \n5.3 (26%) \n5.9 (41%) \n7.2 (71%) \nSpring \n3.4 \n4.5 (30%) \n4.6 (33%) \n5.1 (50%) \n6.3 (85%) \nSummer \n7.4 \n8.5 (14%) \n8.6 (16%) \n9.2 (23%) \n10.4 (40%) \nAutumn \n8.3 \n9.3 (12%) \n9.4 (13%) \n10.1 (21%) \n11.3 (35%) \nTable 5. Average seasonal water temperature (in °C) in the Curonian Lagoon and Baltic Sea. Percentage in the brackets shows the \nchange of the scenario runs compared with the historical period (1975–2005). Table 5. Average seasonal water temperature (in °C) in the Curonian Lagoon and Baltic Sea. Percentage in the brackets shows the \nchange of the scenario runs compared with the historical period (1975–2005). Table 5. Average seasonal water temperature (in °C) in the Curonian Lagoon and Baltic Sea. Percent\nchange of the scenario runs compared with the historical period (1975–2005). is under RCP8.5 scenario, compared with the historical period. The smallest change throughout the analysis periods is projected \n305 \nin summer and autumn. There is no apparent difference between the RCP scenarios in the short-term, contrary to the long-\nterm, when the RCP8.5 scenario displays a higher water temperature increase. in summer and autumn. There is no apparent difference between the RCP scenarios in the short-term, contrary to the long-\nterm, when the RCP8.5 scenario displays a higher water temperature increase. The seasonal dynamics in the coastal area (nearshore temperature) is very similar to the whole south-eastern Baltic Sea. Although, the water temperature is slightly lower during the winter season, on average by 1.5 °C, and higher during other The seasonal dynamics in the coastal area (nearshore temperature) is very similar to the whole south-eastern Baltic Sea. Although, the water temperature is slightly lower during the winter season, on average by 1.5 °C, and higher during other \nseasons, on average by 4.9 °C. 3.6 Water level Data were averaged over south-eastern Baltic Sea and five selected points in the Curonian \nLagoon (CL) as shown in the map insert. Table 6. The average water level during the historical (1975–2005) and the short- (2020–2050) and long-term (2070–2100) periods \nunder RCP4.5 and RCP8.5 scenarios. Data were averaged over south-eastern Baltic Sea and five selected points in the Curonian \nLagoon (CL) as shown in the map insert. The water level in the Baltic Sea, as well as the coastal area, is homogeneous and most of the time staying below the sea level \n330 \nduring the historical period. The water level in the sea is expected to rise by 16–19 cm in the short-term and by 32–39 cm in \nthe long -term, compared with the historical period (Table 6). This change is one half of that in the Curonian Lagoon. At the \ndomains’ connective area – Klaipėda Strait, the water level corresponds mostly to the dynamics of it in the sea. 3.6 Water level The water level in the Curonian Lagoon usually varies between -0.2–0 m during the historical period. In the short-term \nprojections, the model estimates a water level increase by 17–20 cm compared with the historical period (Table 6). The highest \nincrease can be estimated in the long-term, when the water level in the lagoon increases by 32 cm under RCP4.5 and 41 cm \n315 \nunder RCP8.5 scenario, compared with the historical period. For the latter, the highest increase is observed in the Nemunas \nDelta and northern part of the lagoon. 315 15 320 \n \n \n \n \n \n325 \n \nTable 6. The average water level during the historical (1975–2005) and the short- (2020–2050) and long-term (2070–2100) periods \nunder RCP4.5 and RCP8.5 scenarios. Data were averaged over south-eastern Baltic Sea and five selected points in the Curonian \nLagoon (CL) as shown in the map insert. 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \n \nBaltic Sea \n-0.15 \n0.01 \n0.03 \n0.17 \n0.23 \n1 \nKlaipėda Strait \n-0.13 \n0.04 \n0.06 \n0.19 \n0.26 \n2 \nNorthern CL \n-0.06 \n0.11 \n0.13 \n0.26 \n0.35 \n3 \nNemunas Delta \n-0.06 \n0.11 \n0.14 \n0.26 \n0.35 \n4 \nCentral CL \n-0.07 \n0.10 \n0.12 \n0.25 \n0.34 \n5 \nSouthern CL \n-0.08 \n0.09 \n0.11 \n0.24 \n0.33 \nhttps://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. 320 \n \n \n \n \n \n325 \n \nTable 6. The average water level during the historical (1975–2005) and the short- (2020–2050) and long-term (2070–2100) periods \nunder RCP4.5 and RCP8.5 scenarios. Data were averaged over south-eastern Baltic Sea and five selected points in the Curonian \nLagoon (CL) as shown in the map insert. 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \n \nBaltic Sea \n-0.15 \n0.01 \n0.03 \n0.17 \n0.23 \n1 \nKlaipėda Strait \n-0.13 \n0.04 \n0.06 \n0.19 \n0.26 \n2 \nNorthern CL \n-0.06 \n0.11 \n0.13 \n0.26 \n0.35 \n3 \nNemunas Delta \n-0.06 \n0.11 \n0.14 \n0.26 \n0.35 \n4 \nCentral CL \n-0.07 \n0.10 \n0.12 \n0.25 \n0.34 \n5 \nSouthern CL \n-0.08 \n0.09 \n0.11 \n0.24 \n0.33 Table 6. The average water level during the historical (1975–2005) and the short- (2020–2050) and long-term (2070–2100) periods \nunder RCP4.5 and RCP8.5 scenarios. Data were averaged over south-eastern Baltic Sea and five selected points in the Curonian \nLagoon (CL) as shown in the map insert. Table 6. The average water level during the historical (1975–2005) and the short- (2020–2050) and long-term (2070–2100) periods \nunder RCP4.5 and RCP8.5 scenarios. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Table 7. Seasonal water residence time (in days) in the northern, southern, and total lagoon area under RCP4.5 and RCP8.5 \n340 \nscenarios, averaged over historical (1975–2005), short-term (2020–2050), and long-term (2070–2100) periods. Percentage in the \nbrackets shows the change compared with the historical period. Table 7. Seasonal water residence time (in days) in the northern, southern, and total lagoon area under RCP4.5 and RCP8.5 \n340 \nscenarios, averaged over historical (1975–2005), short-term (2020–2050), and long-term (2070–2100) periods. Percentage in the \nbrackets shows the change compared with the historical period. The highest WRT in the Curonian Lagoon is during summer and autumn when the discharge from the surrounding rivers \ndecreases. Seasonal changes compared with the WRT during the historical period revealed that the highest decrease is observed \nduring winter, followed by spring. During summer WRT slightly increases under both RCP scenarios and both analysis periods, \n345 \nwhile in autumn the changes are negligible, apart from RCP8.5 scenario in the long-term, when the WRT decreases by ~10%, \nd t th hi t\ni\nl\ni d Table 7. Seasonal water residence time (in days) in the northern, southern, and total lagoon area under RCP4.5 and RCP8.5 \n340 \nscenarios, averaged over historical (1975–2005), short-term (2020–2050), and long-term (2070–2100) periods. Percentage in the \nbrackets shows the change compared with the historical period. The highest WRT in the Curonian Lagoon is during summer and autumn when the discharge from the surrounding rivers \ndecreases. Seasonal changes compared with the WRT during the historical period revealed that the highest decrease is observed \nduring winter, followed by spring. During summer WRT slightly increases under both RCP scenarios and both analysis periods, \n345 \nwhile in autumn the changes are negligible, apart from RCP8.5 scenario in the long-term, when the WRT decreases by ~10%, \ncompared to the historical period. 3.8 Ice thickness Ice thickness in the Curonian Lagoon is projected to decrease steadily. The average maximum ice thickness during the short-\nterm period can decrease by 25% compared with the historical period. Additionally, there is no apparent difference between \n350 \nthe RCP scenarios (Fig. 9). Adverse case is with long-term projections, for which the average maximum ice thickness under \nRCP4.5 scenario is estimated to decrease by half and by 70% under RCP8.5 scenario, compared with the historical period. These changes indicate a possible severe modification of ice regime in the lagoon, leading to winter seasons with a more likely \nunstable thin ice cover. 55 \nFigure 9. Average ice thickness in the Curonian Lagoon. Black line indicates the average ice thickness of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. Table shows the \naverage maximum ice thickness and a percentage of change of the scenario runs compared with the historical period. Figure 9. Average ice thickness in the Curonian Lagoon. Black line indicates the average ice thickness of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. Table shows the \naverage maximum ice thickness and a percentage of change of the scenario runs compared with the historical period. Figure 9. Average ice thickness in the Curonian Lagoon. Black line indicates the average ice thickness of each analysis period – \nhistorical (1975–2005), short-term (2020–2050), and long-term (2070–2100) under RCP4.5 and RCP8.5 scenarios. Table shows the \naverage maximum ice thickness and a percentage of change of the scenario runs compared with the historical period. 3.7 Water residence time The water residence time (WRT) generally is likely to have a decreasing tendency, apart from the summer season, when it \n335 \nslightly increases (Table 7). In the northern part of the lagoon WRT is always shorter, while the southern part has higher values, \nbecause the water movement in this area is more driven by wind than by the outflowing rivers (e.g., Nemunas). There is no \napparent difference between the climate scenarios in the short-term. The highest decrease, compared with the historical period, \nis observed in the long-term under RCP8.5 scenario. 1975-2005 \n2020-2050 \n2070-2100 \n \n \nHistorical \nRCP45 \nRCP85 \nRCP45 \nRCP85 \nNorth \nWinter \n61 \n49 (-21%) \n48 (-21%) \n46 (-24%) \n41 (-32%) \nSpring \n44 \n39 (-11%) \n40 (-9%) \n41 (-7%) \n39 (-12%) \nSummer \n59 \n63 (7%) \n63 (6%) \n66 (11%) \n61 (4%) \nAutumn \n67 \n66 (-1%) \n66 (-1%) \n67 (0.03%) \n62 (-7%) \nSouth \nWinter \n174 \n123 (-29%) \n119 (-32%) \n112 (-36%) \n88 (-49%) \nSpring \n122 \n96 (-22%) \n99 (-19%) \n101 (-17%) \n94 (-23%) \nSummer \n179 \n196 (9%) \n199 (11%) \n217 (21%) \n192 (7%) \nAutumn \n180 \n178 (-1%) \n178 (-1%) \n176 (-2%) \n158 (-12%) \nTotal \nWinter \n127 \n95 (-25%) \n93 (-27%) \n88 (-30%) \n73 (-43%) \nSpring \n94 \n76 (-18%) \n79 (-16%) \n80 (-14%) \n75 (-20%) \nSummer \n129 \n140 (8%) \n140 (8%) \n147 (14%) \n134 (4%) \nAutumn \n135 \n133 (-1%) \n133 (-1%) \n133 (-2%) \n120 (-11%) 16 The results above give an indication on the variation of the physical parameters that the Nemunas River watershed, Curonian \n360 \nLagoon, and south-eastern part of the Baltic Sea will be subjected to. All parameters indicate a change, more or less strong, \nuntil the end of this century. These changes will be discussed hereafter. 4.1 Local climate insights The results of this study have been achieved with regional climate models that have been downscaled from global models. The \nquality of the model output was generally good, but a comparison with local observed data showed the need to ultimately \n365 \ncorrect the downscaled data by applying a bias correction. This was needed because some variables (precipitation suffices as \none example) were far from being well produced by the regional climate models (Tapiador et al., 2019). g\ng\ng\ncentury (Meier et al., 2022a). Accordingly, our study results show a considerable increase in the projected air temperatures as \nwell (Table 8). These changes will likely result in warmer conditions during winters, less snow and ice cover as well as higher \n370 \nriver discharge in winter, while the increase of air temperature during other seasons might have a severe impact on the crop \nproduction, increase the occurrence of hydrological and/or agricultural droughts and impact changes in the crop calendar of \nthe region. well (Table 8). These changes will likely result in warmer conditions during winters, less snow and ice cover as well as higher \n370 \nriver discharge in winter, while the increase of air temperature during other seasons might have a severe impact on the crop \nproduction, increase the occurrence of hydrological and/or agricultural droughts and impact changes in the crop calendar of \nthe region. Parameter \nDomain \n2020-2050 \n2070-2100 \nRCP4.5 \nRCP8.5 \nRCP4.5 \nRCP8.5 \nAir temperature \nMarine area \n21% \n23% \n34% \n57% \nTerrestrial area \n33% \n35% \n51% \n81% \nPrecipitation \nMarine area \n14% \n15% \n18% \n34% \nTerrestrial area \n14% \n16% \n18% \n35% \nWater inflow \nNemunas River to Lagoon \n13% \n16% \n19% \n37% \nWater outflow \nLagoon to sea \n15% \n18% \n21% \n40% \nSalinity \nCoastal area \n-8% \n-10% \n-15% \n-24% \nSea \n-5% \n-6% \n-11% \n-18% \nSaltwater intrusions (No \nof days year-1) \nJuodkrantė \n-30% \n-34% \n-34% \n-51% \nWater temperature \nLagoon \n11% \n13% \n19% \n26% \nCoastal area \n13% \n14% \n21% \n37% \nSea \n17% \n19% \n30% \n50% \nWater residence time \nNorther part of the lagoon \n-6% \n-6% \n-5% \n-12% \nSouthern part of the lagoon \n-10% \n-9% \n-8% \n-19% \nTotal lagoon area \n-8% \n-8% \n-8% \n-17% \nTable 8. Summary of the percentage change of projected average annual changes (under RCP4.5 and RCP8.5 scenarios) during \nshort-term (2020–2050) and long-term (2070–2100) periods compared with the historical period (1975–2005). 375 \nWith increasing mean temperatures, a distinct increase in the projected monthly and annual average precipitation is also \nexpected. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. 4 Discussion The results above give an indication on the variation of the physical parameters that the Nemunas River watershed, Curonian \n360 \nLagoon, and south-eastern part of the Baltic Sea will be subjected to. All parameters indicate a change, more or less strong, \nuntil the end of this century. These changes will be discussed hereafter. 17 8. Summary of the percentage change of projected average annual changes (under RCP4.5 and RCP With increasing mean temperatures, a distinct increase in the projected monthly and annual average precipitation is also \nexpected. Our results show that by the end of the century an increase of the total annual precipitation by 18 to 34% (~100–200 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. mm) according to RCP4.5 and RCP 8.5 scenarios respectively, is projected. In comparison with a neighbour country, this is \nconsiderably higher than the projected precipitation for the Latvian case, where the total annual increase was estimated to be \n13 to 16% (~80–100 mm) under the same scenarios (Avotniece et al., 2017). 80 mm) according to RCP4.5 and RCP 8.5 scenarios respectively, is projected. In comparison with a neighbour country, this is \nconsiderably higher than the projected precipitation for the Latvian case, where the total annual increase was estimated to be \n13 to 16% (~80–100 mm) under the same scenarios (Avotniece et al., 2017). 380 380 4.2 Water flow The driving force of changes in the lagoon are the boundaries – the Nemunas River inputs and the influence of the Baltic Sea \nthrough the Klaipėda Strait. Historically every year on average Nemunas River supplies three times the volume of water to the \nlagoon (Žilinskas et al., 2012), which is subject to change in the future. The projections indicate a 13–16% (in the short-term) \nand 19–37% (in the long-term) increase in water discharge from the river, which coincides with the projected precipitation \n385 \nincrease (Table 8). This projected change alone will strongly influence the hydrodynamic conditions of the lagoon. The driving force of changes in the lagoon are the boundaries – the Nemunas River inputs and the influence of the Baltic Sea \nthrough the Klaipėda Strait. Historically every year on average Nemunas River supplies three times the volume of water to the \nlagoon (Žilinskas et al., 2012), which is subject to change in the future. The projections indicate a 13–16% (in the short-term) \nand 19–37% (in the long-term) increase in water discharge from the river, which coincides with the projected precipitation \n385 \nincrease (Table 8). This projected change alone will strongly influence the hydrodynamic conditions of the lagoon. Due to the enhanced Nemunas discharge, the fluxes from the Curonian Lagoon into the Baltic Sea will also increase. The \noutflow to the sea is already dominating, especially in the northern part of the lagoon and it will remain so in the future. The \nwater flow will increase in all its path towards the Baltic Sea by 13–18% in the short-term and by ~20% under RCP4.5 scenario Due to the enhanced Nemunas discharge, the fluxes from the Curonian Lagoon into the Baltic Sea will also increase. The \noutflow to the sea is already dominating, especially in the northern part of the lagoon and it will remain so in the future. The \nwater flow will increase in all its path towards the Baltic Sea by 13–18% in the short-term and by ~20% under RCP4.5 scenario in the long-term, while the changes under RCP8.5 scenario can be expected to increase by up to 40% by the end of the century, \n390 \ncompared with the historical period (Table 8). The volume of discharging river water into the Baltic Sea is projected to be dependent on the season and the region (Graham, \n2004; Meier et al., 2022a). 4.1 Local climate insights Our results show that by the end of the century an increase of the total annual precipitation by 18 to 34% (~100–200 18 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. 4.3 Water level dynamics Water levels are the main drivers, together with the Nemunas discharge, of the exchanges through the Klaipėda Strait. In the \n410 \nstrait and particularly in the Curonian Lagoon water levels are higher than in the Baltic Sea, which leads to an increase in \noutflow periods through the strait. Since 1960 until 2008, the average measured water level in the Curonian Lagoon has been \nrising at a rate of approximately 3.0 mm year−1, leading to increase by 18 cm during this period (Čepienė et al., 2022, and \nreferences therein). Our modelling results showed that in the future, water level is projected to increase by 16–20 cm in the short-term and by 32–41 cm in the long-term, compared with the historical period. 415 \nSince the projected water level will be above the mean sea level, it could affect the infrastructure and residents living in the \ncoastal area leading to a significant social and economic damages (Vousdoukas et al., 2020). The low elevation wetland areas \ncovering the southern and eastern part of the lagoon (especially Nemunas Delta area) are more likely of getting flooded. However, the Klaipėda city and the Curonian Spit are located higher and the mean water level rise should not permanently short-term and by 32–41 cm in the long-term, compared with the historical period. 415 \nSince the projected water level will be above the mean sea level, it could affect the infrastructure and residents living in the \ncoastal area leading to a significant social and economic damages (Vousdoukas et al., 2020). The low elevation wetland areas \ncovering the southern and eastern part of the lagoon (especially Nemunas Delta area) are more likely of getting flooded. However, the Klaipėda city and the Curonian Spit are located higher and the mean water level rise should not permanently affect these areas. It is worth mentioning, that the hydrodynamic model has a fixed boundary, thus the increase could be \n420 \noverestimated especially considering the role of coastal wetlands in mediation of global change induced water level changes. affect these areas. It is worth mentioning, that the hydrodynamic model has a fixed boundary, thus the increase could be \n420 \noverestimated especially considering the role of coastal wetlands in mediation of global change induced water level changes. 4.4 Salinity dynamics Our study \nof salinity changes in the south-eastern Baltic Sea demonstrates a decreasing tendency, especially in the coastal area, northward \n430 \nfrom the lagoon outlet, where the sea water is diluted by discharging freshwater from the lagoon (Table 8). of salinity changes in the south-eastern Baltic Sea demonstrates a decreasing tendency, especially in the coastal area, northward \n430 \nfrom the lagoon outlet, where the sea water is diluted by discharging freshwater from the lagoon (Table 8). 4.2 Water flow With this, the increase of riverine nutrient loads is also anticipated (Pihlainen et al., 2020), likewise \nsuggesting that “land based nutrient management will have greater effect on nutrient loads than greenhouse gas emissions” (Climate Change in the Baltic Sea: 2021 Fact Sheet). 395 \nThe seasonal river discharge changes in the Baltic Sea correspond with our study noting that a more likely increase in winter \nand decrease in summer can be expected by the end of this century (Donnelly et al., 2014). The magnitude of changes of \nfreshwater inflow within different subregions of the Baltic Sea has a considerable variation, i.e., an increased outflow in the \nnorthern part of the Baltic Sea and a decreased in the southern part (Graham, 2004). Although the combined uncertainties from the climate models, their bias correction, and applied models make it difficult to draw conclusions about the magnitude of \n400 \nchange (Donnelly et al., 2014). The most alarming are the projected hydrologic changes in the Nemunas River watershed during the winter season in the long-\nterm period. As supported by other studies (Čerkasova et al., 2021; Stonevičius et al., 2017), both RCPs point toward a drastic \nincrease in winter discharges, which might lead to severe flooding of the delta region, which suffers from annual floods of different magnitude (Valiuškevičius et al., 2018). Currently these flood events occur in spring, but will likely occur in winter \n405 \nin the future, and with increased severity, which will pose a serious concern to the local population and stakeholders. The \ncurrently functioning polder system in the Nemunas River delta (Lesutienė et al., 2022) region might not withstand such \nchanges. different magnitude (Valiuškevičius et al., 2018). Currently these flood events occur in spring, but will likely occur in winter \n405 \nin the future, and with increased severity, which will pose a serious concern to the local population and stakeholders. The \ncurrently functioning polder system in the Nemunas River delta (Lesutienė et al., 2022) region might not withstand such \nchanges. 19 4.4 Salinity dynamics When looking at the saltwater intrusion events, it can be clearly seen that the intrusion periods are decreasing and the outflow \nfrom the lagoon becomes dominant, simply because the outflowing water fluxes will stop the salty seawater from entering the \nCuronian Lagoon. This is clearly evident for both climate change scenarios – periods where values higher than the threshold \n425 \nof 2 g kg-1 in Juodkrantė drops by 30–50% compared with the historical period (Table 8). This will lead to some alteration of \nthe ecosystem in the northern part of the Curonian lagoon, which is now considered as oligohaline. When looking at the saltwater intrusion events, it can be clearly seen that the intrusion periods are decreasing and the outflow \nfrom the lagoon becomes dominant, simply because the outflowing water fluxes will stop the salty seawater from entering the As seen and suggested from other Baltic Sea salinity projection studies, the salinity in the Baltic does not have a distinct pattern \nof changing in the future and the projections have a lot of uncertainties (Lehmann et al., 2022; Meier et al., 2022b). Our study \nof salinity changes in the south-eastern Baltic Sea demonstrates a decreasing tendency, especially in the coastal area, northward \n430 \nfrom the lagoon outlet, where the sea water is diluted by discharging freshwater from the lagoon (Table 8). As seen and suggested from other Baltic Sea salinity projection studies, the salinity in the Baltic does not have a distinct pattern \nof changing in the future and the projections have a lot of uncertainties (Lehmann et al., 2022; Meier et al., 2022b). Our study \nof salinity changes in the south-eastern Baltic Sea demonstrates a decreasing tendency, especially in the coastal area, northward \n430 As seen and suggested from other Baltic Sea salinity projection studies, the salinity in the Baltic does not have a distinct pattern \nof changing in the future and the projections have a lot of uncertainties (Lehmann et al., 2022; Meier et al., 2022b). Our study As seen and suggested from other Baltic Sea salinity projection studies, the salinity in the Baltic does not have a distinct pattern \nof changing in the future and the projections have a lot of uncertainties (Lehmann et al., 2022; Meier et al., 2022b). https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. the historical period. Additionally, the coastal average water temperature is lower by 1.5 °C during the winter and higher by \n4.9 °C during the rest of the year than that of the south-eastern Baltic Sea. 440 the historical period. Additionally, the coastal average water temperature is lower by 1.5 °C during the winter and higher by 4.9 °C during the rest of the year than that of the south-eastern Baltic Sea. 440 \nOur study supports the findings of projected increasing water temperature in the Baltic Sea (Meier et al., 2022b). In this article \nwe focused only on the average water temperature of the water column, however, there are many studies revealing the warming \npatterns of the sea surface temperature in the Baltic (Dutheil et al., 2022; Zhu et al., 2022). 4.9 C during the rest of the year than that of the south eastern Baltic Sea. 440 \nOur study supports the findings of projected increasing water temperature in the Baltic Sea (Meier et al., 2022b). In this article \nwe focused only on the average water temperature of the water column, however, there are many studies revealing the warming \npatterns of the sea surface temperature in the Baltic (Dutheil et al., 2022; Zhu et al., 2022). 4.7 Ice thickness in the lagoon Although, the role of ice for \ncontrolling WRT is far less than the increased fluxes from the discharging rivers. modelling studies of future projections (Idzelytė and Umgiesser, 2021; Jakimavičius et al., 2019). The decreasing ice \n460 \nparameters impact on the hydrodynamics of the lagoon suggest a possible increase of water flow and saltwater intrusions, as \nwell as shorter WRT during the ice cover season (Idzelytė et al., 2020; Umgiesser et al., 2016). Although, the role of ice for \ncontrolling WRT is far less than the increased fluxes from the discharging rivers. modelling studies of future projections (Idzelytė and Umgiesser, 2021; Jakimavičius et al., 2019). The decreasing ice \n460 \nparameters impact on the hydrodynamics of the lagoon suggest a possible increase of water flow and saltwater intrusions, as \nwell as shorter WRT during the ice cover season (Idzelytė et al., 2020; Umgiesser et al., 2016). Although, the role of ice for \ncontrolling WRT is far less than the increased fluxes from the discharging rivers. 4.6 Water residence time in the lagoon Water residence time (WRT) is generally decreasing with climate change (Table 8), although the decline is small in the short-\n445 \nterm and in the long-term under RCP4.5 scenario (5–10% less). A general decrease can be found in the whole area of the \nlagoon in the long-term under RCP8.5 (17%). As in the case with saltwater intrusions, this is due to the increased outflow from \nthe lagoon to the sea, where WRT are less varying in the north, whereas it decreased sensibly in the south. When looking at seasonal changes, WRT becomes lower in summer due to decreasing river discharge. Comparing the changes Water residence time (WRT) is generally decreasing with climate change (Table 8), although the decline is small in the short-\n445 \nterm and in the long-term under RCP4.5 scenario (5–10% less). A general decrease can be found in the whole area of the \nlagoon in the long-term under RCP8.5 (17%). As in the case with saltwater intrusions, this is due to the increased outflow from \nthe lagoon to the sea, where WRT are less varying in the north, whereas it decreased sensibly in the south. When looking at seasonal changes, WRT becomes lower in summer due to decreasing river discharge. Comparing the changes When looking at seasonal changes, WRT becomes lower in summer due to decreasing river discharge. Comparing the changes \nfrom the historical period it revealed that the highest decrease is observed during winter and spring, presumably due to the \n450 \nincreased outflowing water fluxes and change of the ice cover regime. When looking at seasonal changes, WRT becomes lower in summer due to decreasing river discharge. Comparing the changes \nfrom the historical period it revealed that the highest decrease is observed during winter and spring, presumably due to the \n450 \nincreased outflowing water fluxes and change of the ice cover regime. from the historical period it revealed that the highest decrease is observed during winter and spring, presumably due to the \n450 \nincreased outflowing water fluxes and change of the ice cover regime. 4.7 Ice thickness in the lagoon Ice thickness projections in the lagoon will become thinner for a longer period during the ice season compared to the historical \nperiod. The projected changes of 25% decreased maximum ice thickness in the short-term and 50–70% in the long-term, could \npresume leading to severe alteration of the ice season regime in the future. This pattern corresponds to the trends in the southern \n455 \nBaltic lagoons (Girjatowicz and Świątek, 2021) and gulf of Riga (Siitam et al., 2017; Kļaviņš et al., 2016;), as well as Baltic \nSea itself (Luomaranta et al 2014) Ice thickness projections in the lagoon will become thinner for a longer period during the ice season compared to the historical \nperiod. The projected changes of 25% decreased maximum ice thickness in the short-term and 50–70% in the long-term, could period. The projected changes of 25% decreased maximum ice thickness in the short term and 50 70% in the long term, could \npresume leading to severe alteration of the ice season regime in the future. This pattern corresponds to the trends in the southern \n455 \nBaltic lagoons (Girjatowicz and Świątek, 2021) and gulf of Riga (Siitam et al., 2017; Kļaviņš et al., 2016;), as well as Baltic \nSea itself (Luomaranta et al., 2014). Ice in the Curonian Lagoon has become a trending topic in the recent years, due to its already evident changes as seen from \nthe historical observations (Jakimavičius et al., 2019), remote sensing (Kozlov et al., 2020; Idzelytė et al., 2019), as well as presume leading to severe alteration of the ice season regime in the future. This pattern corresponds to the trends in the southern \n455 \nBaltic lagoons (Girjatowicz and Świątek, 2021) and gulf of Riga (Siitam et al., 2017; Kļaviņš et al., 2016;), as well as Baltic \nSea itself (Luomaranta et al., 2014). Ice in the Curonian Lagoon has become a trending topic in the recent years, due to its already evident changes as seen from \nthe historical observations (Jakimavičius et al., 2019), remote sensing (Kozlov et al., 2020; Idzelytė et al., 2019), as well as modelling studies of future projections (Idzelytė and Umgiesser, 2021; Jakimavičius et al., 2019). The decreasing ice \n460 \nparameters impact on the hydrodynamics of the lagoon suggest a possible increase of water flow and saltwater intrusions, as \nwell as shorter WRT during the ice cover season (Idzelytė et al., 2020; Umgiesser et al., 2016). 4.5 Water temperature dynamics The mean temperature of the water column in the Klaipėda Strait is less sensitive to changes. In the lagoon water temperature \nis projected to increase under both RCP scenarios compared with the historical period. The highest changes are projected in \nthe long-term under RCP8.5 scenario (Table 8). 435 the long-term under RCP8.5 scenario (Table 8). 435 \nThe increasing temperature patterns are also evident in the south-eastern Baltic Sea, having a higher increase in the long-term \nthan in the short-term. The highest change, compared with the historical period, is mostly occurring during winter and spring. Considering only the coastal area, it represents similar dynamics of the sea, although, having higher changes compared with The increasing temperature patterns are also evident in the south-eastern Baltic Sea, having a higher increase in the long-term \nthan in the short-term. The highest change, compared with the historical period, is mostly occurring during winter and spring. Considering only the coastal area, it represents similar dynamics of the sea, although, having higher changes compared with 20 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. A recent analysis (Ivanauskas et al., 2022) indicated statistical relationship between WRT and salinity and both catches and \npopulations of main commercial fish populations. This is in stark contrast to the predictions for the Baltic Sea fishery, where \n470 \ncatches are expected to be decreasing in quantity, and especially in the quality (Climate Change in the Baltic Sea, 2021). This \ncould be explained by the fact that commercial catches in the Curonian Lagoon already are dominated by eutrophication \ntolerant species like bream and pikeperch, which are also known to prefer higher water temperatures. However, projected ice \ncover decrease will inevitably negatively affect the winter ice fishing practices in the lagoon. The latest studies, revealing the mechanisms of nutrient cycling and eutrophication processes (Bartoli et al., 2018), point \n475 \ntoward the expected reduction of the role of the Curonian Lagoon as a coastal filter. The shortening of the ice cover period \nwill reduce the system denitrification capacity while the decrease of WRT will lead to substantial reduction of nutrient \nretention. So far, the only difference is the more favourable conditions for the cage aquaculture in the lagoon, where the \ndecreased WRT and increased flushing in combination with modern practices of multi trophic aquaculture (e.g. floating \nvegetable gardens) could remove some of the constraints existing at present. 480 The latest studies, revealing the mechanisms of nutrient cycling and eutrophication processes (Bartoli et al., 2018), point \n475 \ntoward the expected reduction of the role of the Curonian Lagoon as a coastal filter. The shortening of the ice cover period \nwill reduce the system denitrification capacity while the decrease of WRT will lead to substantial reduction of nutrient \nretention. So far, the only difference is the more favourable conditions for the cage aquaculture in the lagoon, where the \ndecreased WRT and increased flushing in combination with modern practices of multi trophic aquaculture (e.g. floating \nvegetable gardens) could remove some of the constraints existing at present. 480 4.8 Impacts on the ecosystem structure and functions Despite clear trends for physical indicators, the actual consequences of climate change for the functioning of the Curonian \n465 \nLagoon ecosystem are much more complicated to predict. The foreseen salinity changes in the mostly freshwater ecosystem \nare not of the scale suggesting significant shifts in the ecosystem functioning even in the northern part of the lagoon. However, \nsome decrease in distribution areas of some estuarine species (mysids, barnacles) could be foreseen. Despite clear trends for physical indicators, the actual consequences of climate change for the functioning of the Curonian \n465 \nLagoon ecosystem are much more complicated to predict. The foreseen salinity changes in the mostly freshwater ecosystem \nare not of the scale suggesting significant shifts in the ecosystem functioning even in the northern part of the lagoon. However, \nsome decrease in distribution areas of some estuarine species (mysids, barnacles) could be foreseen. 21 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Acknowledgements This project has received funding from the Research Council of Lithuania (LMTLT), agreement No S-MIP-21-24. This project has received funding from the Research Council of Lithuania (LMTLT), agreement No S-MIP-21-24. Data availability All numerical modelling results created during this study are openly available in the Zenodo open data repository \n(https://doi.org/10.5281/zenodo.7500744) as cited in Idzelytė et al. (2023). 5 Conclusions The setup of the complex Nemunas River–Curonian Lagoon–Baltic Sea model has been successfully used in the climate change \nstudies presented in this paper. The high-resolution hydrological model for the Nemunas River and hydrodynamic model for \nthe Curonian Lagoon and the Baltic Sea were used. The finite element approach has allowed us to achieve a good compromise \nof needed resolution and computer efficiency to carry out all needed hydrodynamic simulations. With the help of downscaled \n485 \nclimate models, which have been ultimately bias-corrected with observations, the tendency of changes was inferred. The study results showed that the Curonian Lagoon will be subject to higher Nemunas discharges that will in turn increase the \noutgoing fluxes into the Baltic Sea. Through these higher fluxes both water residence times and saltwater intrusion events will \ndecrease. Most of these changes will however be only noticeable in the northern part of the lagoon, which are more likely to be influenced by the Nemunas discharge. The southern part of the lagoon will experience much less changes. 490 \nThe foreseen changes in physical characteristics are not of the scale suggesting significant shifts in the ecosystem functioning, \nbut expected rather to manifest in some quantitative alterations in the nutrient retention capacity. However, some ecosystem \nservices such as ice fishing are expected to be completely vanishing due to the physical constraints. Whereas tendencies could be identified in our analysis, it was yet not possible to discuss the uncertainty of the results achieved. The foreseen changes in physical characteristics are not of the scale suggesting significant shifts in the ecosystem functioning, \nbut expected rather to manifest in some quantitative alterations in the nutrient retention capacity. However, some ecosystem \nservices such as ice fishing are expected to be completely vanishing due to the physical constraints. This can be done only through extensive sensitivity analysis, and was not part of the scope of this research article. It will \n495 \nhowever be part of another investigation that will be carried out in the near future. This can be done only through extensive sensitivity analysis, and was not part of the scope of this research article. It will \n495 \nhowever be part of another investigation that will be carried out in the near future. 22 Author contribution \n500 GU and NC initiated the conceptualization and funding acquisition of the research project. RI, NC, and JM performed the \nanalysis and drafted the paper. RI curated the visualization of the results. RI, NC, JM, TD, ARB, AE prepared the original \nmanuscript draft with the assistance of GU. All co-authors reviewed the paper and contributed to the scientific interpretation \nand discussion. GU and NC initiated the conceptualization and funding acquisition of the research project. RI, NC, and JM performed the \nanalysis and drafted the paper. RI curated the visualization of the results. RI, NC, JM, TD, ARB, AE prepared the original \nmanuscript draft with the assistance of GU. All co-authors reviewed the paper and contributed to the scientific interpretation \nand discussion. Competing interests. 505 \nThe authors declare that they have no conflict of interest. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Belkin, \nI. M.: \nRapid \nwarming \nof \nLarge \nMarine \nEcosystems, \nProg. Oceanogr., \n81, \n207–213, \nhttps://doi.org/10.1016/j.pocean.2009.04.011, 2009. Bellafiore, D. and Umgiesser, G.: Hydrodynamic coastal processes in the North Adriatic investigated with a 3D finite element \n525 \nmodel, Ocean Dyn., 60, 255–273, https://doi.org/10.1007/s10236-009-0254-x, 2010. Čepienė, E., Dailidytė, L., Stonevičius, E., and Dailidienė, I.: Sea Level Rise Impact on Compound Coastal River Flood Risk \nin Klaipėda City (Baltic Coast, Lithuania), Water, 14, 414, https://doi.org/10.3390/w14030414, 2022. Čerkasova, N., Umgiesser, G., and Ertürk, A.: Development of a hydrology and water quality model for a large transboundary Bellafiore, D. and Umgiesser, G.: Hydrodynamic coastal processes in the North Adriatic investigated with a 3D finite element \n525 \nmodel, Ocean Dyn., 60, 255–273, https://doi.org/10.1007/s10236-009-0254-x, 2010. Čepienė, E., Dailidytė, L., Stonevičius, E., and Dailidienė, I.: Sea Level Rise Impact on Compound Coastal River Flood Risk \ni\nl i\nd Ci\n(\nl i C\ni h\ni )\n14 414 h\n//d i\n/10 3390/ 14030414 2022 Bellafiore, D. and Umgiesser, G.: Hydrodynamic coastal processes in the North Adriatic investigated with a 3D finite element \n525 \nmodel, Ocean Dyn., 60, 255–273, https://doi.org/10.1007/s10236-009-0254-x, 2010. Čepienė, E., Dailidytė, L., Stonevičius, E., and Dailidienė, I.: Sea Level Rise Impact on Compound Coastal River Flood Risk \nin Klaipėda City (Baltic Coast, Lithuania), Water, 14, 414, https://doi.org/10.3390/w14030414, 2022. river watershed to investigate the impacts of climate change – A SWAT application, Ecol. Eng., 124, 99–115, \n530 \nhttps://doi.org/10.1016/j.ecoleng.2018.09.025, 2018. Čerkasova, N., Umgiesser, G., and Ertürk, A.: Assessing Climate Change Impacts on Streamflow, Sediment and Nutrient \nLoadings of the Minija River (Lithuania): A Hillslope Watershed Discretization Application with High-Resolution Spatial \nInputs, https://doi.org/10.3390/w11040676, 2019. Čerkasova, N., Umgiesser, G., and Ertürk, A.: Modelling framework for flow, sediments and nutrient loads in a large \n535 \ntransboundary river watershed: A climate change impact assessment of the Nemunas River watershed, J. Hydrol., 598, 126422, \nhttps://doi.org/10.1016/j.jhydrol.2021.126422, 2021. Climate Change in the Baltic Sea: 2021 Fact Sheet. Baltic Sea Environment Proceedings n°180. HELCOM/Baltic Earth 2021, \n2021. Climate Change in the Baltic Sea: 2021 Fact Sheet. Baltic Sea Environment Proceedings n°180. HELCOM/Baltic Earth 2021, \n2021. Donnelly, C., Yang, W., and Dahné, J.: River discharge to the Baltic Sea in a future climate, Clim. Change, 122, 157–170, \n540 \nhttps://doi.org/10.1007/s10584-013-0941-y, 2014. References Anthony, A., Atwood, J., August, P., Byron, C., Cobb, S., Foster, C., Fry, C., Gold, A., Hagos, K., Heffner, L., Kellogg, D. 510 \nQ., Lellis-Dibble, K., Opaluch, J. J., Oviatt, C., Pfeiffer-Herbert, A., Rohr, N., Smith, L., Smythe, T., Swift, J., and Vinhateiro, \nN.: Coastal Lagoons and Climate Change: Ecological and Social Ramifications in U.S. Atlantic and Gulf Coast Ecosystems, \nEcol. Soc., 14, art8, https://doi.org/10.5751/ES-02719-140108, 2009. Avotniece, Z., Aniskevich, S., and Malinovskis, E.: Climate Change Scenarios for Latvia. Report summary., Riga, 2017. Anthony, A., Atwood, J., August, P., Byron, C., Cobb, S., Foster, C., Fry, C., Gold, A., Hagos, K., Heffner, L., Kellogg, D. 510 \nQ., Lellis-Dibble, K., Opaluch, J. J., Oviatt, C., Pfeiffer-Herbert, A., Rohr, N., Smith, L., Smythe, T., Swift, J., and Vinhateiro, \nN.: Coastal Lagoons and Climate Change: Ecological and Social Ramifications in U.S. Atlantic and Gulf Coast Ecosystems, \nEcol. Soc., 14, art8, https://doi.org/10.5751/ES-02719-140108, 2009. Avotniece, Z., Aniskevich, S., and Malinovskis, E.: Climate Change Scenarios for Latvia. Report summary., Riga, 2017. Avotniece, Z., Aniskevich, S., and Malinovskis, E.: Climate Change Scenarios for Latvia. Report summary., Riga, 2017. BACC Author Team: Assessment of Climate Change for the Baltic Sea Basin, Springer Berlin Heidelberg, Berlin, Heidelberg, \n515 \nhttps://doi.org/10.1007/978-3-540-72786-6, 2008. BACC II Author Team: Second Assessment of Climate Change for the Baltic Sea Basin, edited by: The BACC II Author \nTeam, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-319-16006-1, 2015. Bartoli, M., Zilius, M., Bresciani, M., Vaiciute, D., Vybernaite-Lubiene, I., Petkuviene, J., Giordani, G., Daunys, D., Ruginis, BACC Author Team: Assessment of Climate Change for the Baltic Sea Basin, Springer Berlin Heidelberg, Berlin, Heidelberg, \n515 \nhttps://doi.org/10.1007/978-3-540-72786-6, 2008. BACC II Author Team: Second Assessment of Climate Change for the Baltic Sea Basin, edited by: The BACC II Author \nTeam, Springer International Publishing, Cham, https://doi.org/10.1007/978-3-319-16006-1, 2015. T., Benelli, S., Giardino, C., Bukaveckas, P. A., Zemlys, P., Griniene, E., Gasiunaite, Z. R., Lesutiene, J., Pilkaitytė, R., and \n520 \nBaziukas-Razinkovas, A.: Drivers of Cyanobacterial Blooms in a Hypertrophic Lagoon, Front. Mar. Sci., 5, \nhttps://doi.org/10.3389/fmars.2018.00434, 2018. 23 Institute, Kaunas, Lithuania, 792, 2001. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Gasiūnaitė, Z. R., Daunys, D., Olenin, S., and Razinkovas, A.: The Curonian Lagoon, in: Ecology of Baltic Coastal Waters, \nedited by: Schiewer, U., Springer, Berlin/Heidelberg, Germany, 197–215, https://doi.org/10.1007/978-3-540-73524-3_9, \n2008. Girjatowicz, J. P. and Świątek, M.: Relationship between Air Temperature Change and Southern Baltic Coastal Lagoons Ice Girjatowicz, J. P. and Świątek, M.: Relationship between Air Temperature Change and Southern Baltic Coastal Lagoons Ice Conditions, Atmosphere (Basel)., 12, 931, https://doi.org/10.3390/atmos12080931, 2021. 560 \nGraham, L. P.: Climate Change Effects on River Flow to the Baltic Sea, Ambio, 33, 235–241, 2004. Gröger, M., Arneborg, L., Dieterich, C., Höglund, A., and Meier, H. E. M.: Summer hydrographic changes in the Baltic Sea, \nKattegat and Skagerrak projected in an ensemble of climate scenarios downscaled with a coupled regional ocean–sea ice–\natmosphere model, Clim. Dyn., 53, 5945–5966, https://doi.org/10.1007/s00382-019-04908-9, 2019. Gupta, R., Bhattarai, R., and Mishra, A.: Development of Climate Data Bias Corrector (CDBC) Tool and Its Application over \n565 \nthe Agro-Ecological Zones of India, Water, 11, 1102, https://doi.org/10.3390/w11051102, 2019. Gupta, R., Bhattarai, R., and Mishra, A.: Development of Climate Data Bias Corrector (CDBC) Tool and Its Application over \n565 \nthe Agro-Ecological Zones of India, Water, 11, 1102, https://doi.org/10.3390/w11051102, 2019. Holt, J., Schrum, C., Cannaby, H., Daewel, U., Allen, I., Artioli, Y., Bopp, L., Butenschon, M., Fach, B. A., Harle, J., \nPushpadas, D., Salihoglu, B., and Wakelin, S.: Potential impacts of climate change on the primary production of regional seas: \nA comparative analysis of five European seas, Prog. Oceanogr., 140, 91–115, https://doi.org/10.1016/j.pocean.2015.11.004, 2016. 570 \nIdzelytė, R. and Umgiesser, G.: Application of an ice thermodynamic model to a shallow freshwater lagoon, Boreal Environ. Res., 26, 61–77, 2021. Idzelytė, R., Kozlov, I. E., and Umgiesser, G.: Remote Sensing of Ice Phenology and Dynamics of Europe’s Largest Coastal \nLagoon (The Curonian Lagoon), Remote Sens., 11, 2059, https://doi.org/10.3390/rs11172059, 2019. Idzelytė, R., Kozlov, I. E., and Umgiesser, G.: Remote Sensing of Ice Phenology and Dynamics of Europe’s Largest Coastal \nLagoon (The Curonian Lagoon), Remote Sens., 11, 2059, https://doi.org/10.3390/rs11172059, 2019. Idzelytė, R., Mėžinė, J., Zemlys, P., and Umgiesser, G.: Study of ice cover impact on hydrodynamic processes in the Curonian \n575 \nLagoon through numerical modeling, Oceanologia, 62, 428–442, https://doi.org/10.1016/j.oceano.2020.04.006, 2020. Donnelly, C., Yang, W., and Dahné, J.: River discharge to the Baltic Sea in a future climate, Clim. Change, 122, 157–170, \n540 \nhttps://doi.org/10.1007/s10584-013-0941-y, 2014. Dutheil, C., Meier, H. E. M., Gröger, M., and Börgel, F.: Understanding past and future sea surface temperature trends in the \nBaltic Sea, Clim. Dyn., 58, 3021–3039, https://doi.org/10.1007/s00382-021-06084-1, 2022. Ferrarin, C. and Umgiesser, G.: Hydrodynamic modeling of a coastal lagoon: The Cabras lagoon in Sardinia, Italy, Ecol. Dutheil, C., Meier, H. E. M., Gröger, M., and Börgel, F.: Understanding past and future sea surface temperature trends in the \nBaltic Sea, Clim. Dyn., 58, 3021–3039, https://doi.org/10.1007/s00382-021-06084-1, 2022. Ferrarin, C. and Umgiesser, G.: Hydrodynamic modeling of a coastal lagoon: The Cabras lagoon in Sardinia, Italy, Ecol. Modell., 188, 340–357, https://doi.org/10.1016/j.ecolmodel.2005.01.061, 2005. 545 Modell., 188, 340–357, https://doi.org/10.1016/j.ecolmodel.2005.01.061, 2005. 545 \nFerrarin, C., Razinkovas, A., Gulbinskas, S., Umgiesser, G., and Bliudžiute, L.: Hydraulic regime-based zonation scheme of \nthe Curonian Lagoon, Hydrobiologia, 611, 133–146, https://doi.org/10.1007/s10750-008-9454-5, 2008. Ferrarin, C., Umgiesser, G., Bajo, M., Bellafiore, D., De Pascalis, F., Ghezzo, M., Mattassi, G., and Scroccaro, I.: Hydraulic \nzonation of the lagoons of Marano and Grado, Italy. A modelling approach, Estuar. Coast. Shelf Sci., 87, 561–572, ,\n,\n,\np\ng\nj\n,\nFerrarin, C., Razinkovas, A., Gulbinskas, S., Umgiesser, G., and Bliudžiute, L.: Hydraulic regime-based zonation scheme of \nthe Curonian Lagoon, Hydrobiologia, 611, 133–146, https://doi.org/10.1007/s10750-008-9454-5, 2008. Ferrarin, C., Umgiesser, G., Bajo, M., Bellafiore, D., De Pascalis, F., Ghezzo, M., Mattassi, G., and Scroccaro, I.: Hydraulic Ferrarin, C., Razinkovas, A., Gulbinskas, S., Umgiesser, G., and Bliudžiute, L.: Hydraulic regime-based zonation scheme of \nthe Curonian Lagoon, Hydrobiologia, 611, 133–146, https://doi.org/10.1007/s10750-008-9454-5, 2008. Ferrarin, C., Umgiesser, G., Bajo, M., Bellafiore, D., De Pascalis, F., Ghezzo, M., Mattassi, G., and Scroccaro, I.: Hydraulic \nzonation of the lagoons of Marano and Grado, Italy. A modelling approach, Estuar. Coast. Shelf Sci., 87, 561–572, \nhttps://doi.org/10.1016/j.ecss.2010.02.012, 2010. 550 https://doi.org/10.1016/j.ecss.2010.02.012, 2010. 550 \nFerrarin, C., Bergamasco, A., Umgiesser, G., and Cucco, A.: Hydrodynamics and spatial zonation of the Capo Peloro coastal \nsystem \n(Sicily) \nthrough \n3-D \nnumerical \nmodeling, \nJ. Mar. Syst., \n117–118, \n96–107, \nhttps://doi.org/10.1016/j.jmarsys.2013.02.005, 2013. Gailiušis, B., Jablonskis, J., and Kovalenkovienė, M.: The Lithuanian rivers, in: Hydrography and Runoff, Lithuanian Energy Gailiušis, B., Jablonskis, J., and Kovalenkovienė, M.: The Lithuanian rivers, in: Hydrography and Runoff, Lithuanian Energy \nInstitute, Kaunas, Lithuania, 792, 2001. 555 Institute, Kaunas, Lithuania, 792, 2001. 555 24 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. Jakimavičius, D., Šarauskienė, D., and Kriaučiūnienė, J.: Influence of climate change on the ice conditions of the Curonian \nLagoon, Oceanologia, 62, 164–172, https://doi.org/10.1016/j.oceano.2019.10.003, 2019. 0 Jakimavičius, D., Šarauskienė, D., and Kriaučiūnienė, J.: Influence of climate change on the ice conditions of the Curonian Lagoon, Oceanologia, 62, 164–172, https://doi.org/10.1016/j.oceano.2019.10.003, 2019. 590 \nKļaviņš, M., Avotniece, Z., and Rodinovs, V.: Dynamics and Impacting Factors of Ice Regimes in Latvia Inland and Coastal \nWaters, Proc. Latv. Acad. Sci. Sect. B. Nat. Exact, Appl. Sci., 70, 400–408, https://doi.org/10.1515/prolas-2016-0059, 2016. Kniebusch, M., Meier, H. E. M., Neumann, T., and Börgel, F.: Temperature Variability of the Baltic Sea Since 1850 and \nAttribution \nto \nAtmospheric \nForcing \nVariables, \nJ. Geophys. Res. Ocean., \n124, \n4168–4187, https://doi.org/10.1029/2018JC013948, 2019. 595 \nKozlov, I., Dailidienė, I., Korosov, A., Klemas, V., and Mingėlaitė, T.: MODIS-based sea surface temperature of the Baltic \nSea Curonian Lagoon, J. Mar. Syst., 129, 157–165, https://doi.org/10.1016/j.jmarsys.2012.05.011, 2014. Kozlov, I. E., Krek, E. V., Kostianoy, A. G., and Dailidienė, I.: Remote Sensing of Ice Conditions in the Southeastern Baltic \nSea and in the Curonian Lagoon and Validation of SAR-Based Ice Thickness Products, Remote Sens., 12, 3754, https://doi.org/10.1029/2018JC013948, 2019. 595 \nKozlov, I., Dailidienė, I., Korosov, A., Klemas, V., and Mingėlaitė, T.: MODIS-based sea surface temperature of the Baltic \nSea Curonian Lagoon, J. Mar. Syst., 129, 157–165, https://doi.org/10.1016/j.jmarsys.2012.05.011, 2014. Kozlov, I. E., Krek, E. V., Kostianoy, A. G., and Dailidienė, I.: Remote Sensing of Ice Conditions in the Southeastern Baltic https://doi.org/10.3390/rs12223754, 2020. 600 \nLehmann, A., Myrberg, K., Post, P., Chubarenko, I., Dailidiene, I., Hinrichsen, H.-H., Hüssy, K., Liblik, T., Meier, H. E. M., \nLips, U., and Bukanova, T.: Salinity dynamics of the Baltic Sea, Earth Syst. Dyn., 13, 373–392, https://doi.org/10.5194/esd-\n13-373-2022, 2022. Lenderink, G., Buishand, A., and Van Deursen, W.: Estimates of future discharges of the river Rhine using two scenario Lenderink, G., Buishand, A., and Van Deursen, W.: Estimates of future discharges of the river Rhine using two scenario \nmethodologies: Direct versus delta approach, Hydrol. Earth Syst. Sci., 11, 1145–1159, https://doi.org/10.5194/hess-11-1145-\n605 \n2007, 2007. Lesutienė, J., Andrašūnas, V., Čerkasova, N., Gasiūnaitė, Z. R., Idzelytė, R., Ivanauskas, E., Kaziukonytė, K., Mėžinė, J., and \nVilkevičiūtė, J.: EcoServe Guidelines (version 2)., Zenodo, https://doi.org/10.5281/zenodo.6546247, 2022. methodologies: Direct versus delta approach, Hydrol. Earth Syst. Sci., 11, 1145–1159, https://doi.org/10.5194/hess-11-1145-\n605 \n2007, 2007. Idzelytė, R., Čerkasova, N., Mėžinė, J., Dabulevičienė, T., Razinkovas-Baziukas, A., Ertürk, A., and Umgiesser, G.: The \ncomputation results of coupled hydrological and hydrodynamic modelling application for the Nemunas River watershed – \nCuronian Lagoon – South-Eastern Baltic Sea continuum, Zenodo [data set], https://doi.org/10.5281/zenodo.7500744, 2023. Ivanauskas, E., Skersonas, A., Andrašūnas, V., Elyaagoubi, S., and Razinkovas-Baziukas, A.: Mapping and Assessing \n580 \nCommercial \nFisheries \nServices \nin \nthe \nLithuanian \nPart \nof \nthe \nCuronian \nLagoon, \nFishes, \n7, \n19, \nhttps://doi.org/10.3390/fishes7010019, 2022. Jakimavičius, D.: Changes of water balance elements of the Curonian Lagoon and their forecast due to anthropogenic and \nnatural factors, Kaunas University of Technology, 2012. Jakimavičius, D. and Kovalenkovienė, M.: Long-term water balance of the Curonian Lagoon in the context of anthropogenic \n585 \nfactors and climate change, Baltica, 23, 33–46, 2010. Jakimavičius, D. and Kovalenkovienė, M.: Long-term water balance of the Curonian Lagoon in the context of anthropogenic \n585 \nfactors and climate change, Baltica, 23, 33–46, 2010. Jakimavičius, D. and Kriaučiūnienė, J.: The Hydrological Changes of the Curonian Lagoon in the Context of Climate Change, \nVilnius Univ. Proc., 10, 36, https://doi.org/10.15388/Klimatokaita.2020.28, 2020. 25 Lesutienė, J., Andrašūnas, V., Čerkasova, N., Gasiūnaitė, Z. R., Idzelytė, R., Ivanauskas, E., Kaziukonytė, K., Mėžinė, J., and \nVilkevičiūtė, J.: EcoServe Guidelines (version 2)., Zenodo, https://doi.org/10.5281/zenodo.6546247, 2022. Lillebø, A. I.: Coastal Lagoons in Europe: Integrated Water Resource Strategies, Water Intell. Online, 14, 2007, 2007. Lesutienė, J., Andrašūnas, V., Čerkasova, N., Gasiūnaitė, Z. R., Idzelytė, R., Ivanauskas, E., Kaziukonytė, K., Mėžinė, J., and \nVilkevičiūtė, J.: EcoServe Guidelines (version 2)., Zenodo, https://doi.org/10.5281/zenodo.6546247, 2022. Lesutienė, J., Andrašūnas, V., Čerkasova, N., Gasiūnaitė, Z. R., Idzelytė, R., Ivanauskas, E., Kaziukonytė, K., Mėžinė, J., and \nVilkevičiūtė, J.: EcoServe Guidelines (version 2)., Zenodo, https://doi.org/10.5281/zenodo.6546247, 2022. Vilkevičiūtė, J.: EcoServe Guidelines (version 2)., Zenodo, https://doi.org/10.5281/zenodo.6546247, 2022. Lillebø, A. I.: Coastal Lagoons in Europe: Integrated Water Resource Strategies, Water Intell. Online, 14, Lillebø, A. I.: Coastal Lagoons in Europe: Integrated Water Resource Strategies, Water Intell. Online, 14, \nhttps://doi.org/10.2166/9781780406299, 2015. 610 https://doi.org/10.2166/9781780406299, 2015. 610 \nLuomaranta, A., Ruosteenoja, K., Jylhä, K., Gregow, H., Haapala, J., and Laaksonen, A.: Multimodel estimates of the changes \nin the Baltic Sea ice cover during the present century, Tellus A Dyn. Meteorol. Oceanogr., 66, 22617, \nhttps://doi.org/10.3402/tellusa.v66.22617, 2014. Meier, H. E. M.: Regional ocean climate simulations with a 3D ice-ocean model for the Baltic Sea. Part 1: model experiments p\ng\nLuomaranta, A., Ruosteenoja, K., Jylhä, K., Gregow, H., Haapala, J., and Laaksonen, A.: Multimodel estimates of the changes \nin the Baltic Sea ice cover during the present century, Tellus A Dyn. Meteorol. Oceanogr., 66, 22617, \nhttps://doi.org/10.3402/tellusa.v66.22617, 2014. in the Baltic Sea ice cover during the present century, Tellus A Dyn. Meteorol. Oceanogr., 66, 22617, \nhttps://doi.org/10.3402/tellusa.v66.22617, 2014. Meier, H. E. M.: Regional ocean climate simulations with a 3D ice-ocean model for the Baltic Sea. Part 1: model experiments Meier, H. E. M.: Regional ocean climate simulations with a 3D ice-ocean model for the Baltic Sea. Part 1: model experiments \nand results for temperature and salinity, Clim. Dyn., 19, 237–253, https://doi.org/10.1007/s00382-001-0224-6, 2002a. 615 and results for temperature and salinity, Clim. Dyn., 19, 237–253, https://doi.org/10.1007/s00382-001-0224-6, 2002a. 615 \nMeier, H. E. M.: Regional ocean climate simulations with a 3D ice–ocean model for the Baltic Sea. Part 2: results for sea ice, \nClim. Dyn., 19, 255–266, https://doi.org/https://doi.org/10.1007/s00382- 001-0225-5, 2002b. Meier, H. E. M., Kniebusch, M., Dieterich, C., Gröger, M., Zorita, E., Elmgren, R., Myrberg, K., Ahola, M. P., Bartosova, A., \nBonsdorff, E., Börgel, F., Capell, R., Carlén, I., Carlund, T., Carstensen, J., Christensen, O. B., Dierschke, V., Frauen, C., and results for temperature and salinity, Clim. Dyn., 19, 237 253, https://doi.org/10.1007/s00382 001 0224 6, 2002a. 615 \nMeier, H. E. M.: Regional ocean climate simulations with a 3D ice–ocean model for the Baltic Sea. Part 2: results for sea ice, \nClim. Dyn., 19, 255–266, https://doi.org/https://doi.org/10.1007/s00382- 001-0225-5, 2002b. Clim. Dyn., 19, 255 266, https://doi.org/https://doi.org/10.1007/s00382 001 0225 5, 2002b. Meier, H. E. M., Kniebusch, M., Dieterich, C., Gröger, M., Zorita, E., Elmgren, R., Myrberg, K., Ahola, M. P., Bartosova, A., \nBonsdorff, E., Börgel, F., Capell, R., Carlén, I., Carlund, T., Carstensen, J., Christensen, O. B., Dierschke, V., Frauen, C., Meier, H. E. M., Kniebusch, M., Dieterich, C., Gröger, M., Zorita, E., Elmgren, R., Myrberg, K., Ahola, M. P., Bartosova, A., \nBonsdorff, E., Börgel, F., Capell, R., Carlén, I., Carlund, T., Carstensen, J., Christensen, O. B., Dierschke, V., Frauen, C., \nFrederiksen, M., Gaget, E., Galatius, A., Haapala, J. J., Halkka, A., Hugelius, G., Hünicke, B., Jaagus, J., Jüssi, M., Käyhkö, \n620 Meier, H. E. M., Kniebusch, M., Dieterich, C., Gröger, M., Zorita, E., Elmgren, R., Myrberg, K., Ahola, M. P., Bartosova, A., \nBonsdorff, E., Börgel, F., Capell, R., Carlén, I., Carlund, T., Carstensen, J., Christensen, O. B., Dierschke, V., Frauen, C., \nFrederiksen, M., Gaget, E., Galatius, A., Haapala, J. J., Halkka, A., Hugelius, G., Hünicke, B., Jaagus, J., Jüssi, M., Käyhkö, \n620 \nJ., Kirchner, N., Kjellström, E., Kulinski, K., Lehmann, A., Lindström, G., May, W., Miller, P. A., Mohrholz, V., Müller-\nKarulis, B., Pavón-Jordán, D., Quante, M., Reckermann, M., Rutgersson, A., Savchuk, O. P., Stendel, M., Tuomi, L., Viitasalo, Frederiksen, M., Gaget, E., Galatius, A., Haapala, J. J., Halkka, A., Hugelius, G., Hünicke, B., Jaagus, J., Jüssi, M., Käyhkö, \n620 \nJ., Kirchner, N., Kjellström, E., Kulinski, K., Lehmann, A., Lindström, G., May, W., Miller, P. A., Mohrholz, V., Müller-\nKarulis, B., Pavón-Jordán, D., Quante, M., Reckermann, M., Rutgersson, A., Savchuk, O. P., Stendel, M., Tuomi, L., Viitasalo, 26 https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. M., Weisse, R., and Zhang, W.: Climate change in the Baltic Sea region: a summary, Earth Syst. Dyn., 13, 457–593, \nhttps://doi.org/10.5194/esd-13-457-2022, 2022a. Meier, H. E. M., Dieterich, C., Gröger, M., Dutheil, C., Börgel, F., Safonova, K., Christensen, O. B., and Kjellström, E.: \n625 \nOceanographic regional climate projections for the Baltic Sea until 2100, Earth Syst. Dyn., 13, 159–199, \nhttps://doi.org/10.5194/esd-13-159-2022, 2022b. Meier, H. E. M., Dieterich, C., Gröger, M., Dutheil, C., Börgel, F., Safonova, K., Christensen, O. B., and Kjellström, E.: \n625 \nOceanographic regional climate projections for the Baltic Sea until 2100, Earth Syst. Dyn., 13, 159–199, \nhttps://doi.org/10.5194/esd-13-159-2022, 2022b. Merkouriadi, I. and Leppäranta, M.: Long-term analysis of hydrography and sea-ice data in Tvärminne, Gulf of Finland, Baltic \nSea, Clim. Change, 124, 849–859, https://doi.org/10.1007/s10584-014-1130-3, 2014. Merkouriadi, I. and Leppäranta, M.: Long-term analysis of hydrography and sea-ice data in Tvärminne, Gulf of Finland, Baltic \nSea, Clim. Change, 124, 849–859, https://doi.org/10.1007/s10584-014-1130-3, 2014. Mėžinė, J., Ferrarin, C., Vaičiutė, D., Idzelytė, R., Zemlys, P., and Umgiesser, G.: Sediment transport mechanisms in a lagoon \n630 \nwith high river discharge and sediment loading, Water (Switzerland), 11, https://doi.org/10.3390/w11101970, 2019. Middelkoop, H., Daamen, K., Gellens, D., Grabs, W., Kwadijk, J. C. ., Lang, H., Parmet, B. W. A. ., Schädler, B., Schulla, J., \nand Wilke, K.: Impact of Climate Change on Hydrological Regimes and Water Resources Management in the Rhine Basin, \nClim. Change, 49, 105–128, https://doi.org/10.1023/A:1010784727448, 2001. Mėžinė, J., Ferrarin, C., Vaičiutė, D., Idzelytė, R., Zemlys, P., and Umgiesser, G.: Sediment transport mechanisms in a lagoon \n630 \nwith high river discharge and sediment loading, Water (Switzerland), 11, https://doi.org/10.3390/w11101970, 2019. ., Gellens, D., Grabs, W., Kwadijk, J. C. ., Lang, H., Parmet, B. W. A. ., Schädler, B., Schulla, J., Olenin, S. and Daunys, D.: Coastal Typology Based on Benthic Biotope and Community Data: The Lithuanian Case Study, \n635 \nCoastline Reports, 4, 65–83, 2004. Omstedt, A., Gustafsson, B., Rodhe, J., and Walin, G.: Use of Baltic Sea modelling to investigate the water cycle and the heat \nbalance in GCM and regional climate models, Clim. Res., 15, 95–108, https://doi.org/10.3354/cr015095, 2000. PAIC: Enhancement of SWAT model. Supplement of 3rd Interim report., SIA „Procesu analīzes un izpētes centrs”, Latvia., Olenin, S. and Daunys, D.: Coastal Typology Based on Benthic Biotope and Community Data: The Lithuanian Case Study, \n635 \nCoastline Reports, 4, 65–83, 2004. p\nOmstedt, A., Gustafsson, B., Rodhe, J., and Walin, G.: Use of Baltic Sea modelling to investigate the water cycle and the heat \nbalance in GCM and regional climate models, Clim. Res., 15, 95–108, https://doi.org/10.3354/cr015095, 2000. PAIC: Enhancement of SWAT model. Supplement of 3rd Interim report., SIA „Procesu analīzes un izpētes centrs”, Latvia., \n42 pp., 2015. 640 42 pp., 2015. 640 \nDe Pascalis, F., Pérez-Ruzafa, A., Gilabert, J., Marcos, C., and Umgiesser, G.: Climate change response of the Mar Menor \ncoastal lagoon (Spain) using a hydrodynamic finite element model, Estuar. Coast. Shelf Sci., 114, 118–129, \nhttps://doi.org/10.1016/j.ecss.2011.12.002, 2011. Pihlainen, S., Zandersen, M., Hyytiäinen, K., Andersen, H. E., Bartosova, A., Gustafsson, B., Jabloun, M., McCrackin, M., Pihlainen, S., Zandersen, M., Hyytiäinen, K., Andersen, H. E., Bartosova, A., Gustafsson, B., Jabloun, M., McCrackin, M., \nMeier, H. E. M., Olesen, J. E., Saraiva, S., Swaney, D., and Thodsen, H.: Impacts of changing society and climate on nutrient \n645 \nloading to the Baltic Sea, Sci. Total Environ., 731, 138935, https://doi.org/10.1016/j.scitotenv.2020.138935, 2020. Rodrigues, M., Oliveira, A., Queiroga, H., Brotas, V., and Fortunato, A. B.: Modelling the effects of climate change in estuarine \necosystems with coupled hydrodynamic and biogeochemical models, 271–288, https://doi.org/10.1016/B978-0-444-63536- Meier, H. E. M., Olesen, J. E., Saraiva, S., Swaney, D., and Thodsen, H.: Impacts of changing society and climate on nutrient \n645 \nloading to the Baltic Sea, Sci. Total Environ., 731, 138935, https://doi.org/10.1016/j.scitotenv.2020.138935, 2020. Rodrigues, M., Oliveira, A., Queiroga, H., Brotas, V., and Fortunato, A. B.: Modelling the effects of climate change in estuarine \necosystems with coupled hydrodynamic and biogeochemical models, 271–288, https://doi.org/10.1016/B978-0-444-63536-\n5.00012-0, 2015. , and Kayser, B.: A high resolution spherical grid topography of the Baltic Sea – 2nd edition, 2001 Seifert, T., Tauber, F., and Kayser, B.: A high resolution spherical grid topography of the Baltic Sea – 2nd edition, 2001. 650 \nSiitam, L., Sipelgas, L., Pärn, O., and Uiboupin, R.: Statistical characterization of the sea ice extent during different winter \nscenarios in the Gulf of Riga (Baltic Sea) using optical remote-sensing imagery, Int. J. Remote Sens., 38, 617–638, \nhttps://doi.org/10.1080/01431161.2016.1268734, 2017. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. von Storch, H., Omstedt, A., Pawlak, J., and Reckermann, M.: Introduction and Summary, 1–22, https://doi.org/10.1007/978-\n3-319-16006-1_1, 2015. Tapiador, F. J., Roca, R., Del Genio, A., Dewitte, B., Petersen, W., and Zhang, F.: Is Precipitation a Good Metric for Model \nPerformance?, Bull. Am. Meteorol. Soc., 100, 223–233, https://doi.org/10.1175/BAMS-D-17-0218.1, 2019. von Storch, H., Omstedt, A., Pawlak, J., and Reckermann, M.: Introduction and Summary, 1–22, https://doi.org/10.1007/978-\n3-319-16006-1_1, 2015. Tapiador, F. J., Roca, R., Del Genio, A., Dewitte, B., Petersen, W., and Zhang, F.: Is Precipitation a Good Metric for Model von Storch, H., Omstedt, A., Pawlak, J., and Reckermann, M.: Introduction and Summary, 1–22, https://doi.org/10.1007/978-\n3-319-16006-1_1, 2015. Tapiador, F. J., Roca, R., Del Genio, A., Dewitte, B., Petersen, W., and Zhang, F.: Is Precipitation a Good Metric for Model \nPerformance?, Bull. Am. Meteorol. Soc., 100, 223–233, https://doi.org/10.1175/BAMS-D-17-0218.1, 2019. Tedesco, L., Vichi, M., Haapala, J., and Stipa, T.: An enhanced sea-ice thermodynamic model applied to the Baltic Sea, Boreal \n0 \nEnviron. Res., 14, 68–80, 2009. 660 Tedesco, L., Vichi, M., Haapala, J., and Stipa, T.: An enhanced sea-ice thermodynamic model applied to the Baltic Sea, Boreal \n660 \nEnviron. Res., 14, 68–80, 2009. Umgiesser, G., Canu, D. M., Cucco, A., and Solidoro, C.: A finite element model for the Venice Lagoon. Development, set \nup, calibration and validation, J. Mar. Syst., 51, 123–145, https://doi.org/10.1016/j.jmarsys.2004.05.009, 2004. Umgiesser, G., Ferrarin, C., Cucco, A., De Pascalis, F., Bellafiore, D., Ghezzo, M., and Bajo, M.: Comparative hydrodynamics ,\n,\n,\nUmgiesser, G., Canu, D. M., Cucco, A., and Solidoro, C.: A finite element model for the Venice Lagoon. Development, set \nup, calibration and validation, J. Mar. Syst., 51, 123–145, https://doi.org/10.1016/j.jmarsys.2004.05.009, 2004. U\ni\nG F\ni\nC C\nA D P\nli\nF B ll fi\nD Gh\nM\nd B j\nM C\nti\nh d\nd\ni of 10 Mediterranean lagoons by means of numerical modeling, J. Geophys. Res. Ocean., 119, 2212–2226, \n665 \nhttps://doi.org/10.1002/2013JC009512, 2014. Umgiesser, G., Zemlys, P., Erturk, A., Razinkovas-Baziukas, A., Mezine, J., and Ferrarin, C.: Seasonal renewal time variability \nin the Curonian Lagoon caused by atmospheric and hydrographical forcing, Ocean Sci., 12, 391–402, \nhttps://doi.org/10.5194/os-12-391-2016, 2016. Valiuškevičius, G., Stonevičius, E., Stankūnavičius, G., and Brastovickytė-Stankevič, J.: Severe floods in Nemunas River \n670 \nDelta, Baltica, 31, 89–99, https://doi.org/10.5200/baltica.2018.31.09, 2018. Stonevičius, E., Rimkus, E., Štaras, A., Kažys, J., and Valiuškevičius, G.: Climate change impact on the Nemunas River Siitam, L., Sipelgas, L., Pärn, O., and Uiboupin, R.: Statistical characterization of the sea ice extent during different winter \nscenarios in the Gulf of Riga (Baltic Sea) using optical remote-sensing imagery, Int. J. Remote Sens., 38, 617–638, \nhttps://doi.org/10.1080/01431161.2016.1268734, 2017. Stonevičius, E., Rimkus, E., Štaras, A., Kažys, J., and Valiuškevičius, G.: Climate change impact on the Nemunas River \nbasinhydrology in the 21st century, Boreal Environ. Res., 22, 49–65, 2017. 655 basinhydrology in the 21st century, Boreal Environ. Res., 22, 49–65, 2017. 655 basinhydrology in the 21st century, Boreal Environ. Res., 22, 49–65, 2017. 655 27 Vohland, K., Rannow, S., and Stagl, J.: Climate Change Impact Modelling Cascade – Benefits and Limitations for \nConservation Management, 63–76, https://doi.org/10.1007/978-94-007-7960-0_5, 2014. Vousdoukas, M. I., Mentaschi, L., Hinkel, J., Ward, P. J., Mongelli, I., Ciscar, J.-C., and Feyen, L.: Economic motivation for Valiuškevičius, G., Stonevičius, E., Stankūnavičius, G., and Brastovickytė-Stankevič, J.: Severe floods in Nemunas River \n670 \nDelta, Baltica, 31, 89–99, https://doi.org/10.5200/baltica.2018.31.09, 2018. Valiuškevičius, G., Stonevičius, E., Stankūnavičius, G., and Brastovickytė-Stankevič, J.: Severe floods in Nemunas River \n670 \nDelta, Baltica, 31, 89–99, https://doi.org/10.5200/baltica.2018.31.09, 2018. Vohland, K., Rannow, S., and Stagl, J.: Climate Change Impact Modelling Cascade – Benefits and Limitations for Vohland, K., Rannow, S., and Stagl, J.: Climate Change Impact Modelling Cascade – Benefits and Limitations for \nConservation Management, 63–76, https://doi.org/10.1007/978-94-007-7960-0_5, 2014. Vousdoukas, M. I., Mentaschi, L., Hinkel, J., Ward, P. J., Mongelli, I., Ciscar, J.-C., and Feyen, L.: Economic motivation for \nraising coastal flood defenses in Europe, Nat. Commun., 11, 2119, https://doi.org/10.1038/s41467-020-15665-3, 2020. 675 Vousdoukas, M. I., Mentaschi, L., Hinkel, J., Ward, P. J., Mongelli, I., Ciscar, J.-C., and Feyen, L raising coastal flood defenses in Europe, Nat. Commun., 11, 2119, https://doi.org/10.1038/s41467-020-15665-3, 2020. 675 \nVybernaite-Lubiene, I., Zilius, M., Saltyte-Vaisiauske, L., and Bartoli, M.: Recent Trends (2012–2016) of N, Si, and P Export \nfrom the Nemunas River Watershed: Loads, Unbalanced Stoichiometry, and Threats for Downstream Aquatic Ecosystems, \nWater, 10, 1178, https://doi.org/10.3390/w10091178, 2018. Vybernaite-Lubiene, I., Zilius, M., Bartoli, M., Petkuviene, J., Zemlys, P., Magri, M., and Giordani, G.: Biogeochemical rnaite-Lubiene, I., Zilius, M., Bartoli, M., Petkuviene, J., Zemlys, P., Magri, M., and Giordani, Budgets of Nutrients and Metabolism in the Curonian Lagoon (South East Baltic Sea): Spatial and Temporal Variations, Water, \n680 \n14, 164, https://doi.org/10.3390/w14020164, 2022. Wang, S., Dieterich, C., Döscher, R., Höglund, A., Hordoir, R., Meier, H. E. M., Samuelsson, P., and Schimanke, S.: \nDevelopment and evaluation of a new regional coupled atmosphere–ocean model in the North Sea and Baltic Sea, Tellus A \nDyn. Meteorol. Oceanogr., 67, 24284, https://doi.org/10.3402/tellusa.v67.24284, 2015. Žaromskis, R.: Oceans, Seas, Estuaries, Debesija, Vilnius, 1996. 685 \nZemlys, P., Ferrarin, C., Umgiesser, G., Gulbinskas, S., and Bellafiore, D.: Investigation of saline water intrusions into the \nCuronian Lagoon (Lithuania) and two-layer flow in the Klaipeda Strait using finite element hydrodynamic model, Ocean Sci., \n9, 573–584, https://doi.org/10.5194/os-9-573-2013, 2013. 28 Zhu, S., Luo, Y., Ptak, M., Sojka, M., Ji, Q., Choiński, A., and Kuang, M.: A hybrid model for the forecasting of sea surface \nwater temperature using the information of air temperature: a case study of the Baltic Sea, All Earth, 34, 27–38, \n690 \nhttps://doi.org/10.1080/27669645.2022.2058689, 2022. \nŽilinskas, G., Jarmalavičius, D., Pupienis, D., Gulbinas, G., Korotkich, P., Palčiauskaitė, R., Pileckas, M., and Raščius, G.: \nCuronian Lagoon coastal management study. Tech. Rep., Nature Heritage Fund, 2012. Zhu, S., Luo, Y., Ptak, M., Sojka, M., Ji, Q., Choiński, A., and Kuang, M.: A hybrid model for the forecasting of sea surface \nt\nt\nt\ni\nth\ni f\nti\nf\ni\nt\nt\nt d\nf th\nB lti\nS\nAll E th 34\n27 38\n90 Žilinskas, G., Jarmalavičius, D., Pupienis, D., Gulbinas, G., Korotkich, P., Palčiauskaitė, R., Pileckas, M., and Raščius, G.: \nCuronian Lagoon coastal management study. Tech. Rep., Nature Heritage Fund, 2012. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. https://doi.org/10.5194/egusphere-2023-303\nPreprint. Discussion started: 24 February 2023\nc⃝Author(s) 2023. CC BY 4.0 License. 690 29"
https://openalex.org/W2969713210	https://hal.science/hal-02532761/document	English	null	Evaluation of global terrestrial evapotranspiration using state-of-the-art approaches in remote sensing, machine learning and land surface modeling	Hydrology and earth system sciences	2,020	cc-by	19,763	Evaluation of global terrestrial evapotranspiration using state-of-the-art approaches in remote sensing, machine learning and land surface modeling Shufen Pan, Naiqing Pan, Hanqin Tian, Pierre Friedlingstein, Stephen Sitch, Hao Shi, Vivek Arora, Vanessa Haverd, Atul Jain, Etsushi Kato, et al. Shufen Pan, Naiqing Pan, Hanqin Tian, Pierre Friedlingstein, Stephen Sitch, Hao Shi, Vivek Arora, Vanessa Haverd, Atul Jain, Etsushi Kato, et al. Shufen Pan, Naiqing Pan, Hanqin Tian, Pierre Friedlingstein, Stephen Sitch, Hao Shi, Vivek Arora, Vanessa Haverd, Atul Jain, Etsushi Kato, et al. To cite this version: Shufen Pan, Naiqing Pan, Hanqin Tian, Pierre Friedlingstein, Stephen Sitch, et al.. Evaluation of global terrestrial evapotranspiration using state-of-the-art approaches in remote sensing, machine learning and land surface modeling. Hydrology and Earth System Sciences, 2020, 24 (3), pp.1485-1509. 10.5194/hess-24-1485-2020. hal-02532761 Distributed under a Creative Commons Attribution 4.0 International License HAL Id: hal-02532761 https://hal.science/hal-02532761v1 Submitted on 24 Aug 2020 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 https://doi.org/10.5194/hess-24-1485-2020 © Author(s) 2020. This work is distributed under the Creative Commons Attribution 4.0 License. Correspondence: Shufen Pan (panshuf@auburn.edu) Correspondence: Shufen Pan (panshuf@auburn.edu) Received: 3 August 2019 – Discussion started: 21 August 2019 Revised: 22 January 2020 – Accepted: 27 January 2020 – Published: 31 March 2020 these three categories of approaches agreed well, with val- ues ranging from 589.6 mm yr−1 (6.56 × 104 km3 yr−1) to 617.1 mm yr−1 (6.87 × 104 km3 yr−1). For the period from 1982 to 2011, both the ensembles of remote-sensing-based physical models and machine-learning algorithms suggested increasing trends in global terrestrial ET (0.62 mm yr−2 with a signiﬁcance level of p < 0.05 and 0.38 mm yr−2 with a sig- niﬁcance level of p < 0.05, respectively). In contrast, the en- semble mean of the LSMs showed no statistically signiﬁcant change (0.23 mm yr−2, p > 0.05), although many of the in- dividual LSMs reproduced an increasing trend. Nevertheless, Abstract. Evapotranspiration (ET) is critical in linking global water, carbon and energy cycles. However, direct mea- surement of global terrestrial ET is not feasible. Here, we ﬁrst reviewed the basic theory and state-of-the-art approaches for estimating global terrestrial ET, including remote-sensing- based physical models, machine-learning algorithms and land surface models (LSMs). We then utilized 4 remote- sensing-based physical models, 2 machine-learning algo- rithms and 14 LSMs to analyze the spatial and temporal vari- ations in global terrestrial ET. The results showed that the ensemble means of annual global terrestrial ET estimated by Evaluation of global terrestrial evapotranspiration using state-of-the-art approaches in remote sensing, machine learning and land surface modeling 10, 07745 Jena, Germany 15Numerical Terradynamic Simulation Group College of Forestry and Conservation p g , p y, , , 14Max Planck Institute for Biogeochemistry, P.O. Box 600164, Hans-Knöll-Str. 10, 07745 Jena, Germany 15Numerical Terradynamic Simulation Group, College of Forestry and Conservation, University of Montana, Missoula, MT 59812, USA Evaluation of global terrestrial evapotranspiration using state-of-the-art approaches in remote sensing, machine learning and land surface modeling Shufen Pan1, Naiqing Pan1,2, Hanqin Tian1, Pierre Friedlingstein3, Stephen Sitch4, Hao Shi1, Vivek K. Arora5, Vanessa Haverd6, Atul K. Jain7, Etsushi Kato8, Sebastian Lienert9, Danica Lombardozzi10, Julia E. M. S. Nabel11, Catherine Ottlé12, Benjamin Poulter13, Sönke Zaehle14, and Steven W. Running15 Vanessa Haverd6, Atul K. Jain7, Etsushi Kato8, Sebastian Lienert9, Danica Lombardozzi10, Julia E. M. S. Nabel11, Catherine Ottlé12, Benjamin Poulter13, Sönke Zaehle14, and Steven W. Running15 1International Center for Climate and Global Change Research, School of Forestry and Wildlife Sciences, Auburn University, Auburn, AL 36832, USA 2State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China 3 y j g 3College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter EX4 4QF, UK 4College of Life and Environmental Sciences, University of Exeter, Exeter EX4 4RJ, UK y j g 3College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter EX4 4QF, UK 4College of Life and Environmental Sciences University of Exeter Exeter EX4 4RJ UK 3College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter EX4 4QF, UK 4 g y Canadian Centre for Climate Modelling and Analysis, Environment Canada, University of Victoria, i BC V8W 2Y2 C d 5Canadian Centre for Climate Modelling and Analysis, Environment Canada, University of Victoria, Victoria, BC V8W 2Y2, Canada 5Canadian Centre for Climate Modelling and Analysis, Environment Canada, University of Victoria, Victoria, BC V8W 2Y2, Canada 6CSIRO Oceans and Atmosphere, GPO Box 1700, Canberra, ACT 2601, Australia 7 6CSIRO Oceans and Atmosphere, GPO Box 1700, Canberra, ACT 2601, Australia 7D t t f At h i S i U i it f Illi i U b IL 61801 USA CSIRO Oceans and Atmosphere, GPO Box 1700, Canberra, ACT 2601, Australia 7Department of Atmospheric Sciences, University of Illinois, Urbana, IL 61801, USA 8Institute of Applied Energy (IAE), Minato-ku, Tokyo 105-0003, Japan nvironmental Physics, Physics Institute, University of Bern, Bern, Switzerland 9Climate and Environmental Physics, Physics Institute, University of Bern, Bern, Switzerland 10 Climate and Global Dynamics Laboratory, National Center for Atmospheric Research, Boulder, C Max Planck Institute for Meteorology, Bundesstr. 53, 20146 Hamburg, Germany LSCE IPSL CNRS O d M i i 91191 Gif Y F LSCE-IPSL-CNRS, Orme des Merisiers, 91191, Gif-sur-Yvette, France 13NASA Goddard Space Flight Center, Biospheric Sciences Laboratory, Greenbelt, MD 20771, USA 14Max Planck Institute for Biogeochemistry, P.O. Box 600164, Hans-Knöll-Str. 1 Introduction Terrestrial evapotranspiration (ET) is the sum of the wa- ter lost to the atmosphere from plant tissues via transpira- tion and that lost from the land surface elements, including soil, plants and open water bodies, through evaporation. Pro- cesses controlling ET play a central role in linking the en- ergy (latent heat), water (moisture ﬂux) and carbon cycles (photosynthesis–transpiration trade-off) in the Earth system. Over 60 % of precipitation on the land surface is returned to the atmosphere through ET (Oki and Kanae, 2006), and the accompanying latent heat (λET, λ is the latent heat of vaporization) accounts for more than half of the solar en- ergy received by the land surface (Trenberth et al., 2009). ET is also coupled with the carbon dioxide (CO2) exchange between the canopy and the atmosphere through vegetation photosynthesis. These linkages make ET an important vari- able in both short-term numerical weather forecasts and long- term climate predictions. Moreover, ET is a critical indicator for ecosystem functioning across a variety of spatial scales. Therefore, in order to enhance our predictive understanding of the Earth system and sustainability, it is essential to accu- rately assess land surface ET in a changing global environ- ment. Conventional techniques, such as lysimeter, eddy co- variance, large aperture scintillometer and the Bowen ra- tio method, are capable of providing ET measurements at point and local scales (Wang and Dickinson, 2012). How- ever, it is impossible to directly measure ET at the global scale because dense global coverage by such instruments is not feasible, and the representativeness of point-scale measurements to comprehensively portray the spatial het- erogeneity of global land surface is also doubtful (Mueller et al., 2011). To address this issue, numerous approaches have been proposed in recent years to estimate global ter- restrial ET and these approaches can be divided into three main categories: (1) remote-sensing-based physical models, (2) machine-learning algorithms and (3) land surface models (Miralles et al., 2011; Mueller et al., 2011; Wang and Dick- inson, 2012). Knowledge of the uncertainties in global ter- However, large uncertainty still exists in quantifying the magnitude of global terrestrial ET and its spatial and tem- poral patterns, despite extensive research (Allen et al., 1998; Liu et al., 2008; Miralles et al., 2016; Mueller et al., 2011; Tian et al., 2010). S. Pan et al.: Evaluation of global terrestrial ET 1486 all 20 models used in this study showed that anthropogenic Earth greening had a positive role in increasing terrestrial ET. The concurrent small interannual variability, i.e., rela- tive stability, found in all estimates of global terrestrial ET, suggests that a potential planetary boundary exists in regu- lating global terrestrial ET, with the value of this boundary being around 600 mm yr−1. Uncertainties among approaches were identiﬁed in speciﬁc regions, particularly in the Ama- zon Basin and arid/semiarid regions. Improvements in pa- rameterizing water stress and canopy dynamics, the utiliza- tion of new available satellite retrievals and deep-learning methods, and model–data fusion will advance our predictive understanding of global terrestrial ET. this is due to the fact that it is modulated not only by sur- face meteorological conditions and soil moisture but also by the physiology and structures of plants. Changes in non- climatic factors such as elevated atmospheric CO2, nitrogen deposition and land covers also serve as inﬂuential drivers of Tv (Gedney et al., 2006; Mao et al., 2015; S. Pan et al., 2018; Piao et al., 2010). As such, the global ratio of transpi- ration to ET (Tv/ET) has long been a matter of debate, with the most recent observation-based estimate being 0.64±0.13 constrained by the global water-isotope budget (Good et al., 2015). Most Earth system models are thought to largely un- derestimate Tv/ET (Lian et al., 2018). Global warming is expected to accelerate the hydrolog- ical cycle (Pan et al., 2015). For the period from 1982 to the late 1990s, ET was reported to have increased by about 7 mm (∼1.2 %) per decade driven by an increase in radia- tive forcing and, consequently, global and regional temper- atures (Douville et al., 2013; Jung et al., 2010; Wang et al., 2010). The contemporary near-surface speciﬁc humidity also increased over both land and ocean (Dai, 2006; Simmons et al., 2010; Willett et al., 2007). More recent studies conﬁrmed that global ET has showed an overall increase since the 1980s (Mao et al., 2015; Yao et al., 2016; Zeng et al., 2018a, 2012, 2016; Zhang et al., 2015; Y. Zhang et al., 2016). However, the magnitude and spatial distribution of such a trend are far from determined. S. Pan et al.: Evaluation of global terrestrial ET Over the past 50 years, pan evaporation has decreased worldwide (Fu et al., 2009; Peterson et al., 1995; Roderick and Farquhar, 2002), implying an increase in ac- tual ET given the pan evaporation paradox. Moreover, the in- crease in global terrestrial ET was found to cease or even be reversed from 1998 to 2008, primarily due to the decreased soil moisture supply in the Southern Hemisphere (Jung et al., 2010). To reconcile the disparity, Douville et al. (2013) argued that the peak ET in 1998 should not be taken as a tip- ping point because ET was estimated to increase in the multi- decadal evolution. More efforts are needed to understand the spatial and temporal variations of global terrestrial ET and the underlying mechanisms that control its magnitude and variability. 2.1.1 Remote-sensing-based physical models Satellite remote sensing has been widely recognized as a promising tool for estimating global ET, because it is capable of providing spatially and temporally continuous measure- ments of critical biophysical parameters affecting ET, includ- ing vegetation states, albedo, the fraction of absorbed photo- synthetically active radiation, land surface temperature and plant functional types (Li et al., 2009). Since the 1980s, a large number of methods have been developed using a vari- ety of satellite observations (K. Zhang et al., 2016). How- ever, some of these methods, such as surface energy bal- ance (SEB) models and surface temperature–vegetation in- dex (Ts–VI), are usually applied at local and regional scales. At the global scale, the vast majority of the existing remote- sensing-based physical models can be categorized into two groups: those based on the Penman–Monteith (PM) equation and those based on the Priestley–Taylor (PT) equation. S. Pan et al.: Evaluation of global terrestrial ET (2007), which potentially makes it attractive for use with remote sensing. However, PM-based models have one intrinsic weakness – temporal upscaling – which is required when translating instantaneous ET estimation into a longer timescale value (Li et al., 2009). This could easily be done at the daily scale under clear-sky conditions but faces chal- lenge at weekly to monthly timescales due to lack of cloud coverage information. In this study, we integrate state-of-the-art estimates of global terrestrial ET, including data-driven and process- based estimates, to assess its spatial pattern, interannual vari- ability, environmental drivers, long-term trend and response to vegetation greening. Our goal is not to compare the var- ious models and choose the best one but to identify the un- certainty sources in each type of estimate and provide sug- gestions for future model development. In the following sec- tions, we ﬁrst give a brief introduction to all of the method- ological approaches and ET datasets used in this study. We then quantify the spatiotemporal variations in global terres- trial ET during the period from 1982 to 2011 by analyzing the results from the current state-of-the-art models. Finally, we discuss some suggested solutions for reducing the identi- ﬁed uncertainties. Remote sensing models based on the Priestley–Taylor equation The Priestley–Taylor (PT) equation is a simpliﬁcation of the PM equation that does not parameterize aerodynamic and surface conductance (Priestley and Taylor, 1972); it can be expressed as follows: 1 Introduction Previous estimates of global land mean annual ET range from 417 to 650 mm yr−1 for the whole or part of the 1982–2011 period (Mu et al., 2007; Mueller et al., 2011; Vinukollu et al., 2011a; Zhang et al., 2010). This large discrepancy among independent studies may be attributed to a lack of sufﬁcient measurements, uncertainty in forcing data, inconsistent spatial and temporal resolutions, ill-calibrated model parameters, and deﬁciencies in model structures. Of the four components of ET (transpiration, soil evaporation, canopy interception and open water evapora- tion), transpiration (Tv) contributes the largest uncertainty; www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1487 S. Pan et al.: Evaluation of global terrestrial ET ﬂux to estimate ET from an open water surface. For vegetated surfaces, canopy resistance was introduced into the Penman equation by Monteith (Monteith, 1965), and the PM equation is formulated as follows: restrial ET estimates from different approaches is a prereq- uisite for future projection and many other applications. In recent years, several studies have compared multiple terres- trial ET estimates (Khan et al., 2018; Mueller et al., 2013; Wartenburger et al., 2018; Y. Zhang et al., 2016); however, most of these studies analyzed multiple datasets of the same approach or focused on investigating similarities and differ- ences among different approaches. Few studies have been conducted to identify uncertainties in multiple estimates of different approaches. λET = 1(Rn −G) + ρaCpVPD/ra 1 + γ (1 + rs/ra) , (1) (1) where 1, Rn, G, ρa, Cp, γ , rs, ra and VPD are the slope of the curve relating saturated water vapor pressure to air tempera- ture, net radiation, soil heat ﬂux, air density, the speciﬁc heat of air, the psychrometric constant, surface resistance, aero- dynamic resistance and the vapor pressure deﬁcit, respec- tively. The canopy resistance term in the PM equation ex- erts a strong control on transpiration. For example, based on the algorithm proposed by Cleugh et al. (2007), the MODIS (Moderate Resolution Imaging Spectroradiometer) ET algo- rithm improved the model performance via the inclusion of environmental stress into the canopy conductance calcula- tion and explicitly accounted for soil evaporation (Mu et al., 2007). Further, Mu et al. (2011) improved the MODIS ET algorithm by considering nighttime ET, adding the soil heat ﬂux calculation, separating the dry canopy surface from the wet, and dividing the soil surface into saturated wet surface and moist surface. Similarly, Zhang et al. (2010) developed a Jarvis–Stewart-type canopy conductance model based on the normalized difference vegetation index (NDVI) to take advantage of the long-term Advanced Very High Resolu- tion Radiometer (AVHRR) dataset. More recently, this model was improved by adding a CO2 constraint function into the canopy conductance estimate (Zhang et al., 2015). Another important revision for the PM approach is proposed by Le- uning et al. (2008). The Penman–Monteith–Leuning method adopts a simple biophysical model for canopy conductance, which can account for the inﬂuences of radiation and the at- mospheric humidity deﬁcit. Additionally, it introduces a sim- pler soil evaporation algorithm than that proposed by Mu et al. Remote sensing models based on the Penman–Monteith equation The Penman equation, derived from the Monin–Obukhov similarity theory and surface energy balance, uses surface net radiation, temperature, humidity, wind speed and ground heat λET = fstress × α × 1 1 + γ × (Rn −G), (2) (2) 2.1.3 Process-based land surface models (LSMs) Although satellite-derived ET products have provided quan- titative investigations of historical terrestrial ET dynamics, they can only cover a limited temporal record of about 4 decades. To obtain terrestrial ET before the 1980s and pre- dict future ET dynamics, LSMs are needed, as they are able to represent a large number of interactions and feedbacks between physical, biological and biogeochemical processes in a prognostic way (Jimenez et al., 2011). ET simulation in LSMs is regulated by multiple biophysical and physio- logical properties or processes, including but not limited to stomatal conductance, leaf area, root water uptake, soil wa- ter, runoff and (sometimes) nutrient uptake (Famiglietti and Wood, 1991; Huang et al., 2016; Lawrence et al., 2007). Al- though almost all current LSMs have these components, dif- ferent parameterization schemes result in substantial differ- ences in ET estimation (Wartenburger et al., 2018). There- fore, in recent years, the multi-model ensemble approach has become popular in quantifying the magnitude, spatiotempo- ral pattern and uncertainty of global terrestrial ET (Mueller et al., 2011; Wartenburger et al., 2018). Yao et al. (2017) showed that a simple model averaging method or a Bayesian model averaging method is superior to each individual model in predicting terrestrial ET. S. Pan et al.: Evaluation of global terrestrial ET 1488 where fstress is a stress factor and is usually computed as a function of environmental conditions. α is the PT param- eter with a value of between 1.2 and 1.3 under water un- stressed conditions and can be estimated using remote sens- ing. Although the original PT equation works well for es- timating potential ET across most surfaces, the Priestley– Taylor coefﬁcient, α, usually needs adjustment to convert potential ET to actual ET (K. Zhang et al., 2016). Thus, Fisher et al. (2008) developed a modiﬁed PT model that keeps α constant but scales down potential ET using eco- physiological constraints and soil evaporation partitioning. The accuracy of their model has been validated against eddy- covariance measurements conducted in a wide range of cli- mates and involving many plant functional types (Fisher et al., 2009; Vinukollu et al., 2011b). Following this idea, Yao et al. (2013) further developed a modiﬁed Priestley–Taylor algorithm that constrains soil evaporation using the index of the soil water deﬁcit derived from the apparent thermal inertia. Miralles et al. (2011) also proposed a novel PT- type model: the Global Land Evaporation Amsterdam Model (GLEAM). GLEAM combines a soil water module, a canopy interception model and a stress module within the PT equa- tion. The key distinguishing features of this model are the use of microwave-derived soil moisture, land surface tem- perature and vegetation density, and the detailed estimation of rainfall interception loss. In this way, GLEAM minimizes the dependence on static variables, avoids the need for pa- rameter tuning and enables the quality of the evaporation es- timates to rely on the accuracy of the satellite inputs (Mi- ralles et al., 2011). Compared with the PM approach, the PT-based approaches avoid the computational complexities of aerodynamic resistance and the accompanying error prop- agation. However, the many simpliﬁcations and semiempir- ical parameterization of physical processes in the PT-based approaches may lower its accuracy. algorithms, such as those based on artiﬁcial neural networks, random forest, and support vector machine algorithms, have been applied in various ecosystems (Antonopoulos et al., 2016; Chen et al., 2014; Feng et al., 2017; Shrestha and Shukla, 2015) and have proved to be more accurate in es- timating ET than simple regression models (Antonopoulos et al., 2016; Chen et al., 2014; Kisi et al., 2015; Shrestha and Shukla, 2015; Tabari et al., 2013). 2.1.2 Vegetation-index-based empirical algorithms and machine-learning methods The principle of empirical ET algorithms is to link observed ET to its controlling environmental factors through various statistical regressions or machine-learning algorithms of dif- ferent complexities. The earliest empirical regression method was proposed by Jackson et al. (1977). At present, the ma- jority of regression models are based on vegetation indices (Glenn et al., 2010), such as the NDVI and the enhanced veg- etation index (EVI), because of their simplicity, resilience in the presence of data gaps, utility under a wide range of con- ditions and connection with vegetation transpiration capacity (Maselli et al., 2014; Nagler et al., 2005; Yuan et al., 2010). As an alternative to statistical regression methods, machine- learning algorithms have been gaining increased attention for ET estimation due to their ability to capture the complex non- linear relationships between ET and its controlling factors (Dou and Yang, 2018). Many conventional machine-learning S. Pan et al.: Evaluation of global terrestrial ET In upscaling FLUXNET ET to the global scale, Jung et al. (2010) used the model tree ensemble method to integrate eddy-covariance measure- ments of ET with satellite remote sensing and surface mete- orological data. In a recent study (Bodesheim et al., 2018), the random forest approach was used to derive global ET at a 30 min timescale. www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET 1489 Table 1. Descriptions of the models used in this study, including their drivers, adopted algorithms, key equations, advantages and limitations, and references. Name Input Algorithm Spatial Temporal Key equations Advantages and limitations References resolution resolution MTE Climate: precipitation, tem- perature, sunshine hour, rel- ative humidity and wet days Vegetation: fAPAR TRIAL + ERROR 0.5◦× 0.5◦ Monthly No speciﬁc equation Limitations: insufﬁcient ﬂux observations in tropical regions; no CO2 effect Jung et al. (2011) RF Climate: land surface temperature, radiation, potential radiation, index of water availability and relative humidity Vegetation: fAPAR, LAI and enhanced vegetation index (EVI) Randomized decision tree 0.5◦× 0.5◦ 30 min No speciﬁc equation Limitations: the same as for MTE Bodesheim et al. (2018) P-LSH Climate: radiation, air tem- perature, vapor pressure, wind speed and CO2 Vegetation: AVHRR NDVI Modiﬁed Penman– Monteith 0.083◦× 0.083◦ Monthly Ev = 1Rn+ρCpVPDga λv 1+ϒ 1+ ga gs Es = RH VPD k 1Rn+ρCpVPDga λv 1+ϒ 1+ ga gs Advantages: more robust physical basis; considers the effects of CO2 Limitations: high meteorological forcing requirements; canopy conductance is based on proxies Zhang et al. (2015) GLEAM Climate: precipitation, net radiation, surface soil moisture, land surface temperature, air tempera- ture and snow depth Vegetation: vegetation opti- cal depth Modiﬁed Priestley– Taylor 0.25◦× 0.25◦ Daily Es = fsSsαs 1 λvρw(1+γ ) Rsn −Gs Esc = fscSscαsc 1 λvρw(1+γ ) Rsc n −Gsc Etc = ftcStcαtc 1 λvρw(1+γ ) Rtcn −Gtc −βEi Advantages: simple; low requirement for meteorological data; well-suited for remote sensing observable variables; soil moisture is considered Limitations: many simpliﬁcations of physi- cal processes; neither VPD nor surface and aerodynamic resistances are explicitly accounted for; strong dependency on net radiation Miralles et al. 2.2 Description of ET models used in this study In this study, we evaluate 20 ET products that are based on remote-sensing-based physical models, machine-learning al- gorithms and LSMs to investigate the magnitudes and spatial patterns of global terrestrial ET over recent decades. Table 1 lists the input data, the ET algorithms adopted, the advan- tages and limitations, and the references for each product. We use a simple model averaging method when calculating the mean value of multiple models. Four physically based remote sensing datasets, includ- ing the Process-based Land Surface Evapotranspiration/Heat Fluxes algorithm (P-LSH), the Global Land Evaporation Amsterdam Model (GLEAM), the Moderate Resolution www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET TRENDY LSMs Advantages: land surface models are process-oriented and physically based. Given their structure almost all models are capable of allowing factorial analysis, where one forcing can be applied at a time. Most models also consider the physiological effect of CO2 on stomatal closure. Limitations: most models typically do not allow integration/assimilation of observation-based vegetation characteristics. Model parameterizations remain uncertain and the same process is modeled in different ways across models. Model parameters may or may not be physically based and, therefore, measurable in the ﬁeld. Models participating in the TRENDYv6 comparison were forced by precipitation, air temperature, speciﬁc humidity, shortwave radiation, longwave radiation and wind speed based on the CRU-NCEPv8 data, as explained in Le Quéré et al. (2018). It is very difﬁcult to list all key equations for all land surface models. Thus, we only list the stomatal conductance equation for each model in this table. Name Algorithm Spatial Temporal Key equations References resolution resolution CABLE Penman–Monteith 0.5◦× 0.5◦ Monthly gs = g0 + g1fwA ca−cp 1 + VPD VPD0 −1 Haverd et al. (2018) CLASS- CTEM Modiﬁed Penman–Monteith 2.8125◦× 2.8125◦ Monthly gc = m Anp (cs−0) 1 (1+VPD/VPD0) + bLAI Melton and Arora (2016) CLM45 Modiﬁed Penman–Monteith 1.875◦× 2.5◦ Monthly gs = g0 + g1A ca RH Oleson et al. (2010) DLEM Penman–Monteith 0.5◦× 0.5◦ Monthly gs = max(gsmaxrcorrbf (ppdf)f (Tmin)f (VPD)f (CO2),gsmin) Pan et al. (2015) ISAM Modiﬁed Penman–Monteith 0.5◦× 0.5◦ Monthly gs = m A Cs/Patm × es ei + btβt Barman et al. (2014) JSBACH Penman–Monteith 1.9◦× 1.9◦ Monthly gs = βw 1.6An,pot ca−ci,pot Knauer et al. (2015) JULES Penman–Monteith 2.5◦× 3.75◦ Monthly Bare soil conductance: gsoil = 1 100 θ1 θc 2 Stomatal conductance is calculated by solving the following two equations: Al = gs (Cs −Ci)/1.6; Ci−0∗ Cc−0∗= f0 1 −1 qc Li et al. (2016) LPJ- GUESS Equations proposed by Monteith (1995) 0.5◦× 0.5◦ Monthly gs = gsmin + 1.6Adt ca(1−λc) Smith (2001) LPJ-wsl Priestley–Taylor 0.5◦× 0.5◦ Monthly gs = gsmin + 1.6Adt ca(1−λc) Sitch et al. (2003) LPX-Bern Modiﬁed equation of Monteith (1995) 1◦× 1◦ Monthly gs = gsmin + 1.6Adt ca(1−λc) Keller et al. (2017) O-CN Modiﬁed Penman–Monteith 1◦× 1◦ Monthly gs = gsmin + 1.6Adt ca(1−λc) Zaehle and Friend (2010) ORCHIDEE Modiﬁed Penman–Monteith 0.5◦× 0.5◦ Monthly gs = g0 + A+Rd ca−cp fvpd gsoil = exp(8.206 −4.255W/Wsat) d’Orgeval et al. S. Pan et al.: Evaluation of global terrestrial ET (2008) ORCHIDEE- MICT Modiﬁed Penman–Monteith 0.5◦× 0.5◦ Monthly gs = g0 + A+Rd ca−cp fvpd Guimberteau et al. S. Pan et al.: Evaluation of global terrestrial ET (2018) VISIT Penman–Monteith 0.5◦× 0.5◦ Monthly gs = g0 + g1fwA ca−cp 1 + VPD VPD0 −1 Ito (2010) Notes: A: net assimilation rate; Adt: total daytime net photosynthesis; An,pot: unstressed net assimilation rate; b: soil moisture factor; bt: stomatal conductance intercept; ca: atmospheric CO2 concentration; cc: critical CO2 concentration; ci: internal leaf concentration of CO2; ci,pot: internal leaf concentration of CO2 for unstressed conditions; cs: leaf surface CO2 concentration; cp: CO2 compensation point; es: vapor pressure at leaf surface; ei: saturation vapor pressure inside the leaf; Es: soil evaporation; Ec: canopy evapotranspiration; Edry: dry canopy evapotranspiration; Ewet: wet canopy evapotranspiration; Ev: canopy transpiration; Ei: canopy interception; Etc: transpiration from tall canopy; Esc: transpiration from short canopy; f : fraction of P to equilibrium soil evaporation; fs: soil fraction; fsc: short canopy fraction; ftc: tall canopy fraction; fvpd: factor of the effect of leaf-to-air vapor pressure difference; fw: a function describing the soil water stress on stomatal conductance; fwet: relative surface wetness parameter; f0: the maximum ratio of internal to external CO2; f (ppdf):limiting factor of photosynthetic photo ﬂux density; f (Tmin): limiting factor of daily minimum temperature; f (VPD): limiting factor of vapor pressure deﬁcit; f (CO2): limiting factor of carbon dioxide; G: ground energy ﬂux; ga: aerodynamic conductance; gm: empirical parameter; gs: stomatal conductance; gsmax: maximum stomatal conductance; gsmin: minimum stomatal conductance; gsoil: bare soil conductance; g0: residual stomatal conductance when the net assimilation rate is 0 ; g1: sensitivity of stomatal conductance to assimilation, ambient CO2 concentration and environmental controls; I: tall canopy interception loss; m: stomatal conductance slope; Patm: atmospheric pressure; PEs: potential soil evaporation; PEcanopy: potential canopy evaporation; qa: speciﬁc air humidity; qc: critical humidity deﬁcit; qs: speciﬁc humidity of saturated air; ra: aerodynamic resistance; rs: stomatal resistance; Rn: net radiation; Rd: day respiration; RH: relative humidity; Ts: actual surface temperature; VPD: vapor pressure deﬁcit; VPD0: the sensitivity of stomatal conductance to VPD; W: top soil moisture; Wcanopy: canopy water; Wsat: soil porosity; α: Priestley–Taylor coefﬁcient; αm: empirical parameter; β: a constant accounting for the times when vegetation is wet; βt: soil water availability factor between 0 and 1; βw: a scaling factor to account for water stress; βs: moisture availability function; ρ: air density; γ : psychrometric constant; λv: latent heat of vaporization; λc: ratio of intercellular to ambient partial pressure of CO2; rcorr: correction factor of temperature and air pressure on conductance; 0∗: CO2 compensation point when leaf day respiration is zero; θ1: parameter of moisture concentration in the top soil layer; θc: parameter of moisture concentration in the spatially varying critical soil moisture; 1: slope of the vapor pressure curve. S. Pan et al.: Evaluation of global terrestrial ET Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/20 S. Pan et al.: Evaluation of global terrestrial ET (2011) MODIS Climate: air temperature, shortwave radiation, wind speed, relative humidity and air pressure Vegetation: LAI and fAPAR, albedo Penman– Monteith– Leuning 0.05◦× 0.05◦ Monthly Ei = fwetfc 1(Rn−G)+ρcp VPD rwc a λvρw 1+γ rwc s rwc a Ev = (1 −fwet)fc 1(Rn−G)+ρcp VPD rta λvρw 1+ϒ rts rta Es = h fwet + (1−fwet)hVPD β i sAsoil+ ρcp(1−fc)VPD ras λvρw S+γ rtot ras Advantages: more robust physical basis; Limitations: requires many variables that are difﬁcult to observe or that are not ob- servable with satellites; canopy conduc- tance is based on proxies; does not consider soil moisture but uses atmospheric humidity as a surrogate; does not consider the effects of CO2 Mu et al. (2011) PML- CSIRO Climate: precipitation, air temperature, vapor pres- sure, shortwave radiation, longwave radiation and wind speed Vegetation: AVHRR LAI, emissivity and albedo Penman– Monteith– Leuning 0.5◦× 0.5◦ Monthly Ev = 1Rn+ρCpVPDga λv 1+ϒ 1+ ga gs Es = f 1As 1+γ Ei: an adapted version of the Gash rainfall intercep- tion model (Van Dijk and Bruijnzeel, 2001) Advantages: more robust physical basis (compared with Priestley–Taylor equa- tion); biophysically based estimation of surface conductance Limitations: high meteorological forcing requirements; canopy conductance is based on proxies; does not consider the effects of CO2 Y. Zhang et al. (2016) www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1490 S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET (2007) improved the model performance through the in- clusion of environmental stress into the canopy conductance calculation and by explicitly accounting for soil evaporation by combining the complementary relationship hypothesis with the PM equation. The MODIS ET product (MOD16A3) used in this study was further improved by considering night- time ET, simplifying the vegetation cover fraction calcula- tion, adding the soil heat ﬂux item, dividing the saturated wet and moist soil, separating the dry and wet canopy, as well as modifying algorithms of aerodynamic resistance, stom- atal conductance and boundary layer resistance (Mu et al., 2011). PML-CSIRO adopts the Penman–Monteith–Leuning algorithm, which calculates surface conductance and canopy conductance using a biophysical model instead of classic em- pirical models. The maximum stomatal conductance is es- timated using the trial-and-error method (Y. Zhang et al., 2016). Furthermore, for each grid covered by natural veg- etation, the PML-CSIRO model constrains ET at the annual scale using the Budyko hydrometeorological model proposed by Fu (1981). The GLEAM ET calculation is based on the PT equation, which requires fewer model inputs than the PM equation, and the majority of these inputs can be di- rectly achieved from satellite observations. Its rationale is to make the most of information about evaporation contained in the satellite-based environmental and climatic observations (Martens et al., 2017; Miralles et al., 2011). Key variables including air temperature, land surface temperature, precipi- tation, soil moisture, vegetation optical depth and snow water equivalent are satellite-observed values. Moreover, the exten- sive usage of microwave remote sensing products in GLEAM ensures the accurate estimation of ET under diverse weather conditions. Here, we use GLEAM v3.2, which has overall better quality than previous versions (Martens et al., 2017). The ﬁrst used machine learning model MTE is based The 14 LSM-derived ET products were from the Trends and Drivers of the Regional Scale Sources and Sinks of Carbon Dioxide (TRENDY) project (including CA- BLE, CLASS-CTEM, CLM45, DLEM, ISAM, JSBACH, JULES, LPJ-GUESS, LPJ-wsl, LPX-Bern, O-CN, OR- CHIDEE, ORCHIDEE-MICT and VISIT). Daily gridded meteorological reanalyzes from the CRU-NCEPv8 dataset (temperature, precipitation, long- and shortwave incoming radiation, wind speed, humidity and air pressure) were used to drive the LSMs. The TRENDY simulations were per- formed in year 2017 and contributed to the global carbon budget reported in Le Quéré et al. (2018). S. Pan et al.: Evaluation of global terrestrial ET 2009). The TRIAL grows model trees from the root node and splits at each node with the criterion of minimizing the sum of squared errors of multiple regressions in both subdomains. ERROR is used to select the model trees that are indepen- dent of one another and have the best performance based on the Schwarz criterion. The canopy fraction of absorbed pho- tosynthetic active radiation (fAPAR), temperature, precipita- tion, relative humidity, sunshine hours and potential radia- tion are used as explanatory variables to train MTE (Jung et al., 2011). The second machine-learning model is the random forest (RF) algorithm, whose rationale is generating a set of independent regression trees by randomly selecting training samples automatically (Breiman, 2001). Each regression tree is constructed using samples selected by a bootstrap sam- pling method. After ﬁxing the individual tree in entity, the ﬁnal result is determined by simple averaging. One merit of the RF algorithm is its capability to handle complicated non- linear problems and high dimensional data (Xu et al., 2018). For the RF product used in this study, multiple explanatory variables including the enhanced vegetation index, fAPAR, the leaf area index, daytime and nighttime land surface tem- perature, incoming radiation, top-of-atmosphere potential ra- diation, the index of water availability and relative humidity were used to train regression trees (Bodesheim et al., 2018). Imaging Spectroradiometer (MODIS) and the PML-CSIRO (Penman–Monteith–Leuning), and two machine-learning datasets, which are based on random forest (RF) and model tree ensemble (MTE), are used in our study. Both machine- learning and physically based remote sensing datasets (six datasets in total) were considered as benchmark products. The ensemble mean of the benchmark products was calcu- lated as the mean value of all machine-learning and physi- cally based satellite estimates as we treated each benchmark dataset equally. Three of the four remote-sensing-based physical models quantify ET using PM approaches. P-LSH adopts a modiﬁed PM approach coupled with biome-speciﬁc canopy conduc- tance determined from NDVI (Zhang et al., 2010). The mod- iﬁed P-LSH model used in this study also accounts for the in- ﬂuences of atmospheric CO2 concentrations and wind speed on canopy stomatal conductance and aerodynamic conduc- tance (Zhang et al., 2015). The MODIS ET model is based on the algorithm proposed by Cleugh et al. (2007). Mu et al. S. Pan et al.: Evaluation of global terrestrial ET We used the re- sults from the S3 simulation from TRENDYv6 (with chang- ing CO2, climate and land use) over the period from 1982 to 2011, which is a time period consistent with other prod- ucts derived from remote-sensing-based physical models and machine-learning algorithms. Table 1. Continued. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1491 2.4 Statistical analysis In performing multiple linear regression, we used the GIMMS LAI for both remote-sensing-based physical mod- els and machine-learning methods as well as the individ- ual TRENDYv6 LAI for each TRENDY model. The grid- ded data of temperature, precipitation and radiation are from CRU-NCEPv8. The signiﬁcance of ET trends was analyzed using the Mann- Kendall (MK) test (Kendall, 1955; Mann, 1945). It is a rank- based nonparametric method that has been widely applied for detecting trends in hydro-climatic time series (Sayemuz- zaman and Jha, 2014; Yue et al., 2002). The Theil–Sen es- timator was applied to estimate the magnitude of the slope. The advantage of this method over an ordinary least squares estimator is that it limits the inﬂuence of the outliers on the slope (Sen, 1968). 3.1 The ET magnitude estimated by multiple models Terrestrial ET interannual variability (IAV) is mainly con- trolled by variations in temperature, precipitation and short- wave solar radiation (Zeng et al., 2018b; Zhang et al., 2015). In this study, we performed partial correlation analyses be- tween ET and these three climatic variables at an annual scale for each grid cell to explore climatic controls on ET IAV. Variability caused by climatic variables was assessed through the square of partial correlation coefﬁcients between ET and temperature, precipitation and radiation. We chose partial correlation analysis because it can quantify the link- age between ET and a single environmental driving factor while controlling the effects of the other remaining environ- mental factors. Partial correlation analysis is a widely ap- plied statistical tool to isolate the relationship between two variables from the confounding effects of many correlated variables (Anav et al., 2015; Jung et al., 2017; Peng et al., 2013). All variables were ﬁrst detrended in the statistical cor- relation analysis, as we focus on the interannual relationship. The study period is from 1982 to 2011 for all models except MODIS and random forest, which have a temporal coverage that is limited to 2001–2011 because of data availability. The multiyear ensemble mean of annual global terrestrial ET from 2001 to 2011 derived by machine-learning methods, remote-sensing-based physical models and the TRENDY models agreed well, with values ranging from 589.6 to 617.1 mm yr−1. However, substantial differences existed among individual models (Fig. 1). LPJ-wsl (455.3 mm yr−1) and LPX-Bern (453.7 mm yr−1) estimated signiﬁcantly lower ET than other models, even in comparison with most previous studies focusing on earlier periods (Table S1 in the Supplement). On the contrary, JULES gave the largest ET estimate (697.3 mm yr−1, which is equal to 7.57 × 104 km3 yr−1) among all models, and showed an obvious in- crease in ET compared with its estimation from 1950 to 2000 (6.5 × 104 km3 yr−1, Table S1). S. Pan et al.: Evaluation of global terrestrial ET After calculating γ LAI ET , γ T ET, γ P ET and γ R ET, the contribu- tion of the trend in factor i (Trend(i)) to the trend in ET (Trend(ET)) can be quantiﬁed as follows: ing N. Pan et al. (2018), grids with an annual mean NDVI of less than 0.1 were assumed to be non-vegetated regions and were, therefore, masked out. NDVI data are from the GIMMS NDVI3gV1 dataset. Temperature, precipitation and radiation are from the CRU-NCEPv8 dataset. Contri(i) = γ i ET × Trend(i) /Trend(ET). (5) (5) 2.3 Description of other datasets To quantify the contributions of vegetation greening to terres- trial ET variations, we used the LAI from the TRENDYv6 S3 simulation. We also used the newest version of the Global Inventory Modeling and Mapping Studies LAI data (GIMMS LAI3gV1) as the satellite-derived LAI. GIMMS LAI3gV1 was generated from AVHRR GIMMS NDVI3g us- ing a model derived from an artiﬁcial neural network (ANN; Zhu et al., 2013). It covers the period from 1982 to 2016 with bimonthly frequency and has a 1/12◦spatial resolu- tion. To achieve a uniform resolution, all data were resam- pled to 1/2◦using the nearest neighbor method. Follow- The ﬁrst used machine-learning model, MTE, is based on the Tree Induction Algorithm (TRIAL) and Evolving Trees with Random Growth (ERROR) algorithm (Jung et al., www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1492 3.2 Spatial patterns of global terrestrial ET Figure 1. Average annual global terrestrial ET estimated by each model during the period from 2001 to 2011. Error bars represent the standard deviation of each model. The four lines indicate the mean value of each category. Figure 1. Average annual global terrestrial ET estimated by each model during the period from 2001 to 2011. Error bars represent the standard deviation of each model. The four lines indicate the mean value of each category. contrast, the IAV of machine-learning-based ET was much weaker. In most regions, the IAV of machine-learning ET is lower than 40 % of the IAV of remote sensing physical ET and LSM ET, and this phenomenon is especially pronounced in tropical regions. Further investigation into the spatial pat- terns of ET IAV for individual models showed that the two machine-learning methods performed equally with respect to estimating spatial patterns of ET IAV (Fig. S4). In con- trast, differences in ET IAV among remote sensing physi- cal estimates and LSM estimates were much larger. LSMs showed the largest differences in the IAV of ET in tropical regions. For example, CABLE and JULES obtained an ET IAV of less than 15 mm yr−1 in most regions of the Amazon Basin, whereas LPJ-GUESS predicted an ET IAV of more than 60 mm yr−1. Figure 5 shows that remote sensing physi- cal ET and LSM ET had comparable IAV north of 20◦S, but the IAV of the machine-learning-based ET was much lower in this region. In the region south of 20◦S, TRENDY ET showed the largest IAV, followed by those of remote sensing physical ET and machine-learning estimates. The three ap- proaches agreed on that ET IAV in the Southern Hemisphere was generally larger than that in the Northern Hemisphere. was also found in the Amazon Basin. MODIS, VISIT and CLASS-CTEM estimated that the annual ET was higher than 1300 mm in the majority of Amazon, whereas JSBACH and LPJ-wsl estimated an ET value lower than 800 mm yr−1 (Fig. S1). As is shown in Fig. S2, the differences in ET estimates among TRENDY models were larger than those among benchmark estimates for tropical and humid regions. The uncertainty of ET estimates from LSMs is particularly large in the Amazon Basin, where the standard deviation of LSM estimates is more than 2 times larger than that of bench- mark estimates. 3.2 Spatial patterns of global terrestrial ET It is noteworthy that, in arid and semiarid re- gions such as western Australia, central Asia, northern China and the western US, the difference in ET estimates among LSMs is signiﬁcantly smaller than those among remote sens- ing models and machine-learning algorithms. 3.2 Spatial patterns of global terrestrial ET As shown in Fig. 2, the spatial patterns of the multiyear average annual ET of different categories were similar. ET was highest in the tropics and low in the northern high lat- itudes and arid regions such as Australia, central Asia, the western US and the Sahel. Compared with remote-sensing- based physical models and LSMs, machine-learning meth- ods obtained a smaller spatial gradient. In general, latitudi- nal proﬁles of ET estimated using different approaches were also consistent (Fig. 3). However, machine-learning meth- ods gave higher ET estimate at high latitudes and lower ET in the tropics compared with other approaches. In the trop- ics, LSMs have signiﬁcantly larger uncertainties than bench- mark products, and the standard deviation of LSMs is about 2 times higher than that of benchmark products (Fig. 3). At other latitudes, LSMs and benchmark ET products generally have comparable uncertainties. The largest difference in ET in the different categories was found in the Amazon Basin (Fig. 2). In most regions of the Amazon Basin, the mean ET of remote sensing physical models is more than 200 mm yr−1 higher than the mean ET of LSMs and machine-learning methods. For individual ET estimates, the largest uncertainty To quantify the contribution of vegetation greening to ter- restrial ET, we separated the trend in terrestrial ET into four components induced by climatic variables and vegetation dy- namics by establishing a multiple linear regression model be- tween global ET and temperature, precipitation, shortwave radiation and LAI (Eqs. 3, 4): δ (ET) = ∂(ET) ∂(LAI)δ (LAI) + ∂(ET) ∂T δ (T) + ∂(ET) ∂(P) δ (P ) + ∂(ET) ∂R δ (R) + ε (3) δ (ET) = γ LAI ET δLAI + γ T ETδT + γ P ETδP + +γ R ETδR + ε, (4) (3) where γ LAI ET , γ T ET, γ P ET and γ R ET are the sensitivities of ET to the leaf area index (LAI), air temperature (T ), precipitation (P ) and radiation (R), respectively. ε is the residual, repre- senting the impacts of other factors. www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET 1493 Figure 1. Average annual global terrestrial ET estimated by each model during the period from 2001 to 2011. Error bars represent the standard deviation of each model. The four lines indicate the mean value of each category. www.hydrol-earth-syst-sci.net/24/1485/2020/ 3.3 Interannual variations in global terrestrial ET The ensemble mean interannual variability (IAV) of remote sensing ET estimates and LSM ET estimates showed sim- ilar spatial patterns (Fig. 4). Both remote sensing physical models and LSMs presented low IAV in ET in the north- ern high latitudes but high IAV in ET in the southwestern US, India, sub-Saharan Africa, the Amazon and Australia. In Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET 1494 Figure 2. Spatial distributions of the mean annual ET derived from (a) remote-sensing-based physical models, (b) machine-learning algo- rithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. Figure 2. Spatial distributions of the mean annual ET derived from (a) remote-sensing-based physical models, (b) machine-learning algo- rithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. Figure 3. Latitudinal proﬁles of the mean annual ET for different categories of models. Each line represents the mean value of the corre- sponding category, and the shading represents the interval of 1 standard deviation. Figure 3. Latitudinal proﬁles of the mean annual ET for different categories of models. Each line represents the mean value of the corre- sponding category, and the shading represents the interval of 1 standard deviation. www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET 1495 Figure 4. Spatial distributions of the interannual variability in ET derived from (a) remote-sensing-based physical models, (b) machine- learning algorithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. The study period used in this study for the interannual variability analysis is from 1982 to 2011. Figure 4. Spatial distributions of the interannual variability in ET derived from (a) remote-sensing-based physical models, (b) machine- learning algorithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. The study period used in this study for the interannual variability analysis is from 1982 to 2011. Figure 5. Latitudinal proﬁles of ET IAV for different categories of models. Each line represents the mean value of the corresponding category, and the shading represents the interval of 1 standard deviation. Figure 5. Latitudinal proﬁles of ET IAV for different categories of models. Each line represents the mean value of the corresponding cate Figure 5. Latitudinal proﬁles of ET IAV for different categories of models. Each line represents the mean value of the corresponding category, and the shading represents the interval of 1 standard deviation. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 3.4 Climatic controls on ET According to the ensemble remote sensing models, temper- ature and radiation dominated ET IAV in northern Eura- sia, northern and eastern North America, southern China, the Congo River basin, and the southern Amazon River basin, while precipitation dominated ET IAV in arid re- gions and semiarid regions (Fig. 6a). The ensemble machine- learning algorithms had a similar pattern, but they suggested a stronger control of radiation in the Amazon Basin and a weaker control of precipitation in several arid regions such as central Asia and northern Australia (Fig. 6b). In com- parison, the ensemble LSMs suggested the strongest control of precipitation on ET IAV (Fig. 6). According to the en- semble LSMs, ET IAV was dominated by precipitation IAV in most regions of the Southern Hemisphere and at north- ern low latitudes. Temperature and radiation only controlled northern Eurasia, eastern Canada and part of the Amazon Basin (Fig. 6d). As is shown in Fig. S6, the majority of LSMs agreed on the dominant role of precipitation in con- trolling ET in regions south of 40◦N. However, the pattern of climatic controls in the ORCHIDEE-MICT model is quite unique and different from all of the other LSMs. Accord- ing to the ORCHIDEE-MICT model, radiation and temper- ature dominate ET IAV in more regions, and precipitation only controls ET IAV in eastern Brazil, northern Russia, cen- tral Europe and a part of tropical Africa. As ORCHIDEE- MICT was developed from ORCHIDEE, the dynamic root parameterization in ORCHIDEE-MICT may explain why ET is less driven by precipitation compared with ORCHIDEE (Haverd et al., 2018). It is noted that two machine-learning algorithms, MTE and RF, showed signiﬁcant discrepancies in the spatial pattern of dominant climatic factors. Accord- ing to the result from MTE, temperature controlled ET IAV in regions north of 45◦N, the eastern US, southern China and the Amazon Basin (Fig. S6e). By contrast, RF suggested that precipitation and radiation dominated ET IAV in these regions (Fig. S6f). According to Fig. 8, the ensemble means of all the three approaches showed increasing trends in ET over western and southern Africa, western India and northern Australia, and decreasing ET over the western US, southern South Amer- ica and Mongolia. Discrepancies in ET trends mainly ap- peared in eastern Europe, eastern India and central China. LSMs also suggested a larger area of decreasing ET in both North America and South America. S. Pan et al.: Evaluation of global terrestrial ET 1496 All remote sensing and machine-learning estimates indi- cate a signiﬁcant increasing trend in ET during the study period (p < 0.05), although the increase rate of P-LSH (1.07 mm yr−2) is more than 3 times as large as that of GLEAM (0.33 mm yr−2). Nevertheless, there is a larger dis- crepancy among LSMs in terms of the ET trend. The ma- jority of LSMs (10 of 14) suggest an increasing trend with an average trend of 0.34 mm yr−2 (p < 0.05), and eight of them are statistically signiﬁcant (see Table 2). However, four LSMs (JSBACH, JULES, ORCHIDEE and ORCHIDEE-MICT) suggest a decreasing trend with an av- erage trend of −0.12 mm yr−2 (p > 0.05). Among the four decreasing trends, only the trend of ORCHIDEE-MICT (−0.34 mm yr−2) is statistically signiﬁcant (p < 0.05). 3.6 Impacts of vegetation changes on ET variations During the period from 1982 to 2011, global LAI trends es- timated from remote sensing data and from the ensemble LSMs are 2.51 × 10−3 m2 m−2 yr−1 (p < 0.01) and 4.63 × 10−3 m2 m−2 yr−1 (p < 0.01), respectively (Table 2). All LSMs suggested a signiﬁcant increasing trend in global LAI (greening). For both benchmark estimates and LSM estimates, it was found that the spatial pattern of trends in ET matched well with that of trends in the LAI (Figs. 8c, d, S5a, b), indicating signiﬁcant effects of veg- etation dynamics on ET variations. According to the re- sults of the multiple linear regression, all models agreed that greening of the Earth since the early 1980s intensiﬁed terrestrial ET (Table 2), although there was a signiﬁcant discrepancy in the magnitude of ET intensiﬁcation which varied from 0.04 to 0.70 mm yr−2. The ensemble LSMs suggested a smaller ET increase (0.23 mm yr−2) than the ensemble remote sensing physical models (0.62 mm yr−2) and machine-learning algorithm (0.38 mm yr−2). Neverthe- less, the greening-induced ET intensiﬁcation estimated by LSMs (0.37 mm yr−2) is larger than that estimated by remote sensing models (0.28 mm yr−2) and machine-learning algo- 3.4 Climatic controls on ET Although the differences in ET trends among individual models were larger than those among the ensemble means of different approaches, the ma- jority of models agreed that ET increased in western and southern Africa, and decreased in the western US and south- ern South America (Fig. S2). For both remote sensing esti- mates and LSM estimates, ET trends in the Amazon Basin had large uncertainty: P-LSH, CLM45 and VISIT suggested a large area of increasing ET, whereas GLEAM, JSBACH and ORCHIDEE suggested a large area of decreasing ET. www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 3.5 Long-term trends in global terrestrial ET All approaches suggested an overall increasing trend in global ET during the period from 1982 to 2011 (Fig. 7), al- though ET decreased from 1998 to 2009. This result is con- sistent with previous studies (Jung et al., 2010; Lian et al., 2018; Zhang et al., 2015). Remote sensing physical mod- els indicated the largest increase in ET (0.62 mm yr−2), fol- lowed by the machine-learning method (0.38 mm yr−2) and land surface models (0.23 mm yr−2). The mean ET of all cat- egories except LSMs signiﬁcantly increased during the study period (p < 0.05). It is noted that the ensemble mean ET values of different categories are statistically correlated with each other (p < 0.001), even if the driving forces of different ET approaches are different. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET 1497 Figure 6. Spatial distributions of climatic controls on the interannual variation of ET derived from the ensemble means of remote-sensing- based physical models (a), machine-learning algorithms (b), benchmark data (c) and the TRENDY LSMs (d). (Red denotes temperature, green denotes precipitation and blue denotes radiation.) Figure 6. Spatial distributions of climatic controls on the interannual variation of ET derived from the ensemble means of remote-sensing- based physical models (a), machine-learning algorithms (b), benchmark data (c) and the TRENDY LSMs (d). (Red denotes temperature, green denotes precipitation and blue denotes radiation.) Figure 7. Interannual variations in global terrestrial ET estimated using different categories of approaches. Figure 7. Interannual variations in global terrestrial ET estimated using different categories of approaches. www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1498 Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET Table 2. Interannual variability (IAV – denoted as standard deviation) and the trend of global terrestrial ET from 1982 to 2011 and the contribution of vegetation greening to the ET trend. (RS refers to remote sensing.) Table 2. Interannual variability (IAV – denoted as standard deviation) and the trend of global terrestrial ET from 1982 to 2011 and the contribution of vegetation greening to the ET trend. (RS refers to remote sensing.) Model ET IAV ET trend Greening-induced Sensitivity of LAI trend (mm yr−1) (mm yr−2) ET change ET to LAI (10−3 m2 m−2 yr−1) (mm yr−2) (mm yr−2 m−2 m−2) Machine MTE 5.93 0.38∗ 0.09 35.86 2.51∗ learning RS models P-LSH 9.95 1.07∗ 0.34 135.46 2.51∗ GLEAM 8.47 0.33∗ 0.14 55.78 2.51∗ PML-CSIRO 7.18 0.41∗ 0.36 143.43 2.51∗ RS model mean 7.98 0.62∗ 0.28 111.55 2.51∗ LSMs CABLE 9.63 0.07 0.35 102.64 3.41∗ CLASS-CTEM 12.22 0.35∗ 0.53 134.52 3.94∗ CLM45 8.68 0.38∗ 0.31 67.54 4.59∗ DLEM 7.21 0.26∗ 0.53 200.76 2.64∗ ISAM 7.50 0.22 0.16 32.26 4.96∗ JSBACH 10.12 −0.05 0.50 217.39 2.30∗ JULES 11.33 −0.02 0.34 85.21 3.99∗ LPJ-GUESS 7.48 0.50∗ 0.28 160.92 1.74∗ LPJ-wsl 4.77 0.24∗ 0.19 31.56 6.02∗ LPX-Bern 4.80 0.20∗ 0.04 4.04 9.90∗ O-CN 10.41 0.32∗ 0.53 89.23 5.94∗ ORCHIDEE 9.28 −0.17 0.21 96.33 2.18∗ ORCHIDEE-MICT 10.70 −0.34∗ 0.50 171.23 2.92∗ VISIT 6.31 0.87∗ 0.70 51.40 13.62∗ LSM mean 7.73 0.23 0.37 79.91 4.63∗ ∗Suggests signiﬁcance at the 95 % conﬁdence level (p < 0.05). ∗Suggests signiﬁcance at the 95 % conﬁdence level (p < 0.05). 4 Discussion and perspectives 4.1 Sources of uncertainty 4.1.1 Uncertainty in the ET estimation of the Amazon Basin 4 Discussion and perspectives 4.1 Sources of uncertainty 4.1.1 Uncertainty in the ET estimation of the Amazon Basin rithms (0.09 mm yr−2), as LSMs suggested a stronger green- ing trend than remote sensing models. The contribution of vegetation greening to ET intensiﬁcation estimated by the ensemble LSMs is larger than 100 %, whereas the contri- butions estimated by the ensemble remote sensing physical models (0.62 mm yr−2) and machine-learning algorithm are smaller than 50 %. Although TRENDY LSMs were driven by the same climate data and remote sensing physical mod- els were driven by varied climate data, TRENDY LSMs still showed a larger discrepancy in terms of the effect of vege- tation greening on terrestrial ET than remote sensing phys- ical models due to the signiﬁcant differences in both LAI trends (1.74–13.63 × 10−3 m2 m−2 yr−1) and the sensitivi- ties of ET to the LAI (4.04–217.39 mm yr−2 m−2 m−2). In comparison, remote sensing physical models had smaller dis- crepancies in terms of the sensitivity of ET to LAI (55.78– 143.43 mm yr−2 m−2 m−2). rithms (0.09 mm yr−2), as LSMs suggested a stronger green- ing trend than remote sensing models. The contribution of vegetation greening to ET intensiﬁcation estimated by the ensemble LSMs is larger than 100 %, whereas the contri- butions estimated by the ensemble remote sensing physical models (0.62 mm yr−2) and machine-learning algorithm are smaller than 50 %. Although TRENDY LSMs were driven by the same climate data and remote sensing physical mod- els were driven by varied climate data, TRENDY LSMs still showed a larger discrepancy in terms of the effect of vege- tation greening on terrestrial ET than remote sensing phys- ical models due to the signiﬁcant differences in both LAI trends (1.74–13.63 × 10−3 m2 m−2 yr−1) and the sensitivi- ties of ET to the LAI (4.04–217.39 mm yr−2 m−2 m−2). In comparison, remote sensing physical models had smaller dis- crepancies in terms of the sensitivity of ET to LAI (55.78– 143.43 mm yr−2 m−2 m−2). S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET 1499 Figure 8. Spatial distributions of ET trends from 1982 to 2011 derived from (a) remote-sensing-based physical models, (b) machine-learning algorithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. Regions with nonsigniﬁcant trends were excluded. Figure 8. Spatial distributions of ET trends from 1982 to 2011 derived from (a) remote-sensing-based physical models, (b) machine-learning algorithms, (c) benchmark datasets and (d) the TRENDY LSMs’ ensemble mean, respectively. Regions with nonsigniﬁcant trends were excluded. timates from the TRENDY LSMs. Nevertheless, it is noted that the congruence across LSM ET estimates does not nec- essarily mean that they are the correct representation of ET. The narrower inter-model range may suggest shared biases. All remote sensing models and machine-learning algorithms except GLEAM do not explicitly take the effects of soil mois- ture into account (Table S1). Given that soil moisture is piv- otal to both canopy conductance and soil evaporation in arid and semiarid regions (A et al., 2019; De Kauwe et al., 2015; Medlyn et al., 2015; Purdy et al., 2018), the lack of soil mois- ture information also increases the bias in ET estimation. In addition, the accuracy of remote sensing data itself is also an uncertainty source. The retrieval of key land surface vari- ables, such as the leaf area index and surface temperature, is inﬂuenced by vegetation architecture, solar zenith angle and satellite observational angle, particularly over heterogeneous surface (Norman and Becker, 1995). (Abramopoulos et al., 1988; Clark et al., 2015; Noilhan and Mahfouf, 1996) and soil moisture’s control on transpiration (Purdy et al., 2018; Szutu and Papuga, 2019) also increase the uncertainty in ET. The abovementioned processes are not independent of each other but interact in complex ways to produce the end result. 4.1.2 Uncertainty in the ET estimation of arid and semiarid regions In arid and semiarid regions, benchmark products show much larger differences in the magnitude of ET than LSMs (Fig. S2). One cause of this phenomenon is the difference in meteorological forcing. Remote sensing and machine- learning datasets used different forcing data. For precipi- tation, RF used the CRUNCEPv6 dataset, MTE used the Global Precipitation Climatology Centre (GPCC) dataset, MODIS used the Global Modeling and Assimilation Ofﬁce (GMAO) dataset, GLEAM used the Multi-Source Weighted- Ensemble Precipitation (MSWEP) dataset, PML-CSIRO used the Princeton Global Meteorological Forcing (PGF) and the WATCH Forcing Data ERA-Interim (WFDEI) datasets, and P-LSH used data derived from four independent sources. As precipitation is the key climatic factor controlling ET in arid and semiarid regions (Fig. 6), discrepancies between dif- ferent forcing precipitation (Sun et al., 2018) may be the main source of large uncertainty there. In comparison, the uniform forcing data reduced the inter-model range in ET es- 4.1.1 Uncertainty in the ET estimation of the Amazon Basin LSMs show large discrepancies in the magnitude and trend of ET in the Amazon Basin (Figs. 3, S3). However, it is challenging to identify the uncertainty sources. Given that the TRENDY LSMs used uniform meteorological inputs, the discrepancies in ET estimates among the participating mod- els mainly arise from the differences in the underlying model structures and parameters. One potential source of uncer- tainty is the parameterization of root water uptake. In the Amazon Basin, a deep root depth has been conﬁrmed using ﬁeld measurements (Nepstad et al., 2004). However, many LSMs have an unrealistically shallow rooting depth (gener- ally less than 2 m), thereby neglecting the existence and sig- niﬁcance of deep roots. The incorrect root distributions en- large the differences in plant available water and root water uptake, producing large uncertainties in ET. In addition, dif- ferences in the parameterization of other key processes per- tinent to ET such as LAI dynamics (Fig. S5), canopy con- ductance variations (Table 1), water movements in the soil Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ 4.1.4 Uncertainty in the global ET trend All three categories of ET models detected an overall increas- ing trend in global terrestrial ET since the early 1980s, which is in agreement with previous studies (Mao et al., 2015; Mi- ralles et al., 2014; Zeng et al., 2018a, b, 2014; Zhang et al., 2015; Y. Zhang et al., 2016). Benchmark products gener- ally suggested stronger ET intensiﬁcation than LSMs. The weaker ET intensiﬁcation in LSMs may be induced by the response of stomatal conductance to the increasing atmo- spheric CO2 concentration. Increasing CO2 affects ET in two ways: on the one hand, increasing CO2 can effectively reduce stomatal conductance and, thus, decrease transpiration (Heij- mans et al., 2001; Leipprand and Gerten, 2006; Swann et al., 2016); on the other hand, it can increase vegetation produc- tivity and, thus, increase LAI. For benchmarks, the second effect could be captured by remotely sensed LAI, NDVI or fAPAR, whereas the ﬁrst effect was neglected by all models except P-LSH (Zhang et al., 2015). In contrast, both effects were modeled in all TRENDY LSMs. LAI dynamics have signiﬁcant inﬂuences on ET. The in- creased LAI trend (greening) since the early 1980s has been reported by previous studies (Mao et al., 2016; Zhu et al., 2016) and has also been conﬁrmed by remote sensing data and all of the TRENDY LSMs used in this study (Table 2, Fig. S5). Zhang et al. (2015) found that the increasing trend of global terrestrial ET from 1982 to 2013 was mainly driven by an increase in the LAI and the enhanced atmosphere wa- ter demand. Using a land–atmosphere coupled global cli- mate model (GCM), Zeng et al. (2018b) further estimated that global LAI increased by about 8 %, resulting in an in- crease of 0.40 ± 0.08 mm yr−2 in global ET (contributing to 55% ± 25 % of the ET increase). This number is close to the estimates of ensemble LSMs (0.37±0.18 mm yr−2). In com- parison, the remote sensing models and machine-learning algorithms used in this study suggested smaller greening- induced ET increases. It is noted that TRENDY LSMs still S. Pan et al.: Evaluation of global terrestrial ET 1500 and has similar ET IAV to LSMs (Fig. S4). This could imply that most of the existing remote sensing models may under- estimate ET IAV in the Southern Hemisphere because the ef- fects of soil moisture are not explicitly considered. Machine- learning algorithms show much lower IAV than other mod- els (Figs. 4, S4). The main reason for this is that ET IAV is partly neglected in the training process, as the magnitude of ET IAV is usually smaller than the spatial and seasonal vari- ability (Anav et al., 2015; Jung et al., 2019). Moreover, the IAV of satellite-based key land surface variables such as the LAI, fAPAR and surface temperature may be not reliable be- cause of the effects of clouds, which, in turn, affects the esti- mation of the IAV of satellite-based ET. It is noted that LSM ET IAV shows large differences at latitudes south of 20◦S (Fig. 5). This divergence in ET IAV indicates that LSMs re- quire better representation of the ET response to climate in the Southern Hemisphere. showed a larger discrepancy in terms of the effect of vege- tation greening on terrestrial ET than remote sensing phys- ical models (Table 2) due to the signiﬁcant differences in the LAI trend (1.74–13.63 × 10−3 m2 m−2 yr−1) and in the sensitivity of ET to LAI (4.04–217.39 mm yr−2 m−2 m−2). Uncertainties in the LAI trend may arise from inappropriate carbon allocations and deﬁciencies in responding to water deﬁcits (Anav et al., 2013; Hu et al., 2018; Murray-Tortarolo et al., 2013; Restrepo-Coupe et al., 2017). Additionally, for machine-learning algorithms, the results from insufﬁcient long-term in situ measurements and sparse observations in tropical, boreal and arid regions imply that there are likely deﬁciencies in representing the temporal variations. 4.1.5 Lack of knowledge of the effects of irrigation Irrigation accounts for about 90 % of human consump- tive water use and largely affects ET in irrigated crop- lands (Siebert et al., 2010). Global water withdrawals for irrigation were estimated to be within the range of 1161– 3800 km3 yr−1 around the year 2000, and they largely in- creased during the period from 2000 to 2014 (Chen et al., 2019). However, none of the remote-sensing-based physi- cal models and machine-learning algorithms explicitly ac- counted for the effects of irrigation on ET, although these effects could be taken into account to some extent by using observed LAI, NDVI or fAPAR to drive the models (Zhang et al., 2015). Considering that annual ET may surpass annual precipitation in cropland areas, Y. Zhang et al. (2016) only used the Budyko hydrometeorological model to constrain the PML-CSIRO model in grids covered by non-crop vegetation. However, the process of irrigation affecting evaporation was still not taken into consideration. For TRENDY LSMs, only 2 of 14 models (DLEM and ISAM) included irrigation pro- cesses (Le Quéré et al., 2018). Therefore, the effects of ir- rigation are largely neglected in existing global ET datasets, which reduces the accuracy of local ET estimates in regions with a large proportion of irrigated cropland. 4.1.3 Uncertainty in the ET IAV in the Southern Hemisphere In regions south of 20◦S (including Australia, southern Africa and southern South America), the ET IAV of remote sensing models and machine-learning algorithms is lower than that of LSMs (Figs. 4, 5), although their spatial patterns are similar. In these regions, GLEAM, the only remote sens- ing model that explicitly considers the effects of soil mois- ture, has larger ET IAV than other remote sensing models www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET gure 9. Interannual variability (a) and the coefﬁcient of variation (b) of ET for each 50 mm interval of mean annual precipitation. Figure 9. Interannual variability (a) and the coefﬁcient of variation (b) of ET for each 50 mm interval of mean annual precipitation. Figure 9. Interannual variability (a) and the coefﬁcient of variation (b) of ET for each 50 mm interval of mean annual precipitation. riability (a) and the coefﬁcient of variation (b) of ET for each 50 mm interval of mean annual precipitation. gure 9. Interannual variability (a) and the coefﬁcient of variation (b) of ET for each 50 mm interval of mean an In comparison, terrestrial ET shows a much lower IAV at the global scale (Table 2), with values ranging from 4.8 to 12.2 mm yr−1 (1 standard deviation), which only equates to 1.0 %–1.8 % of the global annual mean ET. The model results suggest that global terrestrial ET stabilizes at about 6.74×104 km3 yr−1 (603 mm yr−1), which is close to previ- ous estimates (Alton et al., 2009; Mueller et al., 2011; Oki and Kanae, 2006; Zeng et al., 2012). The stability of global terrestrial ET is probably based on partitioning the solar con- stant and suggests that droughts in one place are balanced by excess rain in other places; thus, it all evens out at a global scale. This implies that ET also has a potential plan- etary boundary, a suggestion made by Running (2012) on net primary production (NPP) as a planetary boundary. ET integrates four aspects of the current planetary boundaries deﬁned by Steffen et al. (2015): climate change, freshwater use, land-system change and biochemical ﬂows. Given the importance of ET with respect to linking terrestrial water, carbon, nutrient and energy cycles, more studies on the ET planetary boundary are needed in the context of intensifying global change and increasing anthropogenic perturbations to the Earth system. In short, the multi-model intercomparison indicates that considerable uncertainty exists in both the temporal and spa- tial variations in global ET estimates, although a large por- tion of models adopt similar ET algorithms (Table 1). The major uncertainty source is different for different types of models and regions. The uncertainty is induced by multi- ple factors, including problems pertinent to parameteriza- tion of land processes, lack of in situ measurements, remote sensing acquisition, scaling effects and meteorological forc- ing. 4.1.6 ET variability across the precipitation gradient and its planetary boundary Precipitation is the source of terrestrial ET. According to Fig. 9a, the vast majority of models agree that ET has the largest IAV in regions with annual precipitation between 700 and 1000 mm, although the magnitude of ET IAV has sub- stantial discrepancies among different models. The low ET IAV in arid and semiarid regions does not mean that ET is stable in these regions. In fact, ET has the largest coefﬁcient of variation (CoV – the ratio of ET standard deviation to ET mean value) in arid regions, and all models show a clear neg- ative trend in the CoV with increasing precipitation (Fig. 9b). This is mainly caused by the large CoV of precipitation in arid regions (Fatichi et al., 2012). www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1501 Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ 4.2.2 Machine-learning methods It is well known that the capability of machine-learning algo- rithms to provide accurate ET estimates largely depends on the representativeness of the training datasets with respect to describing ecosystem behaviors (Yao et al., 2017). As a result, machine-learning algorithms may not perform well outside the range of the data used for their training. Unfor- tunately, long-term ﬁeld observations in northern temperate regions are still insufﬁcient. This is an important cause of the small spatial gradient and small IAV of machine-learning ET. Given that remote sensing is capable of providing broad coverage of key biophysical variables at reasonable spatial and temporal resolutions, one way to overcome this chal- lenge is to exclusively use remote sensing observations as training data (Jung et al., 2019; Poon and Kinoshita, 2018). Another simple way to make the IAV of machine-learning ET more realistic is to normalize the yearly anomalies when comparing them with ET estimates from LSMs and remote sensing physical models (Jung et al., 2019). New machine- learning techniques, including the extreme learning machine and the adaptive neuro-fuzzy inference system, can be used to improve the accuracy of ET estimation (Gocic et al., 2016; Ki¸si and Tombul, 2013). Emerging deep-learning methods such as recurrent neural network (RNN) and long short-term memory (LSTM) have the potential to outcompete conven- tional machine-learning methods in modeling ET time series (Reichstein et al., 2018, 2019). Almost all machine-learning datasets used precipitation rather than soil moisture as the explanatory variable when training. However, soil moisture (rather than precipitation) directly controls ET. As more and more global remote-sensing-based soil moisture datasets be- come available, using soil moisture products as input is ex- pected to improve the accuracy of ET estimates, especially for regions with sparse vegetation coverage (Xu et al., 2018). , ; , ; , ) The theories and retrieval algorithms of ET and the related key biophysical variables also need to be further improved. For example, the method for calculating canopy conductance may be improved by integrating remote-sensing-based solar- induced chlorophyll ﬂuorescence (SIF) data. 4.2.1 Remote-sensing-based physical methods Over the past few decades, the development of remote sensing technologies has contributed to a boom in vari- ous ET estimation methods. However, there is still much room for remote sensing technologies to improve (Fisher et al., 2017). The development of new platforms and sensors that have improved global spatiotemporal coverage and us- ing multiband, multisource remote sensing data are the key points. Planned or newly launched satellites, such as NASA’s GRACE Follow-On (GRACE-FO) mission and the ECOsys- tem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) mission, will improve the accuracy of terrestrial ET estimates. ECOSTRESS’s thermal infrared (TIR) multispectral scanner is capable of monitoring diurnal temperature patterns at high resolutions, which provides in- sights into plant response to water stress as well as the means to understand sub-daily ET dynamics (Hulley et al., 2017). GRACE Follow-On observations can be used to constrain subsurface lateral water transfers, which helps to correct soil moisture and subsequently improves the accuracy of ET esti- mates (Rouholahnejad and Martens, 2018). Moreover, build- ing integrated methods that fuse different ET estimates or the upstream satellite-based biophysical variables from different platforms and the other forcing data will be helpful to im- prove the accuracy and spatiotemporal coverage of ET (Ke et al., 2016; Ma et al., 2018; Semmens et al., 2016). 4.2.2 Machine-learning methods SIF data from existing Global Ozone Monitoring Experiment-2 (GOME- 2), Orbiting Carbon Observatory-2 (OCO-2) and TROPO- spheric Monitoring Instrument (TROPOMI) satellite instru- ments as well as the forthcoming OCO-3 and Geostation- ary Carbon Cycle Observatory (GeoCarb) satellites provide a good opportunity to diagnose transpiration and to exam- ine ET partitioning at multiple spatiotemporal scales (Pagán et al., 2019; Stoy et al., 2019; Sun et al., 2017). Theoretical advancements in nonequilibrium thermodynamics and max- imum entropy production (MEP) could be incorporated into the classical ET theories (Xu et al., 2019; K. Zhang et al., 2016). In addition, quantifying the effects of CO2 fertiliza- tion on stomatal conductance is pivotal for remote sensing models to capture the long-term trend of terrestrial ET. S. Pan et al.: Evaluation of global terrestrial ET Based on the results of different approaches, we sug- gest that global terrestrial ET also has a potential planetary boundary, with the value being about 6.74 × 104 km3 yr−1 (603 mm yr−1), which is consistent with previous estimates. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ S. Pan et al.: Evaluation of global terrestrial ET S. Pan et al.: Evaluation of global terrestrial ET 1502 4.2 Recommendations for future development nent with the largest error (Talsma et al., 2018a). Therefore, incorporating the increasingly accessible satellite-based pre- cipitation, soil moisture observations and soil property data will contribute to the improvement of the soil evaporation estimation. Meanwhile, the consideration of soil evaporation under herbaceous vegetation and canopy will also reduce the errors. S. Pan et al.: Evaluation of global terrestrial ET SP initiated this research and was responsi- ble for the integrity of the work as a whole. NP carried out the anal- yses. SP, NP, HT and HS wrote the paper with contributions from all authors. PF, SS, VKA, VH, AKJ, EK, SL, DL, JEMSN, CO, BP, HT and SZ contributed to the TRENDY results. SWR provided the MODIS ET dataset. Author contributions. SP initiated this research and was responsi- ble for the integrity of the work as a whole. NP carried out the anal- yses. SP, NP, HT and HS wrote the paper with contributions from all authors. PF, SS, VKA, VH, AKJ, EK, SL, DL, JEMSN, CO, BP, HT and SZ contributed to the TRENDY results. SWR provided the MODIS ET dataset. S. Pan et al.: Evaluation of global terrestrial ET Code and data availability. TRENDYv6 data are avail- able from Stephen Sitch (s.a.sitch@exeter.ac.uk) upon reasonable request. MODIS ET data are available from http://ﬁles.ntsg.umt.edu/data/NTSG_Products/MOD16/ Code and data availability. TRENDYv6 data are avail- able from Stephen Sitch (s.a.sitch@exeter.ac.uk) upon reasonable request. MODIS ET data are available from http://ﬁles.ntsg.umt.edu/data/NTSG_Products/MOD16/ (Running, 2020). GLEAM ET are available from https://www.gleam.eu/ (Miralles, 2020). Both model tree ensemble and random forest ET data are available from https://www.bgc-jena.mpg.de/geodb/projects/FileDetails.php (Jung, 2020). P-LSH ET data are available from http://ﬁles. ntsg.umt.edu/data/ET_global_monthly/Global_8kmResolution/ (K. E. Zhang, 2020). PML-CSIRO ET data are available from https://data.csiro.au/dap/landingpage?pid=csiro:17375 (Y. Zhang, 2020). CRU-NCEPv8 data are available from Nico- las Viovy upon reasonable request (email address: viovy@dsm- mail.saclay.cea.fr). GIMMS LAI3gV1 data are available from Ranga B. Myneni upon reasonable request (email address: rmyneni@bu.edu). GIMMS NDVI3gV1 data are available from https://ecocast.arc.nasa.gov/data/pub/gimms/3g.v1/ (Pinzon and Tucker, 2020). processes are missing or not being appropriately parameter- ized. Most of the TRENDY LSMs did not simulate processes relevant to human management, including irrigation (Chen et al., 2019) and the application of fertilizers (Mao et al., 2015), or natural disturbances like wildﬁre (Poon and Kinoshita, 2018). Incorporating these processes into present LSMs is critical, although the introduction of new model parameters also potentially leads to an increase in a model’s uncertainty. In light of the importance of soil water availability in con- straining canopy conductance and dynamics, the accurate representation of hydrological processes is a core task for LSMs, particularly in dry regions. Integrating a dynamic root water uptake function and hydraulic redistribution into the LSM can signiﬁcantly improve its performance with respect to estimating seasonal ET and soil moisture (Li et al., 2012). Moreover, other hydrological processes including groundwa- ter (Decker, 2015), lateral ﬂow (Rouholahnejad and Martens, 2018) and water vapor diffusion at the soil surface (Chang et al., 2018) need to be simulated and correctly represented to reproduce the dynamics of soil water and ET. As the canopy LAI plays an important role in regulating ET, correctly simu- lating vegetation dynamics is also critical. One way to do this is to correct the initialization, distribution and parameteriza- tion of vegetation phenology in LSMs (Murray-Tortarolo et al., 2013; Zhang et al., 2019). An appropriate carbon alloca- tion scheme and parameterization of vegetation’s response to water deﬁcits are also important for reproducing vegetation dynamics (Anav et al., 2013). Supplement. The supplement related to this article is available on- line at: https://doi.org/10.5194/hess-24-1485-2020-supplement. Author contributions. 4.2.3 Land surface models In contrast to observation-based methods, LSMs are able to project future changes in ET, and can disentangle the effects of different drivers of ET through factorial analysis. How- ever, results from LSMs are only as good as their parameter- izations of complex land surface processes, which are limited by our incomplete understanding of physical and biological processes (Niu et al., 2011). Although TRENDY LSMs are the state-of-the-art process-based global land surfaces mod- els, improvements are still needed because several important Most existing remote-sensing-based ET studies have fo- cused on total ET; however, the partitioning of ET between transpiration, soil evaporation, and canopy interception may show signiﬁcant divergence even though the total ET is ac- curately estimated (Talsma et al., 2018b). In current remote- sensing-based ET models, soil evaporation, which is sensi- tive to precipitation events and soil moisture, is the compo- www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1503 References Dai, A.: Precipitation characteristics in eighteen coupled climate models, J. Climate, 19, 4605–4630, 2006. Dai, A.: Precipitation characteristics in eighteen coupled climate models, J. Climate, 19, 4605–4630, 2006. Decker, M.: Development and evaluation of a new soil moisture and runoff parameterization for the CABLE LSM including subgrid- scale processes, J. Adv. Model. Earth Sy., 7, 1788–1809, 2015. A, Y., Wang, G., Liu, T., Xue, B., and Kuczera, G.: Spatial variation of correlations between vertical soil water and evapotranspira- tion and their controlling factors in a semi-arid region, J. Hydrol., 574, 53–63, 2019. De Kauwe, M. G., Kala, J., Lin, Y.-S., Pitman, A. J., Medlyn, B. E., Duursma, R. A., Abramowitz, G., Wang, Y.-P., and Miralles, D. G.: A test of an optimal stomatal conductance scheme within the CABLE land surface model, Geosci. Model Dev., 8, 431–452, https://doi.org/10.5194/gmd-8-431-2015, 2015. Abramopoulos, F., Rosenzweig, C., and Choudhury, B.: Im- proved ground hydrology calculations for global climate mod- els (GCMs): Soil water movement and evapotranspiration, J. Cli- mate, 1, 921–941, 1988. d’Orgeval, T., Polcher, J., and de Rosnay, P.: Sensitivity of the West African hydrological cycle in ORCHIDEE to in- ﬁltration processes, Hydrol. Earth Syst. Sci., 12, 1387–1401, https://doi.org/10.5194/hess-12-1387-2008, 2008. Allen, R. G., Pereira, L. S., Raes, D., and Smith, M.: Crop evapotranspiration-Guidelines for computing crop water require- ments – FAO Irrigation and drainage paper 56, Fao, Rome, Italy, 300, D05109, 1998. Dou, X. and Yang, Y.: Evapotranspiration estimation using four different machine learning approaches in different terrestrial ecosystems, Comput. Electron. Agr., 148, 95–106, 2018. Alton, P., Fisher, R., Los, S., and Williams, M.: Simulations of global evapotranspiration using semiempirical and mechanis- tic schemes of plant hydrology, Global Biogeochem. Cy., 23, GB4023, https://doi.org/10.1029/2009GB003540, 2009. Douville, H., Ribes, A., Decharme, B., Alkama, R., and Shefﬁeld, J.: Anthropogenic inﬂuence on multidecadal changes in recon- structed global evapotranspiration, Nat. Clim. Change, 3, 59–62, 2013. Anav, A., Murray-Tortarolo, G., Friedlingstein, P., Sitch, S., Piao, S., and Zhu, Z.: Evaluation of land surface models in reproduc- ing satellite Derived leaf area index over the high-latitude north- ern hemisphere. Part II: Earth system models, Remote Sens., 5, 3637–3661, 2013. Famiglietti, J. and Wood, E. F.: Evapotranspiration and runoff from large land areas: Land surface hydrology for atmospheric general circulation models, Surv. Geophys., 12, 179–204, 1991. Anav, A., Friedlingstein, P., Beer, C., Ciais, P., Harper, A., Jones, C., Murray-Tortarolo, G., Papale, D., Parazoo, N. References C., and Peylin, P.: Spatiotemporal patterns of terrestrial gross primary production: A review, Rev. Geophys., 53, 785–818, 2015. Fatichi, S., Ivanov, V. Y., and Caporali, E.: Investigating Interan- nual Variability of Precipitation at the Global Scale: Is There a Connection with Seasonality?, J. Climate, 25, 5512–5523, 2012. Feng, Y., Cui, N., Gong, D., Zhang, Q., and Zhao, L.: Evaluation of random forests and generalized regression neural networks for daily reference evapotranspiration modelling, Agr. Water Man- age., 193, 163–173, 2017. Antonopoulos, V. Z., Gianniou, S. K., and Antonopoulos, A. V.: Ar- tiﬁcial neural networks and empirical equations to estimate daily evaporation: application to lake Vegoritis, Greece, Hydrolog. Sci. J., 61, 2590–2599, 2016. Fisher, J. B., Tu, K. P., and Baldocchi, D. D.: Global estimates of the land–atmosphere water ﬂux based on monthly AVHRR and ISLSCP-II data, validated at 16 FLUXNET sites, Remote Sens. Environ., 112, 901–919, 2008. Barman, R., Jain, A. K., and Liang, M.: Climate-driven uncertain- ties in modeling terrestrial energy and water ﬂuxes: a site-level to global-scale analysis, Glob. Change Biol., 20, 1885–1900, 2014. Bodesheim, P., Jung, M., Gans, F., Mahecha, M. D., and Reichstein, M.: Upscaled diurnal cycles of land–atmosphere ﬂuxes: a new global half-hourly data product, Earth Syst. Sci. Data, 10, 1327– 1365, https://doi.org/10.5194/essd-10-1327-2018, 2018. Fisher, J. B., Malhi, Y., Bonal, D., Da Rocha, H. R., De Araujo, A. C., Gamo, M., Goulden, M. L., Hirano, T., Huete, A. R., and Kondo, H.: The land–atmosphere water ﬂux in the tropics, Glob. Change Biol., 15, 2694–2714, 2009. Breiman, L.: Random forests, Mach. Learn., 45, 5–32, 2001. Fisher, J. B., Melton, F., Middleton, E., Hain, C., Anderson, M., Allen, R., McCabe, M. F., Hook, S., Baldocchi, D., and Townsend, P. A.: The future of evapotranspiration: Global re- quirements for ecosystem functioning, carbon and climate feed- backs, agricultural management, and water resources, Water Re- sour. Res., 53, 2618–2626, 2017. Chang, L.-L., Dwivedi, R., Knowles, J. F., Fang, Y.-H., Niu, G.-Y., Pelletier, J. D., Rasmussen, C., Durcik, M., Barron-Gafford, G. A., and Meixner, T.: Why Do Large-Scale Land Surface Models Produce a Low Ratio of Transpiration to Evapotranspiration?, J. Geophys. Res.-Atmos., 123, 9109–9130, 2018. Chen, Y., Xia, J., Liang, S., Feng, J., Fisher, J. B., Li, X., Li, X., Liu, S., Ma, Z., and Miyata, A.: Comparison of satellite-based evapotranspiration models over terrestrial ecosystems in China, Remote Sens. Environ., 140, 279–293, 2014. Fu, B. 5 Conclusion Competing interests. The authors declare that they have no conﬂict of interest. In this study, we evaluated 20 global terrestrial ET estimates including 4 from remote-sensing-based physical models, 2 from machine-learning algorithms and 14 from TRENDY LSMs. The ensemble mean values of global terrestrial ET for the three categories agreed well, with values ranging from 589.6 to 617.1 mm yr−1. All three categories detected an overall increasing trend in global ET during the period from 1982 to 2011 and suggested a positive effect of vegetation greening on ET intensiﬁcation. However, the multi-model in- tercomparison indicates that considerable uncertainties still exist in both temporal and spatial variations in global terres- trial ET estimates. LSMs showed signiﬁcant differences in the ET magnitude in tropical regions, especially in the Ama- zon Basin, whereas benchmark ET products showed a larger inter-model range in arid and semiarid regions than LSMs. Trends in ET estimates also showed signiﬁcant discrepan- cies among LSMs. These uncertainties are induced by the parameterization of land processes, meteorological forcings, the lack of in situ measurements, remote sensing acquisi- tion and scaling effects. Model developments and observa- tional improvements provide two parallel pathways towards improving the accuracy of global terrestrial ET estimation. Acknowledgements. We thank all of the contributors who provided the data used in this study, particularly the TRENDY modeling groups. Additional details on funding support for the participating 14 land surface models is provided by the TRENDY Project. Financial support. This research has been supported by the Na- tional Science Foundation (grant nos. 1903722, 1243232 and AGS 12-43071), the US Department of Energy (grant no. DE- SC0016323), the OUC-AU Joint Center program, the Auburn Uni- versity IGP program, the National Institute of Food and Agricul- ture/US Department of Agriculture (grant nos. 2015-67003-23489 and 2015-67003-23485), EC H2020 (CCiCC; grant no. 821003), SNSF (grant no. 20020_172476), and the Earth Systems and Cli- mate Change Hub (funded by the Australian Government’s National Environmental Science Program). Review statement. This paper was edited by Pierre Gentine and re- viewed by two anonymous referees. www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 1504 S. Pan et al.: Evaluation of global terrestrial ET J., and Berendse, F.: Effects of elevated CO2 and vascular plants on evapotranspiration in bog vegetation, Glob. Change Biol., 7, 817–827, 2001. Hu, Z., Shi, H., Cheng, K., Wang, Y. P., Piao, S., Li, Y., Zhang, L., Xia, J., Zhou, L., and Yuan, W.: Joint structural and physiological control on the interannual variation in productivity in a temper- ate grassland: A data-model comparison, Glob. Change Biol., 24, 2965–2979, 2018. Keller, K. M., Lienert, S., Bozbiyik, A., Stocker, T. F., Chu- rakova (Sidorova), O. V., Frank, D. C., Klesse, S., Koven, C. D., Leuenberger, M., Riley, W. J., Saurer, M., Siegwolf, R., Weigt, R. B., and Joos, F.: 20th century changes in carbon isotopes and water-use efﬁciency: tree-ring-based evaluation of the CLM4.5 and LPX-Bern models, Biogeosciences, 14, 2641– 2673, https://doi.org/10.5194/bg-14-2641-2017, 2017. Huang, S., Bartlett, P., and Arain, M. A.: Assessing nitrogen con- trols on carbon, water and energy exchanges in major plant func- tional types across North America using a carbon and nitrogen coupled ecosystem model, Ecol. Model., 323, 12–27, 2016. Kendall, M. G.: Rank correlation methods, Charles Grifﬁn, London, UK, 196 pp., 1955. Hulley, G., Hook, S., Fisher, J., and Lee, C.: ECOSTRESS, A NASA Earth-Ventures Instrument for studying links between the water cycle and plant health over the diurnal cycle, 2017 IEEE International Geoscience and Remote Sensing Sympo- sium (IGARSS), 23–28 July 2017, Fort Worth, USA, 5494–5496, 2017. Khan, M. S., Liaqat, U. W., Baik, J., and Choi, M.: Stand-alone uncertainty characterization of GLEAM, GLDAS and MOD16 evapotranspiration products using an extended triple collocation approach, Agr. Forest Meteorol., 252, 256–268, 2018. Ki¸si, Ö. and Tombul, M.: Modeling monthly pan evaporations using fuzzy genetic approach, J. Hydrol., 477, 203–212, 2013. Ito, A.: Evaluation of the impacts of defoliation by tropical cy- clones on a Japanese forest’s carbon budget using ﬂux data and a process-based model, J. Geophys. Res.-Biogeo., 115, G04013, https://doi.org/10.1029/2010JG001314, 2010. Kisi, O., Sanikhani, H., Zounemat-Kermani, M., and Niazi, F.: Long-term monthly evapotranspiration modeling by several data- driven methods without climatic data, Comput. Electron. Agr., 115, 66–77, 2015. Jackson, R., Reginato, R., and Idso, S.: Wheat canopy temperature: a practical tool for evaluating water requirements, Water Resour. Res., 13, 651–656, 1977. Knauer, J., Werner, C., and Zaehle, S.: Evaluating stomatal models and their atmospheric drought response in a land surface scheme: A multibiome analysis, J. Geophys. Res.-Biogeo., 120, 1894– 1911, 2015. Jimenez, C., Prigent, C., Mueller, B., Seneviratne, S. S. Pan et al.: Evaluation of global terrestrial ET 1505 Jung, M., Reichstein, M., Ciais, P., Seneviratne, S. I., Shefﬁeld, J., Goulden, M. L., Bonan, G., Cescatti, A., Chen, J., and De Jeu, R.: Recent decline in the global land evapotranspiration trend due to limited moisture supply, Nature, 467, 951–954, 2010. elling by extreme learning machine, Comput. Electron. Agr., 127, 56–63, 2016. elling by extreme learning machine, Comput. Electron. Agr., 127, 56–63, 2016. Good, S. P., Noone, D., and Bowen, G.: Hydrologic connectivity constrains partitioning of global terrestrial water ﬂuxes, Science, 349, 175–177, 2015. Jung, M., Reichstein, M., Margolis, H. A., Cescatti, A., Richard- son, A. D., Arain, M. A., Arneth, A., Bernhofer, C., Bonal, D., and Chen, J.: Global patterns of land-atmosphere ﬂuxes of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and meteorologi- cal observations, J. Geophys. Res.-Biogeo., 116, G00J07, https://doi.org/10.1029/2010JG001566, 2011. Guimberteau, M., Zhu, D., Maignan, F., Huang, Y., Yue, C., Dantec- Nédélec, S., Ottlé, C., Jornet-Puig, A., Bastos, A., Laurent, P., Goll, D., Bowring, S., Chang, J., Guenet, B., Tifaﬁ, M., Peng, S., Krinner, G., Ducharne, A., Wang, F., Wang, T., Wang, X., Wang, Y., Yin, Z., Lauerwald, R., Joetzjer, E., Qiu, C., Kim, H., and Ciais, P.: ORCHIDEE-MICT (v8.4.1), a land surface model for the high latitudes: model description and validation, Geosci. Model Dev., 11, 121–163, https://doi.org/10.5194/gmd-11-121- 2018, 2018. Jung, M., Reichstein, M., Schwalm, C. R., Huntingford, C., Sitch, S., Ahlström, A., Arneth, A., Camps-Valls, G., Ciais, P., and Friedlingstein, P.: Compensatory water effects link yearly global land CO2 sink changes to temperature, Nature, 541, 516–520, 2017. Haverd, V., Smith, B., Nieradzik, L., Briggs, P. R., Woodgate, W., Trudinger, C. M., Canadell, J. G., and Cuntz, M.: A new version of the CABLE land surface model (Subversion revision r4601) incorporating land use and land cover change, woody vegetation demography, and a novel optimisation-based approach to plant coordination of photosynthesis, Geosci. Model Dev., 11, 2995– 3026, https://doi.org/10.5194/gmd-11-2995-2018, 2018. Jung, M., Koirala, S., Weber, U., Ichii, K., Gans, F., Camps-Valls, G., Papale, D., Schwalm, C., Tramontana, G., and Reichstein, M.: The FLUXCOM ensemble of global land-atmosphere energy ﬂuxes, Scientiﬁc Data, 6, 74, https://doi.org/10.1038/s41597- 019-0076-8, 2019. Ke, Y., Im, J., Park, S., and Gong, H.: Downscaling of MODIS One kilometer evapotranspiration using Landsat-8 data and machine learning approaches, Remote Sens., 8, 215, https://doi.org/10.3390/rs8030215, 2016. Heijmans, M. M., Arp, W. References P.: On the calculation of the evaporation from land surface, Sci. Atmos. Sin., 5, 23–31, 1981. Fu, G., Charles, S. P., and Yu, J.: A critical overview of pan evapora- tion trends over the last 50 years, Climatic Change, 97, 193–214, 2009. Chen, Y., Feng, X., Fu, B., Shi, W., Yin, L., and Lv, Y.: Recent global cropland water consumption constrained by observations, Water Resour. Res., 55, 3708–3738, 2019. Gedney, N., Cox, P., Betts, R., Boucher, O., Huntingford, C., and Stott, P.: Detection of a direct carbon dioxide effect in continental river runoff records, Nature, 439, 835–838, 2006. Clark, M. P., Fan, Y., Lawrence, D. M., Adam, J. C., Bolster, D., Gochis, D. J., Hooper, R. P., Kumar, M., Leung, L. R., and Mackay, D. S.: Improving the representation of hydrologic pro- cesses in Earth System Models, Water Resour. Res., 51, 5929– 5956, 2015. Glenn, E. P., Nagler, P. L., and Huete, A. R.: Vegetation index meth- ods for estimating evapotranspiration by remote sensing, Surv. Geophys., 31, 531–555, 2010. Gocic, M., Petkoviæ, D., Shamshirband, S., and Kamsin, A.: Com- parative analysis of reference evapotranspiration equations mod- Cleugh, H. A., Leuning, R., Mu, Q., and Running, S. W.: Regional evaporation estimates from ﬂux tower and MODIS satellite data, Remote Sens. Environ., 106, 285–304, 2007. www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET J., Zaehle, S., and Zhu, D.: Global Carbon Budget 2017, Earth Syst. Sci. Data, 10, 405–448, https://doi.org/10.5194/essd-10-405-2018, 2018. Melton, J. R. and Arora, V. K.: Competition between plant functional types in the Canadian Terrestrial Ecosystem Model (CTEM) v. 2.0, Geosci. Model Dev., 9, 323–361, https://doi.org/10.5194/gmd-9-323-2016, 2016. Leuning, R., Zhang, Y., Rajaud, A., Cleugh, H., and Tu, K.: A simple surface conductance model to estimate re- gional evaporation using MODIS leaf area index and the Penman-Monteith equation, Water Resour. Res., 44, W10419, https://doi.org/10.1029/2007WR006562, 2008. Miralles, D. G.: GLEAM v3, available at: https://www.gleam.eu/, last access: 27 March 2020. Li, J., Wang, Y. P., Duan, Q., Lu, X., Pak, B., Wiltshire, A., Robert- son, E., and Ziehn, T.: Quantiﬁcation and attribution of errors in the simulated annual gross primary production and latent heat ﬂuxes by two global land surface models, J. Adv. Model. Earth Sy., 8, 1270–1288, 2016. Miralles, D. G., Holmes, T. R. H., De Jeu, R. A. M., Gash, J. H., Meesters, A. G. C. A., and Dolman, A. J.: Global land-surface evaporation estimated from satellite-based observations, Hydrol. Earth Syst. Sci., 15, 453–469, https://doi.org/10.5194/hess-15- 453-2011, 2011. Miralles, D. G., van den Berg, M., Gash, J., Parinussa, R., Jeu, R., Beck, H., Holmes, T., Jimenez, C., Verhoest, N., and Dorigo, W.: El Nino-La Nina cycle and recent trends in continental evapora- tion, Nat. Clim. Change, 4, 122–126, 2014. Li, L., Wang, Y. P., Yu, Q., Pak, B., Eamus, D., Yan, J., Gorsel, E., and Baker, I. T.: Improving the responses of the Australian community land surface model (CABLE) to seasonal drought, J. Geophys. Res.-Biogeo., 117, G04002, https://doi.org/10.1029/2012JG002038, 2012. Miralles, D. G., Jiménez, C., Jung, M., Michel, D., Ershadi, A., Mc- Cabe, M. F., Hirschi, M., Martens, B., Dolman, A. J., Fisher, J. B., Mu, Q., Seneviratne, S. I., Wood, E. F., and Fernández-Prieto, D.: The WACMOS-ET project – Part 2: Evaluation of global ter- restrial evaporation data sets, Hydrol. Earth Syst. Sci., 20, 823– 842, https://doi.org/10.5194/hess-20-823-2016, 2016. Li, Z.-L., Tang, R., Wan, Z., Bi, Y., Zhou, C., Tang, B., Yan, G., and Zhang, X.: A review of current methodologies for regional evap- otranspiration estimation from remotely sensed data, Sensors, 9, 3801–3853, 2009. Lian, X., Piao, S., Huntingford, C., Li, Y., Zeng, Z., Wang, X., Ciais, P., McVicar, T. R., Peng, S., and Ottlé, C.: Partitioning global land evapotranspiration using CMIP5 models constrained by ob- servations, Nat. Clim. Change, 8, 640–646, 2018. S. Pan et al.: Evaluation of global terrestrial ET I., Mc- Cabe, M., Wood, E., Rossow, W., Balsamo, G., Betts, A., and Dirmeyer, P.: Global intercomparison of 12 land surface heat ﬂux estimates, J. Geophys. Res.-Atmos., 116, D02102, https://doi.org/10.1029/2010JD014545, 2011. Lawrence, D. M., Thornton, P. E., Oleson, K. W., and Bonan, G. B.: The partitioning of evapotranspiration into transpiration, soil evaporation, and canopy evaporation in a GCM: Impacts on land–atmosphere interaction, J. Hydrometeorol., 8, 862–880, 2007. Jung, M.: Latent heat ﬂux on land, available at: https:// www.bgc-jena.mpg.de/geodb/projects/FileDetails.php, last ac- cess: 27 March 2020. Leipprand, A. and Gerten, D.: Global effects of doubled atmo- spheric CO2 content on evapotranspiration, soil moisture and runoff under potential natural vegetation, Hydrolog. Sci. J., 51, 171–185, 2006. Jung, M., Reichstein, M., and Bondeau, A.: Towards global empirical upscaling of FLUXNET eddy covariance obser- vations: validation of a model tree ensemble approach using a biosphere model, Biogeosciences, 6, 2001–2013, https://doi.org/10.5194/bg-6-2001-2009, 2009. Le Quéré, C., Andrew, R. M., Friedlingstein, P., Sitch, S., Pongratz, J., Manning, A. C., Korsbakken, J. I., Peters, G. P., Canadell, Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ S. Pan et al.: Evaluation of global terrestrial ET 1506 J. G., Jackson, R. B., Boden, T. A., Tans, P. P., Andrews, O. D., Arora, V. K., Bakker, D. C. E., Barbero, L., Becker, M., Betts, R. A., Bopp, L., Chevallier, F., Chini, L. P., Ciais, P., Cosca, C. E., Cross, J., Currie, K., Gasser, T., Harris, I., Hauck, J., Haverd, V., Houghton, R. A., Hunt, C. W., Hurtt, G., Ily- ina, T., Jain, A. K., Kato, E., Kautz, M., Keeling, R. F., Klein Goldewijk, K., Körtzinger, A., Landschützer, P., Lefèvre, N., Lenton, A., Lienert, S., Lima, I., Lombardozzi, D., Metzl, N., Millero, F., Monteiro, P. M. S., Munro, D. R., Nabel, J. E. M. S., Nakaoka, S., Nojiri, Y., Padin, X. A., Peregon, A., Pfeil, B., Pierrot, D., Poulter, B., Rehder, G., Reimer, J., Rödenbeck, C., Schwinger, J., Séférian, R., Skjelvan, I., Stocker, B. D., Tian, H., Tilbrook, B., Tubiello, F. N., van der Laan-Luijkx, I. T., van der Werf, G. R., van Heuven, S., Viovy, N., Vuichard, N., Walker, A. P., Watson, A. J., Wiltshire, A. J., Zaehle, S., and Zhu, D.: Global Carbon Budget 2017, Earth Syst. Sci. Data, 10, 405–448, https://doi.org/10.5194/essd-10-405-2018, 2018. induced greening of the northern extratropical land surface, Nat. Clim. Change, 2016, 959–963, 2016. Martens, B., Miralles, D. G., Lievens, H., van der Schalie, R., de Jeu, R. A. M., Fernández-Prieto, D., Beck, H. E., Dorigo, W. A., and Verhoest, N. E. C.: GLEAM v3: satellite-based land evapora- tion and root-zone soil moisture, Geosci. Model Dev., 10, 1903– 1925, https://doi.org/10.5194/gmd-10-1903-2017, 2017. Maselli, F., Papale, D., Chiesi, M., Matteucci, G., Angeli, L., Raschi, A., and Seufert, G.: Operational monitoring of daily evapotranspiration by the combination of MODIS NDVI and ground meteorological data: Application and evaluation in Cen- tral Italy, Remote Sens. Environ., 152, 279–290, 2014. S., Nakaoka, S., Nojiri, Y., Padin, X. A., Peregon, A., Pfeil, B., Pierrot, D., Poulter, B., Rehder, G., Reimer, J., Rödenbeck, C., Schwinger, J., Séférian, R., Skjelvan, I., Stocker, B. D., Tian, H., Medlyn, B. E., Zaehle, S., De Kauwe, M. G., Walker, A. P., Dietze, M. C., Hanson, P. J., Hickler, T., Jain, A. K., Luo, Y., and Parton, W.: Using ecosystem experiments to improve vegetation models, Nat. Clim. Change, 5, 528–534, 2015. Tilbrook, B., Tubiello, F. N., van der Laan-Luijkx, I. T., van der Werf, G. R., van Heuven, S., Viovy, N., Vuichard, N., Walker, A. P., Watson, A. J., Wiltshire, A. S. Pan et al.: Evaluation of global terrestrial ET Piao, S., Ciais, P., Huang, Y., Shen, Z., Peng, S., Li, J., Zhou, L., Liu, H., Ma, Y., and Ding, Y.: The impacts of climate change on water resources and agriculture in China, Nature, 467, 43–51, 2010. I.: Benchmark products for land evapotranspiration: LandFlux- EVAL multi-data set synthesis, Hydrol. Earth Syst. Sci., 17, 3707–3720, https://doi.org/10.5194/hess-17-3707-2013, 2013. Murray-Tortarolo, G., Anav, A., Friedlingstein, P., Sitch, S., Piao, S., Zhu, Z., Poulter, B., Zaehle, S., Ahlström, A., and Lomas, M.: Evaluation of land surface models in reproducing satellite- derived LAI over the high-latitude northern hemisphere. Part I: Uncoupled DGVMs, Remote Sens., 5, 4819–4838, 2013. Pinzon, J. E. and Tucker, C. J.: GIMMS NDVI3gV1, available at: https://ecocast.arc.nasa.gov/data/pub/gimms/3g.v1/, last access: 27 March 2020. Poon, P. and Kinoshita, A.: Estimating Evapotranspiration in a Post- Fire Environment Using Remote Sensing and Machine Learning, Remote Sens., 10, 1728, https://doi.org/10.3390/rs10111728, 2018. Nagler, P. L., Scott, R. L., Westenburg, C., Cleverly, J. R., Glenn, E. P., and Huete, A. R.: Evapotranspiration on western US rivers es- timated using the Enhanced Vegetation Index from MODIS and data from eddy covariance and Bowen ratio ﬂux towers, Remote Sens. Environ., 97, 337–351, 2005. Priestley, C. and Taylor, R.: On the assessment of surface heat ﬂux and evaporation using large-scale parameters, Mon. Weather Rev., 100, 81–92, 1972. Nepstad, D., Lefebvre, P., Lopes da Silva, U., Tomasella, J., Schlesinger, P., Solórzano, L., Moutinho, P., Ray, D., and Guer- reira Benito, J.: Amazon drought and its implications for for- est ﬂammability and tree growth: A basin-wide analysis, Glob. Change Biol., 10, 704–717, 2004. Purdy, A. J., Fisher, J. B., Goulden, M. L., Colliander, A., Halver- son, G., Tu, K., and Famiglietti, J. S.: SMAP soil moisture im- proves global evapotranspiration, Remote Sens. Environ., 219, 1–14, 2018. Reichstein, M., Besnard, S., Carvalhais, N., Gans, F., Jung, M., Kraft, B., and Mahecha, M.: Modelling Landsurface Time-Series with Recurrent Neural Nets, IGARSS 2018-2018 IEEE Interna- tional Geoscience and Remote Sensing Symposium, 22–27 July 2018, Valencia, Spain, 7640–7643, 2018. Niu, G. Y., Yang, Z. L., Mitchell, K. E., Chen, F., Ek, M. B., Bar- lage, M., Kumar, A., Manning, K., Niyogi, D., and Rosero, E.: The community Noah land surface model with multiparameteri- zation options (Noah-MP): 1. Model description and evaluation with local-scale measurements, J. Geophys. Res.-Atmos., 116, D12109, https://doi.org/10.1029/2010JD015139, 2011. S. Pan et al.: Evaluation of global terrestrial ET Reichstein, M., Camps-Valls, G., Stevens, B., Jung, M., Denzler, J., Carvalhais, N., and Prabhat: Deep learning and process un- derstanding for data-driven Earth system science, Nature, 566, 195–204, 2019. Noilhan, J. and Mahfouf, J.-F.: The ISBA land surface parameteri- sation scheme, Global Planet. Change, 13, 145–159, 1996. Norman, J. M. and Becker, F.: Terminology in thermal infrared re- mote sensing of natural surfaces, Remote Sens. Rev., 12, 159– 173, 1995. Restrepo-Coupe, N., Levine, N. M., Christoffersen, B. O., Albert, L. P., Wu, J., Costa, M. H., Galbraith, D., Imbuzeiro, H., Martins, G., and da Araujo, A. C.: Do dynamic global vegetation models capture the seasonality of carbon ﬂuxes in the Amazon basin? A data-model intercomparison, Glob. Change Biol., 23, 191–208, 2017. Oki, T. and Kanae, S.: Global hydrological cycles and world water resources, Science, 313, 1068–1072, 2006. Oleson, K. W., Lawrence, D. M., Gordon, B., Flanner, M. G., Kluzek, E., Peter, J., Levis, S., Swenson, S. C., Thornton, E., and Feddema, J.: Technical description of version 4.0 of the Commu- nity Land Model (CLM), NCAR technical note, National Center for Atmospheric Research, Boulder, CO, USA, 257 pp., 2010. Roderick, M. L. and Farquhar, G. D.: The cause of decreased pan evaporation over the past 50 years, Science, 298, 1410–1411, 2002. Rouholahnejad, E. and Martens, B.: Improving global estimates of terrestrial evaporation by integrating remotely-sensed obser- vations of changes in terrestrial water storage (GRACE) into a global evaporation model (GLEAM), AGU Fall Meeting, 10– 14 December 2018, Washington, D.C., USA, 2018. Pagán, R. B., Maes, H. W., Gentine, P., Martens, B., and Miralles, G. D.: Exploring the Potential of Satellite Solar-Induced Fluo- rescence to Constrain Global Transpiration Estimates, Remote Sens., 11, 413, https://doi.org/10.3390/rs11040413, 2019. Pan, N., Feng, X., Fu, B., Wang, S., Ji, F., and Pan, S.: Increasing global vegetation browning hidden in overall vegetation green- ing: Insights from time-varying trends, Remote Sens. Environ., 214, 59–72, 2018. Running, S. W.: A measurable planetary boundary for the bio- sphere, Science, 337, 1458–1459, 2012. Running, S. W.: MOD16A_MONTHLY.MERRA_GMAO_1kmALB, available at: http://ﬁles.ntsg.umt.edu/data/NTSG_Products/ MOD16/, last access: 27 March 2020. Pan, S., Tian, H., Dangal, S. R., Yang, Q., Yang, J., Lu, C., Tao, B., Ren, W., and Ouyang, Z.: Responses of global terrestrial evapo- transpiration to climate change and increasing atmospheric CO2 in the 21st century, Earth’s Future, 3, 15–35, 2015. MOD16/, last access: 27 March 2020. Sayemuzzaman, M. and Jha, M. S. Pan et al.: Evaluation of global terrestrial ET Monteith, J. L.: Evaporation and environment, Symposia of the society for experimental biology, Cambridge University Press (CUP), Cambridge, UK, 19, 205–234, 1965. Monteith, J. L.: Accommodation between transpiring vegetation and the convective boundary layer, J. Hydrol., 166, 251–263, 1995. Liu, M., Tian, H., Chen, G., Ren, W., Zhang, C., and Liu, J.: Effects of Land-Use and Land-Cover Change on Evapotranspiration and Water Yield in China During 1900–2000, J. Am. Water Resour. As., 44, 1193–1207, 2008. Mu, Q., Heinsch, F. A., Zhao, M., and Running, S. W.: Development of a global evapotranspiration algorithm based on MODIS and global meteorology data, Remote Sens. Environ., 111, 519–536, 2007. Ma, Y., Liu, S., Song, L., Xu, Z., Liu, Y., Xu, T., and Zhu, Z.: Estimation of daily evapotranspiration and irrigation water ef- ﬁciency at a Landsat-like scale for an arid irrigation area using multi-source remote sensing data, Remote Sens. Environ., 216, 715–734, 2018. Mu, Q., Zhao, M., and Running, S. W.: Improvements to a MODIS global terrestrial evapotranspiration algorithm, Remote Sens. En- viron., 115, 1781–1800, 2011. Mann, H. B.: Nonparametric tests against trend, Econometrica, 1945, 245–259, 1945. Mueller, B., Seneviratne, S. I., Jimenez, C., Corti, T., Hirschi, M., Balsamo, G., Ciais, P., Dirmeyer, P., Fisher, J., and Guo, Z.: Eval- uation of global observations-based evapotranspiration datasets and IPCC AR4 simulations, Geophys. Res. Lett., 38, L06402, https://doi.org/10.1029/2010GL046230, 2011. Mao, J., Fu, W., Shi, X., Ricciuto, D. M., Fisher, J. B., Dickin- son, R. E., Wei, Y., Shem, W., Piao, S., and Wang, K.: Disen- tangling climatic and anthropogenic controls on global terres- trial evapotranspiration trends, Environ. Res. Lett., 10, 094008, https://doi.org/10.1088/1748-9326/10/9/094008, 2015. Mueller, B., Hirschi, M., Jimenez, C., Ciais, P., Dirmeyer, P. A., Dolman, A. J., Fisher, J. B., Jung, M., Ludwig, F., Maignan, F., Miralles, D. G., McCabe, M. F., Reichstein, M., Shefﬁeld, Mao, J., Ribes, A., Yan, B., Shi, X., Thornton, P. E., Séférian, R., Ciais, P., Myneni, R. B., Douville, H., and Piao, S.: Human- J., Wang, K., Wood, E. F., Zhang, Y., and Seneviratne, S. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ 1507 S. Pan et al.: Evaluation of global terrestrial ET B.: Global estimates of evapotranspiration for climate studies using multi-sensor remote sensing data: Evaluation of three process- based approaches, Remote Sens. Environ., 115, 801–823, 2011b. Wang, K. and Dickinson, R. E.: A review of global ter- restrial evapotranspiration: Observation, modeling, climatol- ogy, and climatic variability, Rev. Geophys., 50, RG2005, https://doi.org/10.1029/2011RG000373, 2012. Stoy, P. C., El-Madany, T. S., Fisher, J. B., Gentine, P., Gerken, T., Good, S. P., Klosterhalfen, A., Liu, S., Miralles, D. G., Perez- Priego, O., Rigden, A. J., Skaggs, T. H., Wohlfahrt, G., Anderson, R. G., Coenders-Gerrits, A. M. J., Jung, M., Maes, W. H., Mam- marella, I., Mauder, M., Migliavacca, M., Nelson, J. A., Poyatos, R., Reichstein, M., Scott, R. L., and Wolf, S.: Reviews and syn- theses: Turning the challenges of partitioning ecosystem evapo- ration and transpiration into opportunities, Biogeosciences, 16, 3747–3775, https://doi.org/10.5194/bg-16-3747-2019, 2019. Wang, K., Dickinson, R. E., Wild, M., and Liang, S.: Evidence for decadal variation in global terrestrial evapotranspiration between 1982 and 2002: 1. Model development, J. Geophys. Res.-Atmos., 115, D20112, https://doi.org/10.1029/2009JD013671, 2010. Wartenburger, R., Seneviratne, S. I., Hirschi, M., Chang, J., Ciais, P D D Elli J F lb h C G li S N d Wartenburger, R., Seneviratne, S. I., Hirschi, M., Chang, J., Ciais, P., Deryng, D., Elliott, J., Folberth, C., Gosling, S. N., and Gudmundsson, L.: Evapotranspiration simulations in ISIMIP2a- Evaluation of spatio-temporal characteristics with a comprehen- sive ensemble of independent datasets, Environ. Res. Lett., 13, 075001, https://doi.org/10.1088/1748-9326/aac4bb, 2018. Sun, Q., Miao, C., Duan, Q., Ashouri, H., Sorooshian, S., and Hsu, K. L.: A review of global precipitation data sets: Data sources, estimation, and intercomparisons, Rev. Geophys., 56, 79–107, 2018. Willett, K. M., Gillett, N. P., Jones, P. D., and Thorne, P. W.: Attri- bution of observed surface humidity changes to human inﬂuence, Nature, 449, 710–712, 2007. Sun, Y., Frankenberg, C., Wood, J. D., Schimel, D. S., Jung, M., Guanter, L., Drewry, D. T., Verma, M., Porcar-Castell, A., Grifﬁs, T. J., Gu, L., Magney, T. S., Köhler, P., Evans, B., and Yuen, K.: OCO-2 advances photosynthesis observation from space via solar-induced chlorophyll ﬂuorescence, Science, 358, eaam5747, https://doi.org/10.1126/science.aam5747, 2017. Xu, D., Agee, E., Wang, J., and Ivanov, V. Y.: Estimation of Evap- otranspiration of Amazon Rainforest Using the Maximum En- tropy Production Method, Geophys. Res. Lett., 46, 1402–1412, 2019. Swann, A. L., Hoffman, F. M., Koven, C. D., and Randerson, J. S. Pan et al.: Evaluation of global terrestrial ET 1508 sub-tropical environment, Agr. Forest Meteorol., 200, 172–184, 2015. Components in Remote Sensing-Based Models, Remote Sens., 10, 1601, https://doi.org/10.3390/rs10101601, 2018a. Siebert, S., Burke, J., Faures, J. M., Frenken, K., Hoogeveen, J., Döll, P., and Portmann, F. T.: Groundwater use for irrigation – a global inventory, Hydrol. Earth Syst. Sci., 14, 1863–1880, https://doi.org/10.5194/hess-14-1863-2010, 2010. Talsma, C. J., Good, S. P., Jimenez, C., Martens, B., Fisher, J. B., Miralles, D. G., McCabe, M. F., and Purdy, A. J.: Partitioning of evapotranspiration in remote sensing-based models, Agr. Forest Meteorol., 260–261, 131–143, 2018b. Simmons, A., Willett, K., Jones, P., Thorne, P., and Dee, D.: Low- frequency variations in surface atmospheric humidity, tempera- ture, and precipitation: Inferences from reanalyses and monthly gridded observational data sets, J. Geophys. Res.-Atmos., 115, D01110, https://doi.org/10.1029/2009JD012442, 2010. Tian, H., Chen, G., Liu, M., Zhang, C., Sun, G., Lu, C., Xu, X., Ren, W., Pan, S., and Chappelka, A.: Model estimates of net primary productivity, evapotranspiration, and water use efﬁciency in the terrestrial ecosystems of the southern United States during 1895– 2007, Forest Ecol. Manage., 259, 1311–1327, 2010. Sitch, S., Smith, B., Prentice, I. C., Arneth, A., Bondeau, A., Cramer, W., Kaplan, J., Levis, S., Lucht, W., and Sykes, M. T.: Evaluation of ecosystem dynamics, plant geography and terres- trial carbon cycling in the LPJ dynamic global vegetation model, Glob. Change Biol., 9, 161–185, 2003. Trenberth, K. E., Fasullo, J. T., and Kiehl, J.: Earth’s global energy budget, B. Am. Meteorol. Soc., 90, 311–324, 2009. Van Dijk, A. and Bruijnzeel, L. A.: Modelling rainfall intercep- tion by vegetation of variable density using an adapted analyt- ical model. Part 1. Model description, J. Hydrol., 247, 230–238, 2001. Smith, B.: LPJ-GUESS-an ecosystem modelling framework, De- partment of Physical Geography and Ecosystems Analysis, INES, Sölvegatan, 12, 22362, 2001. Vinukollu, R. K., Meynadier, R., Shefﬁeld, J., and Wood, E. F.: Multi-model, multi-sensor estimates of global evapotranspira- tion: Climatology, uncertainties and trends, Hydrol. Process., 25, 3993–4010, 2011a. Steffen, W., Richardson, K., Rockström, J., Cornell, S. E., Fetzer, I., Bennett, E. M., Biggs, R., Carpenter, S. R., de Vries, W., de Wit, C. A., Folke, C., Gerten, D., Heinke, J., Mace, G. M., Persson, L. M., Ramanathan, V., Reyers, B., and Sörlin, S.: Planetary bound- aries: Guiding human development on a changing planet, Sci- ence, 347, 1259855, https://doi.org/10.1126/science.1259855, 2015. Vinukollu, R. K., Wood, E. F., Ferguson, C. R., and Fisher, J. S. Pan et al.: Evaluation of global terrestrial ET K.: Seasonal and annual precipita- tion time series trend analysis in North Carolina, United States, Atmos. Res., 137, 183–194, 2014. Pan, S., Chen, G., Ren, W., Dangal, S. R. S., Banger, K., Yang, J., Tao, B., and Tian, H.: Responses of global terrestrial water use efﬁciency to climate change and rising atmospheric CO2 concen- tration in the twenty-ﬁrst century, Int. J. Digit. Earth, 11, 558– 582, 2018. Semmens, K. A., Anderson, M. C., Kustas, W. P., Gao, F., Alﬁeri, J. G., McKee, L., Prueger, J. H., Hain, C. R., Cammalleri, C., Yang, Y., Xia, T., Sanchez, L., Mar Alsina, M., and Vélez, M.: Monitor- ing daily evapotranspiration over two California vineyards using Landsat 8 in a multi-sensor data fusion approach, Remote Sens. Environ., 185, 155–170, 2016. Peng, S., Piao, S., Ciais, P., Myneni, R. B., Chen, A., Cheval- lier, F., Dolman, A. J., Janssens, I. A., Penuelas, J., and Zhang, G.: Asymmetric effects of daytime and night-time warming on Northern Hemisphere vegetation, Nature, 501, 88–92, 2013. Sen, P. K.: Estimates of the regression coefﬁcient based on Kendall’s tau, J. American Stat. Assoc., 63, 1379–1389, 1968. Peterson, T. C., Golubev, V. S., and Groisman, P. Y.: Evaporation losing its strength, Nature, 377, 687–688, 1995. Shrestha, N. and Shukla, S.: Support vector machine based mod- eling of evapotranspiration using hydro-climatic variables in a www.hydrol-earth-syst-sci.net/24/1485/2020/ Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 S. Pan et al.: Evaluation of global terrestrial ET 1509 Zhang, K.: P-LSH ET, available at: http://ﬁles.ntsg.umt.edu/ data/ET_global_monthly/Global_8kmResolution/, last access: 27 March 2020. and surface eddy covariance observations, Agr. Forest Meteorol., 223, 151–167, 2016. and surface eddy covariance observations, Agr. Forest Meteorol., 223, 151–167, 2016. Yao, Y., Liang, S., Li, X., Chen, J., Liu, S., Jia, K., Zhang, X., Xiao, Z., Fisher, J. B., and Mu, Q.: Improving global terrestrial evap- otranspiration estimation using support vector machine by inte- grating three process-based algorithms, Agr. Forest Meteorol., 242, 55–74, 2017. Zhang, K., Kimball, J. S., Nemani, R. R., and Running, S. W.: A continuous satellite-derived global record of land surface evapotranspiration from 1983 to 2006, Water Resour. Res., 46, W09522, https://doi.org/10.1029/2009WR008800, 2010. Yuan, W., Liu, S., Liu, H., Randerson, J. T., Yu, G., and Tieszen, L. L.: Impacts of precipitation seasonality and ecosystem types on evapotranspiration in the Yukon River Basin, Alaska, Water Resour. Res., 46, W02514, https://doi.org/10.1029/2009WR008119, 2010. Zhang, K., Kimball, J. S., Nemani, R. R., Running, S. W., Hong, Y., Gourley, J. J., and Yu, Z.: Vegetation greening and climate change promote multidecadal rises of global land evapotranspi- ration, Sci. Rep., 5, 15956, https://doi.org/10.1038/srep15956, 2015. Yue, S., Pilon, P., and Cavadias, G.: Power of the Mann–Kendall and Spearman’s rho tests for detecting monotonic trends in hy- drological series, J. Hydrol., 259, 254–271, 2002. Zhang, K., Kimball, J. S., and Running, S. W.: A review of remote sensing based actual evapotranspiration estimation, WIRES: Wa- ter, 3, 834–853, 2016. Zaehle, S. and Friend, A.: Carbon and nitrogen cycle dy- namics in the O-CN land surface model: 1. Model de- scription, site-scale evaluation, and sensitivity to param- eter estimates, Global Biogeochem. Cy., 24, GB1005, https://doi.org/10.1029/2009GB003521, 2010. Zhang, L., Lei, H., Shen, H., Cong, Z., Yang, D., and Liu, T.: Eval- uating the Representation of Vegetation Phenology in the Com- munity Land Model 4.5 in a Temperate Grassland, J. Geophys. Res.-Biogeo., 124, 187–210, 2019. Zhang, Y.: Monthly global observation-driven Penman-Monteith- Leuning (PML) evapotranspiration and components, available at: https://data.csiro.au/dap/landingpage?pid=csiro:17375, last ac- cess: 27 March 2020. Zeng, Z., Piao, S., Lin, X., Yin, G., Peng, S., Ciais, P., and Myneni, R. B.: Global evapotranspiration over the past three decades: estimation based on the water balance equation com- bined with empirical models, Environ. Res. Lett., 7, 014026, https://doi.org/10.1088/1748-9326/7/1/014026, 2012. Zhang, Y., Peña-Arancibia, J. L., McVicar, T. R., Chiew, F. H., Vaze, J., Liu, C., Lu, X., Zheng, H., Wang, Y., and Liu, Y. S. Pan et al.: Evaluation of global terrestrial ET T.: Plant responses to increasing CO2 reduce estimates of cli- mate impacts on drought severity, P. Natl. Acad. Sci. USA, 113, 10019–10024, 2016. Xu, T., Guo, Z., Liu, S., He, X., Meng, Y., Xu, Z., Xia, Y., Xiao, J., Zhang, Y., Ma, Y., and Song, L.: Evaluating Different Machine Learning Methods for Upscaling Evapotranspiration from Flux Towers to the Regional Scale, J. Geophys. Res.-Atmos., 123, 8674–8690, 2018. Szutu, D. J. and Papuga, S. A.: Year-Round Transpiration Dynamics Linked With Deep Soil Moisture in a Warm Desert Shrubland, Water Resour. Res., 55, 5679–5695, 2019. Yao, Y., Liang, S., Cheng, J., Liu, S., Fisher, J. B., Zhang, X., Jia, K., Zhao, X., Qin, Q., and Zhao, B.: MODIS-driven estimation of terrestrial latent heat ﬂux in China based on a modiﬁed Priestley– Taylor algorithm, Agr. Forest Meteorol., 171, 187–202, 2013. Tabari, H., Martinez, C., Ezani, A., and Talaee, P. H.: Applicability of support vector machines and adaptive neurofuzzy inference system for modeling potato crop evapotranspiration, Irrigation Sci., 31, 575–588, 2013. Yao, Y., Liang, S., Li, X., Liu, S., Chen, J., Zhang, X., Jia, K., Jiang, B., Xie, X., and Munier, S.: Assessment and simulation of global terrestrial latent heat ﬂux by synthesis of CMIP5 climate models Talsma, C. J., Good, S., Miralles, D., Fisher, J., Martens, B., Jimenez, C., and Purdy, A.: Sensitivity of Evapotranspiration Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ S. Pan et al.: Evaluation of global terrestrial ET Y.: Multi-decadal trends in global terrestrial evap- otranspiration and its components, Sci. Rep., 6, 19124, https://doi.org/10.1038/srep19124, 2016. Zeng, Z., Wang, T., Zhou, F., Ciais, P., Mao, J., Shi, X., and Piao, S.: A worldwide analysis of spatiotemporal changes in water balance-based evapotranspiration from 1982 to 2009, J. Geo- phys. Res.-Atmos., 119, 1186–1202, 2014. Zhu, Z., Bi, J., Pan, Y., Ganguly, S., Anav, A., Xu, L., Samanta, A., Piao, S., Nemani, R. R., and Myneni, R. B.: Global data sets of vegetation leaf area index (LAI) 3g and Fraction of Photosyn- thetically Active Radiation (FPAR) 3g derived from Global In- ventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI3g) for the period 1981 to 2011, Remote Sens., 5, 927–948, 2013. Zeng, Z., Zhu, Z., Lian, X., Li, L. Z., Chen, A., He, X., and Piao, S.: Responses of land evapotranspiration to Earth’s greening in CMIP5 Earth System Models, Environ. Res. Lett., 11, 104006, https://doi.org/10.1088/1748-9326/11/10/104006, 2016. Zeng, Z., Peng, L., and Piao, S.: Response of terrestrial evapotran- spiration to Earth’s greening, Curr. Opin. Env. Sust., 33, 9–25, 2018a. Zhu, Z., Piao, S., Myneni, R. B., Huang, M., Zeng, Z., Canadell, J. G., Ciais, P., Sitch, S., Friedlingstein, P., and Arneth, A.: Green- ing of the Earth and its drivers, Nat. Clim. Change, 2016, 791– 795, 2016. Zeng, Z., Piao, S., Li, L. Z., Wang, T., Ciais, P., Lian, X., Yang, Y., Mao, J., Shi, X., and Myneni, R. B.: Impact of Earth greening on the terrestrial water cycle, J. Climate, 31, 2633–2650, 2018b. Hydrol. Earth Syst. Sci., 24, 1485–1509, 2020 www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/ www.hydrol-earth-syst-sci.net/24/1485/2020/
https://openalex.org/W2148163334	https://www.scielo.br/j/abb/a/6HYsbJK9nkbJ65fN4H7GBPH/?lang=en&format=pdf	English	null	Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum (Apocynaceae)	Acta Botanica Brasílica/Acta Botânica Brasílica	2,010	cc-by	4,868	Recebido em 20/08/2009. Aceito em 1/06/2010 RESUMO – (Relações fenológicas entre duas galhas induzidas por insetos e suas plantas hospedeiras: Aspidosperma australe e A. spruceanum (Apocynace- ae)). Embora a diversidade de galhas na região neotropical seja grande, poucos são os estudos fenológicos comparando a fenologia das espécies hospedeiras com aquela das galhas. O desenvolvimento de galhas geralmente requer alta sincronia fenológica entre os organismos associados. A relação entre a fenologia de duas galhas foliares induzidas por um Cecidomyiidae e Aspidosperma spruceanum Benth. ex Müell. Arg. e de uma espécie de Pseudophacopteron sp. e A. australe Müell. Arg. foram investigadas. O trabalho foi realizado em dez indivíduos de cada espécie em intervalos quinzenais, levando em consideração a percentagem de folhas galhadas. Durante um ano, foram observadas três fenofases distintas para as galhas foliares e quatro fenofases para a hospedeira. A maior percentagem de folhas galhadas em A. australe (80%) ocorreu após o pico de brotação, com uma boa correlação entre a brotação e a indução de novas galhas. Em A. spruceanum, a percentagem de folhas galhadas foi sempre acima de 50%, o que pode ser relacionado à contínua produção de folhas e indução das galhas nesta espécie. Em ambas as espécies, as galhas em desenvolvimento foram observadas durante todo o ano, indicando multivoltinismo. A capacidade de induzir galhas em tecidos jovens e maduros parece ser uma boa estratégia para a sobrevivência destes galhadores. Palavras chaves: abundância de galhas, fenologia, interação inseto-planta, sincronismo ABSTRACT – (Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum (Apocynaceae)). Although gall diversity in the Neotropical region is immense, comparative studies on the phenology of host plants and their galls are scarce. Gall systems generally require high levels of phenological synchrony between the associated organisms. The relationships between the phenology of two leaf galls induced by an unidentiﬁ ed Cecidomyiidae in Aspidosperma spruceanum Benth. ex Müell. Arg. and by Pseudophacopteron sp. in A. australe Müell. Arg. were investigated. The investigation was performed on ten individuals per species in 15-day intervals taking into consideration the percentage of galled leaves. In a one-year study, three distinct phenophases for the leaf galls and four phenophases for host plants were observed. The maximum percentage of leaf galls (80%) on A. australe occurred just after the peak of leaf sprouting. In A. 1 Universidade Federal de Minas Gerais, Instituto de Ciências Biológicas, Departamento de Botânica, Belo Horizonte, MG, Brasil 2 Universidade Federal de Goiás, Instituto de Ciências Biológicas, Departamento de Biologia Geral, Goiânia, Brasil 3 Corresponding author: lemos@icb.ufmg.br Recebido em 20/08/2009. Aceito em 1/06/2010 spruceanum, the percentage of leaf galls was always over 50%, which can be related to continuous leaf production and gall induction in this species. In both species, developing galls were observed over the entire year, indicating multivoltinism. The ability to induce galls at young and mature sites seems to be a good strategy for galling species survivorship. Key words: gall abundance, plant interaction, phenology, synchronism Acta bot. bras. 24(3): 727-733. 2010. Acta bot. bras. 24(3): 727-733. 2010. Introduction dependent on the oviposition site. Even though a preference for oviposition in young tissues was recorded (Rohfritsch 1992), some herbivores may induce galls either in young or mature leaves (Souza et al. 2000; Arduin & Kraus 2001; Oliveira & Isaias 2009). When gall induction occurs in young tissues, phenological synchronization between the host plant and the galling herbivore is necessary for the reproductive success of the latter (Yukawa 2000; Weis et al. 1988). Also, because of this requirement, there is a strong tendency among galling insects towards univoltinism (Weis et al. 1988). Considering the importance of this adjustment in the life cycles of the two organisms involved, the current study aimed to describe leaf gall phenology in two species of Aspidosperma over a year’s time, relating the correspondence of initial gall phases to the availability of host plant tissues in responsive state. As evergreen plants are common in the Neotropics, we hypothesize that the availability of responsive tissues during the entire year is a major inﬂ uence towards multivoltinism. In order to verify this statement, we analyzed two systems involving A. australe and an undescribed species of Pseudophacopteron (Phacopteronidae), and A. spruceanum and an unidentiﬁ ed species of Cecidomyiidae. Phenological studies deal with recurrent biological events particular to one or among several species, with the interference of biotic and abiotic factors (Lieth 1974). Moreover, they deal with seasonal regulation of events in life histories of organisms (Rathcke & Lacey 1985). For plants, in particular, triggering of the phenophases may be related to variations in abiotic or biotic environment, and so a particular phenophase may be affected by selective pressures imposed by animals (van Schaik et al. 1993). Vegetative phenophases are dependent on climate seasonality, with strong variation in strategies for the production of new leaves. In the tropics, leaf sprouting may be continuous or seasonal, depending on limiting factors such as water availability (Wright & van Schaik 1994). The occurrence of peaks of leaf sprouting in a year’s time, low herbivore abundance at the dry season or asynchronic production of leaves may enable some leaves or ﬂ owers to escape from predation (Coley & Barone 1996). Gall inducers are plant parasites (Price 1984), which usually have strict, speciﬁ c relationships either with their host plant or with organs of this plant. Introduction This speciﬁ city may be a consequence of limited movements during the larval phase, and to feeding sites, namely, the cells and tissues of the gall (Yamazaki & Ohsaki 2006), which is eventually Due to the great abundance of galls in these two host plants, the gall systems have been the focus of Campos et al.: Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... 728 Campos et al.: Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... morphological, ultrastructural, chemical and histochemical studies. The anatomical aspects of the galls of A. australe were presented by Christiano (2002), while Formiga et al. (2009) related the structural aspects of the galls of A. spruceanum to the variation in phenolic contents over a year’s time. Oliveira and Isaias (2010) and Oliveira et al. (2010) related the ultrastructural and histochemical aspects of both galls to tissue nutritional metabolism. Therefore, the study of phenological events in galling insects and host- plant life cycles should contribute to the knowledge of these woody Apocynaceae species, and also generate complete models involving gall systems in the Neotropics. leaf lamina (Fig. 2, 3), developmental phase was indicated by leaf lamina eruption (Fig. 4-7), and senescent phase by the presence of an aperture on the gall surface (Fig. 8, 9). For gall phenology, percentage of the three distinct phases was visually evaluated for each sample. All data were expressed as an average of the two evaluations per month. Numerical data were submitted to Spearman correlation using Graph Pad Prisma software (Motulsky 1992-2009). Results Leaf galls were observed in both Aspidosperma species during the entire study period (Fig. 10, 11). For A. australe, levels of gall infestation ﬂ uctuated over the year, with 27% in September 2006. The highest percentage of galled leaves for this species was recorded in December 2006 (80%), just after the peak of leaf sprouting which occurred in September 2006 (40%). At this time of the year, A. australe had the highest activity index and percentage of phenology intensity. A high percentage of galled leaves occurred again in July 2007 (77%), independently of new leaf production. After that time, the number of mature leaves reaches its maximum for all plant individuals (Fig. 12, 14). In A. spruceanum, gall infestation was superior to 50% all year long, reaching 75% in July 2007 (Fig. 11). Also, there is a continuous production of new leaves and the maintenance of a great proportion of mature leaves in the canopy (Fig. 13-15). The maximum percentage of galled leaves in A. spruceanum occurred nearly two months after leaf sprouting (July, 76%), similarly to A. australe. In the present study, galls in developmental phase were recorded over the entire year (Fig. 18, 19). Materials and methods Phenological events of Aspidosperma australe and A. spruceanum were analyzed at Pampulha Campus of the Universidade Federal de Minas Gerais (UFMG) (19º52’S; 43º57’W) between September 2006 and September 2007. Meteorological data were obtained from INMET at Belo Horizonte-Minas Gerais, Brazil, and clearly set a dry and a rainy season. During the study period, higher values of maximum temperature were recorded in March (31.7°C), and lower values of minimum temperature in June and July (15.7°C and 9.3°C, respectively) (Fig. 1). The wet season began in September 2006, with higher accumulated precipitation (350 mm) in January 2007. The dry season bega n in May 2007, with no precipitation in June and August 2007. For vegetative phenology, 10 individuals of each species were observed at 15-day intervals. Leaf sprouting, presence of mature (dark green) and senescent leaves (yellowish green), and leaf fall were evaluated by the visual percentage of areas occupied by the crown. The activity index (percentage of individuals in a speciﬁ c phenophase), and the percentage of phenophase intensity, the Fournier index (Fournier 1974), were evaluated as described by Bencke & Morellato (2002), using a scale of ﬁ ve categories (0-4 with intervals of 25%). The results correspond to an average of the two evaluations per month. The activity of the gall inducing Phacopteronidae was not recorded in A. australe in March, May, and from July to September 2007 (Fig. 18), exactly in the dry season (Fig. 1). For this insect, maximum induction activity was recorded in September 2006, concomitantly with the annual peak in the production of new leaves (Fig. 12). The Cecidomyiidae Gall phenophases (induction, development, and senescence) were evaluated in 400 leaves per species. Each plant individual (n = 10) crown was divided into four quarters according to the cardinal points from where 40 leaves per individual were sampled, at each 15-day-interval observation. Induction phase was deﬁ ned as the presence of distinct light green spots on Figure 1. Maximum (▪▪▲▪▪), medium (—■—), and minimum (▪▪▼▪▪) annual temperature recorded for Belo Horizonte from August 2006 to September 2007. Bars represent the monthly accumulated precipitation during the same period (Data from INMET, Belo Horizonte, Minas Gerais, Brazil). Figure 1. Maximum (▪▪▲▪▪), medium (—■—), and minimum (▪▪▼▪▪) annual temperature recorded for Belo Horizonte from August 2006 to September 2007. Bars represent the monthly accumulated precipitation during the same period (Data from INMET, Belo Horizonte, Minas Gerais, Brazil). 729 Acta bot. Discussion On the other hand, even though senescence and leaf fall have been recorded for A. spruceanum all year long (Fig. 17), continuous leaf sprouting (Fig. 13) compensates this lost, and did not result in signiﬁ cant reduction of total leaf area in the crown. New leaf production was distinct between the two Aspidosperma host species. While A. australe had a deﬁ nite period of leaf sprouting, A. spruceanum produced leaves almost continuously all year long, except in November and December 2006, and July 2007 (Fig. 12, 13). Nevertheless, both species had mature leaves over the entire year, which were optional oviposition sites for both galling herbivores. These herbivores had great reproductive success and are multivoltines, with the presence of adult forms (mature and senescent gall phases) and oviposition (induction gall phases) in almost all the months. In fact, the two Aspidosperma species had a high index of gall infestation all year long (Fig. 10, 11), and gall induction occurred either in young or mature leaves (Fig. 12-15, 18, 19). Aspidosperma australe shows striking phenophases, and the gall inducing activity accomplishes the events of the plant life cycle. The correlation between the induction of galls and the presence of mature leaves all year long demonstrated the availability of reactive sites for gall inducing. On the other hand, A. spruceanum presented leaf sprouting and mature leaves all year long, with no expressive, distinct phenophases. The negative correlation between the induction phase of the galls and leaf senescence and abscission denoted the necessity of the galling insect for reactive sites for oviposition, and nutrient and water availability for the development of the gall, corroborating the nutritional hypothesis (Weiss et al. 1988; Bronner 1992). In the galls of A. australe, there was some correlation among plant phenophases and gall developmental phases. There was an increase in the percentage of gall induction in mature leaves (r = 0.562, p ≤ 0.05), galls in development phase concomitantly with mature leaves (r = 0.824, p ≤ 0.001), and gall eclosion occurred when senescent leaves were being shed (r = 0.958, p ≤ 0.001 and r = 0.821, p ≤ 0.001) (Tab. 1). Discussion Generally, in the tropics, insect populations decrease in the dry season. New life cycles restart at the beginning of the rainy season, followed by a gradual increment until the next dry season (Coley & Barone 1996; Bale et al. 2002). In the present study, galls in the developmental phase were recorded over the entire year, which apparently contradicts the general premise, but reinforces the strategy of using distinct reactive sites for gall induction all year long. g y g In A. australe, there was a low percentage of mature leaves from November 2006 to August 2007 as a consequence of no leaﬁ ng, and senescence and abscission slowness, which also resulted in a reduction of total leaf area during the dry season. On the other hand, continuous leaf sprouting in A. spruceanum compensates leaf senescence, and did not result in signiﬁ cant reduction of total leaf area. These phenological events are generally related to seasonal variation in plant water status, mainly inﬂ uenced by the availability of water in the soil, as well as by the capacity of water storage in plant tissues, as pointed out by Borchert (1994). Nevertheless, leaf area maintenance in the two species studied, even in the dry season, indicated that, even after three months without precipitation, there was no signiﬁ cant limitation in the availability of water in the soil for these plants, similarly to the observation of Lemos-Filho & Mendonça-Filho (2000) for three woody legumes in the Atlantic Forest. Although maintenance of leaf area may also be related to an increase in transpiration during the dry season, A. australe or A. spruceanum presented a reduction in stomatal conductance at times of higher air vapor pressure deﬁ cit (Lemos-Filho et al. 2007). For the Phacopteronidae, maximum induction activity occurred during leaf sprouting, while for the Cecidomyiidae, induction activity occurred over the entire year. In these two species, the availability of induction sites distinctly inﬂ uenced the behavior of galling herbivores. At the end of the rainy season, a low percentage of mature leaves on A. australe occurred, as a consequence of an increase in leaf senescence and abscission (Fig. 16). The lack of new leaf production from November 2006 to August 2007, as well as the slow senescence and abscission patterns, resulted in a reduction of total leaf area during the dry season (32.5%). Materials and methods bras. 24(3): 727-733. 2010. Figures 2-9. Developmental phases of galls induced by Phacopteronidae in Aspidosperma australe (2, 4, 5 and 8), and by Cecidomyiidae in Aspidosperma spruc anum (3, 6, 7 and 9). 2-3. Galls in induction phase. 4-7. Galls in developmental phase. 8-9. Senescent galls. The arrows indicate the corresponding developmen phase. Bars = 1 mm. Figures 2-9. Developmental phases of galls induced by Phacopteronidae in Aspidosperma australe (2, 4, 5 and 8), and by Cecidomyiidae in Aspidosperma spruce- anum (3, 6, 7 and 9). 2-3. Galls in induction phase. 4-7. Galls in developmental phase. 8-9. Senescent galls. The arrows indicate the corresponding developmental phase. Bars = 1 mm. Figures 10-11. Percentage of galled leaves of Aspidosperma austral and A. spruceanum from September 2006 to September 2007. Figures 10-11. Percentage of galled leaves of Aspidosperma austral and A. spruceanum from September 2006 to September 2007. Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... 730 Campos et al.: Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... Figures 12-17. Phenological events in Aspidosperma australe and A. spruceanum according to the activity index (--■--), and Fournier percentage of intensity (—▲—) (n=10). 12-13. Leaf sprouting. 14-15. Mature leaves. 16-17. Senescence. Figures 12-17. Phenological events in Aspidosperma australe and A. spruceanum according to the activity index (--■--), and Fournier percentage of intensity (—▲—) (n=10). 12-13. Leaf sprouting. 14-15. Mature leaves. 16-17. Senescence. Figures 18-19. Percentage of galls in induction (□), development (■) and senescent (■) phases (n=10) in Aspidosperma australe and A. spruceanum, from September 2006 to September 2007. Figures 18-19. Percentage of galls in induction (□), development (■) and senescent (■) phases (n=10) in Aspidosperma australe and A. spruceanum, from September 2006 to September 2007. 731 Acta bot. bras. 24(3): 727-733. 2010. induced galls in A. spruceanum during two periods, namely, in November-December 2006, and in April-May 2007. The former corresponded to the vegetative phenological phase when no leaf sprouting occurred (Fig. 13), and the latter coincided with the presence of many young leaves which had been permanently produced since January 2007. The end of the gall cycles, i.e., gall senescence, was recorded over most of the year, except in September, October, and December, 2006 for A. australe (Fig. 18), while for A. spruceanum, the Cecidomyiidae ecloded from their galls all year long. Acknowledgements p g p In both Aspidosperma species, gall induction occurred either in young or mature leaves, a behavior previously recorded for tropical galls of Ficus microcarpa induced by Gynaikothrips ﬁ corum (Souza et al. 2000), for unidentiﬁ ed species of galling herbivores in Baccharis concina, B. dracunculifolia (Arduin & Kraus 2001), and in Copaifera lansgdorfﬁ i (Oliveira & Isaias 2009). Even though galls should generally be induced in young tissues (Rohfritsch 1992), the galling inducers of Aspidosperma species trigger galling stimuli in differentiated cells. This behavior generates a high speciﬁ city to the host plants in situations where no responsive tissues are available; the galling herbivores may use alternative inducing sites, the mature tissues, as observed by Oliveira & Isaias (2009). Due to the short duration of the adult phase of galling insects, the females that do not locate susceptible structures must deposit their eggs at non- preferential sites (Weis et al. 1988; Eliason & Potter 2000). The synchronization of the herbivore life cycle with phenological phases of the host plant determines the quality and quantity of feeding resources, which are critical for galling insects (Yukawa 2000). As pointed out by Mendonça Jr. (2001), the most important step in the life history of galling insects is the discovery of adequate reactive sites for gall induction. In A. australe and A. spruceanum, the analyses of vegetative phenological phases revealed a continuous availability of responsive tissues. The responsiveness of either mature or immature leaf tissues was evidenced by signals of developing galls during the one-year period analyzed. It was evident that the galling herbivores of these two Aspidosperma species might use mature tissues as alternative oviposition sites. These facts explained the high level of infestation during the year, and denoted that there was a perfect adjustment between the resources of Aspidosperma host plants and the life cycles In both Aspidosperma species, gall induction occurred either in young or mature leaves, a behavior previously recorded for tropical galls of Ficus microcarpa induced by Gynaikothrips ﬁ corum (Souza et al. 2000), for unidentiﬁ ed species of galling herbivores in Baccharis concina, B. dracunculifolia (Arduin & Kraus 2001), and in Copaifera lansgdorfﬁ i (Oliveira & Isaias 2009). Even though galls should generally be induced in young tissues (Rohfritsch 1992), the galling inducers of Aspidosperma species trigger galling stimuli in differentiated cells. Discussion Negative correlations were observed between gall induction and senescent leaves (r = -0.868, p ≤ 0.001 and r = -0.783, p ≤ 0.01), between gall development and leaf abscission (r = -0.629, p ≤ 0.05), and between gall eclosion and the percentage of mature leaves (r = -0.661, p ≤ 0.05). In A. spruceanum, there was no signiﬁ cant correlation among host plant phenophases and developmental phases of the gall, except between gall eclosion and leaf senescence (r = 0.656, p ≤ 0.05). Even though gall senescence was recorded during most of the year, either for A. australe or A. spruceanum, this data should be analyzed with care because senescent galls should have been counted more than once, and consequently were overestimated. In the current study, both species of galling herbivores are multivoltines, which seems to be common at subtropical and tropical latitudes, where insect life cycles are often synchronized with host plant phenology (Burckhardt 2005). Even with signiﬁ cant leaf fall, mature leaf maintenance all year long made it possible to classify these species as evergreen, according to the proposal of Duff et al. (1997). Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... 732 732 Campos et al.: Phenological relationships between two insect galls and their host plants: Aspidosperma australe and A. spruceanum... Table 1. Spearman correlation among plant phenophases (leaf sprouting, mature and senescent leaves, and leaf abscission) and phases of gall development (indu tion, development and post-eclosion) in Aspidosperma australe and A, spruceanum, R-values with an asterisk are signiﬁ cant at * P≤0,05; P≤0,01; * P≤0,00 Table 1. Spearman correlation among plant phenophases (leaf sprouting, mature and senescent leaves, and leaf abscission) and phases of gall development (induc- tion, development and post-eclosion) in Aspidosperma australe and A, spruceanum, R-values with an asterisk are signiﬁ cant at * P≤0,05; P≤0,01; * P≤0,001. Leaf sprouting (%) Mature leaves (%) Senescent leaves (%) Leaf abscission(%) Aspidosperma austral Induction phase (%) 0,338 0,562* -0,868* -0,783 Development phase (%) -0,308 0,824*** -0,450 -0,629* Post-eclosion (%) -0,382 -0,661* 0,958* 0,821* Aspidosperma spruceanum Induction phase (%) -0,091 -0,199 0,270 0,039 Development phase (%) -0,081 0,426 -0,476 -0,105 Post-eclosion (%) 0,018 -0,512 0,656* 0,086 of their associated herbivores, which is responsible for the multivoltinism established in these two systems. References Arduin, M. & Kraus, J.E. 2001. Anatomia de galhas de ambrosia em folhas de Baccharis conccina e Baccharis dracunculifolia (Asteraceae). g de Baccharis conccina e Baccharis dracunculifolia (Asteraceae). Revista Brasileira de Botânica 24: 63-72. de Baccharis conccina e Baccharis dracunculifolia (Asteraceae). Revista Brasileira de Botânica 24: 63-72. Revista Brasileira de Botânica 24: 63-72. Bale, J.S.; Marsters, G.J.; Hodkinson, I.D.; Awmack, C.; Bezemer, M.; Brown, V.; Butterﬁ eld, J.; Buse, A.; Coulson, J.C.; Farrar, J.; Good, J.E.G.; Harrington, R.; Hartley, S.; Jones, T.H.; Lindroth, R.L.; Press, M.C.; Symrnioudis, I.; Watt, A.D. & Writtaker, J.B. 2002. Herbivory in global climate change research: direct effects of rising temperature on insect herbivores. Global Change Biology 8: 1-16Bencke, C.S.C. & Morellato, P.C. 2002. Comparação de dois métodos de avaliação da fenologia de plantas, sua interpretação e representação. Revista Brasileira de Botânica 25: 269-275. Borchert, R. 1994. Soil and stem water storage determine phenology and distribution of tropical dry forest trees. Ecology 75: 1437-1449. Bronner, R. 1992. The role of nutritive cells in the nutrition of cynipids and cecidomyiids. Pp. 118-140. In: Shorthouse, J.D. & Rohfritsch, O. (eds.). Biology of insect induced galls. Oxford University Press, New York. The synchronization of the herbivore life cycle with phenological phases of the host plant determines the quality and quantity of feeding resources, which are critical for galling insects (Yukawa 2000). As pointed out by Mendonça Jr. (2001), the most important step in the life history of galling insects is the discovery of adequate reactive sites for gall induction. In A. australe and A. spruceanum, the analyses of vegetative phenological phases revealed a continuous availability of responsive tissues. The responsiveness of either mature or immature leaf tissues was evidenced by signals of developing galls during the one-year period analyzed. It was evident that the galling herbivores of these two Aspidosperma species might use mature tissues as alternative oviposition sites. These facts explained the high level of infestation during the year, and denoted that there was a perfect adjustment between the resources of Aspidosperma host plants and the life cycles Burckhardt, D. 2005. Biology, ecology, and evolution of gall-inducing Psyllids (Hemiptera: Psylloidea). Pp. 243-258. In: Shaefer, C.W.& Withers, T.M. (eds.). Biology, ecology, and evolution of gall-inducing arthropods. Enﬁ eld, Science Publishers. Christiano, J.C.S. 2000. Respostas cecidogênicas no modelo Aspidosperma australe (Apocynaceae) – Pseudophacopteron sp. (Homoptera: Phacopteronidae). Dissertação do Mestrado. Universidade de São Paulo, São Paulo. Coley, P.D. & Barone, J.A. Discussion On the other hand, in temperate and subarctic or subantarctic regions many species are univoltine with winter dormancy or inactivity. Overwintering occurs as egg, larva, or adult depending on the species. Acknowledgements This behavior generates a high speciﬁ city to the host plants in situations where no responsive tissues are available; the galling herbivores may use alternative inducing sites, the mature tissues, as observed by Oliveira & Isaias (2009). Due to the short duration of the adult phase of galling insects, the females that do not locate susceptible structures must deposit their eggs at non- preferential sites (Weis et al. 1988; Eliason & Potter 2000). The authors are gratefully thankful to G.W. Fernandes and D. Burckhardt for the insect identiﬁ cation, and FAPEMIG for ﬁ nancial support (Proc. CRA-465/04). J.P. Lemos-Filho and Rosy M. S. Isaias also thank CNPq. References 1996. Herbivory and plant defenses in tropical forests. Annual Review of Ecology and Systematics 27: 305-335. Duff, G.A.; Myers, B.A.; Willians, R.J.; Eamus, D.; O`Grady, A. & Fordyce, I.R. 1997. Seasonal patterns in soil moisture, vapor pressure deﬁ cit, tree canopy cover and pre-dawn water potential in a Northern Australia savannah. Australian Journal of Botany 45: 211-224. Eliason, E.A. & Potter, D.A. 2000. Budburst phenology, plant vigor, and host genotype effects on the leaf-galling generation in Callirhytis cornigera (Hymenoptera: Cynipidae) on pin oak. Environmental Entomology 29: 1999-1207. 733 Acta bot. bras. 24(3): 727-733. 2010. Oliveira D.C.; Magalhães, T.A.; Carneiro, R.G.S.; Alvim, M.N. & Isaias, R.M.S. 2010. Do Cecidomyiidae galls of Aspidosperma spruceanum (Apocynaceae) ﬁ t the pre-established cytological and histochemical patterns? Protoplasma 242: 81-93. Formiga, A.T.; Gonçalves, S.J.M.R.; Soares, G.L.G. & Isaias, R.M.S. 2009. Relação entre o teor de fenóis totais e o cilco das galhas de Cecidomyiidae em Aspidosperma spruceanum Müll. Arg. (Apocynaceae). Acta Botanica Brasilica 23: 93-99. Fournier, L.A. 1974. Um método cuantitativo para la medición de características fenológicas em árbores. Turrialba 24: 422-423. Price, P.W. 1984. Insect ecology. New York, Academic Press. Rathcke, B. & Lacey, E.P. 1985. Phenological patterns of terrestrial plants. Annual Review of Ecology and Systematics 16: 179–214. Rathcke, B. & Lacey, E.P. 1985. Phenological patterns of terrestrial plants. A l R i f E l d S t ti 16 179 214 Lemos-Filho, J.P. & Mendonça-Filho, C.V. 2000. Seasonal changes in the water status of three woody legumes from the Atlantic forest, Caratinga, Brazil. Journal of Tropical Ecology 16: 21-32. Annual Review of Ecology and Systematics 16: 179–214 Rohfritsch, O. 1992. Patterns in gall development. Pp. 60-86. In: Shorthouse, J.D. & Rohfritsch, O. (eds.). Biology of insect induced galls. New York, Oxford University Press. Lemos-Filho, J.P.; Christiano, J.C.S. & Isaias, R.M.S. 2007. Efeitos da infestação de insetos galhadores na condutância e taxa relativa de transporte de elétrons em folhas de Aspidosperma australe Müell. Arg. e de A. spruceanum Benth. ex Müell. Revista Brasileira de Biociências 5: 1152-1154. Souza, S.C.P.M.; Kraus, J.E.; Isaias, R.M.S. & Neves, L.J. 2000. Anatomical and ultrastructural aspects of leaf galls in Ficus microcarpa L.F. (Moraceae) induced by Gynaikothrips ﬁ corum Marchal (Thysanoptera). Acta Botanica Brasilica 14: 57-69. Acta Botanica Brasilica 14: 57-69. Acta Botanica Brasilica 14: 57-69. Acta Botanica Brasilica 14: 57-69. Lieth, H. 1974. Purpose of a phenology book. Pp. 3-19. In: Lieth, H. (ed.). Phenology and seasonality modeling. Berlin, Springer. van Schaik, C.P.; Terbough, J.W. & Wright, S.J. 1993. The phenology of tropical forests: adaptive signiﬁ cance and consequences for primary van Schaik, C.P.; Terbough, J.W. & Wright, S.J. 1993. The phenology of tropical forests: adaptive signiﬁ cance and consequences for primary consumers. Annual Review of Ecology and Systematics 24: 353-377. tropical forests: adaptive signiﬁ cance and consequences for primary consumers. Annual Review of Ecology and Systematics 24: 353-377. Mendonça Jr., M.S. 2001. Galling insect diversity patterns: the resource synchronization hypothesis. Oikos 95: 171-176. Weis, A.E.; Walton, R. & Crego, C.L. 1988. Reactive plant tissue sites and the population biology of gall makers. Annual Review of Entomology 33: 467-486. y yp Motulsky, H. 1992-2009. Analyzing data with Graph Pad Prism software. San Diego. GraphPad Software Inc. Oliveira, D.C. & Isaias, R.M.S. 2009. Inﬂ uence of leaﬂ et age in anatomy and possible adaptive values of the midrib gall of Co paifera langsdorfﬁ i (Fabaceae: Caesalpinioideae). Revista de Biologia Tropical 57: 293-302. Wright, S.J. & van Schaik, C.P. 1994. Light and the phenology of tropical trees. American Naturalist 143: 102-199. Yamazaki, K. & Ohsaki, N. 2006. Willow leaf traits affecting host use by lead-gall-forming sawﬂ y. Population Ecology 48: 363-371. Oliveira, D.C. & Isaias, R.M.S. 2010. Cytological and histochemical gradients induced by a sucking insect in galls of Aspidosperma australe Arg. Muell (Apocynaceae). Plant Science 178: 350-358. Yukawa, J. 2000. Synchronization of gallers with host plant phenology. Population Ecology 42: 105-113. Versão eletrônica do artigo em www.scielo.br/abb e htpp://www.botanica.org.br/acta/ojs Versão eletrônica do artigo em www.scielo.br/abb e htpp://www.botanica.org.br/acta/ojs
https://openalex.org/W4292714700	https://www.aanda.org/10.1051/0004-6361/202141917/pdf	English	null	ConKer: An algorithm for evaluating correlations of arbitrary order	Astronomy & astrophysics	2,022	cc-by	12,637	1. Introduction CenterFinder algorithm (Brown et al. 2021), designed to ﬁnd locations in space likely to be the centers of the baryon acous- tic oscillations (BAOs). CenterFinder counts the number of galaxies removed from a particular location by a given distance by convolving spherical kernels with the matter distribution. ConKer uses the same functionality to evaluate npcfs. A sim- ilar approach was suggested in Slepian & Eisenstein (2016), and developed to evaluate the Legendre expansion of the 3pcf for continuous matter tracers in Portillo et al. (2018). In addi- tion to implementing this approach for higher order correlations, ConKer introduces spatial partitioning deﬁned with respect to the light of sight (LOS). This partitioning minimizes memory usage, allows for parallel computing, and enables an easy calcu- lation of npcfs in the µ-slices, where the angle θ in the deﬁnition of µ = cos θ is measured with respect to the LOS. Understanding the dynamics of inﬂation in the early universe is linked with the study of primordial density ﬂuctuations, in par- ticular with their deviations from a Gaussian distribution (see, e.g., Maldacena 2003; Bartolo et al. 2004; Acquaviva et al. 2003). High order correlations have been shown to be sensitive to non-Gaussian density ﬂuctuations (see Meerburg et al. 2019 and references within). However, a brute force approach leads to pro- hibitively expensive O(Nn) computations of correlations, where N is the number of tracer objects and n is the correlation order. This problem has been studied extensively (March 2013). Sev- eral approaches to mitigate it were suggested for the calculation of three-point correlations; for example, Yuan et al. (2018) used the small angle assumption, while Slepian & Eisenstein (2015) evaluated the Legendre expansion of the three-point correlation function (3pcf). The second approach was recently generalized for n-point correlation functions (npcfs) in Philcox et al. (2021) resulting in an O(N2) algorithm. ConKer is applicable to discrete matter tracers, such as galax- ies, and to continuous tracers, such as Lyman-α and 21 cm line intensity, or matter maps derived from weak lenses. The method can be applied to evaluate autocorrelations, and cross- correlations between different matter tracers. Here we present an alternative and computationally efﬁcient way of evaluating these correlations. Similarly to March (2013), here this algorithm exploits spatial proximity, and similarly to Zhang & Yu (2011) and Slepian & Eisenstein (2016) it uses a fast Fourier transform (FFT) to speed up the calculation. 1 The python3 implementation of this algorithm can be downloaded at https://github.com/zbrown89/divide_conker ABSTRACT Context. High order correlations in the cosmic matter density have become increasingly valuable in cosmological analyses. However, computing these correlation functions is computationally expensive. Aims. We aim to circumvent these challenges by developing a new algorithm called ConKer for estimating correlation functions. Methods. This algorithm performs convolutions of matter distributions with spherical kernels using FFT. Since matter distributions and kernels are deﬁned on a grid, it results in some loss of accuracy in the distance and angle deﬁnitions. We study the algorithm setting at which these limitations become critical and suggest ways to minimize them. Results. ConKer is applied to the CMASS sample of the SDSS DR12 galaxy survey and corresponding mock catalogs, and is used to compute the correlation functions up to correlation order n = 5. We compare the n = 2 and n = 3 cases to traditional algorithms to verify the accuracy of the new algorithm. We perform a timing study of the algorithm and ﬁnd that three of the four distinct processes within the algorithm are nearly independent of the catalog size N, while one subdominant component scales as O(N). The dominant portion of the calculation has complexity of O(N4/3 c log Nc), where Nc is the of cells in a three-dimensional grid corresponding to the matter density. ﬁnd ConKer to be a fast and accurate method of probing high order correlations in the cosmic matter density, then tion to upcoming surveys of large-scale structure. Conclusions. We ﬁnd ConKer to be a fast and accurate method of probing high order correlations in the cosmic m discuss its application to upcoming surveys of large-scale structure. ds. cosmology: observations – large-scale structure of Universe – dark energy – dark matter – methods: statistical – Key words. cosmology: observations – large-scale structure of Universe – dark energy – dark matter – methods: sta inﬂation Received 30 July 2021 / Accepted 21 July 2022 Received 30 July 2021 / Accepted 21 July 2022 Astronomy & Astrophysics Astronomy & Astrophysics A&A 667, A129 (2022) https://doi.org/10.1051/0004-6361/202141917 © Z. Brown et al. 2022 Astronomy & Astrophysics A&A 667, A129 (2022) https://doi.org/10.1051/0004-6361/202141917 © Z. Brown et al. 2022 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication. 1. Introduction These char- acteristics combined with implementation facilities help achieve a notable reduction in computational time and complexity. A129, page 1 of 12 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication. ConKer: An algorithm for evaluating correlations of arbitrary order Z. Brown , G. Mishtaku, and R. Demina nt of Physics and Astronomy, University of Rochester, 500 Joseph C. Wilson Boulevard, Rochester, NY 14627, USA brown5@ur.rochester.edu Department of Physics and Astronomy, University of Rochester, 500 Joseph C. Wilson Boulevard, Rochester, NY 14 e-mail: zbrown5@ur.rochester.edu Received 30 July 2021 / Accepted 21 July 2022 2.1. Strategy The values of mi are scanned from −li to +li, resulting in 2(2l + 1) con- volutions for each value of l, since both real and imaginary parts of Ylm must be used. Fig. 2. Cartoon illustrating a three-point correlation (0 −1 −2) with dif- ferent scales s1 on the side (0 −1) and s2 on the side (0 −2). Integration over spherical shells K1 and K2 is equivalent to counting all possible triangles that have one vertex at point 0 and the other two anywhere on the spheres. For the calculation of ξiso 3 (s1, s2), the kernels are uniformly populated, while for the evaluation of the Legendre expansion coefﬁ- cients ξl1,l2 3 (s1, s2) they are populated with the values of Ylimi (i = 1, 2). The values of mi are scanned from −li to +li, resulting in 2(2l + 1) con- volutions for each value of l, since both real and imaginary parts of Ylm must be used. Fig. 2. Cartoon illustrating a three-point correlation (0 −1 −2) with dif- ferent scales s1 on the side (0 −1) and s2 on the side (0 −2). Integration over spherical shells K1 and K2 is equivalent to counting all possible triangles that have one vertex at point 0 and the other two anywhere on the spheres. For the calculation of ξiso 3 (s1, s2), the kernels are uniformly populated, while for the evaluation of the Legendre expansion coefﬁ- cients ξl1,l2 3 (s1, s2) they are populated with the values of Ylimi (i = 1, 2). The values of mi are scanned from −li to +li, resulting in 2(2l + 1) con- volutions for each value of l, since both real and imaginary parts of Ylm must be used. as n-pletes. We consider all possible n-pletes with one vertex at point 0, characterized by a vector r with the other vertices deﬁned by vectors ri, (i = 1, ...n −1). For each point we deﬁne a vector connecting it with point 0: si = ri −r. We refer to a unit vector corresponding to any vector u as ˆu. We let ρ(r) be the density of the matter tracer (e.g., galaxy count per unit volume) at a location r, with ¯ρ(r) being the density of expected observations from tracers randomly distributed over the surveyed volume. 2.1.1. Isotropic npcf ξn(s1, ...s(n−1)) = X L ξL n (s1, ...s(n−1))ΠL(cos θi). (4) (4) The isotropic (averaged over the orientations of si) n-point correlation function is deﬁned as The coefﬁcients ξL n in this expansion are functions of the dis- tances si, but not the angles θi. ξiso n (s1, ..s(n−1)) = 1 R0n Z ∆(r)∆(r1)..∆(r(n−1))dˆs1..dˆs(n−1)dr, (2) Following the example of Slepian & Eisenstein (2015) and its generalization (Philcox et al. 2021), we use the spherical harmonic addition theorem: where the integration over r implies all possible positions of point 0 in the surveyed volume. The normalization R0 n is eval- uated by performing the same integration over the random distribution of tracers: Pl(cos θi) = 4π 2l + 1 lX m=−l Y∗ lm(ˆr)Ylm( ˆsi). (5) (5) R0 n = Z ¯ρ(r)¯ρ(r1)..¯ρ(r(n−1))dˆs1..dˆs(n−1)dr. (3) (3) The evaluation of ξL n is then reduced to The evaluation of ξL n is then reduced to ConKer computes the integral of the matter density ﬁeld over ˆsi by placing a spherical kernel Ki of radius si on point 0 and taking its inner product with the ∆-ﬁeld. For correlation order n, this implies a convolution of n −1 kernels. This is illustrated in Fig. 2 for the three-point correlation. The sum of the product of the kernel with the matter density ﬁeld is computed for a given location of point 0. Then the kernel is moved to a different loca- tion, thus scanning the entire surveyed volume and creating n −1 maps of convolutions of kernels with the ∆-ﬁeld. After this, the products of these maps and the original map of density ﬂuctu- ations ∆(r) are summed over to produce the integral over r in Eq. (2). To evaluate R0 n the same procedure is performed on the random distribution of tracers. ξL n (s1, ...s(n−1)) = 1 R0n Z dr∆(r) X m1 .. X mn−1 CLMa∗ l1m1..al(n−1)m(n−1), (6 (6) where coupling coefﬁcients CLM with M = (m1, m2, ...m(n−1)) are deﬁned in terms of Wigner 3-j symbols (see Appendix A). The values of mi are scanned from −li to +li. The calculation of the coefﬁcients, alm(r, si) = Z ∆(r + si)Ylm( ˆsi)d ˆsi, (7) (7) 2.1. Strategy We deﬁne the deviation from the expected density as L = (l1, l2, ...l(n−1)) represents orders in the Legendre expansion. The angular dependence of the npcf can be characterized via a decomposition in this basis: ∆(r) = ρ(r) −¯ρ(r). (1) ∆(r) = ρ(r) −¯ρ(r). (1) 2.1. Strategy The two-point correlation can be visualized as an excess (or deﬁcit) of sticks of a given length over a random combination of two points distributed over space. The three-point correlation corresponds to an excess of triangles, the four-point correlation to an excess of pyramids (the four points do not necessarily lie in one plane), and so on (see Fig. 1). We refer to these ﬁgures p p y The developed algorithm convolves kernels with matter maps, hence it is named ConKer1. It is an extension of the 1 The python3 implementation of this algorithm can be downloaded at https://github.com/zbrown89/divide_conker A129 (2022) Fig. 2. Cartoon illustrating a three-point correlation (0 −1 −2) with dif- ferent scales s1 on the side (0 −1) and s2 on the side (0 −2). Integration over spherical shells K1 and K2 is equivalent to counting all possible triangles that have one vertex at point 0 and the other two anywhere on the spheres. For the calculation of ξiso 3 (s1, s2), the kernels are uniformly populated, while for the evaluation of the Legendre expansion coefﬁ- cients ξl1,l2 3 (s1, s2) they are populated with the values of Ylimi (i = 1, 2). The values of mi are scanned from −li to +li, resulting in 2(2l + 1) con- volutions for each value of l, since both real and imaginary parts of Ylm must be used. A&A 667, A129 (2022) Fig. 1. Cartoon illustrating the two-point correlation (0 −1), three-point correlation (0 −1 −2), four-point correlation (0 −1 −2 −3), and n-point correlation (0 −1 −2 −3 −... −(n −1)). Fig. 1. Cartoon illustrating the two-point correlation (0 −1), three-point correlation (0 −1 −2), four-point correlation (0 −1 −2 −3), and n-point correlation (0 −1 −2 −3 −... −(n −1)). Fig. 2. Cartoon illustrating a three-point correlation (0 −1 −2) with dif- ferent scales s1 on the side (0 −1) and s2 on the side (0 −2). Integration over spherical shells K1 and K2 is equivalent to counting all possible triangles that have one vertex at point 0 and the other two anywhere on the spheres. For the calculation of ξiso 3 (s1, s2), the kernels are uniformly populated, while for the evaluation of the Legendre expansion coefﬁ- cients ξl1,l2 3 (s1, s2) they are populated with the values of Ylimi (i = 1, 2). 2 Technically, in this procedure each galaxy is counted twice, once when it is in the center cell and once when it is in the spherical shell, resulting in the overall factor of 2. However, since the same is also true for the random galaxies used for normalization, this factor cancels out. 2.1.2. Legendre expansion implies an integration over all possible orientations of vector ˆsi. It is equivalent to convolving the matter density with a sphere of radius si centered on point 0 and populated with the values of Ylm. We let θi indicate the angle between a vector si and r (see Fig. 2). We deﬁne the basis as a product of the Legendre polynomi- als, ΠL(cos θi) = Pl1(cos θ1)Pl2(cos θ2)..Pl(n−1)(cos θ(n−1)), where A129, page 2 of 12 Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Each ( jk)th partition is populated by galaxies with angular coordinates within the following limits: Each ( jk)th partition is populated by galaxies with angular coordinates within the following limits: N(X, Y, Z) = D(X, Y, Z) −R(X, Y, Z). A similar ﬁeld NMP(X, Y, Z) is deﬁned using the extended grids. A similar ﬁeld NMP(X, Y, Z) is deﬁned using the extended grids. The default mass assignment scheme used in ConKer is a three-dimensional histogram, or nearest grid point (NGP) method. However, Jing (2005) and Cui et al. (2008) showed that the galaxy power spectrum measured using FFT algorithms is sensitive to the choice of mass assignment scheme. The ConKer algorithm includes the option to use the cloud in cell (CIC) method when deﬁning the density ﬁelds. Of the two methods, CIC is more computationally expensive since it maps each tracer 2.1.3. Edge correction Having the same deﬁnition of the LOS for all the objects in the partition introduces some inaccuracy in angles, especially for objects near the boundary. However, the maximum deviation in the LOS deﬁnition is on the order of θ2 p/2. Hence, this inaccu- racy can be minimized by the proper choice of the partition size θp. where the matrix MLL′ is deﬁned as MLL′ = X K (−1) P l′ iGLKL′ fK MLL′ = X K (−1) P l′ iGLKL′ fK. (9) (9) For the deﬁnition of the Gaunt integral GLKL′, see Appendix A. For the deﬁnition of the Gaunt integral GLKL′, see Appendix A. In each partition we deﬁne a grid with spacing gS , such that the volume of each cubic grid cell is g3 S . The default value of gS is set to be equal to the bin width ∆s. However, the user can select a ﬁner resolution in the density ﬁeld and kernels. In this case, gS is set to a desired fraction of the radial bin size ∆s. During the ﬁnal steps of the algorithm, correlation functions are resampled to the appropriate s-bins. 2.2. Input The inputs to the algorithm are catalogs of the observed number count of tracers D, with the total number being ND, and R, which represents a number count of randomly distributed points within the same ﬁducial volume and corresponding number count NR. Most surveys provide the angular coordinates: right ascension α and declination δ, and the redshift z of each tracer. The rela- tionship between the redshift and the comoving radial distance is cosmology dependent: On the grid we deﬁne three-dimensional histograms, D(X, Y, Z) and R(X, Y, Z), which represent tracer counts in the cell (X, Y, Z) from data and random catalogs, respectively. These histograms may be populated by the raw count or the weighted count of tracers from the input catalogs. In every partition two additional grids are constructed, DMP(X, Y, Z) and RMP(X, Y, Z), which contain an extended map of objects within an additional θp in the declination direction and an additional θp/min(cos δk) in the right ascension direction of the LOS. During the convolution step the center of the ker- nel is placed on each cell of D(X, Y, Z) and R(X, Y, Z), while the convolution is performed with the extended maps DMP(X, Y, Z) and RMP(X, Y, Z). This procedure ensures that the entire survey region is covered, but that double-counting is avoided2. r(z) = c H0 Z z 0 dz′ p ΩM(z′ + 1)3 + Ωk(z′ + 1)2 + ΩΛ . (10) r(z) = c H0 Z z 0 dz′ p ΩM(z′ + 1)3 + Ωk(z′ + 1)2 + ΩΛ . (10) (10) Here ΩM, Ωk, and ΩΛ are the relative present-day matter, curva- ture, and cosmological constant densities, respectively; H0 is the present-day Hubble constant; and c is the speed of light. These user-deﬁned parameters represent the ﬁducial cosmology. In this study we used the following values for the cosmological parame- ters: c = 300 000 km s−1, H0 = 100h km s−1 Mpc−1, ΩM = 0.29, ΩΛ = 0.71, and Ωk = 0. The integral in Eq. (10) is evaluated numerically in ConKer. Cartesian coordinates of a tracer labeled (X,Y,Z) are evaluated based on r, α, and δ. Here ΩM, Ωk, and ΩΛ are the relative present-day matter, curva- ture, and cosmological constant densities, respectively; H0 is the present-day Hubble constant; and c is the speed of light. These user-deﬁned parameters represent the ﬁducial cosmology. 2.1.3. Edge correction Irregular survey boundaries and nonuniformities in the redshift selection function can introduce anisotropies in an otherwise isotropic distribution. The formalism to correct for the edge effects developed in Slepian & Eisenstein (2015) and Philcox et al. (2021) is ideal for implementation using the kernel con- volution functionality. The procedure involves the evaluation of Legendre moments of the random distribution f L n according to Eqs. (6) and (7) with the ¯ρ-ﬁeld used instead of ∆. In ConKer this is realized by convolving spherical kernels populated with the values of Ylm with the random distribution of tracers. α j−1 →α j = α j−1 + 2θp min(cos δk), (12) δk−1 →δk = δk−1 + 2θp. (13) (12) (13) Here min(cos δk) is the minimum value of cos δk in this partition. This factor is introduced for each region to have an approxi- mately square span of 2smax in the azimuthal and polar directions at the smallest comoving radius. The lowest boundaries are determined by the survey coverage. Here min(cos δk) is the minimum value of cos δk in this partition. This factor is introduced for each region to have an approxi- mately square span of 2smax in the azimuthal and polar directions at the smallest comoving radius. The lowest boundaries are determined by the survey coverage. A set of edge-corrected Legendre coefﬁcients ˆξLn are cal- culated based on uncorrected ξL n evaluated using Eq. (6) and coefﬁcients f L n by solving a system of linear equations The deﬁnition of the Cartesian (ˆx, ˆy, ˆz) system is unique to each partition with the ˆx-axis pointing to its center cell. The transformation from global sky to local Cartesian coordinates is given in Appendix B. ˆξLn = M−1 LL′ξL′ n , (8) ˆξLn = M−1 LL′ξL′ n , (8) The LOS in each partition is deﬁned as pointing along the ˆx-axis. Having the same deﬁnition of the LOS for all the objects in the partition introduces some inaccuracy in angles, especially for objects near the boundary. However, the maximum deviation in the LOS deﬁnition is on the order of θ2 p/2. Hence, this inaccu- racy can be minimized by the proper choice of the partition size θp. The LOS in each partition is deﬁned as pointing along the ˆx-axis. 2.2. Input In this study we used the following values for the cosmological parame- ters: c = 300 000 km s−1, H0 = 100h km s−1 Mpc−1, ΩM = 0.29, ΩΛ = 0.71, and Ωk = 0. The integral in Eq. (10) is evaluated numerically in ConKer. Cartesian coordinates of a tracer labeled (X,Y,Z) are evaluated based on r, α, and δ. A three-dimensional local density variation histogram N(X, Y, Z), which is a discretized representation of the ∆ﬁeld, is deﬁned on the grid to represent the difference in counts between D and R, normalized to ND: The algorithm computes correlation functions over a given range of scales or separation distances si. The range of distances from smin to smax is divided into Nb bins, which sets the bin width ∆s. (14) N(X, Y, Z) = D(X, Y, Z) −R(X, Y, Z). (14) 2.4. Kernels We construct spherical kernels Klm i (X, Y, Z) on a cube, just large enough to encompass a sphere of radius si + ∆s/2. The grid deﬁned on this cube has the same spacing gS as that used to construct the density ﬁelds. The cube is extended to have an odd number of cells in each coordinate direction. This ensures that it has a well-deﬁned center cell, the center of which deﬁnes the center of the kernel. We construct a spherical shell of the inner radius si −∆s/2 and the outer radius si + ∆s/2 centered on the center of the ker- nel. A numeric value is assigned to each kernel cell. If none of a cell’s volume falls within the shell, it is assigned 0. If, however, any part of the cell is within the shell, it is assigned the value that corresponds to the fraction of its volume within the shell. This is realized by considering a user deﬁned number of random positions within the each cell. We refer to this kernel as “ﬂat”. If the user has elected to employ a grid resolution gS ﬁner than the desired radial binning, an integer number of extra ker- nels are constructed between the limits of each s-bin, where the integer gives the rate of resampling. For example, to resam- ple correlation functions at twice the desired radial binning, 2gS = ∆s and a factor of two additional kernels are constructed for each s-bin. We construct a spherical shell of the inner radius si −∆s/2 and the outer radius si + ∆s/2 centered on the center of the ker- nel. A numeric value is assigned to each kernel cell. If none of a cell’s volume falls within the shell, it is assigned 0. If, however, any part of the cell is within the shell, it is assigned the value that corresponds to the fraction of its volume within the shell. This is realized by considering a user deﬁned number of random positions within the each cell. We refer to this kernel as “ﬂat”. If the user has elected to employ a grid resolution gS ﬁner than the desired radial binning, an integer number of extra ker- nels are constructed between the limits of each s-bin, where the integer gives the rate of resampling. 2.5. Convolution We construct a three-dimensional histogram Wlm i (X, Y, Z) with entries equal to the inner product of kernel Klm i centered on a cell with coordinates (X, Y, Z) and matter density variation ﬁeld NMP(X, Y, Z): 2.3. Partitioning and mapping The initial step of the algorithm divides data and random cata- logs into partitions based on the angular variables (α, δ). Each partition spans the entire range of redshifts from zmin to zmax corresponding to the co-moving radii rmin and rmax, evaluated according to Eq. (10). The angular size θp is determined by the angle subtended by smax at rmin: θp = smax rmin . θp = smax rmin . (11) (11) A129, page 3 of 12 A&A 667, A129 (2022) Fig. 3. Two-dimensional slices of spherical kernels, radius 108 h−1 Mpc, gS = 8 h−1 Mpc. The slices are transverse to the LOS direction moving outward (top to bottom). The value of kernel cells is shown for four conﬁgurations: l = 0, m = 0 (real); l = 2, m = 0 (real); l = 2, m = 1 (real); l = 2, m = 1 (imaginary). to multiple grid cells. The stage of the algorithm that places matter tracers in grid cells is referred to as mapping. 2.4. Kernels For example, to resam- ple correlation functions at twice the desired radial binning, 2gS = ∆s and a factor of two additional kernels are constructed for each s-bin. Fig. 3. Two-dimensional slices of spherical kernels, radius 108 h−1 Mpc, gS = 8 h−1 Mpc. The slices are transverse to the LOS direction moving outward (top to bottom). The value of kernel cells is shown for four conﬁgurations: l = 0, m = 0 (real); l = 2, m = 0 (real); l = 2, m = 1 (real); l = 2, m = 1 (imaginary). Fig. 4. Two-dimensional slices of spherical kernels, radius 108 h−1 Mpc, gS = 8 h−1 Mpc. The slices are along the LOS direction. The value of kernel cells is shown for three conﬁgurations corresponding to a com- plete wedge (0.0 < µ < 1.0), ξ⊥(0.0 < µ < 0.5), and ξ∥(0.5 < µ < 1.0). 2.4.1. Kernels for Legendre expansion We let lmax be the largest Legendre multipole probed in the eval- uation of the npcf. For each l from 0 to lmax, and each m from −l to +l, we compute the spherical harmonic Ylm(θK, φK), where θK is the polar angle with respect to the kernel’s ˆx-axis, and φK is the azimuthal angle. In each nonzero cell the kernel Klm i (X, Y, Z) is assigned values equal to the product of the ﬂat kernel with the average of Ylm, evaluated over the region of the kernel cell corresponding to the radial limits. The averaging is done during the same procedure of random sampling as used in the construc- tion of the ﬂat kernel. For cases when m , 0, real and imaginary kernels are constructed using ℜ[Ylm] and ℑ[Ylm]. Kernels of var- ious conﬁgurations are shown in slices along the LOS direction in Fig. 3. Fig. 4. Two-dimensional slices of spherical kernels, radius 108 h−1 Mpc, gS = 8 h−1 Mpc. The slices are along the LOS direction. The value of kernel cells is shown for three conﬁgurations corresponding to a com- plete wedge (0.0 < µ < 1.0), ξ⊥(0.0 < µ < 0.5), and ξ∥(0.5 < µ < 1.0). Blm i (X, Y, Z) ≡RMP(X, Y, Z) ∗∗∗Klm i (X, Y, Z). 2.4.2. Kernels for µ-slices The execution time of this step is about two orders of magnitude lower than the convolution step in the case of the 3pcf, and trivial in the case of the 2pcf. Only at very large corre- lation orders does the execution time of the summation process begin to approach that of convolution. At the summation stage, computing the npcf requires that the already computed grids deﬁned by s, l, and m be combined and summed over. The execution time of this step is about two orders of magnitude lower than the convolution step in the case of the 3pcf, and trivial in the case of the 2pcf. Only at very large corre- lation orders does the execution time of the summation process begin to approach that of convolution. For the calculation of the normalization factor R0 n, only convolu- tion with the isotropic kernel (l = 0, m = 0) is required. However, for the edge correction the convolution of Klm with the random catalog for all values of l and m is needed. Often a large ensem- ble of simulated catalogs, or mocks, each having the same survey boundaries is considered. In this case the convolution of all ker- nels with the random catalog needs only be performed once. This allows for a signiﬁcant reduction in computation time and disk- space allocation. In the case corresponding to the µ-slice kernels, the convolution step is nearly identical, replacing only Klm i with the appropriate kernel, Kµ,i, and repeating for each desired bin in µ. This stage of the algorithm is referred to as convolution. Of particular interest are correlations between objects with a deﬁned scale, such as those arising from spherical sound waves in the primordial plasma, also known as BAOs. In this case one tracer taken as a starting point is displaced from the other (n −1) points by the same distance s1 (e.g., the three-point correla- tion corresponds to isosceles triangles randomly distributed in space). This equidistant case corresponds to the diagonal of the n-point correlation function (hence referred to as the diagonal npcf) for l = 0, which is calculated as 2.6. Memory management The mapping and especially convolution steps produce multiple large three-dimensional grids, Wlm i and Blm i ; however, it is not feasible to store them in their entirety in the RAM of almost any computer. Thus, during the convolution step of the algorithm, ConKer writes each unique grid to a temporary ﬁle. The user can chose to save the entire grid as a three-dimensional array or to sparcify it. Sparciﬁcation is marginally more time consuming, but saves a considerable amount of disk space. ξ0,diag n (s) = P grid W0(W0 1)n−1 P grid B0(B0 1)n−1 . (21) (21) One of the advantages of the ConKer algorithm is that regard- less of the desired correlation order, n, no new convolution operations need take place. Thus, the time-consuming step is only performed once per catalog, which facilitates subsequent calculations of correlation functions to arbitrary order, n. During this step of the algorithm, referred to as ﬁle opera- tions, the grids are temporarily stored for the following summa- tion step. 2.7. Evaluation of n-point correlation functions A schematic ﬂowchart of the ConKer algorithm is shown in Fig. 5. Beginning with a deﬁnition of the cosmological param- eters and binning, the partitioning is performed on the random catalog. If that catalog has already been used, the previous par- titioning scheme is employed. During the convolution step, the user can choose whether or not to perform the convolution with the catalog of random tracers for edge correction. It is more time consuming, but only needs to be performed once per random cat- alog. During the summation step the user is able to compute an arbitrary number of correlation functions of a desired order n and lmax. Since ﬁles corresponding to the convolved grids are often large, the user can choose to delete them upon completion of the summation step. For each ( jk)th partition we concatenate every grid Wlm i and Blm i such that they now represent a convolution of the kernel with the entire density ﬁeld. A map of Wlm i represents discretized coef- ﬁcients alm i in Eq. (6). According to this equation to evaluate coefﬁcients ξL n these maps must be convolved with the mat- ter distribution ∆(r), a discretized representation of which is N(X, Y, Z), and summed over the entire grid. Hence, ξL n (s1, ...s(n−1)) = P grid W0 P m1 .. P m(n−1) CLMW∗l1m1 1 ..Wln−1mn−1 n−1 P grid B0B0 1..B0 n−1 , (17) (17) where W0 = N(X, Y, Z) and B0 = R(X, Y, Z). This step in the algorithm is referred to as summation. If the user wishes to employ the µ-wedge kernels, the pro- cedure is nearly identical; however, each step of the calculation must be repeated for each slice. This increases the relevant com- putational parameters such as run time and ﬁle sizes, but does not affect memory considerations as the calculations in each slice are performed independently. where W0 = N(X, Y, Z) and B0 = R(X, Y, Z). This step in the algorithm is referred to as summation. 2.7.1. Cases The speciﬁc case of lth multipole of the 2pcf is calculated as ξl 2(s) = 1 √ 2l + 1 P grid W0Wl0 1 P grid B0B0 1 . (18) 3 W 2.4.2. Kernels for µ-slices There is a class of problems (e.g., redshift distortion stud- ies) where it is important to evaluate npcfs in several µ-slices (Sánchez et al. 2017). The parameter µ = cos θ is deﬁned with respect to the LOS. The deﬁnition of LOS is the same through the entire partition, as discussed in Sect. 2.3, which introduces some inaccuracy for objects near the region’s boundary. For example, a maximum scanned distance of smax = 150 Mpc h−1 and a minimum redshift of zmin = 0.4 result in an upper bound- ary on the inaccuracy in the LOS deﬁnition of 0.8%. For interior objects and higher redshifts it is even smaller. The application of µ-slices is naturally implemented via a deﬁnition of kernel Kµ,i that is not populated over the entire sphere, but rather in a section corresponding to the desired range of µ, as shown in Fig. 4. They are normalized such that a complete µ-wedge kernel from µ = 0 to µ = 1 is identical to the l = 0 kernel used in the Legendre expansion. (15) Wlm i (X, Y, Z) ≡NMP(X, Y, Z) ∗∗∗Klm i (X, Y, Z). (15) Here ∗∗∗denotes a three-dimensional discrete convolution per- formed using FFT and Wlm i is a discretized representation of coefﬁcients alm calculated according to Eq. (7). The procedure is performed in each partition, for each bin in s and for all pos- sible values of l and m according to a given lmax, resulting in 2(lmax +1)2Nb convolutions. Once these maps of Wlm i (X, Y, Z) are created, they can be used to calculate the correlation functions of arbitrary order n. For normalization purposes we perform the same procedure on the ﬁeld of random counts. The result of the convolution of random density ﬁeld RMP with kernel Klm i is referred to as Blm i (X, Y, Z): (16) Blm i (X, Y, Z) ≡RMP(X, Y, Z) ∗∗∗Klm i (X, Y, Z). A129, page 4 of 12 A129, page 4 of 12 Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order At the summation stage, computing the npcf requires that the already computed grids deﬁned by s, l, and m be combined and summed over. 3.2. Comparison to existing methods The primary advantage of ConKer is in the behavior of the execution time as a function of the total number of objects N = ND + NR shown in top plot of Fig. 6 for the four stages of the algorithm: mapping, convolution, ﬁle operations, and sum- mation. The surveyed volume, V is kept ﬁxed. As expected, the execution time of convolution, ﬁle operations, and summation are nearly independent of N, and all three scale below O(N1/2). Mapping is an O(N) calculation, and only starts to dominate for catalogs signiﬁcantly larger than 100M objects. We present comparisons of the 2pcf and 3pcf evaluated using ConKer to well-established methods. In Cuesta et al. (2016) the monopole and quadrupole terms of the Landy and Szalay esti- mator of the 2pcf (Landy & Szalay 1993; Hamilton 1993) are computed for the combined SGC and NGC catalogs of the SDSS DR12 CMASS survey galaxies. We performed the calculation using ConKer with the same binning, and compared it to the Cuesta et al. (2016) results in Fig. 7. For the monopole and the quadrupole terms, we note good agreement between the two methods. At low scales (s < 30 h−1 Mpc), we ﬁnd the largest deviation between the two. The differences expectedly arise due to the discretization of the density ﬁeld and kernel in ConKer. The kernel represents a spherical shell of width gS mapped onto a three-dimensional Cartesian grid. Thus, once the kernel size becomes comparable with the grid spacing a resolution in the distance determination is degraded. This occurs, when the kernel size is less than approximately 5gS . This does not mean, how- ever, that we are unable to probe correlations at small scales. Instead, this simply requires a ﬁner grid, and resampling. By reducing the grid spacing (blue and red points in Fig. 7) we recovered the agreement down to lower scales. Based on this, we recommend setting the smin parameter larger than 5∆s if using the default sampling. Any differences in the size of the errors results from the fact that we use two separate mock ensembles to estimate the covariance. The main parameter that determines the computation time of ConKer is the grid spacing gS , which is set by default to be equal to the bin width ∆s. The maximum distance probed smax determines the number of steps in s: Nb = (smax −smin)/∆s. (18) (18) We evaluated the performance of ConKer using SDSS DR12 CMASS galaxies (Ross et al. 2017), their associated random catalogs, and an ensemble of MultiDark-Patchy mocks (Kitaura et al. 2016; Rodríguez-Torres et al. 2016). We applied ConKer to the SGC and NGC catalogs for data, randoms, and 20 mocks. For this study, which highlights the algorithm’s ability to probe cor- relations near the clustering and the BAO scales, we computed correlation functions for a distance range of 8–176 h−1 Mpc in 21 bins of width of 8 h−1 Mpc. In all cases, the standard system- atic (Ross et al. 2017) as well as FKP weights (Feldman et al. 1994) were used to create the density ﬁeld map. The default NGP mass assignment scheme was used and the grid spacing of gS = 8 h−1 Mpc unless speciﬁed otherwise. The lth multipole of the 3pcf is then calculated as The lth multipole of the 3pcf is then calculated as ξl 3(s1, s2) = 4π √ 2l + 1 P grid W0 P m W∗lm 1 Wlm 2 P grid B0B0 1B0 2 . (19) (19) This procedure is repeated on the random ﬁeld in order to con- struct the terms necessary for edge-correction (see Philcox et al. 2021): This procedure is repeated on the random ﬁeld in order to con- struct the terms necessary for edge-correction (see Philcox et al. 2021): f l 3(s1, s2) = 4π √ 2l + 1 P grid B0 P m B∗lm 1 Blm∗ 2 P grid B0B0 1B0 2 . (20) (20) A129, page 5 of 12 A129, page 5 of 12 A&A 667, A129 (2022) Fig. 5. ConKer ﬂowchart. Parameters or settings chosen by the user are in blue, internal processes are in green, and decisions made by the user are in red. A&A 667, A129 (2022) Fig. 5. ConKer ﬂowchart. Parameters or settings chosen by the user are in blue, internal processes are in green, and decisions made by the user are in red. 3.1. Timing study The ﬁle operations step scales more favorably with gS ; how- ever, it dominates the execution time for gS above approximately 5 h−1 Mpc. We demonstrate the efﬁcient nature of our algorithm with the fol- lowing timing study, performed using a personal computer with a 10 CPU core Apple M1 Pro chip and 32 GB of memory. All execution times are in units of CPU seconds. A129, page 6 of 12 3.2. Comparison to existing methods Bottom: dependence on the grid resolution gS for 12.5M objects. The points represent the measured CPU time. The dashed lines are the results of the ﬁt to a power law, with the scaling given in the ﬁgure. The solid line (bottom) is the time of the convolution step ﬁt to N4/3 c log Nc. Fig. 7. A comparison of the 2pcf as measured by the ConKer algorithm versus existing clustering algorithms. Top: 2pcf monopole ξ0 2(s) com- puted using ConKer with three different values of grid spacing gS = 8 (grey), 4 (blue), 2 (red) h−1 Mpc, and results from Cuesta et al. (2016) (purple line). Finer grids were resampled to ∆s = 8 h−1 Mpc bins. The error bars on the ConKer points were extracted from the covariance matrix of the Patchy mock ensemble. The errors in Cuesta et al. (2016; shaded region) are from an ensemble of quick particle mesh mocks (White et al. 2014). The 2pcf is multiplied by s2 to emphasize features at large scales. The lower subpanel shows the residual between the two methods. Bottom: same as the top panel, but for the 2pcf quadrupole ξ2 2(s). Here, the 2pcf is multiplied by s instead of s2 by convention. the same behavior at small scales as in the case of the Legen- dre expansion, where the agreement is recovered by reducing the grid spacing. In Figs. 7 and 8 comparing the l = 0, 2 and ξ⊥,∥cases to exist- ing clustering algorithms, the uncertainties on the survey data measurements were derived from the mock ensemble. The size of the error bar corresponding to point-i is σi = √Cii, where C is the mock covariance matrix. More importantly, the distribution over the residuals is cen- tered at approximately zero, meaning our estimator is not biased compared to the nbodykit implementation. The largest devia- tions between the algorithms again arise at smaller scales where the kernel resolution is degraded. The 3pcf calculation using ConKer was faster by a factor of ∼3, and scales more favorably with N, since nbodykit is an O(N2) algorithm. The 3pcf algorithm implemented in nbodykit is based on the work of Slepian & Eisenstein (2015). Using the two meth- ods, we computed the edge-corrected 3pcf up to l = 3 of the same subsample of NGC galaxies used in the timing study (see caption of Fig. 6 for details). 3.2. Comparison to existing methods Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Fig. 6. Execution time of the four processes in the ConKer algorithm, as applied to subsamples of the SDSS DR12 CMASS NGC galaxies (150◦< αg < 210◦, 0◦< δg < 60◦). Each point represents a calculation of the 2pcf (l = 0, 2, 4), 3pcf (l = 0–5), and diagonal npcf up to n = 5. Top: dependence on the number of combined data and random objects, using a grid resolution gS = 8 h−1 Mpc. Bottom: dependence on the grid resolution gS for 12.5M objects. The points represent the measured CPU time. The dashed lines are the results of the ﬁt to a power law, with the scaling given in the ﬁgure. The solid line (bottom) is the time of the convolution step ﬁt to N4/3 c log Nc. the same behavior at small scales as in the case of the Legen- dre expansion, where the agreement is recovered by reducing the Fig. 7. A comparison of the 2pcf as measured by the ConKer algorithm versus existing clustering algorithms. Top: 2pcf monopole ξ0 2(s) com- puted using ConKer with three different values of grid spacing gS = 8 (grey), 4 (blue), 2 (red) h−1 Mpc, and results from Cuesta et al. (2016) (purple line). Finer grids were resampled to ∆s = 8 h−1 Mpc bins. The error bars on the ConKer points were extracted from the covariance matrix of the Patchy mock ensemble. The errors in Cuesta et al. (2016; shaded region) are from an ensemble of quick particle mesh mocks (White et al. 2014). The 2pcf is multiplied by s2 to emphasize features at large scales. The lower subpanel shows the residual between the two methods. Bottom: same as the top panel, but for the 2pcf quadrupole ξ2 2(s). Here, the 2pcf is multiplied by s instead of s2 by convention. Fig. 6. Execution time of the four processes in the ConKer algorithm, as applied to subsamples of the SDSS DR12 CMASS NGC galaxies (150◦< αg < 210◦, 0◦< δg < 60◦). Each point represents a calculation of the 2pcf (l = 0, 2, 4), 3pcf (l = 0–5), and diagonal npcf up to n = 5. Top: dependence on the number of combined data and random objects, using a grid resolution gS = 8 h−1 Mpc. 3.2. Comparison to existing methods The 3pcf, using both estimators, is shown in Fig. 9, as well as the distributions of the differ- ences between them. For this comparison, we computed 3pcf in 11 bins of width of 10 h−1 Mpc for s from 45 to 155 h−1 Mpc. Over this range of scales, we ﬁnd a good agreement between the two methods. By convention, the diagonal terms of the 3pcf are excluded in this comparison (see Slepian & Eisenstein 2015). 3.2. Comparison to existing methods The total number of grid cells Nc depends on the surveyed volume: Nc = V/g3 S . Mapping is independent of gS . For gS below approx- imately 5 h−1 Mpc, the dominant process is convolution of cubic volumes, on which the kernels are deﬁned, containing Nk cells: Nk = (s/gS )3 < (smax/gS )3. It is repeated Nc times with each grid cell being the center of the kernel. Since the convolution is performed using FFT with a typical complexity of N log N, the complexity of each step in s is O(Nc log Nk). Thus, the time complexity of the convolution is tconv ∝NbNc log Nk< smax gS V g3 S log Vk g3 S ∝g−4 S log gS ∝N4/3 c log Nc. (22) (22) In addition to the Legendre expansion, we also compute the 2pcf of the NGC sample in two µ-slices, corresponding to the transverse (ξ⊥) and parallel (ξ∥) cases. To compare, we repeated the calculation using nbodykit, an open source cosmology toolkit (Hand 2018). The results are shown in Fig. 8. We ﬁnd The observed scaling of the convolution step as g−4.3 S is in good agreement with this analytic prediction as depicted by the solid brown line in Fig. 6 (bottom). The observed scaling of the convolution step as g−4.3 S is in good agreement with this analytic prediction as depicted by the solid brown line in Fig. 6 (bottom). A129, page 6 of 12 r evaluating correlations of arbitrary order Fig. 7. A comparison of the 2pcf as measured by the ConKer algorithm versus existing clustering algorithms. Top: 2pcf monopole ξ0 2(s) com- puted using ConKer with three different values of grid spacing gS = 8 (grey), 4 (blue), 2 (red) h−1 Mpc, and results from Cuesta et al. (2016) (purple line). Finer grids were resampled to ∆s = 8 h−1 Mpc bins. The error bars on the ConKer points were extracted from the covariance matrix of the Patchy mock ensemble. The errors in Cuesta et al. (2016; shaded region) are from an ensemble of quick particle mesh mocks (White et al. 2014). The 2pcf is multiplied by s2 to emphasize features at large scales. The lower subpanel shows the residual between the two methods. Bottom: same as the top panel, but for the 2pcf quadrupole ξ2 2(s). Here, the 2pcf is multiplied by s instead of s2 by convention. Z. 3.3. ConKer npcf Top: edge-corrected 3pcf, ξl 3(s1, s2), computed using ConKer (left column) and nbodykit (right column) up to l = 3. The color scale is applicable to all panels. Bottom: residual between the two methods, shown as a distribution over all (s1, s2) pairs for l = 0 (blue), l = 1 (yellow), l = 2 (green), and l = 3 (red). combining this measure over the whole space of N objects allows us to calculate the space’s n-point correlation function. The algorithm exploits an intrinsic spatial proximity char Fig. 9. A comparison of the 3pcf estimated using ConKer compared to existing methods. Top: edge-corrected 3pcf, ξl 3(s1, s2), computed using ConKer (left column) and nbodykit (right column) up to l = 3. The color scale is applicable to all panels. Bottom: residual between the two methods, shown as a distribution over all (s1, s2) pairs for l = 0 (blue), l = 1 (yellow), l = 2 (green), and l = 3 (red). Fig. 8. Points showing the transverse ξ⊥(s) and parallel ξ∥(s) compo- nents of the 2pcf computed using ConKer with three values of grid spacing gS = 8 (red–yellow), 4 (purple–green), 2 (black–grey) h−1 Mpc. Solid lines in the corresponding colors show the results from the nbodykit 2pcf algorithm. This ﬁgure corresponds only to NGC galax- ies. The error bars and shaded regions are determined from the ensem- ble of Patchy mocks. The lower subpanels show the residual between the two methods. Fig. 8. Points showing the transverse ξ⊥(s) and parallel ξ∥(s) compo- nents of the 2pcf computed using ConKer with three values of grid spacing gS = 8 (red–yellow), 4 (purple–green), 2 (black–grey) h−1 Mpc. Solid lines in the corresponding colors show the results from the nbodykit 2pcf algorithm. This ﬁgure corresponds only to NGC galax- ies. The error bars and shaded regions are determined from the ensem- ble of Patchy mocks. The lower subpanels show the residual between the two methods. Fig. 8. Points showing the transverse ξ⊥(s) and parallel ξ∥(s) compo- nents of the 2pcf computed using ConKer with three values of grid spacing gS = 8 (red–yellow), 4 (purple–green), 2 (black–grey) h−1 Mpc. Solid lines in the corresponding colors show the results from the nbodykit 2pcf algorithm. This ﬁgure corresponds only to NGC galax- ies. The error bars and shaded regions are determined from the ensem- ble of Patchy mocks. 3.3. ConKer npcf The lower subpanels show the residual between the two methods. The diagonal elements of the npcfs are shown in Fig. 11 for n = 2, 3, 4, and 5. We observe the expected features of the npcfs. The diagonal elements of the npcfs are shown in Fig. 11 for n = 2, 3, 4, and 5. We observe the expected features of the npcfs. These include an increase in magnitude at small scales present for n = 2 and n = 3, corresponding to galaxy clustering, and a well-deﬁned “bump” at the BAO scale for all. The npcfs based on the random catalog ﬂuctuate about ξn = 0 at several orders of magnitude below the signal observed in mocks. In each case, the covariance matrix of the diagonal npcf is estimated on the set of 20 Patchy mocks using the following pro- cedure. For the ith bin of the npcf, we let the average value across the mock ensemble be deﬁned as < ξ >i. If the number of mocks in the ensemble is Nq, then the covariance matrix elements are Ci j = 1 Nq −1 X q ξ0,diag n (si)−< ξ >i ξ0,diag n (sj)−< ξ >j . (23 Fig. 9. A comparison of the 3pcf estimated using ConKer compared to existing methods. Top: edge-corrected 3pcf, ξl 3(s1, s2), computed using ConKer (left column) and nbodykit (right column) up to l = 3. The color scale is applicable to all panels. Bottom: residual between the two methods, shown as a distribution over all (s1, s2) pairs for l = 0 (blue), l = 1 (yellow), l = 2 (green), and l = 3 (red). (23) We show an example of the reduced covariance matrix for the n = 2, l = 0 diagonal case in Fig. 12. Off-diagonal elements of the edge-corrected 3pcf are shown in Fig. 13 for the NGC sample. Emphasizing for large-scale fea- tures, we observe strong indicators of BAO in both the l = 0 and l = 1 cases. combining this measure over the whole space of N objects allows us to calculate the space’s n-point correlation function. The algorithm exploits an intrinsic spatial proximity char- acteristic in the objective of querying structures of negligible dimensions in a much bigger space. 3.3. ConKer npcf This spatial proximity factor leads to space partitioning algorithms targeting a nearest neigh- bor query approach (see, e.g., the tree-based npcf algorithms in March 2013). However, ConKer uses this factor as a heuristic in limiting its query space immediately to only the deﬁned separa- tion for each point in the space. We note how this is in contrast with the former technique family. In a nearest neighbor approach to the npcf problem, the query space is grown at each point in the greater embedding space, aggregating the n-point statistic until 3.3. ConKer npcf We used ConKer to compute the diagonal elements of the npcf as deﬁned in Eq. (21) for n = 2, 3, 4, 5 and l = 0, and off-diagonal elements of the 3pcf for an ensemble of MultiDark- Patchy mocks. For this calculation, the NGC/SGC catalogs were divided into 27/14 partitions, as shown in Fig. 10. A129, page 7 of 12 A129, page 7 of 12 A&A 667, A129 (2022) Fig. 8. Points showing the transverse ξ⊥(s) and parallel ξ∥(s) compo- nents of the 2pcf computed using ConKer with three values of grid spacing gS = 8 (red–yellow), 4 (purple–green), 2 (black–grey) h−1 Mpc. Solid lines in the corresponding colors show the results from the nbodykit 2pcf algorithm. This ﬁgure corresponds only to NGC galax- ies. The error bars and shaded regions are determined from the ensem- ble of Patchy mocks. The lower subpanels show the residual between the two methods. The diagonal elements of the npcfs are shown in Fig. 11 for n = 2, 3, 4, and 5. We observe the expected features of the npcfs. These include an increase in magnitude at small scales present for n = 2 and n = 3, corresponding to galaxy clustering, and a well-deﬁned “bump” at the BAO scale for all. The npcfs based on the random catalog ﬂuctuate about ξn = 0 at several orders of magnitude below the signal observed in mocks. In each case, the covariance matrix of the diagonal npcf is estimated on the set of 20 Patchy mocks using the following pro- cedure. For the ith bin of the npcf, we let the average value across the mock ensemble be deﬁned as < ξ >i. If the number of mocks in the ensemble is Nq, then the covariance matrix elements are Ci j = 1 Nq −1 X q ξ0,diag n (si)−< ξ >i ξ0,diag n (sj)−< ξ >j . (23) We show an example of the reduced covariance matrix for the n = 2, l = 0 diagonal case in Fig. 12. Off-diagonal elements of the edge-corrected 3pcf are shown in Fig. 13 for the NGC sample. Emphasizing for large-scale fea- tures, we observe strong indicators of BAO in both the l = 0 and l = 1 cases. Fig. 9. A comparison of the 3pcf estimated using ConKer compared to existing methods. A129, page 8 of 12 Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Fig. 10. Angular footprint of the NGC (top) and SGC (bottom) galaxies (grey dots), divided into partitions (dashed red lines), each with a unique LOS (blue marker). The solid red line highlights one such partition. During the convolution step, the center of the kernel is positioned within these boundaries, while the convolution is performed over the region bound by the solid green line. Fig. 11. Diagonal npcf, ξn(s), from n = 2 to n = 5 (from top to bot- tom), for the combined SGC+NGC sample of the ensemble of MD Patchy mocks (green) and the associated randoms (red dashed line). The shaded green region around the mock average gives the errors com- puted from the diagonal elements of the mock co-variance matrix, C. The npcf is multiplied by a proxy for the total kernel volume, where ˜s = s/(100 h−1 Mpc). This convention highlights features at large scales. Fig. 10. Angular footprint of the NGC (top) and SGC (bottom) galaxies (grey dots), divided into partitions (dashed red lines), each with a unique LOS (blue marker). The solid red line highlights one such partition. During the convolution step, the center of the kernel is positioned within these boundaries, while the convolution is performed over the region bound by the solid green line. Fig. 10. Angular footprint of the NGC (top) and SGC (bottom) galaxies (grey dots), divided into partitions (dashed red lines), each with a unique LOS (blue marker). The solid red line highlights one such partition. During the convolution step, the center of the kernel is positioned within these boundaries, while the convolution is performed over the region bound by the solid green line. the greater space is fully queried, whereas ConKer aggregates the statistic over all embedding space, and then grows the query space before repeating. Fig. 11. Diagonal npcf, ξn(s), from n = 2 to n = 5 (from top to bot- tom), for the combined SGC+NGC sample of the ensemble of MD Patchy mocks (green) and the associated randoms (red dashed line). The shaded green region around the mock average gives the errors com- puted from the diagonal elements of the mock co-variance matrix, C. The npcf is multiplied by a proxy for the total kernel volume, where ˜s = s/(100 h−1 Mpc). This convention highlights features at large scales. Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order This design choice realized in ConKer’s core subroutine, a convolution of the query space with the embedding space per- formed by an FFT algorithm, distills the complexity from the brute force. This approach, combined with the heuristic above, lets the dominant components of ConKer achieve independence of the number of objects, as shown in Fig. 6. There is certainly a trade-off between the sparsity of the whole space and the bias toward linear complexity in number of objects, as expected from an FFT-based algorithm, but even for very dense catalogs, we expect the scaling in the number of objects to be capped by O(N), where N is the total number of objects. Ultimately, ConKer is a hybrid algorithm that draws from both computational geometry and signal processing to achieve linear complexity in the number of objects. Fig. 12. Reduced covariance matrix Cij/ pCiiC jj for the n = 2 diago- nal isotropic correlation function. It is computed from the ensemble of Patchy mocks. 4.1. Algorithmic framework The query for a pattern in matter distributions may prompt the employment of machine learning techniques. ConKer, being a spatial statistics algorithm, offers an alternative to such an approach that is fast and transparent. It exploits the fact that the full set of equidistant points from any given point makes a sphere, with its surface density being a direct measure of how spherically structured this subspace of points is. Aggregating and A129, page 8 of 12 Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order 4.2. ConKer versus other methods The idea of convolving spherically symmetric kernels with the density ﬁelds to evaluate the number of objects removed from a certain point by a given distance was originally proposed in Zhang & Yu (2011). In this work the Legendre expansion of npcf was not considered. In Slepian & Eisenstein (2015) and Philcox et al. (2021) a KdTree algorithm was used and the spherical function decomposition was evaluated for each galaxy pair, resulting in O(N2) calculation. This method works well for smaller scales or sparse surveys. In the same papers using FFT-based convolution for the Legendre expansion was also suggested. This approach has an advantage for denser surveys or continuous tracers since the computational time depends on the volume but not on density. The idea was later realized in Fig. 12. Reduced covariance matrix Cij/ pCiiC jj for the n = 2 diago- nal isotropic correlation function. It is computed from the ensemble of Patchy mocks. A129, page 9 of 12 A&A 667, A129 (2022) Fig. 14. Distribution, W0W1/⟨B0B1⟩at s1 = 108 h−1 Mpc, for the CMASS galaxy sample of the SDSS DR12 survey (black), one of the MD Patchy mocks (green), and the associated randoms (red). This plot includes NGC galaxies only. Fig. 13. Three-point correlation function ξl 3(s1, s2) of the NGC mock ensemble, computed up to l = 5. Multiplication by ˜s2 1 ˜s2 2 emphasizes the same large-scale features as in Fig. 11. The diagonal elements are excluded from these plots. Fig. 14. Distribution, W0W1/⟨B0B1⟩at s1 = 108 h−1 Mpc, for the CMASS galaxy sample of the SDSS DR12 survey (black), one of the MD Patchy mocks (green), and the associated randoms (red). This plot includes NGC galaxies only. normalized by the average B0B1 is presented in Fig. 14 for a kernel size of 108 h−1 Mpc for data, mock, and random catalogs. normalized by the average B0B1 is presented in Fig. 14 for a kernel size of 108 h−1 Mpc for data, mock, and random catalogs. Moreover, W0 and W1 as well as their product are maps. While in the npcf the location information is entirely lost, in the maps produced by ConKer it is preserved and can be used for cross-correlation studies between different tracers, such as weak lensing, Lyα, and CMB. Fig. 13. Three-point correlation function ξl 3(s1, s2) of the NGC mock ensemble, computed up to l = 5. Acknowledgements. The authors would like to thank Z. Slepian for his inter- est and insightful comments, F. Weisenhorn for assistance with the nbodykit 3pcf calculations, A. Ross for providing numeric values of SDSS DR12 CMASS 2pcf, and S. BenZvi, K. Douglass and S. Gontcho A Gontcho for useful dis- cussions. The authors acknowledge support from the US Department of Energy under the grant DE-SC0008475.0. Funding for SDSS-III has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, and the US Department of Energy Ofﬁce of Science. The SDSS-III web site is http://www.sdss3.org/. SDSS-III is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS-III Collaboration including the University of Arizona, the Brazilian Par- ticipation Group, Brookhaven National Laboratory, Carnegie Mellon University, University of Florida, the French Participation Group, the German Participation Group, Harvard University, the Instituto de Astroﬁsica de Canarias, the Michi- gan State/Notre Dame/JINA Participation Group, Johns Hopkins University, Lawrence Berkeley National Laboratory, Max Planck Institute for Astrophysics, 5. Conclusion ConKer extends the approach to all n > 1. ConKer computes the integral in Eq. (7) by convolving a spherical kernel Ki of radius si populated with the values of Ylm with the matter den- sity ﬁeld. The deﬁnition of a kernel on the grid necessarily leads to some loss of precision in the distance deﬁnition. This is par- tially mitigated in ConKer by weighting the grid cells with the fraction of the cell’s volume contained within a given spherical shell and averaging Ylm’s over this volume. Convolving the entire surveyed volume (as in Portillo et al. 2018) leads to large arrays that need to be stored resulting in signiﬁcant memory require- ments. In light of anticipated large volume surveys such as DESI (DESI Collaboration 2016), this limitation becomes particularly stringent. ConKer convolves a cubic volume just large enough to encompass a sphere of the speciﬁed radius, thus limiting the memory requirements for kernel storage. Additionally, ConKer introduces a partitioning scheme, as discussed in Sect. 2.3. As a result, the array size is limited by the partition’s volume. This scheme has an additional beneﬁt; it allows for the evaluation of npcfs in µ-slices as, discussed in Sect. 2.4.2, which is particu- larly relevant for parallel versus transverse to the LOS analysis. Finally, partitioning naturally allows for parallelized computing processes. We presented ConKer, an algorithm that convolves spherical kernels with matter maps allowing for fast evaluation of the n-point correlation functions, its expansion in Legendre polyno- mials, and its µ-slices. The algorithm can be broken into three stages: mapping, convolution, and summation. The execution time of convolution and summation are independent of the cat- alog size N, while mapping is a O(N) calculation, which starts dominating for catalogs larger than 100M objects. The dominant part of the convolution is with complexity O(N4/3 c log Nc), where Nc is the number of grid cells. A comparison to the standard techniques shows good agree- ment. We study the performance using SDSS DR12 CMASS galaxies, their associated random catalogs, and an ensemble of MultiDark-Patchy mocks. The results up to n = 5 are presented. Further metrics that may offer additional sensitivity to primor- dial non-Gaussianities are also suggested such as the distribution over weights Wi and their products. 4.2. ConKer versus other methods Multiplication by ˜s2 1 ˜s2 2 emphasizes the same large-scale features as in Fig. 11. The diagonal elements are excluded from these plots. Moreover, W0 and W1 as well as their product are maps. While in the npcf the location information is entirely lost, in the maps produced by ConKer it is preserved and can be used for cross-correlation studies between different tracers, such as weak lensing, Lyα, and CMB. (Portillo et al. 2018) for the evaluation of the 3pcf of the continuous tracer. Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Cui, W., Liu, L., Yang, X., et al. 2008, ApJ, 687, 738 Max Planck Institute for Extraterrestrial Physics, New Mexico State University, New York University, Ohio State University, Pennsylvania State University, Uni- versity of Portsmouth, Princeton University, the Spanish Participation Group, University of Tokyo, University of Utah, Vanderbilt University, University of Virginia, University of Washington, and Yale University. The massive produc- tion of all MultiDark-Patchy mocks for the BOSS Final Data Release has been performed at the BSC Marenostrum supercomputer, the Hydra cluster at the Instituto de Fısica Teorica UAM/CSIC, and NERSC at the Lawrence Berke- ley National Laboratory. We acknowledge support from the Spanish MICINNs Consolider-Ingenio 2010 Programme under grant MultiDark CSD2009-00064, MINECO Centro de Excelencia Severo Ochoa Programme under grant SEV- 2012-0249, and grant AYA2014-60641-C2-1-P. The MultiDark-Patchy mocks was an effort led from the IFT UAM-CSIC by F. Prada’s group (C.-H. Chuang, S. Rodriguez-Torres and C. Scoccola) in collaboration with C. Zhao (Tsinghua U.), F.-S. Kitaura (AIP), A. Klypin (NMSU), G. Yepes (UAM), and the BOSS galaxy clustering working group. Cui, W., Liu, L., Yang, X., et al. 2008, ApJ, 687, 738 DESI Collaboration (Aghamousa, A., et al.) 2016, The DESI Experiment Part I: Science,Targeting, and Survey Design [arXiv:1611.00036] Feldman, H. A., Kaiser, N., & Peacock, J. A. 1994, AJ, 426, 2 Hamilton, A. J. S. 1993, ApJ, 417, 19 Hand, N. E. A. 2018, AJ, 156, 4 Jing, Y. 2005, ApJ, 620, 559 Jing, Y. 2005, ApJ, 620, 559 Kitaura, F.-S., Rodríguez-Torres, S., Chuang, C.-H., et al. 2016, MNRAS, 456, 4156 Landy, S. D., & Szalay, A. S. 1993, ApJ, 412, 64 Landy, S. D., & Szalay, A. S. 1993, ApJ, 412, 64 Maldacena, J. M. 2003, JHEP, 05, 013 Landy, S. D., & Szalay, A. S. 1993, ApJ, 412, 64 Maldacena, J. M. 2003, JHEP, 05, 013 y, , y, , p , Maldacena, J. M. 2003, JHEP, 05, 013 March, W. B. 2013, Ph.D. Thesis, Georgia Institute of Technology, USA March, W. B. 2013, Ph.D. Thesis, Georgia Institute of Technology, USA Meerburg P D Green D Flauger R et al 2019 BAAS 51 107 Meerburg, P. D., Green, D., Flauger, R., et al. 2019, BAAS, 51, 107 Philcox, O. H. E., Slepian, Z., Hou, J., et al. 2021, MNRAS, 509, 2457 Portillo, S. K. N., Slepian, Z., Burkhart, B., Kahraman, S., & Finkbeiner, D. P. 2018, ApJ, 862, 119 Rodríguez-Torres, S. Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order A., Chuang, C.-H., Prada, F., et al. 2016, MNRAS, 460, 1173 4.3. Applications beyond correlation functions Though traditionally the order of correlation n is viewed as an important parameter, for the diagonal npcf, all the information is entirely encoded by the weights W0 and W1. It was pointed out in Carron & Neyrinck (2012) that the npcf is inadequate in captur- ing the tails of non-Gaussianities. The distributions over W0 and W1 (as opposed to their sum over the sample, as is used in the npcf) could recover that sensitivity, which is a subject for future studies. The distribution of the product of two weights, W0W1 A129, page 10 of 12 Z. Brown et al.: ConKer: An algorithm for evaluating correlations of arbitrary order Acquaviva, V., Bartolo, N., Matarrese, S., & Riotto, A. 2003, Nucl. Phys. B, 667, 119 Bartolo, N., Komatsu, E., Matarrese, S., & Riotto, A. 2004, Phys. Rep., 402, 103 Brown, Z., Mishtaku, G., Demina, R., Liu, Y., & Popik, C. 2021, A&A, 647, A196 Carron, J., & Neyrinck, M. 2012, ApJ, 750, 28 Cuesta, A. J., Vargas-Magana, M., Beutler, F., et al. 2016, MNRAS, 457, 1770 References (B.11) GLL′L′′ = 1 (4π)3/2 h 3 Y i=1 p 2li + 1 q 2l′ i + 1 q 2l′′ i + 1 li l′ i l′′ i 0 0 0 ! i  l1 l′ 1 l′′ 1 l2 l′ 2 l′′ 2 l3 l′ 3 l′′ 3 , (A.4) X = r(z) Y = r(z) Z = r(z) X = r(z)n′′ x , (B.9) Y = r(z)n′′ y , (B.10) Z = r(z)n′′ z . (B.11) X = r(z)n′′ x , Y = r(z)n′′ y , Z = r(z)n′′ z . for n = 5 GLL′L′′ = 1 (4π)2 p 2l12 + 1 q 2l′ 12 + 1 q 2l′′ 12 + 1 h 4 Y i=1 p 2li + 1 q 2l′ i + 1 q 2l′′ i + 1 li l′ i l′′ i 0 0 0 ! i  l1 l2 l12 l′ 1 l′ 2 l′ 12 l′′ 1 l′′ 2 l′′ 12   l12 l3 l4 l′ 12 l′ 3 l′ 4 l′′ 12 l′′ 3 l′′ 4 . (A.5) (A.5) Here three-dimensional matrices represent Wigner 9-j symbols. Here three-dimensional matrices represent Wigner 9-j symbols. References Ross, A. J., Beutler, F., Chuang, C.-H., et al. 2017, MNRAS, 464, 1168 Sá h A G S i R C M l 2017 MNRAS 46 , , , , g, , , , , Sánchez, A. G., Scoccimarro, R., Crocce, M., et al. 2017, MNRAS, 464, 1640 Acquaviva, V., Bartolo, N., Matarrese, S., & Riotto, A. 2003, Nucl. Phys. B, 667, 119 Bartolo, N., Komatsu, E., Matarrese, S., & Riotto, A. 2004, Phys. Rep., 402, 103 Brown, Z., Mishtaku, G., Demina, R., Liu, Y., & Popik, C. 2021, A&A, 647, A196 Carron, J., & Neyrinck, M. 2012, ApJ, 750, 28 Cuesta, A. J., Vargas-Magana, M., Beutler, F., et al. 2016, MNRAS, 457, 1770 Acquaviva, V., Bartolo, N., Matarrese, S., & Riotto, A. 2003, Nucl. Phys. B, 667, 119 Slepian, Z., & Eisenstein, D. J. 2015, MNRAS, 454, 4142 Slepian, Z., & Eisenstein, D. J. 2016, MNRAS, 455, L31 p White, M., Tinker, J. L., & McBride, C. K. 2014, MNRAS, 437, 2594 Yuan, S., Eisenstein, D. J., & Garrison, L. H. 2018, MNRAS, 478, 2019 Zhang, X., & Yu, C. 2011, Third IEEE International Conference on Cl White, M., Tinker, J. L., & McBride, C. K. 2014, MNRAS, 437, 2594 Yuan S Eisenstein D J & Garrison L H 2018 MNRAS 478 2019 Yuan, S., Eisenstein, D. J., & Garrison, L. H. 2018, MNRAS, 478, 2019 Zhang, X., & Yu, C. 2011, Third IEEE International Conference on Cloud Computing Technology and Science, 634 Carron, J., & Neyrinck, M. 2012, ApJ, 750, 28 Cuesta, A. J., Vargas-Magana, M., Beutler, F., et al. 2016, MNRAS, 457, 1770 A129, page 11 of 12 A129 (2022) Fig. B.1. Transformation from global coordinate system (z-axis pointing to zenith) to local coordinate system (x′′ pointing along the LOS). 129 (2022) A&A 667, A129 (2022) Appendix A: Useful formulae For L = (l1, l2, ..l(n−1)) and M = (m1, m2, ..m(n−1)) with each −li ≤ mi ≤li we can deﬁne the coupling coefﬁcient CLM in terms of Wigner 3-j symbols (3x2 matrices) as CLM = (−1) P li p 2l12 + 1.. p 2l12..(n−3) + 1 X m12.. (−1)κ l1 l2 l12 m1 m2 −m12 ! l12 l3 l123 m12 m3 −m123 ! .. l12..(n−3) l(n−2) l(n−1) m12..(n−3) m(n−2) −m(n−1) ! , Fig. B.1. Transformation from global coordinate system (z-axis pointing to zenith) to local coordinate system (x′′ pointing along the LOS). Fig. B.1. References Transformation from global coordinate system (z-axis pointing to zenith) to local coordinate system (x′′ pointing along the LOS). (A.1) Rotation by the polar angle δLOS deﬁnes coordinate system F′′, in which the galaxy’s unit vector is n′′ x = n′ x cos δLOS + n′ z sin δLOS , (B.4) n′′ y = n′ y, (B.5) n′′ z = −n′ x sin δLOS + n′ z cos δLOS . (B.6) where κ = l12 −m12 + l123 −m123 + ..l(n−1) −m(n−1). where κ = l12 −m12 + l123 −m123 + ..l(n−1) −m(n−1). where κ = l12 −m12 + l123 −m123 + ..l(n−1) −m(n−1). The Gaunt integral GL L L used in the edge correct ( ) ( ) The Gaunt integral GL1L2L3 used in the edge correction procedure is deﬁned as (B.6) GL1L2L3 = Z dˆrΠL1(ˆr)ΠL2(ˆr)ΠL3(ˆr). (A.2) (A.2) (A.2) In coordinate system F′′, the ˆx′-axis is directed along the LOS. The angular coordinates in F′′ are In coordinate system F′′, the ˆx′-axis is directed along the LOS. The angular coordinates in F′′ are For n = 3 it is For n = 3 it is α′′ = α −αLOS , (B.7) δ′′ = sin−1(n′′ z ). (B.8) For n = 3 it is Gl1l2l3 = 1 4π p 2l1 + 1 p 2l2 + 1 p 2l3 + 1 l1 l2 l3 0 0 0 !2 , (A.3) for n = 4 α′′ = α −αLOS , (B.7) δ′′ = sin−1(n′′ z ). (B.8) Thus, Cartesian coordinates in the local coordinate system are deﬁned as For n 3 it is Gl1l2l3 = 1 4π p 2l1 + 1 p 2l2 + 1 p 2l3 + 1 l1 l2 l3 0 0 0 !2 , (A.3) α′′ = α −αLOS , δ′′ = sin−1(n′′ z ). (B.7) (B.8) (B.7) (B.8) Thus, Cartesian coordinates in the local coordinate system are deﬁned as GLL′L′′ = 1 (4π)3/2 h 3 Y i=1 p 2li + 1 q 2l′ i + 1 q 2l′′ i + 1 li l′ i l′′ i 0 0 0 ! i  l1 l′ 1 l′′ 1 l2 l′ 2 l′′ 2 l3 l′ 3 l′′ 3 , (A.4) X = r(z)n′′ x , (B.9) Y = r(z)n′′ y , (B.10) Z = r(z)n′′ z . Appendix B: Coordinate transformation A global coordinate system is deﬁned so that the ˆz-axis is point- ing to the zenith, corresponding to declination angle δ = 90o = π/2. For each partition we deﬁne the local coordinate system with the ˆx′-axis pointing to its center cell, along the LOS, deﬁned by angles (αLOS , δLOS ). Transformation from the global coordinate system F to the coordinate system F′ is obtained by the rotation by the azimuthal angle αLOS . Then for a galaxy with angles (α, δ) the unit vector in system F′ is given by n′ x = cos δ cos(α −αLOS ), (B.1) n′ y = cos δ sin(α −αLOS ), (B.2) n′ z = sin δ. (B.3) A129, page 12 of 12 n′ x = cos δ cos(α −αLOS ), (B.1) n′ y = cos δ sin(α −αLOS ), (B.2) n′ z = sin δ. (B.3) A129, page 12 of 12 (B.1) (B.2) (B.3)
https://openalex.org/W4210735278	https://hal.sorbonne-universite.fr/hal-03582460/document	English	null	Single-Cell Technologies for the Study of Antibody-Secreting Cells	Frontiers in immunology	2,022	cc-by	8,283	To cite this version: Matteo Broketa, Pierre Bruhns. Single-Cell Technologies for the Study of Antibody-Secreting Cells. Frontiers in Immunology, 2022, 12, pp.821729. 10.3389/fimmu.2021.821729. hal-03582460 Distributed under a Creative Commons Attribution 4.0 International License Single-Cell Technologies for the Study of Antibody-Secreting Cells Matteo Broketa 1,2* and Pierre Bruhns 1* 1 Institut Pasteur, Universite´ de Paris, INSERM UMR 1222, Unit of Antibodies in Therapy and Pathology, Paris, France, 2 Sorbonne Universite´ , Collège doctoral, Paris, France Matteo Broketa 1,2* and Pierre Bruhns 1* 1 Institut Pasteur, Universite´ de Paris, INSERM UMR 1222, Unit of Antibodies in Therapy and Pathology, Paris, France, 2 Sorbonne Universite´ , Collège doctoral, Paris, France Antibody-secreting cells (ASC), plasmablasts and plasma cells, are terminally differentiated B cells responsible for large-scale production and secretion of antibodies. ASC are derived from activated B cells, which may differentiate extrafollicularly or form germinal center (GC) reactions within secondary lymphoid organs. ASC therefore consist of short-lived, poorly matured plasmablasts that generally secrete lower-afﬁnity antibodies, or long-lived, highly matured plasma cells that generally secrete higher- afﬁnity antibodies. The ASC population is responsible for producing an immediate humoral B cell response, the polyclonal antibody repertoire, as well as in parallel building effective humoral memory and immunity, or potentially driving pathology in the case of autoimmunity. ASC are phenotypically and transcriptionally distinct from other B cells and further distinguishable by morphology, varied lifespans, and anatomical localization. Single cell analyses are required to interrogate the functional and transcriptional diversity of ASC and their secreted antibody repertoire and understand the contribution of individual ASC responses to the polyclonal humoral response. Here we summarize the current and emerging functional and molecular techniques for high- throughput characterization of ASC with single cell resolution, including ﬂow and mass cytometry, spot-based and microﬂuidic-based assays, focusing on functional approaches of the secreted antibodies: speciﬁcity, afﬁnity, and secretion rate. HAL Id: hal-03582460 https://hal.sorbonne-universite.fr/hal-03582460v1 Submitted on 21 Feb 2022 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License MINI REVIEW MINI REVIEW published: 31 January 2022 doi: 10.3389/fimmu.2021.821729 MINI REVIEW published: 31 January 2022 doi: 10.3389/fimmu.2021.821729 Edited by: Ludvig M. Sollid, University of Oslo, Norway Edited by: Ludvig M. Sollid, University of Oslo, Norway Reviewed by: Ida Lindeman, Oslo University Hospital, Norway George Georgiou, University of Texas at Austin, United States Correspondence: Matteo Broketa matteo.broketa@pasteur.fr Pierre Bruhns bruhns@pasteur.fr Specialty section: This article was submitted to Systems Immunology, a section of the journal Frontiers in Immunology Keywords: plasma cell (PC), high-throuput technique, antibody secreting cell, droplet microﬂuidics, antibodies, B cells, functional bioassay INTRODUCTION Antibody-secreting cells (ASC) are B cells that have differentiated following activation to secrete various soluble isotypes of their immunoglobulin receptor with the purpose of binding their target antigen throughout the body (1). ASC are predominantly generated within the germinal center reactions of secondary lymphoid organs (2), although extra-follicular responses may also generate ASC (3, 4). Following antigen exposure, parallel downregulation of major regulatory genes of activated B cells and upregulation of a unique ASC transcriptional program drives differentiation of B cells into early-ASC or plasmablasts. Several mechanisms have been proposed to govern ASC fate determination, but a unifying model has not yet been determined (1, 5). Plasmablasts are an unstable ASC intermediate that require input from a survival niche to persist long term. The bone marrow has been extensively studied as an ASC niche for its role in harboring ASC following Received: 24 November 2021 Accepted: 29 December 2021 Published: 31 January 2022 Edited by: Ludvig M. Sollid, University of Oslo, Norway Reviewed by: Ida Lindeman, Oslo University Hospital, Norway George Georgiou, University of Texas at Austin, United States Correspondence: Matteo Broketa matteo.broketa@pasteur.fr Pierre Bruhns bruhns@pasteur.fr Citation: Broketa M and Bruhns P (2022) Single-Cell Technologies for the Study of Antibody-Secreting Cells. Front. Immunol. 12:821729. doi: 10.3389/fimmu.2021.821729 January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org High-Throughput Characterization of ASC Broketa and Bruhns infection and immunization, however the majority of ASC are located in the gut-associated lymphoid tissue and produce IgA (6); the thymus has also recently emerged as an ASC niche (7, 8). The factors involved in early-ASC homing to survival niches are not completely understood. Plasmablasts may also be drawn towards sites of inﬂammation, where they act locally and acutely without persistence (9). Plasma cells are a long-lived ASC subset, characterized by reinforced expression of genes within the ASC differentiation network, responsible for secreting large quantities of antibodies from within their survival niches. heterogeneity, particularly in autoimmune disorders. High throughput (HT) analyses of single cells are becoming more accessible, affordable, and common in literature concerning adaptive immune responses. This review will outline current and emerging HT techniques to characterize single ASC (Figure 1), with discussion of recent applications of these techniques to study the role of ASC in various pathologies as well as to expand understanding of fundamental ASC biology. FUNCTIONAL ANALYSES ASC are often considered the “apex” of B cell differentiation as they actualize antibody-mediated humoral immunity and are terminally differentiated. ASC contribute to both the acute humoral response to infection by rapidly generating early antibodies at sites of infection as well as later secreting higher afﬁnity antibodies produced by germinal center reactions to aid in pathogen clearance and protective immunity. While the ASC response is advantageous during infection and when co-opted for immunization, emergence of ASC secreting antibodies towards self-antigens is a deleterious factor in many autoimmune disorders (10). Despite their importance, much is still unknown regarding ASC differentiation, selection, and The primary effector functions of ASC are mediated by their secreted antibodies, and thereby are characterized by a secreted phenotype. Polyclonal antibody responses of ASC are routinely assessed by sampling the serum or ASC-containing organs for binding antibodies and their global potency (11). Such methods provide an overview of the cumulative ASC response but are unable to interrogate ASC diversity. Techniques with single cell resolution are required to uncover the contributions of individual clones to the polyclonal ASC pool, the prevalence of speciﬁc groups of ASC, and the relationships between the factors FIGURE 1 \| Overview of common techniques for single ASC characterization. Antibody secreting cells (ASC) may be characterized functionally (red region) or molecularly (green region). Functional methods include microﬂuidic approaches (top), including stationary and ﬂowed droplet-based systems and microwell systems (far right: Berkeley Lights Beacon setup), cytometry-based approaches, and spot-based assays. The red streak represents a laser beam; the yellow and dark red bulb shapes indicate a positive microwell). Molecular methods (bottom) may assess the VH (red band) and VL (blue band) antibody-encoding mRNA transcripts or non-antibody related mRNA transcripts (green bands). Molecular approaches commonly amplify VH chains only (e.g., VH-seq), VH and VL chains with the addition of barcodes (grey bands) or linkage (e.g., BCR-seq), or all mRNA within the cell (e.g., scRNAseq). FIGURE 1 \| Overview of common techniques for single ASC characterization. Antibody secreting cells (ASC) may be characterized functionally (red region) or molecularly (green region). Functional methods include microﬂuidic approaches (top), including stationary and ﬂowed droplet-based systems and microwell systems (far right: Berkeley Lights Beacon setup), cytometry-based approaches, and spot-based assays. The red streak represents a laser beam; the yellow and dark red bulb shapes indicate a positive microwell). Spot-Based Assays Spot-based assays spatially distribute ASC by dilution and rely on localized membrane capture of secreted antibody or cytokine from ASC, followed by visualization with enzyme- or ﬂuorophore-linked secondary antibodies (16, 17), termed enzyme-linked immunospot (ELISpot) and Fluorospot, respectively. The number of “spots” formed is indicative of the prevalence of ASC and/or antigen-speciﬁc ASC from a given sample. The strength of spot-based assays are their ease of use, robustness, and versatility between models (18, 19). Spot-based assays also directly observe secretion and therefore robustly identify ASC, in contrast to indirect-functional or molecular techniques where the ASC deﬁnition is inferred. However, ELISpot and Fluorospot are limited to the detection of only 1- 2 or 1-4 soluble analytes, respectively, cannot precisely quantify secretion rates, and the ASC identiﬁed cannot be recovered for further analysis. ELISpot has seen extensive use in characterizing the ASC response during pathogenic infections (19, 20), in autoimmune disorders (21–23), and following immunization (24). Bonezi et al (25) recently used ELISpot to demonstrate the contribution of altered tryptophan metabolism to the hyper- abundance of ASC frequently seen during ﬂavivirid infections, Dengue virus being the most common example. Flow cytometry has identiﬁed ASC and ASC subsets in the context of autoimmune disorders and their contributions to acute and chronic disease states (42–45), and the ASC response to natural infection, as recently shown by Woodruff et al (46) in the context of COVID-19. The authors showed by ﬂow cytometry a predominance of extrafollicularly-activated ASC in circulation, based on varied expression of CD11c, CXCR5, and CD21, and demonstrate its effect on infection morbidity. Noticeably, distinction between ASC derived from germinal centers or extrafollicularly is not possible with most other functional assays and is often overlooked. Application of FACS to in vitro models of ASC differentiation (47) and survival (48, 49) also highlights the ability of ﬂow cytometry to distinguish ASC subpopulations. FACS isolation of populations of interest may be followed by a spot-based assay to query the subsequent prevalence of ASC within those populations (50–52). FUNCTIONAL ANALYSES It should be noted that many other soluble effectors besides antibodies are also known to be secreted from ASC (12, 13), including various interleukins and transforming growth factor-beta one (TGF-b1), and HT functional assays may also be readily adapted to study the role of these soluble factors. and markers that differentiate ASC. It should be noted that many other soluble effectors besides antibodies are also known to be secreted from ASC (12, 13), including various interleukins and transforming growth factor-beta one (TGF-b1), and HT functional assays may also be readily adapted to study the role of these soluble factors. The topic of techniques for measuring single-cell protein secretion in immunology and their origins have been previously described (14, 15). Three general strategies are available for characterizing the secreted effectors of immune cells, each with applicability to ASC: spot-based assays, cytometry-based assays, and microﬂuidic-assays. These approaches globally share the strategy of isolating ASC, either by distance or compartmentalization, to ensure that “positive” loci or compartments reﬂect secretions or markers of single cells. Here we will focus on what these techniques offer to describe ASC behavior and highlight notable demonstrations of their use. FUNCTIONAL ANALYSES Molecular methods (bottom) may assess the VH (red band) and VL (blue band) antibody-encoding mRNA transcripts or non-antibody related mRNA transcripts (green bands). Molecular approaches commonly amplify VH chains only (e.g., VH-seq), VH and VL chains with the addition of barcodes (grey bands) or linkage (e.g., BCR-seq), or all mRNA within the cell (e.g., scRNAseq). January 2022 \| Volume 12 \| Article 821729 2 Frontiers in Immunology \| www.frontiersin.org High-Throughput Characterization of ASC Broketa and Bruhns Human ASC are often deﬁned in ﬂow cytometry as larger cells with CD27 and CD38high expression among CD3- CD20- cells, and may be subdivided by expression of CD138, CD19, CD45, CD81, HLA-DR, and immunoglobulin isotypes (30–33). Plasma cells secreting human IgA or IgM, but not IgG, express surface immunoglobulin and may be identiﬁed by direct surface staining (34, 35). Local capture of IgG antibodies secreted by ASC onto their surface has been proposed to identify human IgG-secreting ASC by ﬂow cytometry, using CD45 as a membrane anchor onto which IgG antibodies are captured by their constant region (Fc) (36), but has not yet been widely used. Surface markers and soluble effectors, namely antibodies and cytokines, not expressed at the cell membrane may be assessed using cell permeabilization and ﬁxation to allow labelling reagents access within the cell; such intracellular staining normally precludes downstream assessment of cells by other functional or molecular methods. However, Price et al (37) demonstrated the feasibility to molecularly characterize ASC following intracellular staining for both antibody isotype as well as antibody speciﬁcity using tetramer constructs of antigens bound to ﬂuorophore-conjugated streptavidin. Alternatively, early IgG-expressing ASC (IgG-ASC) often continue to bear their antibodies at the cell membrane, and surface staining for antibody isotype and antigen-speciﬁcity with tetramers offers a simpler method for isolating such IgG-ASC, though represents a restricted view of the whole IgG-ASC population (38, 39). Flow cytometry is also a valuable tool for scouting potential novel markers of ASC subsets (40). The use of intracellular tracing dyes further allows for assessment of ASC proliferation, as Scharer et al (41) have applied to ASC emergence from activated B cells. and markers that differentiate ASC. Cytometry-Based Assays Flow cytometry and its derivative, ﬂuorescence-activated cell sorting (FACS), have become a staple in investigations of B cells and ASC, allowing for relatively straightforward and rapid evaluation of multiple samples or experimental conditions (26, 27). Multicolor ﬂow cytometry allows for immense diversity and depth of information gathered through the customization of antibody panels to target surface and intracellular markers. In contrast to ELISpot, ﬂow cytometry can precisely identify the proportions of different ASC, B cell, and lymphocyte populations in each sample. Flow cytometry and ELISpot are frequently used in concert to measure global changes in the proportion of ASC and other B cells and the robustness and antigen-speciﬁcity of the ASC response, respectively (28, 29). It should be noted that current cytometric-gating strategies for ASC populations have not been demonstrated to completely deﬁne all ASC without exclusion of rarer phenotypes. Strategies often vary considerably, utilize a unique marker that may not be ASC-identifying in all contexts, or rely on intracellular staining of transcription factors (31, 53). ASC are distinct in their transcriptional regulation compared to other B cells, and the transcription factors IRF4, BLIMP-1, and XBP1 are ideal for complete identiﬁcation of ASC (1, 53). Indeed, BLIMP1-YFP January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org 3 High-Throughput Characterization of ASC Broketa and Bruhns technical demonstrations and antibody discovery (66–68), mainly due to their cost and limited throughput. mice have been invaluable to studying ASC differentiation, behavior, and transcriptional regulation (54, 55). IRF4-based labelling has similarly been used to identify differentiating human plasma cells following inﬂuenza vaccination (56). Droplet microﬂuidics achieves single cell compartmentalization using two immiscible ﬂuids, an aqueous phase containing the cell and assay reagents and an oil phase that separates the aqueous phase into droplets with single cells (14). Large numbers of droplets can be quickly produced for each new sample and the assay composition can be easily changed. Secretion by encapsulated ASC is assessed by beads or other cells within the droplet that act as capture surfaces for secreted cytokine or antibody, to which ﬂuorescent reagents in the droplet can localize for measurement. Flow-based or ﬂowed droplet assays can characterize ASC with high throughput by passing droplets at high frequency through a laser for rapid detection of ﬂuorescence relocalization within droplets (69–71). Microﬂuidics Microﬂuidic single cell assays essentially miniaturize existing techniques to assess ASC secretions, to increase the scale and throughput of analyses and enable single cell resolution via compartmentalization into either wells or droplets (14, 15). Miniaturization also allows for greatly reduced reagent usage and less cells are needed per experiment. The functional measurements within microﬂuidic ASC assays largely share the concept of capturing secreted antibody or cytokine onto a physical surface followed by visualization with ﬂuorescent detection reagents. The advantages of microﬂuidic-based functional analyses of ASC include their direct identiﬁcation of ASC via antibody secretion, a high throughput, and the ability to absolutely quantify ASC secretions over time. y p q An alternative approach to study ASC in droplets is to collect droplets in a horizontal plane and measure changes in droplet ﬂuorescence over multiple timepoints, termed a stationary droplet-assay (14) or DropMap (73, 74). DropMap measures ﬂuorescence relocation to a central line of antibody-capturing beads aligned using a magnetic ﬁeld within each droplet. Both secreted antibody relocation and soluble antigen relocation to the beadline are measured. The ability to measure relocation over time within DropMap allows for the determination of both antibody afﬁnity and secretion rate, in addition to determining the proportion of antigen-speciﬁc and total ASC within a sample as with other microﬂuidic techniques. DropMap has been applied to investigate the physiology of ASC in viral infection (75), autoimmunity (74), and fundamental ASC biology (73, 76). In our view, DropMap offers a major advancement in current capabilities to functionally characterize ASC, particularly for deﬁning the speciﬁcity and afﬁnity repertoire of ASC. An example alternative strategy to DropMap for afﬁnity repertoire mapping is to sort large numbers of B cells or ASC for their BCR sequences, followed by gene synthesis or direct cloning, re- expression, and kinetic analysis of these antibodies by a technique such as bio-layer interferometry (BLI) (77–79) (Figure 2); this approach is signiﬁcantly more complex, costly, and time consuming than DropMap. Recently however, we reported antigen-speciﬁc single cell memory B cell in vitro differentiation into ASC that allowed for fast and large scale (~400) afﬁnity measurements by BLI from culture supernatants without costly antibody re-expression (80). Initial sorting for Micro- and nano-well microﬂuidic approaches use microfabrication techniques to create thousands of wells into which individual ASC can be introduced (60–62). Cytometry-Based Assays Droplet measurement may also be paired with dielectrophoretic sorting, which pushes or pulls droplets into separate channels by manipulating electric ﬁelds, allowing for further assessment of directly identiﬁed ASC or antigen-speciﬁc ASC (69, 72). We reported in Gerard et al (69) the CelliGo assay using a double ﬂuorescent sandwich ELISA in microﬂuidic droplets for the identiﬁcation, sorting, and VH-VL sequencing of antigen-speciﬁc IgG antibodies produced by ASC from immunized mice. In this study, we demonstrated screening of a bacterial antigen (tetanus toxoid), an autoantigen linked to Rheumatoid Arthritis (Glucose-6-phosphate Isomerase), and an insoluble, membrane-expressed antigen (tetraspanin-8; TSPAN8); antigen-speciﬁc ASC against all 3 model antigens were able to be sorted by this ﬂowed droplet microﬂuidic technique. Mass cytometry is based on a similar principle as ﬂow cytometry, using heavy metal ions instead of ﬂuorophores to distinguish the antibodies in the cytometry panel. Glass et al (57) utilized multiple mass cytometry panels to characterize over 350 B cell surface markers, combined with analyses of isotype usage, BCR sequence, metabolic proﬁle, biosynthesis activity, and signaling response, which together constitute an expansive single- cell atlas of human B cells. An additional mass cytometry panel targeting enzymes associated with different metabolic pathways (58) found plasma cells to be highly metabolically active, and further subdivisible based on transcriptional activity; such metabolic distinctions have largely been restricted to molecular studies or classical metabolomics (49, 59). Although mass cytometry can simultaneously detect ~40 parameters per panel, the vaporization of analyzed cells prevents further study of the precise subsets identiﬁed, which remains the advantage of FACS. Frontiers in Immunology \| www.frontiersin.org Microﬂuidics FIGURE 2 \| Comparison of DropMap vs a classical pipeline to deﬁne afﬁnity repertoires. The afﬁnity repertoire of ASC towards a given antigen is an important metric for the quality of the ASC response. ASC are commonly isolated from the spleen, bone marrow, or blood of donor/patients or experimental animals. The DropMap assay (top) offers a platform that within 1 hour can return the afﬁnity repertoire in a single assay, as well as the IgG secretion rate and frequency of ASC according to direct ex vivo measurement (refer to main text for details); this approach is limited by requiring ASC to secrete antibodies at the time of data acquisition, data can only be acquired once, and afterwards cells of interest are lost. An alternative strategy (bottom) to yield a similar afﬁnity repertoire ﬁrst requires cell isolation by FACS, followed by VH and VL targeted RT-PCR (& BCR-seq if required) and cloning into an expression vector, transfection of the vector into an expression system and cell culture, puriﬁcation and ﬁnally assessment of the recombinant expressed antibody by BLI or SPR; this pipeline requires 1½ -2 weeks to complete, is signiﬁcantly more complex and costly, and can only assess VH-VL pairs that could be successfully ampliﬁed, cloned and expressed as recombinant antibodies. FACS into 96 or 384 well plates, or microﬂuidic droplet-based barcoded scRNAseq (e.g., 10xGenomics Chromium). Importantly, BCR sequences can be computationally inferred from scRNA-seq data using open-source libraries (e.g. BraCeR (84), VDJPuzzle (85)), linking antibody sequence (i.e., genotype, clonal information) to transcriptome phenotype. Recent advances on ASC characterization using RNAseq, ATAC-seq and ChIP-seq have been reviewed elsewhere (86), as well as molecular mechanisms leading to plasma cell differentiation from the germinal center reaction (87). Glaros et al (38) highlight the capacity for scRNAseq to identify B cell subsets and observe shifts in differentiation, identifying antigen availability as a key regulator of the plasmablast response. ASC may also bias the antibody repertoire through restrictive FACS gating strategies unless a direct functional assay is used prior to directly select for ASC, whereas DropMap offers an unbiased screen of PBMCs or an enriched B cell pool for ASC. Another novel approach involving BCR sequencing paired with liquid chromatography–tandem mass spectrometry proteomics (81, 82) allows for unbiased molecular assessment of antibodies from the serum, but requires infrastructure and specialist supervision, and is currently largely restricted to the most represented antibodies in the serum. Microﬂuidics The nature of the wells and coating and reagent strategy used can be readily customized to assay different antibodies, antigens, or secretions. The small volumes of these wells allow for absolute quantiﬁcation of secreted molecules, in contrast to ELISpot where only relative measurements can be made (14). Well-based microﬂuidic assays also have lower detection thresholds than ELISpot, making them more sensitive and more capable of wholly representing ASC (63). Importantly, as ASC are identiﬁed and sustained within the static well array, cells can be recovered following the assay for additional functional or molecular assessment. The Berkeley Lights “nanopen” platform introduces considerable advancement to well-based ASC assays using optoﬂuidics, offering integrated workﬂows for antibody screening of ASC followed by targeted recovery of speciﬁc cells for molecular analysis (64). This rapid pipeline for functional antibody assessment was recently used to characterize antibody repertoires following SARS-COV2 infection (65). Despite their advantages over ELISpot, well-based microﬂuidic assays have seen limited application beyond January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org Broketa and Bruhns High-Throughput Characterization of ASC FIGURE 2 \| Comparison of DropMap vs a classical pipeline to deﬁne afﬁnity repertoires. The afﬁnity repertoire of ASC towards a given antigen is an important metric for the quality of the ASC response. ASC are commonly isolated from the spleen, bone marrow, or blood of donor/patients or experimental animals. The DropMap assay (top) offers a platform that within 1 hour can return the afﬁnity repertoire in a single assay, as well as the IgG secretion rate and frequency of ASC according to direct ex vivo measurement (refer to main text for details); this approach is limited by requiring ASC to secrete antibodies at the time of data acquisition, data can only be acquired once, and afterwards cells of interest are lost. An alternative strategy (bottom) to yield a similar afﬁnity repertoire ﬁrst requires cell isolation by FACS, followed by VH and VL targeted RT-PCR (& BCR-seq if required) and cloning into an expression vector, transfection of the vector into an expression system and cell culture, puriﬁcation and ﬁnally assessment of the recombinant expressed antibody by BLI or SPR; this pipeline requires 1½ -2 weeks to complete, is signiﬁcantly more complex and costly, and can only assess VH-VL pairs that could be successfully ampliﬁed, cloned and expressed as recombinant antibodies. Microﬂuidics The major limitation to widespread adoption of functional, droplet-based assays of ASC is their bespoke nature. Although droplet microﬂuidics is a growing ﬁeld, with most reagents commercially available, labs must produce their own consumables for droplet production and the apparatus for assay observation, which require specialized equipment and expertise. A common strategy for molecular ASC clonal identiﬁcation is bulk sequencing of VH regions (37), which allows for large numbers of cells to be assayed without the cost and limitations imposed by a need for single cell approaches. However, the VL sequence information is lost, and it is not possible to re-express the original antibody recombinantly. Alternatively, Price et al (37) recently used bulk RNAseq to identify unique transcriptional proﬁles for IgG, IgM, and IgA ASC subsets, with the ability to assess both clonality and gene expression. Such approaches offer insight into the clonal diversity of ASC responses but have less resolution than paired VH-VL chain sequencing and have no possibility for functional assessment of clones within the antibody repertoire. When the full VH-VL sequence is required, single-cell approaches are required, and currently available techniques for HT BCRseq have been reviewed by Curtis and Lee (88), who highlight advances in VH MOLECULAR ANALYSES The molecular basis for the behavior of ASC is equally important to understanding ASC physiology. The molecular aspects of ASC most investigated are (i) the heavy and light chain variable regions (V H and V L) sequences, formed by gene rearrangements and sequence diversiﬁcation mechanisms (83), encoding the antigen binding domains of antibodies and (ii) the transcriptional proﬁle of ASCs; these aspects may be probed using bulk VH-seq or RNAseq, BCR-seq, and increasingly now single-cell RNAseq (scRNAseq) following either single-cell ASC Frontiers in Immunology \| www.frontiersin.org January 2022 \| Volume 12 \| Article 821729 5 High-Throughput Characterization of ASC Broketa and Bruhns and VL chain pairing by barcoding (69) or linkage as well as increased cell-throughput through the use microﬂuidics. Moreover, techniques such as LIBRA-seq (89) or CelliGo (69) allow for assessment of antigen speciﬁcity or antigen speciﬁcity coupled to antibody secretion, respectively, integrated within their BCRseq pipelines. Wang et al (90) used microﬂuidics to compartmentalize human plasmablasts as single cells and generate paired VH-VL repertoires for direct Fab display on yeast and functional assessment. Notably, Jiang et al (91) utilized a combination of bulk VH-seq, BCRseq, and full scRNAseq to identify autoreactive ASC persisting after treatment with rituximab. Such an approach leverages the complexity and throughput of each technique with the depth of information required to address the authors’ underlying question, saving considerable time, and lowering costs overall. Using the 10x genomics technology, integrating emulsion-based single cell separation with barcoded RNAseq (Chromium), plasmablast-derived mAbs from individuals who received SARS-CoV-2 spike mRNA vaccine were characterized for antigen speciﬁcity, epitope mapping and neutralization potential (92). undetectable. As we showed in Eyer et al. (73), antibody secretion by a seemingly homogenous population of ASC may vary by several orders of magnitude. The causes of such secretory diversity remain poorly understood. Current techniques for direct measurement of single cell immunoglobulin secretion (spot- and microﬂuidic-based assays) are currently unable to distinguish precise B cell populations alone. ASC identiﬁed and isolated from direct functional assays have not been assessed afterwards with higher parameter techniques such as ﬂow and mass cytometry or scRNAseq, to interrogate the source of this secretory diversity. Ultimately, investigations of ASC physiology should strive to employ a combination of the techniques discussed. In the context of antibodies, with only a repertoire of VH-VL sequences the contribution and quality of a given antibody to the humoral immune response is difﬁcult to appreciate. 2. Shlomchik MJ, Weisel F. Germinal Center Selection and the Development of Memory B and Plasma Cells. Immunol Rev (2012) 247(1):52–63. doi: 10.1111/ j.1600-065X.2012.01124.x 3. Di Niro R, Lee SJ, Vander Heiden JA, Elsner R, Trivedi N, Bannock J. SalmOnella Infection Drives Promiscuous B Cell Activation Followed By MOLECULAR ANALYSES Likewise, knowing the afﬁnity or neutralization repertoire of antibodies from ASC without knowing their clonality and molecular basis limits insight into ASC population dynamics and distribution. Evolution of the existing and emerging methods for HT ASC phenotypic characterization to readily integrate with the various single cell molecular techniques should be an immediate goal to overcome current limitations. HT single cell molecular (scRNA- seq) techniques are becoming more accessible and affordable, with great efforts to increase the ﬁdelity and ease of analysis. HT single cell assays will be central to unravelling how ASC phenotypic markers relate to their developmental stage, antibody afﬁnity, and antibody or cytokine secretion rate, and to what degree these elements are interrelated. A ﬁnal molecular approach worth mentioning is the use of mass spectrometry to assess the relative abundance of speciﬁc antibodies in serum or a tissue sample, known as Ab-seq (93). Pairing Ab-seq with any of the RNA-seq approaches above allows for inferences to be made regarding the contribution of speciﬁc ASC clones to the polyclonal ASC response, providing invaluable insight into the immunological relevance of ASC clonal diversity. Lee et al (94) utilized Ab-seq to identify potently neutralizing antibodies and ASC clones persisting across multiple exposures to inﬂuenza. Ab-seq has also been applied in similar studies for antibodies against norovirus (95) and HIV-1 (96). FUNDING The availability of high throughput methods for the study of ASC at the single cell level has progressed immensely for both functional and molecular characterization. Considerable progress has been made in understanding ASC development and differentiation, but with limited information regarding the relationships between population surface marker expression, transcriptional and metabolic activity, and the functional “quality” of the ASC (afﬁnity, speciﬁcity, secretion rate). PB acknowledges funding from the French National Research Agency grant ANR-14-CE16-0011 project DROPmAbs and ANR-18-CE15-0001 project Autoimmuni-B, by the Institut Carnot Pasteur Microbes et Santé (ANR 11 CARN 0017-01), the Institut Pasteur and the Institut National de la Santé et de la Recherche Médicale (INSERM). MB is a recipient of a CIFRE fellowship from the French Association Nationale de la Recherche et de la Technologie (ANRT). None of the sources of funding have an interest in the subject matter or materials discussed in the submitted manuscript. The presence of immunoglobulins within all plasmablasts or plasma cells does not exclude the possibility for signiﬁcant stratiﬁcation of secretion rates from high to nearly AUTHOR CONTRIBUTIONS All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication. Multiplexing Is the Future for High Throughput Single ASC Analyses Multiplexing Is the Future for High Throughput Single ASC Analyses 1. Nutt SL, Hodgkin PD, Tarlinton DM, Corcoran LM. The Generation of Antibody-Secreting Plasma Cells. Nat Rev Immunol (2015) 15(3):160–71. doi: 10.1038/nri3795 REFERENCES Hauser AE, Debes GF, Arce S, Cassese G, Hamann A, Radbruch A, et al. Chemotactic Responsiveness Toward Ligands for CXCR3 and CXCR4 Is Regulated on Plasma Blasts During the Time Course of a Memory Immune Response. J Immunol (2002) 169(3):1277–82. doi: 10.4049/jimmunol.169.3.1277 30. Carrell J, Groves CJ. OMIP-043: Identiﬁcation of Human Antibody Secreting Cell Subsets. Cytom Part A. (2018) 93(2):190–3. doi: 10.1002/cyto.a.23305 31. Sanz I, Wei C, Jenks SA, Cashman K, Tipton C, Woodruff M, et al. Challenges and Opportunities for Consistent Classiﬁcation of Human B Cell and Plasma Cell Populations. Front Immunol (2019) 10:2458(OCT). doi: 10.3389/ ﬁmmu.2019.02458 10. Chang HD, Tokoyoda K, Hoyer B, Alexander T, Khodadadi L, Mei H, et al. Pathogenic Memory Plasma Cells in Autoimmunity. Curr Opin Immunol (2019) 61:86–91. doi: 10.1016/j.coi.2019.09.005 32. Landsverk OJB, Snir O, Casado RB, Richter L, Mold J, Réu P, et al. Antibody- Secreting Plasma Cells Persist for Decades in Human Intestine. J Exp Med (2017) 214(2):309–17. doi: 10.1084/JEM.20161590 11. Galipeau Y, Greig M, Liu G, Driedger M, Langlois MA. Humoral Responses and Serological Assays in SARS-CoV-2 Infections. Front Immunol (2020) 11:610688. doi: 10.3389/ﬁmmu.2020.610688 12. Pioli PD. Plasma Cells, the Next Generation: Beyond Antibody Secretion. Front Immunol (2019) 10:2768. doi: 10.3389/ﬁmmu.2019.02768 33. Paiva B, Puig N, Cedena MT, de Jong B, Ruiz Y, Rapado I, et al. Differentiation Stage of Myeloma Plasma Cells: Biological and Clinical Signiﬁcance. Leukemia (2017) 31(2):382–92. doi: 10.1038/leu.2016.211 13. Upasani V, Id IR, Id TC, Rodenhuis-Zybert I, Cantaert T. Antibody- Independent Functions of B Cells During Viral Infections. PloS Pathog (2021) 17(7):e1009708. doi: 10.1371/journal.ppat.1009708 34. Blanc P, Moro-Sibilot L, Barthly L, Jagot F, This S, de Bernard S, et al. Mature IgM-Expressing Plasma Cells Sense Antigen and Develop Competence for Cytokine Production Upon Antigenic Challenge. Nat Commun (2016) 7:13600. doi: 10.1038/ncomms13600 14. Bucheli OTM, Sigvaldadóttir I, Eyer K. Measuring Single-Cell Protein Secretion in Immunology: Technologies, Advances, and Applications. Eur J Immunol (2021) 51(6):1334–47. doi: 10.1002/eji.202048976 35. Pinto D, Montani E, Bolli M, Garavaglia G, Sallusto F, Lanzavecchia A, et al. A Functional BCR in Human IgA and IgM Plasma Cells. Blood (2013) 121 (20):4110–4. doi: 10.1182/BLOOD-2012-09-459289 15. Jammes FC, Maerkl SJ. How Single-Cell Immunology is Beneﬁting From Microﬂuidic Technologies. Microsystems Nanoeng (2020) 6(1):45. doi: 10.1038/s41378-020-0140-8 36. Pinder CL, Kratochvil S, Cizmeci D, Muir L, Guo Y, Shattock R, et al. Isolation and Characterization of Antigen-Speciﬁc Plasmablasts Using a Novel Flow Cytometry–Based Ig Capture Assay. REFERENCES J Immunol (2017) 199(12):4180–8. doi: 10.4049/jimmunol.1701253 16. Tzeng SJ. The Isolation, Differentiation, and Quantiﬁcation of Human Antibody-Secreting B Cells From Blood: Elispot as a Functional Readout of Humoral Immunity. J Vis Exp (2016) 2016(118):e54582. doi: 10.3791/54582 17. Janetzki S, Rueger M, Dillenbeck T. Stepping Up ELISpot: Multi-Level Analysis in FluoroSpot Assays. Cells (2014) 3(4):1102–15. doi: 10.3390/ cells3041102 37. Price MJ, Hicks SL, Bradley JE, Randall TD, Boss JM, Scharer CD. IgM, IgG, and IgA Inﬂuenza-Speciﬁc Plasma Cells Express Divergent Transcriptomes. J Immunol (2019) 203(8):2121–9. doi: 10.4049/JIMMUNOL.1900285 18. Meyer Sauteur PM, Seiler M, Trück J, Unger WWJ, Paioni P, Relly C, et al. Diagnosis of Mycoplasma Pneumoniae Pneumonia With Measurement of Speciﬁc Antibody-Secreting Cells. Am J Respir Crit Care Med (2019) 200 (8):1066–9. doi: 10.1164/rccm.201904-0860LE 38. Glaros V, Rauschmeier R, Artemov AV, Reinhardt A, Ols S, Emmanouilidi A, et al. Limited Access to Antigen Drives Generation of Early B Cell Memory While Restraining the Plasmablast Response. Immunity (2021) 54(9):2005– 2023.e10. doi: 10.1016/J.IMMUNI.2021.08.017 19. Carter MJ, Mitchell RM, Sauteur PMM, Kelly DF, Trück J. The Antibody- Secreting Cell Response to Infection: Kinetics and Clinical Applications. Front Immunol (2017) 8:630. doi: 10.3389/ﬁmmu.2017.00630 39. Franz B, May KF, Dranoff G, Wucherpfennig K. Ex Vivo Characterization and Isolation of Rare Memory B Cells With Antigen Tetramers. Blood (2011) 118 (2):348. doi: 10.1182/BLOOD-2011-03-341917 20. Lima-Junior J, Morgado F, Conceição-Silva F. How Can Elispot Add Information to Improve Knowledge on Tropical Diseases? Cells (2017) 6 (4):31. doi: 10.3390/cells6040031 40. Llinàs L, Lázaro A, de Salort J, Matesanz-Isabel J, Sintes J, Engel P. Expression Proﬁles of Novel Cell Surface Molecules on B-Cell Subsets and Plasma Cells as Analyzed by Flow Cytometry. Immunol Lett (2011) 134(2):113–21. doi: 10.1016/j.imlet.2010.10.009 21. Faresjö M. Enzyme Linked Immuno-Spot; a Useful Tool in the Search for Elusive Immune Markers in Common Pediatric Immunological Diseases. Cells (2012) 1(2):141–52. doi: 10.3390/cells1020141 41. Scharer CD, Patterson DG, Mi T, Price MJ, Hicks SL, Boss JM. Antibody- Secreting Cell Destiny Emerges During the Initial Stages of B-Cell Activation. Nat Commun (2020) 11:3989. doi: 10.1038/s41467-020-17798-x 22. Mahévas M, Michel M, Weill JC, Reynaud CA. Long-Lived Plasma Cells in Autoimmunity: Lessons From B-Cell Depleting Therapy. Front Immunol (2013) 4:494(DEC). doi: 10.3389/ﬁmmu.2013.00494 42. Mahévas M, Patin P, Huetz F, Descatoire M, Cagnard N, Bole-Feysot C, et al. B Cell Depletion in Immune Thrombocytopenia Reveals Splenic Long-Lived Plasma Cells. J Clin Invest (2013) 123(1):432–42. doi: 10.1172/JCI65689 23. REFERENCES 1. Nutt SL, Hodgkin PD, Tarlinton DM, Corcoran LM. The Generation of Antibody-Secreting Plasma Cells. Nat Rev Immunol (2015) 15(3):160–71. doi: 10.1038/nri3795 1. Nutt SL, Hodgkin PD, Tarlinton DM, Corcoran LM. The Generation of Antibody-Secreting Plasma Cells. Nat Rev Immunol (2015) 15(3):160–71. doi: 10.1038/nri3795 January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org 6 High-Throughput Characterization of ASC Broketa and Bruhns Extrafollicular Afﬁnity Maturation. Immunity (2015) 43(1):120–131. doi: 10.1016/j.immuni.2015.06.013 25. Bonezi V, Cataneo AHD, Branquinho MSF, Silva M, Gonzalez-Dias P, Pereira S, et al. Flavivirus-Mediating B Cell Differentiation Into Antibody-Secreting Cells in Humans Is Associated With the Activation of the Tryptophan Metabolism. Front Immunol (2020) 11:20. doi: 10.3389/ﬁmmu.2020.00020 4. Allman D, Wilmore JR, Gaudette BT. The Continuing Story of T-Cell Independent Antibodies. Immunol Rev (2019) 288(1):128–35. doi: 10.1111/ imr.12754 26. Cossarizza A, Chang HD, Radbruch A, Acs A, Adam D, Adam-Klages S, et al. Guidelines for the Use of Flow Cytometry and Cell Sorting in Immunological Studies (Second Edition). Eur J Immunol (2019) 49(10):1457–973. doi: 10.1002/eji.201970107 5. Shaffer AL, Lin KI, Kuo TC, Yu X, Hurt E, Rosenwald A, et al. Blimp-1 Orchestrates Plasma Cell Differentiation by Extinguishing the Mature B Cell Gene Expression Program. Immunity (2002) 17(1):51–62. doi: 10.1016/S1074-7613(02)00335-7 j 27. Boonyaratanakornkit J, Taylor JJ. Techniques to Study Antigen-Speciﬁc B Cell Responses. Front Immunol (2019) 10:1694. doi: 10.3389/ﬁmmu.2019.01694 6. Lightman SM, Utley A, Lee KP. Survival of Long-Lived Plasma Cells (LLPC): Piecing Together the Puzzle. Front Immunol (2019) 10:965. doi: 10.3389/ ﬁmmu.2019.00965 28. Ellebedy AH, Jackson KJL, Kissick HT, Nakaya H, Davis C, Roskin K, et al. Deﬁning Antigen-Speciﬁc Plasmablast and Memory B Cell Subsets in Blood Following Viral Infection and Vaccination of Humans. Nat Immunol (2016) 17(10):1226. doi: 10.1038/NI.3533 7. Cordero H, King RG, Dogra P, Dufeu C, See S, Chong A, et al. Intrathymic Differentiation of Natural Antibody-Producing Plasma Cells in Human Neonates. Nat Commun (2021) 12(1):5761. doi: 10.1038/s41467-021-26069-2 29. Quách TD, Rodrı́guez-Zhurbenko N, Hopkins TJ, Guo X, Hernández A, Li W, et al. Distinctions Among Circulating Antibody-Secreting Cell Populations, Including B-1 Cells, in Human Adult Peripheral Blood. J Immunol (2016) 196 (3):1060–9. doi: 10.4049/jimmunol.1501843 8. Nuñez S, Moore C, Gao B, Rogers K, Hidalgo Y, Del Nido P, et al. The Human Thymus Perivascular Space is a Functional Niche for Viral-Speciﬁc Plasma Cells. Sci Immunol (2016) 1(6). doi: 10.1126/sciimmunol.aah4447 9. REFERENCES Kuwana M, Okazaki Y, Kaburaki J, Kawakami Y, Ikeda Y. Spleen Is a Primary Site for Activation of Platelet-Reactive T and B Cells in Patients With Immune Thrombocytopenic Purpura. J Immunol (2002) 168(7):3675–82. doi: 10.4049/ jimmunol.168.7.3675 43. Xu M, Jiang Y, Wang J, Liu J, Liu C, Liu D, et al. Distinct Variations of Antibody Secreting Cells and Memory B Cells During the Course of Kawasaki Disease. BMC Immunol (2019) 20(1):16. doi: 10.1186/s12865-019-0299-7 24. Saletti G, Ç uburu N, Yang JS, Dey A, Czerkinsky C. Enzyme-Linked Immunospot Assays for Direct Ex Vivo Measurement of Vaccine-Induced Human Humoral Immune Responses in Blood. Nat Protoc (2013) 8(6):1073– 87. doi: 10.1038/nprot.2013.058 44. Tipton CM, Fucile CF, Darce J, Chida A, Ichikawa T, Gregoretti I, et al. Diversity, Cellular Origin and Autoreactivity of Antibody-Secreting Cell Population Expansions in Acute Systemic Lupus Erythematosus. Nat Immunol (2015) 16(7):755–65. doi: 10.1038/ni.3175 January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org Broketa and Bruhns High-Throughput Characterization of ASC 45. Jin W, Luo Z, Yang H. Review Article Peripheral B Cell Subsets in Autoimmune Diseases: Clinical Implications and Effects of B Cell-Targeted Therapies. J Immunol Res (2020) 2020:9518137. doi: 10.1155/2020/9518137. Published online. 65. Zost SJ, Gilchuk P, Chen RE, Case J, Reidy J, Trivette A, et al. Rapid Isolation and Proﬁling of a Diverse Panel of Human Monoclonal Antibodies Targeting the SARS-CoV-2 Spike Protein. Nat Med (2020) 26(9):1422. doi: 10.1038/ S41591-020-0998-X 66. Ogunniyi AO, Story CM, Papa E, Guillen E, Love JC. Screening Individual Hybridomas by Microengraving to Discover Monoclonal Antibodies. Nat Protoc (2009) 4(5):767–82. doi: 10.1038/NPROT.2009.40 46. Woodruff MC, Ramonell RP, Nguyen DC, Cashman K, Saini A, Haddad N, et al. Extrafollicular B Cell Responses Correlate With Neutralizing Antibodies and Morbidity in COVID-19. Nat Immunol 2020 2112 (2020) 21(12):1506–16. doi: 10.1038/s41590-020-00814-z 67. Ogunniyi AO, Thomas BA, Politano TJ, Varadarajan N, Landais E, Poignard P, et al. Proﬁling Human Antibody Responses by Integrated Single-Cell Analysis. Vaccine (2014) 32(24):2866–73. doi: 10.1016/J.VACCINE.2014.02.020 47. Ramadani F, Bowen H, Upton N, Hobson P, Chan Y, Chen J, et al. Ontogeny of Human IgE-Expressing B Cells and Plasma Cells. Allergy Eur J Allergy Clin Immunol (2017) 72(1):66–76. doi: 10.1111/all.12911 68. Story CM, Papa E, Hu CC, Ronan JL, Herlihy K, Ploegh HL, et al. Proﬁling Antibody Responses by Multiparametric Analysis of Primary B Cells. Proc Natl Acad Sci U S A (2008) 105(46):17902–7. doi: 10.1073/PNAS.0805470105 48. REFERENCES Kallies A, Hasbold J, Tarlinton DM, Dietrich W, Corcoran L, Hodgkin P, et al. Plasma Cell Ontogeny Deﬁned by Quantitative Changes in Blimp-1 Expression. J Exp Med (2004) 200(8):967–77. doi: 10.1084/jem.20040973 75. Kräutler NJ, Yermanos A, Pedrioli A, Welten S, Lorgé D, Greczmiel U, et al. Quantitative and Qualitative Analysis of Humoral Immunity Reveals Continued and Personalized Evolution in Chronic Viral Infection. Cell Rep (2020) 30(4):997–1012.e6. doi: 10.1016/j.celrep.2019.12.088 56. Arumugakani G, Stephenson SJ, Newton DJ, Rawstron A, Emery P, Doody G, et al. Early Emergence of CD19-Negative Human Antibody-Secreting Cells at the Plasmablast to Plasma Cell Transition. J Immunol (2017) 198(12):4618– 28. doi: 10.4049/jimmunol.1501761 76. Eyer K, Castrillon C, Chenon G, Bibette J, Bruhns P, Grifﬁths A, et al. The Quantitative Assessment of the Secreted IgG Repertoire After Recall to Evaluate the Quality of Immunizations. J Immunol (2020) 205(4):1176–84. doi: 10.4049/jimmunol.2000112 57. Glass DR, Tsai AG, Oliveria JP, Hartmann F, Kimmey S, Calderon A, et al. An Integrated Multi-Omic Single-Cell Atlas of Human B Cell Identity. Immunity (2020) 53(1):217–32.e5. doi: 10.1016/J.IMMUNI.2020.06.013 77. Alanine DGW, Quinkert D, Kumarasingha R, Gilson PR, Higgins MK, Draper Correspondence SJ. Human Antibodies That Slow Erythrocyte Invasion Potentiate Malaria-Neutralizing Antibodies. Cell (2019) 178:216–228.e21. doi: 10.1016/j.cell.2019.05.025 58. Hartmann FJ, Mrdjen D, McCaffrey E, Glass D, Greenwald N, Bharadwaj A, et al. Single-Cell Metabolic Proﬁling of Human Cytotoxic T Cells. Nat Biotechnol 2020 392 (2020) 39(2):186–97. doi: 10.1038/s41587-020-0651-8 78. Viant C, Weymar GHJ, Escolano A, Cipolla M, Gazumyan A, Correspondence MCN. Antibody Afﬁnity Shapes the Choice Between Memory and Germinal Center B Cell Fates. Cell (2020) 183:1298–1311.e11. doi: 10.1016/ j.cell.2020.09.063 59. Guo M, Price MJ, Patterson DG, Barwick B, Haines R, Kania A, et al. EZH2 Represses the B Cell Transcriptional Program and Regulates Antibody- Secreting Cell Metabolism and Antibody Production. J Immunol (2018) 200 (3):1039–52. doi: 10.4049/jimmunol.1701470 79. Bornholdt ZA, Turner HL, Murin CD, Li W, Sok D, Souders CA, et al. Isolation of Potent Neutralizing Antibodies From a Survivor of the 2014 Ebola Virus Outbreak. Science (2016) 351(6277):1078–83. doi: 10.1126/ SCIENCE.AAD5788 60. Han Q, Bradshaw EM, Nilsson B, Haﬂercd DA, Love JC. Multidimensional Analysis of the Frequencies and Rates of Cytokine Secretion From Single Cells by Quantitative Microengraving. Lab Chip (2010) 10(11):1391–400. doi: 10.1039/B926849A 80. Sokal A, Barba-Spaeth G, Fernández I, Broketa M, Azzaoui I, de la Selle A, et al. REFERENCES Nguyen DC, Lewis HC, Joyner C, Warren V, Xiao H, Kissick H, et al. Extracellular Vesicles From Bone Marrow-Derived Mesenchymal Stromal Cells Support Ex Vivo Survival of Human Antibody Secreting Cells. J Extracell Vesicles (2018) 7(1):1463778. doi: 10.1080/20013078.2018.1463778 69. Gérard A, Woolfe A, Mottet G, Reichen M, Castrillon C, Menrath V, et al. High-Throughput Single-Cell Activity-Based Screening and Sequencing of Antibodies Using Droplet Microﬂuidics. Nat Biotechnol (2020) 38(6):715–21. doi: 10.1038/s41587-020-0466-7 49. Nguyen DC, Joyner CJ, Sanz I, Lee FEH. Factors Affecting Early Antibody Secreting Cell Maturation Into Long-Lived Plasma Cells. Front Immunol (2019) 10:2138. doi: 10.3389/ﬁmmu.2019.02138 70. Matutis L, Gilbert J, Ung WL, Weitz DA, Grifﬁths AD, Heyman JA. Single- Cell Analysis and Sorting Using Droplet-Based Microﬂuidics. Nat Protoc (2013) 8(5):870–91. doi: 10.1038/NPROT.2013.046 50. Auladell M, Nguyen TH, Garcillán B, Mackay F, Kedzierska K, Fox A. Distinguishing Naive- From Memory-Derived Human B Cells During Acute Responses. Clin Transl Immunol (2019) 8(11):e01090. doi: 10.1002/cti2.1090 71. Ding R, Hung KC, Mitra A, Ung L, Lightwood D, Tu R, et al. Rapid Isolation of Antigen-Speciﬁc B-Cells Using Droplet Microﬂuidics. RSC Adv (2020) 10 (45):27006–13. doi: 10.1039/d0ra04328a 51. Doucett VP, Gerhard W, Owler K, Curry D, Brown L, Baumgarth N. Enumeration and Characterization of Virus-Speciﬁc B Cells by Multicolor Flow Cytometry. J Immunol Methods (2005) 303(1-2):40–52. doi: 10.1016/ j.jim.2005.05.014 72. Shembekar N, Hu H, Eustace D, Merten CA. Single-Cell Droplet Microﬂuidic Screening for Antibodies Speciﬁcally Binding to Target Cells. Cell Rep (2018) 22(8):2206–15. doi: 10.1016/j.celrep.2018.01.071 52. Di Niro R, Mesin L, Raki M, Zheng N, Lund-Johansen F, Lundin K, et al. Rapid Generation of Rotavirus-Speciﬁc Human Monoclonal Antibodies From Small-Intestinal Mucosa. J Immunol (2010) 185(9):5377–83. doi: 10.4049/ jimmunol.1001587 73. Eyer K, Doineau RCL, Castrillon CE, Briseño-Roa L, Menrath V, Mottet G, et al. Single-Cell Deep Phenotyping of IgG-Secreting Cells for High- Resolution Immune Monitoring. Nat Biotechnol (2017) 35(10):977–82. doi: 10.1038/nbt.3964 53. Tellier J, Nutt SL. Standing Out From the Crowd: How to Identify Plasma Cells. Eur J Immunol (2017) 47(8):1276–9. doi: 10.1002/eji.201747168 74. Canales-Herrerias P, Crickx E, Broketa M, Sokal A, Chenon G, Azzaoui I, et al. Single-Cell Analyses of Immune Thrombocytopenic Patients Reveal Multiorgan Dissemination of High-Afﬁnity Autoreactive Plasma Cells. bioRxiv (2021) 2021.06.29.450355. doi: 10.1101/2021.06.29.450355 54. Fooksman DR, Schwickert TA, Victora GD, Dustin ML, Nussenzweig MC, Skokos D. Development and Migration of Plasma Cells in the Mouse Lymph Node. Immunity (2010) 33(1):118–27. doi: 10.1016/j.immuni.2010.06.015 55. REFERENCES mRNA Vaccination of Naive and COVID-19-Recovered Individuals Elicits Potent Memory B Cells That Recognize SARS-CoV-2 Variants. Immunity (2021) 54(12):2893–907.e5. doi: 10.1016/J.IMMUNI.2021.09.011 61. Zhu H, Stybayeva G, Silangcruz J, Yan J, Ramanculov E, Dandekar S, et al. Detecting Cytokine Release From Single T-Cells. Anal Chem (2009) 81 (19):8150–6. doi: 10.1021/AC901390J 62. Torres AJ, Hill AS, Love JC. Nanowell-Based Immunoassays for Measuring Single-Cell Secretion: Characterization of Transport and Surface Binding. Anal Chem (2014) 86(23):11562–9. doi: 10.1021/AC4030297 81. Williams LD, Ofek G, Schätzle S, McDaniel J, Lu X, Nicely N, et al. Potent and Broad HIV-Neutralizing Antibodies in Memory B Cells and Plasma. Sci Immunol (2017) 2(7):27. doi: 10.1126/SCIIMMUNOL.AAL2200 63. Rybczynska M, Baudry J, Klaus E. The Impact of Frost-Damage on the Quality and Quantity of the Secreted Antigen-Speciﬁc IgG Repertoire. Vaccine (2020) 38(33):5337–42. doi: 10.1016/j.vaccine.2020.05.066 82. Lee J, Boutz DR, Chromikova V, Joyce M, Vollmers C, Leung K, et al. Molecular-Level Analysis of the Serum Antibody Repertoire in Young Adults Before and After Seasonal Inﬂuenza Vaccination. Nat Med 2016 2212. (2016) 22(12):1456–64. doi: 10.1038/nm.4224 64. Winters A, McFadden K, Bergen J, Landas J, Berry K, Gonzalez A, et al. Rapid Single B Cell Antibody Discovery Using Nanopens and Structured Light. MAbs (2019) 11(6):1025. doi: 10.1080/19420862.2019.1624126 83. Rees AR. Understanding the Human Antibody Repertoire. mAbs (2020) 12: (1):e1729683. doi: 10.1080/19420862.2020.1729683 January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org Broketa and Bruhns High-Throughput Characterization of ASC 94. Lee J, Paparoditis P, Horton AP, Frühwirth A, McDaniel J, Jung J, et al. Persistent Antibody Clonotypes Dominate the Serum Response to Inﬂuenza Over Multiple Years and Repeated Vaccinations. Cell Host Microbe (2019) 25 (3):367–76.e5. doi: 10.1016/j.chom.2019.01.010 84. Lindeman I, Emerton G, Mamanova L, Snir O, Polanski K, Qiao S, et al. BraCeR: B-Cell-Receptor Reconstruction and Clonality Inference From Single-Cell RNA- Seq. Nat Methods 2018 158 (2018) 15(8):563–5. doi: 10.1038/s41592-018-0082-3 85. Rizzetto S, Koppstein DNP, Samir J, Singh M, Reed J, Cai C, et al. B-Cell Receptor Reconstruction From Single-Cell RNA-Seq With VDJPuzzle. Bioinformatics (2018) 34(16):2846–7. doi: 10.1093/BIOINFORMATICS/BTY203 95. Lindesmith LC, McDaniel JR, Changela A, Verardi R, Kerr S, Costantini V, et al. Sera Antibody Repertoire Analyses Reveal Mechanisms of Broad and Pandemic Strain Neutralizing Responses After Human Norovirus Vaccination. Immunity (2019) 50(6):1530–41.e8. doi: 10.1016/j.immuni.2019.05.007 86. Delaloy C, Schuh W, Jäck H-M, Bonaud A, Espéli M. Single Cell Resolution of Plasma Cell Fate Programming in Health and Disease. Eur J Immunol (2021). REFERENCES doi: 10.1002/EJI.202149216 96. Sajadi MM, Dashti A, Rikhtegaran Tehrani Z, Tolbert W, Seaman M, Ouyang X, et al. Identiﬁcation of Near-Pan-Neutralizing Antibodies Against HIV-1 by Deconvolution of Plasma Humoral Responses. Cell (2018) 173(7):1783– 95.e14. doi: 10.1016/j.cell.2018.03.061 87. Ise W, Kurosaki, Tomohiro, Kurosaki T. Plasma Cell Differentiation During the Germinal Center Reaction. Immunol Rev (2019) 288:64–74. doi: 10.1111/ imr.12751 88. Curtis NC, Lee J. Beyond Bulk Single-Chain Sequencing: Getting at the Whole Receptor. Curr Opin Syst Biol (2020) 24:93–9. doi: 10.1016/j.coisb.2020.10.008 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. PB is a paid consultant to Regeneron Pharmaceuticals. MB declares that he has no relevant conﬂicts of interest. 89. Setliff I, Shiakolas AR, Pilewski KA, Murji A, Mapengo R, Janowska K, et al. High-Throughput Mapping of B Cell Receptor Sequences to Antigen Speciﬁcity. Cell (2019) 179(7):1636–1646.e15. doi: 10.1016/j.cell.2019.11.003 90. Wang B, Dekosky BJ, Timm MR, Lee J, Normandin E, Misasi J, et al. Functional Interrogation and Mining of Natively Paired Human VH:VL Antibody Repertoires. Nature Biotechnol (2018) 36:152–5. doi: 10.1038/nbt.4052 Publisher’s Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. 91. Jiang R, Fichtner ML, Hoehn KB, Pham M, Stathopoulos P, Nowak R, et al. Single-Cell Repertoire Tracing Identiﬁes Rituximab-Resistant B Cells During Myasthenia Gravis Relapses. JCI Insight (2020) 5(14):e136471. doi: 10.1172/ JCI.INSIGHT.136471 Copyright © 2022 Broketa and Bruhns. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 92. Amanat F, Thapa M, Lei T, Ahmed S, Adelsberg D, Carreño J, et al. SARS- CoV-2 mRNA Vaccination Induces Functionally Diverse Antibodies to NTD, RBD, and S2. Cell (2021) 184(15):3936–48.e10. doi: 10.1016/j.cell.2021.06.005 93. January 2022 \| Volume 12 \| Article 821729 REFERENCES Snapkov I, Chernigovskaya M, Sinitcyn P, Lê Quý K, Nyman TA, Greiff V. Progress and Challenges in Mass Spectrometry-Based Analysis of Antibody Repertoires. Trends Biotechnol (2021). doi: 10.1016/j.tibtech.2021.08.006 January 2022 \| Volume 12 \| Article 821729 Frontiers in Immunology \| www.frontiersin.org 9
https://openalex.org/W2800734406	https://ejournal2.undip.ac.id/index.php/ijb/article/download/2373/1436	English	null	Value stock and growth stock on Indonesia stock exchange after global crisis	DIJB (Diponegoro International Journal of Business)	2,018	cc-by-sa	4,203	Diponegoro International Journal of Business Vol. 1, No. 1,2018, pp.8-13 pp Published by Department of Management, Faculty of Economics and Business, Universitas Diponegoro (p-ISSN: 2580-4987; e-ISSN: 2580-4995) Published online in http://ejournal2.undip.ac.id/index.php/ijb/index Abstract value stock, growth stock, returns, Sharpe ratio Keywords Correspondence to : hardjum@gmail.com MUHAMMAD FADHIL RABBANI1, HARJUM MUHARAM1 MUHAMMAD FADHIL RABBANI1, HARJUM MUHARAM1 1Department of Management, Faculty of Economics and Business, Universitas Diponegoro This study was conducted to determine whether there are differences between the stock return of value stocks and growth stock in Indonesia before and after the world financial crisis that occurred in 2008. To investigate the difference, the stocks formed into a portfolio that is based on the 2002 calculated in 2002 and 2009 when the world financial crisis has ended. The formation of the portfolio based on stocks that have gone public before 2000 and have the complete data during the study period. For the determination of the categories of stocks used Price-to-Earnings ratio, price-to-book ratio and price-to-cash flow ratio. Shares of stock that has a very high ratio will be eliminate to avoid bias that may occur if the stocks are still included. Similarly, the stocks of which are negative because they do not meet the criteria as a value stock. Then ANOVA test conducted to determine differences in returns and Sharpe ratio on the portfolio which was formed in 2002 and in 2009. Results from this studies are not found differences in returns and Sharpe ratio on both the portfolio. This indicates that the formation of the portfolio by value stocks and growth stock can not be used as a guide to get a high return. INTRODUCTION compared to it’s fundamentals (e.g. earnings, book value, cash flow, dividends) whereby growth stocks are those stocks that traded with higher prices compared it’s fundamentals. While Jenn Yaw Yen, Qian Sun and Yuxing Yan (2004) stated that value (growth) stocks described as something that has to do with the relatively low (high) market price in relation to some estimates of intrinsic value, like price to book value (P/B), price to earnings (P/E), and price to cashflow (P/C). Many people are still in doubt and confused to do an investment in capital market. However, it is easy to invest in capital market, and there are many ways to pick a stock. One of them is to make a portfolio. It means that an investor can easily spread (make a difersivication) it’s investment in any opportunity (Suad Husnan: 43). In this case investor make a diversivication to reduce their accounted risk (Suad Husnan: 45). So does Markowitz (1954) who said that a portfolio must be diversified to minimize the risk. Every investor have their own preference and different risk level that can be taken. The higher the expected profit, the bigger the risk that must be accounted. g ( ) p ( ) Many studies conclusively record that value stocks outperform growth stocks (value premium) in US and Japan. Fama and French (1998;2007) and Black and Millian (2004;2006) studied about value stocks and growth stocks with return, risk and the whole performance. The results are value stocks outperform growth stocks. Basu (1977) also shows that stocks in US with lower price to earnings (P/E) tend to have higher rate return than stocks with higher (P/E) (e.g. growth stocks). Chan, Hamao and Lakonishok (1991) also found the same thing in Japan stock market. This research is also proven by Fama and French (1992, 1993, 1996), Lakonishok, Shleifer and Vishny (1994) and Chan and Lakonishok (2004) US and Europe There are many ways to pick a stock, one of them is by classification. Often times stocks are classified in value and growth categories by investors (Jenn Yaw Yen, Qian Sun dan Yuxing Yan 2004). Value stock and growth stock was first introduced by Graham and Dodd (1934). Classifying stocks as value or growth There are many ways to classify stocks as a value stocks or growth stocks. But, there are 3 ratios that mostly used by scholars, which are price-to-earnings (P/E), price-to-book (P/B), and price-to-cashflow (P/C) or equivalent from these ratios, such as market- to-book, book-to-market, earning-to-price, and cash flow-to-price. These ratios are commonly used because that ratios create a stable results (Fama and French, 1998). This research aims to find if there is a difference between portfolio return and sharpe portfolio of value stocks and growth stocks in Indonesia stock exchange and if value stocks have a higher return from growth stocks in Indonesia stock exchange. INTRODUCTION The simple definition from value stocks and growth stocks according to Fama and French, (1993) are: value stocks are those stocks that traded with lower prices Received: 18thJanuary 2018 Revised: 11th February 2018 Accepted: 8th March 2018 8 8 Diponegoro International Journal of Business, Vol. 1, No. 1,2018, pp. 8-13 stock market, Australia and EAFE. However, the invention may also be a problem anomaly for rational expectations, because it is based on conventional wisdom, options for growth depends on the economic conditions in the future and to be more risky than the assets in place (Zhang, 2005). Graham and Dodd (1934) stated that this exaltation happen due to poor performance in the past in wich the expectation arises that this performance will continue in the future. However, poor perfomance does not have to refer in particular towards default. It could also be a signal that the company reached its maturity in which the company’s growth becomes stable and does not give any indication anymore of excessive growth. Based on history, a stock with higher book to market ratio get a higher return rate than stocks with lower book to market ratio. The difference in return rate of value and growth stocks, that commonly known as value premium, is about 6% every year, which known to give a threat as a serious challenge towards standard asset pricing model like capital asset pricing model (CAPM) (Hengjie Ai and Dana Kiku: 2013). Value premium Graham and Dodd (1994) were one of the first shcolars to acknowledge that value stocks and growth stocks are each other’s enemy. They also gave definition about those two stocks. While according to Jenn Yaw Yen, Qian Sun and Yuxing Yan (2004) value (growth) stocks described as something that has to do with the relatively its low (high) market price in relation to some estimates of intrinsic value, like price to book value (P/B), price to earnings (P/E), dan price to cashflow (P/C). Value premium or value-growth spread exist when value stocks outperform growth stocks in certain condition (Capaul et al, 1993). This value premium is very important since the results lead to whether investors are more confidence to buy value or growth stocks. The higher the value premium, the more likely it is that investors to choose value stocks due to the higher returns compared to growth stocks (Bird and Casavvechia, 2007). When this figure is between zero, it will indicate an ignorance in buying value or growth stocks. When this figure is below zero, it will indicate the existence of discount value, which is growth stocks give a higher return from value stocks. When value premium is significantly and substantially higher than market return, then a potential bubble is shaped (Brown et al, 2008). Growth stocks Graham and Dodd (1934) defined growth stocks as a stock that traded in a relative high price compared to it’s fundamental. Growth stocks can be characterized as a stocks with expected return and growth substantially higher than the market average and a probability to rise further (Bourguignon and De Jong, 2003). Investors believe that this rise are known as growth (or glamour) stocks. A global crisis that occured in 2008 weakened all the capital market in the world. But Indonesia can overcome it very well so that IHSG price not declining too many. Strangely, IHSG price rise and reach the second highest in ASEAN after previously in third position. So it is possible to affect the value of investor’s stock portfolio. Value stocks Value stocks are stocks with it’s price to earnings (P/E), price to book (P/B), and or price to cashflow (P/C) are lower compared to market average (Graham and Dodd 1934; Fama and French, 1998; Chan and Lakonishok, 2004; Athanassakos, 2009). 9 Diponegoro International Journal of Business, Vol. 1, No. 1,2018, pp. 8-13 Portfolio theory Markowitz (1954) said that a portfolio must be diversified to reduce the risk. Ross (1976) in APT (Arbitrage Pricing Theory) said that if an investment have the same characteristic cannot be sold with a different price, if those investments sold with different price then there will be a chance to get return or profit without risk. H3: There are differences in portfolio sharpe ratio value stocks and growth stocks on Indonesia Stock Exchange during 2002-2015. Sharpe ratio Capaul et al (1993) stated that sharpe ratio is a measure of the reward obtained per unit risk. For an investment, reward is measured by the average excess return (return minus riskless rate) and risk by the standard deviation of excess returns. cess etu s RI is the average rate of return for portfolio i during a specified time period, RFR is the average rate of return on risk-free assets during the same time period, this research use BI rate as risk-free rate and i as the standard deviation of the rate of return for portfolio i during the time period. H1: There are differences in portfolio return between value stocks and growth stocks in Indonesia Stock Exchange during 2002-2015. H2: Value stocks has a bigger return than growth stocks in Indonesia stock Exchange during 2002-2015. The higher the sharpe ratio, the better and higher the performance and return invesment. In every case, the value index provided the best results (Capau et al, 1993). Because of that, this research will evaluate whether value stocks in Indonesia have a higher sharpe ratio than growth stocks on Indonesia Stock Exchange during the time period of research. The next two hypotheses are: Value premium in emerging market that already happen called actual return. While expected return is a return that expected. This return is used to determine risk in the future. This two return used by investors as a tool to forecast stock return that they owned (Halim, 2003). Fama and French (1998) analyzed possible value premiums in 16 emerging markets. From the observation, found evidence of a value premium that was remarkably high (14,13 percent) compared to developed international markets. Huang Yang (2008) also observed a positive value premiums in the China stok market. Brown et al (2008) who researched emerging market in Asia record an existence of value premium in Hong Kong (0,72 percent), Korea (0,42 percent) and Singapore (0,42 percent) but a value discount in Taiwan (1,26 percent). ed ( a , 003) ed ( a , 003) Ri is the return for stock i, P1 is price for time 1, P0 is price for time 0, D1 is dividend for time 1. To calculate portfolio, return, this research use this formula: Rpy = i Ri py Rpy is the monthly portfolio return in month y, wi is the individual weight of a stock in a portfolio, Ri is the return of stock i. At first, IHSG is kown as a low index in ASEAN, however since 2004 IHSG grow exceeding Thailand and Malaysia, even higher than Singapore after the financial crisis in 2008 till now. An interesting phenomenom where IHSG grow dramatically after the financial crisis. It can be concluded that Indonesia stock market also have many value and growth stocks and very interesting to investors. So does a research about value premium phenomenom in Indonesia stock exchange. That is why two hypotheses can be made: Variables The price-earnings or P/E is the ratio of the current stock price to last year’s earnings per share. P/E ratio tells us how much stock purchasers must pay per dollar of earnings that the firm generates (Bodie et al. 2011). The lower (higher) rates of P/E give the perception that the expectation on future earnings will also be lower (higher) (Bodie et al, 2009). consequently, stocks with a low P/E ratio are categorized as value stocks and stocks with a high P/E ratio are categorized as growth stocks. P/E = Where Py is the daily average closing price of a company’s stock in fiscal year y, NOCFf is net operating cash flowat FYE, and TS is total shares. Where P y is the daily average closing price of a company’s stock in a fiscal year y, EPSf is Earning per share at fical year end (FYE)f. Price-to-book ratio or P/B is a fianncial ratio used to compare a copany’s current market price to its book value. It is also sometimes known as a Book-to-Market ratio (Wikipedia). Fama and French (1998) have used this ratio to separate value and growth stocks. A higher (lower) market price of a stock gives an indication that investors have assigned additional (no) value to a company (Bodie et al, 2009). Stocks with a low P/B ratio are categorized as value stocks and stocks with a high P/B ratio are categorized as growth stocks. Stocks and portfolio return H4 : Value stocks have a higher sharpe ratio than growth stocks on Indonesia Stock Exchange during 2002-2015. Stock return is a result gained from invesment or rate of profit gained by investor from an investment that have been done (Hartono 2000: 107). Many investor make an investment to get a maximum return. Stock return is divided into two, which are return 10 Diponegoro International Journal of Business, Vol. 1, No. 1,2018, pp. 8-13 Tabel 1. Post Hoc Test for portfolio 2002 Based on (I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Tukey HSD Sig. Bonferroni Return P/E 2002 value Growth -42,37833 51,25083 0,6933 1,000 P/B 2002 value Growth -18,97167 63,70368 0,952 1,000 P/C 2002 value Growth 9,50167 64,48845 0,988 1,000 Sharpe ratio P/E 2002 value Growth -103,93167 129,16255 0,706 1,000 P/B 2002 value Growth 1,28833 86,39743 1,000 1,000 P/C 2002 value Growth 87,76333 87,33983 0,585 0,933 METHODS Variables The price-earnings or P/E is the ratio of the current stock price to last year’s earnings per share. P/E ratio tells us how much stock purchasers must pay per dollar of earnings that the firm generates (Bodie et al. 2011). The lower (higher) rates of P/E give the perception that the expectation on future earnings will also be lower (higher) (Bodie et al, 2009). consequently, stocks with a low P/E ratio are categorized as value stocks and stocks with a high P/E ratio are categorized as growth stocks. P/E = Price-to-cash flow ratio is used by investors to evaluate the invesment attractiveness from a standpoint of a company’s stock. P/C ratio compares the stock’s market price to the amount of cash flow the company generates on a per-share basis. Chan and Lakonishok (2004) argue that P/C has become extremelypopular to classify value and growth stocks since it views the company’s performance from a different point of cash in- and outflows as compared to earnings. ). Stocks with a low P/C ratio are categorized as value stocks and stocks with a high P/C ratio are categorized as growth stocks. P/C = Where Py is the daily average closing price of Tabel 1. Post Hoc Test for portfolio 2002 Based on (I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Tukey HSD Sig. Stocks and portfolio return Bonferroni Return P/E 2002 value Growth -42,37833 51,25083 0,6933 1,000 P/B 2002 value Growth -18,97167 63,70368 0,952 1,000 P/C 2002 value Growth 9,50167 64,48845 0,988 1,000 Sharpe ratio P/E 2002 value Growth -103,93167 129,16255 0,706 1,000 P/B 2002 value Growth 1,28833 86,39743 1,000 1,000 P/C 2002 value Growth 87,76333 87,33983 0,585 0,933 Tabel 1. Post Hoc Test for portfolio 2002 Price-to-cash flow ratio is used by investors to evaluate the invesment attractiveness from a standpoint of a company’s stock. P/C ratio compares the stock’s market price to the amount of cash flow the company generates on a per-share basis. Chan and Lakonishok (2004) argue that P/C has become extremelypopular to classify value and growth stocks since it views the company’s performance from a different point of cash in- and outflows as compared to earnings. ). Stocks with a low P/C ratio are categorized as value stocks and stocks with a high P/C ratio are categorized as growth stocks. RESULTS AND ANALYSIS growth is not significant in 0,05 (p<0,05), it means that H0 is accepted and H1 is rejected. Then statistically there is no difference in portfolio return between all classifications in Indonesia, eventhought in some of the classifications found that there are some categories with positive results, during research period time for portfolio construction based on P/E, P/B and P/C in year 2002 and 2009. Therefore, first hypothesis is rejected. This research used ANOVA and conducted four tests, which are: test of homogeneity variance, test of between subjects, post hoc test and homogenous subset. It was used to test relationship between one independent variabel with one or more other independent variabels (Ghozali, 2013). ( ) Variable used in this research are Price to Earning (P/E) ratio, Price to Book (P/B) ratio, and Price to cash flow (P/C) ratio that used to separate value and growth stock and return portfolio. P/E ratio is a ratio from current stock price with Earning Per Share. P/E ratio give information about how much stock buyer must pay for every dollar revenue that produced by the company. High (low) P/E give a perception about revenue in the future is also high (low) (Bodie et. al, 2011). P/B ratio is a financial ratio that commonly used to compare current stock market price of a company with it’s book value. High (low) P/B ratio can be explained that the company is overvalued (undervalued). P/C ratio is a ratio which used to compare a company's market value to its own cash flow. This ratio is also used to measure the market prospects of the company's assets in the future from financial point of view (Bragg, 2007). value (growth) stock categorization is based from P/E, P/B, and P/C ratio, where a stock with the lowest (highest) ratio will be categorized as value (growth) stock. Second hypothesis stated that value stocks have a bigger return than growth stocks on Indonesia stock exchange during 2002-2015. However, result from mean difference shows that not every return rate disparity between value stocks with growth stocks have a positive score. So does with the result from research above stated not significant in 0,05 (p<0,05). Hence, it can be concluded that H0 is accepted and H2 is rejected. Sampling Population in this research is all companies that listed in Indonesia Stock Exchange during 2001-2015. Samples taken in this research are stocks that already listed in Indonesia stock exchange since 2000 and can provide data that needed completely during 2002-2015. For companies that are new-listed after year 2000 or delisted between those period will be excluded. So does a financial institutions will be excluded because it can trigger biases when making a decision about value premium because of it’s leverage and financial multiples are not equally the same as for non financial institutions (Fama and French (1993). P/B ratio = Where Py is the daily average closing price of a company’s stock in fiscal year y, TAf is total assets at FYE f, IAf is intangible assets at FYE f, adn TLf is total liabilities at FYE f. 11 Diponegoro International Journal of Business, Vol. 1, No. 1,2018, pp. 8-13 Table 2. Post Hoc Test for portfolio 2009 Based on (I) Category (J) Category Mean Difference (I-J) Std. Error Sig. Tukey HSD Sig. Bonferroni Return P/E 2009 value Growth -16,15333 23,51092 0,774 1,000 P/B 2009 value Growth 32,81000 18,78367 0,211 0,303 P/C 2009 value Growth 11,85167 23,87833 0,874 1,000 Sharpe ratio P/E 2009 value Growth -30,74333 50,07911 0,815 1,000 P/B 2009 value Growth 45,92667 70,66577 0,795 1,000 P/C 2009 value Growth -12,16333 39,23126 0,949 1,000 Table 2. Post Hoc Test for portfolio 2009 CONCLUSION The majority result from Post Hoc test are more than 0,05 which means that there is no difference in return portfolio in all categories. This means that H1 which stated that there is a difference between portfolio return in Indonesia from year 2002 to 2015 rejected. So does with the majority results from Post Hoc test are more than 0,05 which means that there is no difference in sharpe ratio portfolio in all caegories. This means that H2 which stated that there is a difference between sharpe ratio in Indonesia from year 2003 to 2015 rejected. Cronqvist, H., Siegel, S. & Yu, F. (2015). Value versus growth investing: Why do different investors have different styles?. Journal of Financial Economics, Vol. 117, p. 333-349. Financial Economics, Vol. 117, p. 333-349. Fama, E.F. & French, K.R. (2007). The anatomy of value and growth stock returns, Financial Analysis Journal, Vol. 63 (6), p. 44-54. Ghozali, Imam. (2013). Aplikasi Analisis Multivariate dengan Program SPSS 20. Semarang: BP UNDIP. g Graham, B. & Dodd, D.L. (1934). Security Analysis NY, United States. McGraw Hill New York. Hoekjan, R.M. (2011). The Performance of Value VS Growth Stocks during The Financial Crisis. Turkey: Master Thesis University of Twente. Although the result from hypothesis test mathematically support the hypothesis, however, statistically based on ANOVA test, there is no meaningful difference in return and sharpe ratio in this research. Based on statistic data, it can be concluded that return from each stock in portfolio are not affected by classifications (growth, medium dan value) and those classifications cannot explain return from each stock in portfolio. And every growth, medium and value classification statistically cannot become a foundation to make a portfolio. So does a risk that measured with sharpe ratio for value stocks does not have a difference with growth stocks in Indonesia. Huang, Y. & Yang, J. (2008). Value premium in the Chinese stock market: free lunch or paid lunch?, working paper, p. 1-31. Husnan S. (2003). Dasar-Dasar Teori Portofolio dan Analisis Sekuritas. Yogyakarta: AMP YKPN. In, F., Kim, S. & Gencay, R. Investment horizon effect on asset allocation between value and growth strategies. Journal Economy Modelling, Vol 28, p. 1489-1497. Olin, Tomi. (2011). Value Investing in the Finnish Stock Market. Master Thesis Aalto University School of Economics. Santoso, Singgih. (2015). Menguasai SPSS 22 From Basic to Expert Skills. Jakarta: PT Elex Media Komputindo. Our research has some limitations. REFERENCES Athanassakos, G. (2009). Value versus growth stock returns and the value premium: the Canadian experience 1985-2005. Canadian journal of Administrative Sciences, Vol. 26, p. 109-121. Bird, R. & Casavecchia, L. (2007). Sentiment and financial health indicators for value and growth stocks: The European experience. The European Journal of Finance, Vol. 13 (8), p. 769-793. RESULTS AND ANALYSIS Third hypothesis stated that there are differences in sharpe ratio portfolio of value stocks and growth stocks on Indonesia Stock Exchange during 2002-2015.result from However, test above shows that the difference between sharpe ratio classification of value and growth is not significant in 0,05 (p>0,05), it means that H0 is accepted and H3 is rejected. Afterwards, statistically there is no difference in sharpe ratio between all classifications in Indonesia, eventhought in some of the classifications found that there are some categories with positive results, during research period time for portfolio construction based on P/E, P/B and P/C from 2002 to 2009. Therefore, third hypothesis is rejected. In the first hypothesis stated that there are differences in portfolio return of value stocks and growth stocks on Indonesia Stock Exchange during 2002-2015. However, result from test above shows that the difference between classifications return of value and 12 Diponegoro International Journal of Business, Vol. ..., No. ..., ... Year, pp. ...-... Fourth hypothesis stated that value stocks have a bigger sharpe ratio than growth stocks on Indonesia Stock Exchange during 2002-2015. However, results from mean difference shows that not every disparity sharpe ratio rate between value stocks with growth stocks have a positive score. As well as the result from research above stated not significant in 0,05 (p<0,05). Hence, it can be concluded that H0 is accepted and H4 is rejected. a higher return in the future. For future research, it is expected tha researcher increase variabel, sampel size, other measurements and countries than Indonesia CONCLUSION A small sample caused by limited data. Independent variabel consist only return and sharpe ratio. Then the risk measurement using only Sharpe Ratio. The Brandes Institutes. (2012). Value vs. Glamour: A Global Phenomenon. Sandiego: W ld Th b d I i The Brandes Institutes. (2012). Value vs. Glamour: A Global Phenomenon. Sandiego: Worldscope: The brandes Institutes. ( ) A Global Phenomenon. Worldscope: The brandes Institutes. Yen, J.Y., Sun, Q., & Yan, Y. (2004). Value versus growth stocks in Singapore. Journal of Multi Finance Management,Vol 14, p. 19-34. Based on the research above, it is found no difference between value and growth stock in Indonesia during research period time. Hence, classification of stock into value and growth stock cannot become a guidance for investor to create a portfolio for achieving Yesica, E., (2014). Value vs growth stocks returns on the Indonesia stock exchange study to companies listed on Indonesia Stock Exchange 2003-2013 period. Semarang: Universitas Diponegoro. 13
https://openalex.org/W4324385712	https://www.mdpi.com/2571-6255/6/3/122/pdf?version=1679038508	English	null	Effects of Rossby Waves Breaking and Atmospheric Blocking Formation on the Extreme Forest Fire and Floods in Eastern Siberia 2019	Fire	2,023	cc-by	20,514	Article Effects of Rossby Waves Breaking and Atmospheric Blocking Formation on the Extreme Forest Fire and Floods in Eastern Siberia 2019 Olga Yu. Antokhina 1, Pavel N. Antokhin 1 , Boris D. Belan 1 , Alexander V. Gochakov 1,2 , Yuliya V. Martynova 3 , Konstantin N. Pustovalov 3,4 , Lena D. Tarabukina 5 and Elena V. Devyatova 6,* Olga Yu. Antokhina 1, Pavel N. Antokhin 1 , Boris D. Belan 1 , Alexander V. Gochakov 1,2 , Yuliya V. Martynova 3 , Konstantin N. Pustovalov 3,4 , Lena D. Tarabukina 5 and Elena V. Devyatova 6,* 1 V.E. Zuev Institute of Atmospheric Optics of SB RAS, Tomsk 663055, Russia; antokhina@iao.ru (O.Y.A.) 2 Siberian Regional Hydrometeorological Research Institute, Novosibirsk 630099, Russia 3 Institute of Monitoring of Climatic and Ecological Systems of SB RAS, Tomsk 663055, Russia 4 Department of Meteorology and Climatology, National Research Tomsk State University, Tomsk 634028 Russia 5 Yu.G. Shafer Institute of Cosmophysical Research and Aeronomy, SB RAS, Yakutsk 677980, Russia 6 Institute of Solar-Terrestrial Physics of SB RAS, Irkutsk 664033, Russia * Correspondence: devyatova@iszf.irk.ru; Tel.: +7-914-876-2328 Abstract: In 2019, the southern region of Eastern Siberia (located between 45◦N and 60◦N) experi- enced heavy ﬂoods, while the northern region (between 60◦N and 75◦N) saw intense forest ﬁres that lasted for almost the entire summer, from 25 June to 12 August. To investigate the causes of these natural disasters, we analyzed the large-scale features of atmospheric circulation, speciﬁcally the Rossby wave breaking and atmospheric blocking events. In the summer of 2019, two types of Rossby wave breaking were observed: a cyclonic type, with a wave breaking over Siberia from the east (110◦E–115◦E), and an anticyclonic type, with a wave breaking over Siberia from the west (75◦E–90◦E). The sequence of the Rossby wave breaking and extreme weather events in summer, 2019 are as follows: 24–26 June (cyclonic type, extreme precipitation, ﬂood), 28–29 June and 1–2 July (anticyclonic type, forest ﬁres), 14–17 July (both types of breaking, forest ﬁres), 25–28 July (cyclonic type, extreme precipitation, ﬂood), 2 and 7 August (anticyclonic type, forest ﬁres). Rossby wave breaking occurred three times, resulting in the formation and maintenance of atmospheric blocking over Eastern Siberia: 26 June–3 July, 12–21 July and 4–10 August. In general, the scenario of the summer events was as follows: cyclonic Rossby wave breaking over the southern part of Eastern Siberia (45◦N–60◦N) caused extreme precipitation (ﬂoods) and led to low gradients of potential vorticity and potential temperature in the west and east of Lake Baikal. Citation: Antokhina, O.Y.; Antokhin, P.N.; Belan, B.D.; Gochakov, A.V.; Martynova, Y.V.; Pustovalov, K.N.; Tarabukina, L.D.; Devyatova, E.V. Effects of Rossby Waves Breaking and Atmospheric Blocking Formation on the Extreme Forest Fire and Floods in Eastern Siberia 2019. Fire 2023, 6, 122. https://doi.org/10.3390/ﬁre6030122 Academic Editors: Grant Williamson and John Abatzoglou Received: 18 January 2023 Revised: 3 March 2023 Accepted: 10 March 2023 Published: 15 March 2023 Keywords: forest ﬁres; precipitation; Siberia; Rossby wave breaking; atmospheric blocking; wave activity ﬂux; temperature Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Article Effects of Rossby Waves Breaking and Atmospheric Blocking Formation on the Extreme Forest Fire and Floods in Eastern Siberia 2019 The increased wave activity ﬂux from the Europe–North Atlantic sector caused the anticyclonic-type Rossby wave breaking to occur west of the area of a low potential vorticity gradient and north of 60◦N. This, in turn, contributed to the maintenance of blocking anticyclones in the north of Eastern Siberia, which led to the intensiﬁcation and expansion of the area of forest ﬁres. These events were preceded by an increase in the amplitude of the quasi-stationary wave structure over the North Atlantic and Europe during the ﬁrst half of June. fire fire fire fire 1. Introduction g g g In the summer of 2019, Eastern Siberia (ES) experienced record-breaking floods with peaks at the end of June and the end of July, caused by extreme rainfall [10,11], https://tass.com/floods-in-irkutsk-region, accessed on 11 March 2023, and extreme, long-duration forest fires from the end of June to mid-August caused by dry thunder- storms and high air temperature [https://tass.ru/proisshestviya/6703544, accessed on 11 March 2023, https://www.nasa.gov/image-feature/goddard/2019/huge-wildfires- in-russias-siberian-province-continue, accessed on 11 March 2023]. The search for the large- and synoptic-scale atmospheric conditions that contributed to the occurrence of both forest fires and floods in Siberia in 2019 is crucial in order to better understand and potentially predict such weather extremes. Many scientific studies have shown that extreme weather events are caused by upper tropospheric ridges and troughs associated with the propagation, stationarity, and breaking of Rossby waves [12–19]. According to the findings of Moore et al. 2019 [20] and Liu 2017 [21], Rossby wave breaking (RWB) plays a crucial role in both horizontal and vertical large-scale transport and mixing, as demonstrated in observations and idealized general circulation models. The most ex- traordinary example demonstrating how the same large-scale event (RWB) caused both wildfire and flood is the atmospheric pattern in July–August 2010, the Rossby wave breaking caused the Russian heatwave (fire) and Pakistan flood [22]. Of course, high summer rainfall and floods in continental extratropics are frequently related to regional convective storms [23]. However, regional storms are often also forced by synoptic- and large-scale atmospheric dynamics. Thus, in papers [24,25] the importance of cyclones and their associated frontal systems, for the occurrence of regional-scale precipitation extremes is quantified. The current state of knowledge regarding large-scale meteorological pat- terns associated with short-duration (less than 1 week) extreme precipitation events over North America is considered in [26]. Bosart and Moore, 2017 [27] highlight how the large- and synoptic-scale flow can evolve to facilitate multiple geographically separated but dynamically linked extreme weather events in North America in October 2007. The study presented in [28] analyzed precipitation events in the Selenga river basin and atmospheric blocking over Eurasia during July from 1979 to 2017. The results showed that when there was joint blocking over Europe and the Russian Far East (RFE), it led to aridity over the southern (Mongolian) part of the Selenga basin and increased precipitation over the northern (Russian) part of the basin. 1. Introduction The increase in weather extremes remains a crucial question linked to global climate change [1,2]. The air temperature and precipitation are among the most critical climate indicators, and they are closely linked to extreme weather events such as droughts, wildﬁres, and ﬂoods during boreal summer. According to research conducted by Groisman et al. in 2017 [2], numerous studies have indicated a rise in precipitation intensity in Northern https://www.mdpi.com/journal/ﬁre Fire 2023, 6, 122. https://doi.org/10.3390/ﬁre6030122 Fire 2023, 6, 122 2 of 25 Eurasia. Additionally, extended periods of no rainfall have been accompanied by summer droughts and an increase in unusual temperature patterns. These changes have led to an increase in the occurrence of forest ﬁres and ﬂoods [3,4]. Both ﬁres and ﬂoods are extremely dangerous for the Russian economy and human health. Eurasia. Additionally, extended periods of no rainfall have been accompanied by summer droughts and an increase in unusual temperature patterns. These changes have led to an increase in the occurrence of forest ﬁres and ﬂoods [3,4]. Both ﬁres and ﬂoods are extremely dangerous for the Russian economy and human health. Furthermore, forest ﬁres may affect regional air quality and human health and feed- back processes between the climate and the biosphere due to the emission of atmospheric carbon dioxide and aerosols [5–7]. In the past, the extent of Siberia’s boreal ﬁres was un- derestimated in terms of their contribution to global ﬁre emissions. Currently, the Siberian forest ﬁres have been the focus of many scientiﬁc papers, recognized as one of the most dramatic phenomena. Soja et al. 2004 [6] showed that boreal ﬁre is signiﬁcant to the global carbon budget. Siberia’s boreal forests are estimated to contain two-thirds of the world’s total boreal forests [5]. According to [3], more than 70% and up to 90% of the total area burned in Russia occurs in Siberia. The Eastern Siberian and Far Eastern regions are partic- ularly prone to intense forest ﬁres [8], which can be hazardous due to their contribution to Arctic ice melting through the settling of black carbon [9]. 1. Introduction p Li and Ruan, 2018 [29] and Xu et al., 2019 [30] suggested two teleconnections over northern Eurasia between North Atlantic and the Eurasian continent in the summertime. The ﬁrst pattern termed the Atlantic–Eurasian (AEA) teleconnection [29] has ﬁve action centers in the middle troposphere: subtropical North Atlantic Ocean, Eastern Europe, Mongolia–north China, the northeastern North Atlantic Ocean, and the Kara Sea–Northern Siberia. According to Li and Ruan (2018) [29], the AEA is a large-scale Rossby wave train that originates in the subtropical North Atlantic Ocean. The second pattern, the Fire 2023, 6, 122 3 of 25 British–Baikal Corridor (BBC), proposed in [30], consists of four geographically ﬁxed centers located in the upper troposphere: over the west of the British Isles, the Baltic Sea, western Siberia, and Lake Baikal. The authors of [30] suggest that the BBC pattern is related to the summertime upper-tropospheric polar front jet. Here, we focus on these patterns as potential sources of wave energy that may contribute to the breaking of Rossby waves and the formation of blockings over Siberia. The variability of the Asian summer monsoon anticyclone (ASMA) has been a recent focus of research, as reported in [31]. The ASMA is located between the subtropical westerly jet to the north and easterly jets to the south. Therefore, it can be viewed as a potential source of excitation or modulation of Rossby waves, which could impact the atmospheric circulation over Siberia during the summer season. The changes in the trend of RWB in the Northern hemisphere during the last four decades were revealed by Bowley et al. 2019 [32] and Jing and Banerjee, 2018 [33]. Jing and Banerjee, 2018 [33] discovered that the frequency of both types of breaking (AWB and CWB) has increased since 1981 above 320 K, and their mean latitude has shifted poleward. They also revealed that such changes in AWB frequency and latitudinal area are more signiﬁcant in summer than in winter [33]. Bowley and his co-authors in their work [32] suggest that the increase in the frequency of AWB in summer is likely due to the Asian monsoon. The identiﬁed trends in RWB may indicate a more frequent occurrence of blockings in Siberia during the summer, and a poleward shift of these events. g p The RWBs are typically studied using isentropic potential vorticity (PV) as a diag- nostic tool [34–36]. 1. Introduction The rapid and irreversible deformation of PV contours is observed during the amplification and breaking of Rossby waves [37]. The RWB is often identi- fied by the reversal of the latitudinal gradient of PV on isentropic surfaces (or potential temperature on the dynamical tropopause—PV-Θ) [38,39]. According to the direction of PV/PV-Θ contour deformation, the types of RWB are divided into cyclonic and an- ticyclonic (CWB and AWB) [21,37,40,41]. Several studies have demonstrated that RWB plays a crucial role in the formation and persistence of blocking patterns, including in the Siberian region [15,38,42,43]. Additionally, Chyi et al. 2019 [43] reported that the frequency of blocking and RWB in Siberia is high during summer, particularly in late July. This phenomenon can be explained by the “Northward jump of the Asian jet stream” in summer [44]. It is known that midlatitude circulation is predominant for the boreal forest area. In turn, mid-latitude circulation is characterized by the strong dependence of surface temperature on circulation patterns (cyclones, anticyclones, blocking, high amplitude ridges, and troughs). In discussions related to heatwaves and long-lasting droughts in mid-latitude regions, particular attention is paid to the propagation of Rossby waves (RWP) and their breaking, as well as atmospheric blocking (AB) in the middle and upper troposphere [13,17,45–48]. The positive feedback between heat waves associated with RW/AB and soil moisture has also been discussed [13,49,50]. The effect of RW on forest ﬁres in both the northern and southern hemispheres was discussed by Hayasaka et al. 2019 [51] and Reeder et al. 2015 [52]. Both papers have concluded that forest ﬁre occurrences are associated with the presence of warm and dry air masses, which can be facilitated by the propagation of the Rossby waves [52] and the meandering of westerly ﬂow [51]. p p g y g y Chyi et al. 2019 [43] showed that both AWB and CWB types are associated with precipitation in Central Siberia. They demonstrated the dynamic processes for AWB and CWB events and how they lead to different precipitation patterns in the region. The deepening of the trough from the sub-Arctic region was found to be associated with both types of breaking by Chyi et al. 2019 [43]. Antokhina et al. 2019 [53] found a statistically signiﬁcant relationship between blocking frequency and precipitation in Siberia during July. It has been shown that the north–south precipitation anomalies dipole is associated with atmospheric blocking. 2. Data and Methods 2.1. Data The scale and intensity of forest ﬁres, as well as the background air quality, were estimated using carbon monoxide emissions from biomass burning, wildﬁre hotspot data, aerosol concentrations recorded from 1 June to 31 August 2019. g Carbon monoxide emission (CObb) from biomass burning data was obtained from the global ﬁre assimilation system (GFAS) [55] (CAMS global ﬁre assimilation system: http://apps.ecmwf.int/datasets/data/cams-gfas/, accessed on 12 December 2021). GFAS is based on the ﬁre radiative power (FRP) from the MODIS instrument onboard the National Aeronautics and Space Administration (NASA) Terra and Aqua satellites (daily averaged FRP with 0.1◦resolution) [55]. The clustering procedure was applied to gridded GFAS data. The DBSCAN (density-based spatial clustering of applications with noise) clustering algorithm was used [56]. We use the following parameters: Eps = 0.6 grad. (“the maximum distance between two samples for one to be considered as in the neighborhood of the other”, https://scikit-learn.org, accessed on 12 December 2021) and MinPts = 7 (“the number of samples (or total weight) in a neighborhood for a point to be considered as a core point. This includes the point itself”, https://scikit-learn.org, accessed on 12 December 2021). p p g For analyzing wildﬁre hotspots, the ﬁre information for resource management sys- tem (FIRMS) (https://modaps.modaps.eosdis.nasa.gov/services/about/products/c6-nrt/ MOD14.html, accessed on 12 December 2021) and the hotspots visualizer Worldview based on satellite images (http://worldview.earthdata.nasa.gov, accessed on 12 December 2021) were used. The hotspots data in FIRMS [57,58] are based on “Fires and Thermal Anoma- lies” (MOD14/MYD14) obtained by the spectroradiometer MODIS on Channels 4 and 11 µm from the Terra and Aqua satellites. The algorithm for hotspot detection is based on recognizing thermal anomalies on the surface. g g To demonstrate the effect of forest ﬁres on the background air quality, we utilized aerosol data with a diameter of 0.25 µm, which was derived from the Fonovaya station [59], http://lop.iao.ru/EN/fon/diffbat/, accessed on 12 December 2021, https://peexhq.home. blog/2019/09/13/siberian-aerosol-measurements-at-fonovaya-station/, accessed on 12 December 2021 located in the border between Western and Eastern Siberia. For the analysis of the daily number of lightning strokes, we used data obtained from the Worldwide Lightning Location Network [60], WWLLN, http://wwlln.net/, ac- cessed on 12 December 2021 with horizontal resolution 0.25◦× 0.25◦. The WWLLN records approximately 10–20% of lightning strokes that have the highest charge, which are often associated with positive “cloud-ground” lightning, including so-called “dry lightning”. 1. Introduction It was explained by the RWB direction and the blocking formation. During RWB, cold air masses (high PV) are advected to the south, while warm (low PV) air masses are advected to the north. Hence, the RWB and blocking formation can promote Fire 2023, 6, 122 4 of 25 the stable vertical anticyclone structure in north Siberia (forest ﬁre) and the unstable cutoff cyclone structure in the south (ﬂood). The process associated with extreme precipitation is also related to the mixing of air masses from the high-latitude regions of the stratosphere into the low-latitude regions of the troposphere and vice versa, due to RWB [34,54]. The aim of this work is to conduct a process-oriented analysis and evaluation of the Rossby wave propagation, their breaking, blocking formation, ﬂoods, and forest ﬁres in Siberia during the summer of 2019. First, we present the chronology for all processes in June–July–August 2019. Second, we estimate air masses’ properties (horizontal exchange) and dynamical tropopause properties (instability, vertical mixing) for RWB events. 2.2. Atmospheric Blocking (AB, Blocks) There are several ways to detect blocking. We use the blocking index as the Barriopedro et al. 2006 [65]. This index is based on the GHGS criterion proposed by Lejenäs and Økland 1980 [66] and the GHGN criterion suggested by Tibaldi and Molteni 1991 [67]. The 500 hPa geopotential height gradients (GHG) north and south (GHGN and GHGS, respectively) are calculated. Barriopedro et al. 2006 [65] used the ﬁve values of ∆. GHGS = Z(ϕ0) −Z(ϕs) ϕ0 −ϕs (1) GHGN = Z(ϕn) −Z(ϕ0) ϕn −ϕ0 (2) (1) (2) potential height, ϕn= 77.5◦N + ∆, ϕ0= 60◦N + ∆, ϕs = 40◦N ± ∆, ∆= −5.0◦, −2.5◦, 0◦, 2.5◦or 5.0◦. Z −500 hPa geopotential height, ϕn= 77.5◦N + ∆, ϕ0= 60◦N + ∆, ϕs = 40◦N ± ∆, ∆= −5.0◦, −2.5◦, 0◦, 2.5◦or 5.0◦. A longitude is considered blocked when GHGS > 0, GHGN < 10 m/deg for at least one of the ﬁve ∆values. 2. Data and Methods 2.1. Data Therefore, for the study of forest ﬁres, the WWLLN data is appropriate for lightning analysis. For precipitation analysis, we used the ﬁrst guess daily product with a spatial resolution of 1◦from the global precipitation climatology centre (Deutscher Wet- terdienst) (GPCC) [61], https://opendata.dwd.de/climate_environment/GPCC/html/ gpcc_ﬁrstguess_daily_doi_download.html, accessed on 12 December 2021 from 1 June to 31 August. The GPCC ﬁrst guess is particularly useful for monitoring extreme weather events [61,62]. This set represents ground observations of daily precipitation derived from the quality-controlled stations. This was important for the study because the paper focuses Fire 2023, 6, 122 5 of 25 on extreme precipitation. The GPCC dataset is in good agreement with the weather station data for the Siberia area [63]. Four synoptic hour (0, 6, 12, 18 UTC) data (potential temperature at the 2 potential vorticity units (PVU) level (dynamic tropopause), potential vorticity at 350 K level, geopo- tential height, temperature, u,v wind components at 500 hPa, 10 m temperature and u,v wind components, convective available potential energy and total column water) from the ECMWF ERA-Interim reanalysis datasets [64], https://apps.ecmwf.int/datasets/data/ interim-full-daily/levtype=sfc/, accessed on 12 December 2021 were used in this study for 1 June–31 August 2019. The horizontal resolution is 1.5◦× 1.5◦and 2.5◦× 2.5◦. 3. Results 3.1. The Scenario RWB, Blocks, Precipitation, and Forest Fires with a Synoptic View 3.1. The Scenario RWB, Blocks, Precipitation, and Forest Fires with a Synoptic View In order to conduct a process-oriented analysis and evaluation of the Rossby wave propagation, their breaking, blocking formation, ﬂoods, and forest ﬁres in Siberia during the summer of 2019 we started our study by tracing the chronology for all the characteristics we have that describe the event. The results are shown in Figures 1–3 and in Table S1 (in Supplementary Materials). Figure 1. (a) Total CO emission (90◦E–120◦E, 55◦N–65◦N) (kg/s), A0.25 aerosol concertation in Western Siberia (Fonovaya) (N), HS—the total number of hotspots (90◦E–120◦E, 55◦N–65◦N); the total number of the lightning strikes (LS); WS—surface wind speed; tcw—total column water; precip— atmospheric precipitation. tcw, ws, precip_north for the center of forest ﬁre area 105◦E–60◦N, precip south: for June 54.5◦N–97.5◦E, for July the average amount for two grid points: 51.5◦N, 103.5◦E and 104.5◦E; (b) Time-longitude cross-sections of GHGS (m/◦lat). GHGS > 0 corresponds blocking, GHGS > −2—near blocking, Era-Interim data, red and blue ﬁll—surface temperatures anomalies. Light blue vertical line—high precipitation events; yellow vertical line—start and ﬁnish of forest ﬁre period. Figure 1. (a) Total CO emission (90◦E–120◦E, 55◦N–65◦N) (kg/s), A0.25 aerosol concertation in Western Siberia (Fonovaya) (N), HS—the total number of hotspots (90◦E–120◦E, 55◦N–65◦N); the Figure 1. (a) Total CO emission (90◦E–120◦E, 55◦N–65◦N) (kg/s), A0.25 aerosol concertation in Western Siberia (Fonovaya) (N), HS—the total number of hotspots (90◦E–120◦E, 55◦N–65◦N); the total number of the lightning strikes (LS); WS—surface wind speed; tcw—total column water; precip— atmospheric precipitation. tcw, ws, precip_north for the center of forest ﬁre area 105◦E–60◦N, precip south: for June 54.5◦N–97.5◦E, for July the average amount for two grid points: 51.5◦N, 103.5◦E and 104.5◦E; (b) Time-longitude cross-sections of GHGS (m/◦lat). GHGS > 0 corresponds blocking, GHGS > −2—near blocking, Era-Interim data, red and blue ﬁll—surface temperatures anomalies. Light blue vertical line—high precipitation events; yellow vertical line—start and ﬁnish of forest ﬁre period. Figure 1. (a) Total CO emission (90◦E–120◦E, 55◦N–65◦N) (kg/s), A0.25 aerosol concertation in Western Siberia (Fonovaya) (N), HS—the total number of hotspots (90◦E–120◦E, 55◦N–65◦N); the total number of the lightning strikes (LS); WS—surface wind speed; tcw—total column water; precip— atmospheric precipitation. 2.3. Rossby Wave Breaking (RWB) and Wave Activity Flux (WAF) propagation is described by the horizontal wave activity ﬂux (WAF) at 250 hPa. The wave activity ﬂux proposed by Plumb in 1985 is utilized in this study [69]. 2.3. Rossby Wave Breaking (RWB) and Wave Activity Flux (WAF) The detection of RWB was performed by using isentropic potential vorticity (PV) [37]. RWB are characterized by a poleward intrusion of low potential vorticity (or high potential temperature) air and an equatorward intrusion of high potential vorticity (or low potential temperature) [68]. The detection of RWB events in this paper is based on the overturn- ing contour identiﬁcation technique developed by Strong and Magnusdottir, 2008 [40]. The overturning technique can be applied to contour on the dynamical tropopause (DT), isentropic surface (Θ), or pressure surface. For DT and Θ, the techniques are dynamically consistent; low potential temperature streamers on the DT are generally equivalent to high potential vorticity streamers on isentropic surfaces. Jing and Banerjee, 2018 used the isentropic surfaces 320, 350, and 370 K. Bowley et al. 2019 [32] used DT (potential temperature on the DT–PV-Θ). It was discovered that the area with a high frequency of CWB is located eastward of Lake Baikal. We applied the identiﬁcation and analysis of RWB the following way: We applied the identiﬁcation and analysis of RWB the following way: 1. For the isentropic surface at 350 K (which is used to reveal the exchange along the subtropical tropopause [37]), an automatic algorithm was used to search for the overturning contour from 1 to 9 PVU with an interval of 0.5 PVU. For the automatic detection of centers and squares of overturning areas, we used the identiﬁcation technique developed by Barnes and Hartmann, 2012 [39]; 2. For each day between 1 June and August 31, we conducted a synoptic analysis of the potential temperature on the DT (PV-Θ) maps. We utilized PV-Θ to analyze air masses transformation and blocking formation in the mid-latitude area [42,43]; 3. For a 3D visualization of RWB processes, we calculated the 3D surface of DT (2PVU) in three dimensions: longitude, latitude, and geopotential height. The wave activity ﬂux (WAF) that indicates the propagation of planetary waves can usually be used to localize regions of wave activity sources and sinks. Wave energy The wave activity ﬂux (WAF) that indicates the propagation of planetary waves can usually be used to localize regions of wave activity sources and sinks. Wave energy Fire 2023, 6, 122 6 of 25 6 of 25 propagation is described by the horizontal wave activity ﬂux (WAF) at 250 hPa. The wave activity ﬂux proposed by Plumb in 1985 is utilized in this study [69]. 3. Results tcw, ws, precip_north for the center of forest ﬁre area 105◦E–60◦N, precip south: for June 54.5◦N–97.5◦E, for July the average amount for two grid points: 51.5◦N, 103.5◦E and 104.5◦E; (b) Time-longitude cross-sections of GHGS (m/◦lat). GHGS > 0 corresponds blocking, GHGS > −2—near blocking, Era-Interim data, red and blue ﬁll—surface temperatures anomalies. Light blue vertical line—high precipitation events; yellow vertical line—start and ﬁnish of forest ﬁre period. 7 of 25 Fire 2023, 6, 122 ure 2. The forest fire clusters are obtained using the DBSCAN algorithm based on GFAS data m 1 June–31 August. Blue—groups existed from 1 to 15 July; red—16–31 July; green—1–12 gust. Area 1st—58° N–63° N, 95° E–102° E, 2nd—56° N–61° N, 104° E–115° E, 3rd—61° N–64° 02° E–108° E Figure 2. The forest ﬁre clusters are obtained using the DBSCAN algorithm based on GFAS data from 1 June–31 August. Blue—groups existed from 1 to 15 July; red—16–31 July; green—1–12 August. Area 1st—58◦N–63◦N, 95◦E–102◦E, 2nd—56◦N–61◦N, 104◦E–115◦E, 3rd—61◦N–64◦N, 102◦E–108◦E. 8 of 27 e 2. The forest fire clusters are obtained using the DBSCAN algorithm based on GFAS data 1 June–31 August. Blue—groups existed from 1 to 15 July; red—16–31 July; green—1–12 st. Area 1st—58° N–63° N, 95° E–102° E, 2nd—56° N–61° N, 104° E–115° E, 3rd—61° N–64° 2° E 108° E Figure 2. The forest ﬁre clusters are obtained using the DBSCAN algorithm based on GFAS data from 1 June–31 August. Blue—groups existed from 1 to 15 July; red—16–31 July; green—1–12 August. Area 1st—58◦N–63◦N, 95◦E–102◦E, 2nd—56◦N–61◦N, 104◦E–115◦E, 3rd—61◦N–64◦N, 102◦E–108◦E. 8 of 27 02° E–108° E Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50° N–70° N. The black vertical line is the western border of Eastern Siberia. Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the first flood; 3—first forest fire large cluster formation; 4—second and third forest fire cluster formation and forest fire amplification; 5—the second flood; 6 –forest fire period, amplification forest fires activity. Era-Interim data. Figure 1 provides visual representations of the dynamics of forest fire intensity, Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50◦N–70◦N. The black vertical line is the western border of Eastern Siberia. 3. Results Speciﬁcally, Figure 1a includes the following graphs: (1) Bottom graphs: the day-to-day variation of total CO emission (90◦E–120◦E, 55◦N– 65◦N) (grey ﬁll), aerosol concertation (A0.25) in Western Siberia at Fonovaya station (black line), the total number of hotspots (HS) (90◦E–120◦E, 55◦N–65◦N) (red line); p (2) Middle graphs: the day-to-day variation of total column water (tcw, purple line) and surface wind speed (ws, green line) for the center of forest ﬁre area 105◦E, 60◦N; (3) Upper graphs: the day-to-day variation of the total number of the lightning strikes (LS, yellow color), precipitation for the center of the forest ﬁre area 105◦E, 60◦N (precip_north, light-blue color); precipitation: for June in point 54.5◦N, 97.5◦E, for July—the average amount for two grid points: 51.5◦N, 103.5◦E and 51.5◦N, 104.5◦E (precip_south, pink color). Time-longitude cross-section in Figure 1b shows GHGS (black line) and surface tem- perature anomalies average for 55◦N–65◦N (red and blue ﬁll). GHGS > 0 corresponds to blocking, GHGS > −2—near blocking. Light blue vertical line—high precipitation events. Yellow vertical line—start and ﬁnish of forest ﬁre period. p Table S1 (in Supplementary Materials) is the collection of the chronology of precipi- tation (which has led to the ﬂoods), forest ﬁres, blocks over Eastern Siberia, and RWB for Eastern Siberia and western border territory (60◦E–120◦E). Figure 2 displays the forest ﬁre clusters (FFCs), which were identiﬁed using the DBSCAN algorithm based on GFAS data from 1 June to 31 August. We combined the groups based on the time of FFCs existence in the following manner. Blue color—groups existed during 1–15 July, red—16–31 July. green—1–12 August. g y y g g Based on Figures 1 and 2 and Table S1 (Supplementary Materials), we divided the period from 1 June to 15 August into six time intervals: 1–19 June (preliminary period), 20–25 June (the ﬁrst ﬂood), 26 June–5 July (ﬁrst forest ﬁre large cluster formation), 6–21 July (second and third forest ﬁre cluster formation and forest ﬁre ampliﬁcation), 22–28 July (the second ﬂood), 29 July–15 August (forest ﬁre period, ampliﬁcation forest ﬁres activity). In the analysis of the ﬁrst period, which preceded the extreme precipitation and forest ﬁres, we relied on published results regarding anomalies in Eurasia during June 2019. p g g g First (1–19 June)—Period preceding ﬁrst blocking over the ES, extreme precipitation, and forest ﬁres. 3. Results Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the ﬁrst ﬂood; 3—ﬁrst forest ﬁre large cluster formation; 4—second and third forest ﬁre cluster formation and forest ﬁre ampliﬁcation; 5—the second ﬂood; 6—forest ﬁre period, ampliﬁcation forest ﬁres activity. Era-Interim data. Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50° N–70° N. The black vertical line is the western border of Eastern Siberia. Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the first flood; 3—first forest fire large cluster formation; 4—second and third forest fire cluster formation and forest fire amplification; 5—the second flood; 6 –forest fire period, amplification forest fires activity. Era-Interim data. Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50◦N–70◦N. The black vertical line is the western border of Eastern Siberia. Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the ﬁrst ﬂood; 3—ﬁrst forest ﬁre large cluster formation; 4—second and third forest ﬁre cluster formation and forest ﬁre ampliﬁcation; 5—the second ﬂood; 6—forest ﬁre period, ampliﬁcation forest ﬁres activity. Era-Interim data. Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50° N–70° N. The black vertical line is the western border of Eastern Siberia. Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the first flood; 3—first forest fire large cluster formation; 4—second and third forest fire cluster formation and forest fire amplification; 5—the second flood; 6 –forest fire period, amplification forest fires activity. Era-Interim data. Figure 3. Horizontal (S) component of the WAF (m2 s−2), according to Plumb equation (1985) for 50◦N–70◦N. The black vertical line is the western border of Eastern Siberia. Black horizontal lines are six-time interval boundaries: 1—preliminary period; 2—the ﬁrst ﬂood; 3—ﬁrst forest ﬁre large cluster formation; 4—second and third forest ﬁre cluster formation and forest ﬁre ampliﬁcation; 5—the second ﬂood; 6—forest ﬁre period, ampliﬁcation forest ﬁres activity. Era-Interim data. Fire 2023, 6, 122 8 of 25 8 of 25 Figure 1 provides visual representations of the dynamics of forest ﬁre intensity, precip- itation, and atmospheric circulation characteristics during the summer of 2019. Speciﬁcally, Figure 1a includes the following graphs: Figure 1 provides visual representations of the dynamics of forest ﬁre intensity, precip- itation, and atmospheric circulation characteristics during the summer of 2019. 3. Results It is worth noting that during the ﬁrst half of June, the distribution of atmospheric blocking over Eurasia exhibited the “one by one” type of the “Europe + Russian Far East” pattern described in [28] (the longitude time cross-section of the GHGS blocking index for summer 2019 is in Supplementary Materials, Figure S1). pp y g Figure 6 in Zhao et al. 2020 [50] shows that the June AEA index in 2019 was the most positive since 1979. A positive AEA index is characterized by the high geopotential height over the North Atlantic Ocean, Eastern Europe, Mongolia–north China, and low geopotential height over the northeastern North Atlantic Ocean and the Kara Sea–Northern Siberia. We have supposed that the extreme AEA teleconnection in June 2019 caused the pressure pattern of “the deep tropospheric trough in Western Siberia/high amplitude ridge over Eastern Siberia” and as a result, a strong baroclinic zone appeared between Western and Eastern Siberia Figure 2 in [29]. The AEA index remained positive up to the 20th June 2019 [50]. The wave activity ﬂux (WAF) was pronounced in the North Atlantic-European sector from 1 June to 15 June (Figure 3, Interval 1). 10 of 27 terval 2, −50 W–0 E) and deepening of the trough were observed (Video V1 in Supple- mentary Materials). On 21–22 June, a part of the trough was cut off and started to move towards the southeast (Video V1 in Supplementary Materials); on 24–25 June, a cutoff ow was formed (Figure 4) with the center—53° N 83° E The weak rainfall in the front Second—Period of the ﬁrst ﬂood wave (20–25 June). Up to 20 June, the circulation pattern over the Atlantic–Eurasia region corresponded to the positive AEA index, with a trough over the Kara Sea [50]. Between 20 and 21 June, a strengthening of WAF (Figure 3, Interval 2, −50 W–0 E) and deepening of the trough were observed (Video S1 in Supplementary Materials). On 21–22 June, a part of the trough was cut off and started to move towards the southeast (Video S1 in Supplementary Materials); on 24–25 June, a cutoff low was formed (Figure 4), with the center—53◦N, 83◦E. The weak rainfall in the front part of the cutoff low started on 23 June (Video S2 in Supplementary Materials), and on 24 June, the precipitation became extreme. 3. Results The quasi-stationary rainfall belt was located over the Eastern Sayan Mountains during 24–26 June (Figure 5a, Video S2 in Supplementary Materials, Figure 6a,b). It caused an extreme ﬂood in the Irkutsk oblast (Tulun, Nizhneudinsk, Shitkino) [10]. The cyclonic overturning of PV = 4–8 PVU was revealed for 25 June. Still, the ﬁrst signs of PV contours deformation were observed on 24 June and maintained until 26 June. The overturning region is located near the border between the East Asia summer monsoon (EASM) and the polar vortex intrusion [71]. In the cyclonic part of the breaking, there were ascending vertical motions of 1.3 hPa/s (as shown in Figure 7a), while in the anticyclonic (warm) part, descending motions were observed. In the area located under the warm part of the breaking, water vapor ﬂuxes were intensiﬁed (according to Era-Interim date, Figure not shown). The EASM jump was observed on 25 June, which is evident in Figure 5 and Video S2 based on the streamline pattern, and in Figure 8a,b based on the change of the total column water between 23 and 25 June. ow was formed (Figure 4), with the center 53 N, 83 E. The weak rainfall in the front part of the cutoff low started on 23 June (Video V2 in Supplementary Materials), and on 24 June, the precipitation became extreme. The quasi-stationary rainfall belt was located over the Eastern Sayan Mountains during 24–26 June (Figure 5a, Video V2 in Supple- mentary Materials, Figure 6a,b). It caused an extreme flood in the Irkutsk oblast (Tulun, Nizhneudinsk, Shitkino) [10]. The cyclonic overturning of PV = 4–8 PVU was revealed for 25 June. Still, the first signs of PV contours deformation were observed on 24 June and maintained until 26 June. The overturning region is located near the border between the East Asia summer monsoon (EASM) and the polar vortex intrusion [71]. In the cyclonic part of the breaking, there were ascending vertical motions of 1.3 hPa/s (as shown in Figure 7a), while in the anticyclonic (warm) part, descending motions were observed. In the area located under the warm part of the breaking, water vapor fluxes were intensified (according to Era-Interim date, Figure not shown). 3. Results The most striking large-scale weather event over northern Eurasia in June 2019 was the record-breaking heat in Europe [50,70], https://climate.copernicus.eu/surface-air- temperature-june-2019, accessed on 11 March 2023. We draw attention to the record-breaking high temperatures over Europe preceding the large-scale atmospheric events that we are studying over Siberia, which could possibly be associated with them. We do not exclude the possibility of such an association, relying on previous works describing teleconnection patterns over Northern Eurasia, such as AEA [29] and BBC [30], as well as previous works describing the record-breaking high temperature over Europe in June 2019 [50,70]. In these works, the record-breaking heat in Europe is analyzed in detail and associated with the AEA and BBC, respectively. Zhao et al. 2020 [50] based on NCEP–NCAR reanalysis data showed that a strong anomalous anticyclone appeared over Europe in June of 2019. The southerly wind anomalies to the west side of the anomalous anticyclone transported warmer air from lower latitudes towards Europe, contributing to the increase in surface air temperature there (Figure 2a in Zhao et al. 2020 [50]). The wave train with a barotropic vertical structure extended eastward from high latitudes in the North Atlantic across Europe to the Russian Far East (Figure 2b in Zhao et al. 2020 [50]). Positive anomalies in geopotential height were observed over Eastern Europe and Eastern Siberia, while negative anomalies were present over the Kara Sea East. (Figure 2b in Zhao et al. 2020 [50]). The authors have attributed this wave train to the AEA teleconnection pattern. Xu et al., 2020 [70] using monthly mean Japanese 55-year reanalysis (JRA-55) demon- strated an anomalous anticyclone in the upper troposphere and associated it with the BBC Fire 2023, 6, 122 9 of 25 pattern. The authors demonstrate that the anomalous anticyclone results from an unusually intensiﬁed British–Baikal corridor (BBC) pattern and a synoptic Rossby wave breaking (RWB) event over Europe. The authors describe three sub-monthly heat waves during June 2019. The ﬁrst two were associated with the BBC pattern, and the third was related to the combination of the previous BBC pattern activity and the RWB event, Figure 5 (in [70]). 3. Results The EASM jump was observed on 25 June, which is evident in Figure 5 and Video V2 based on the streamline pattern, and in Figure 8a,b based on the change of the total column water between 23 and 25 June. So we have concluded that the CWB and precipitation were caused by the deep in- trusion of polar air masses and were observed simultaneously. The extreme rainfall can be attributed to the wave breaking. We believe that two main factors contributed to the scenario of intrusion and breaking on 20–25 June 2019: an intense wave flux associated with an unusual positive AEA teleconnection throughout June and the characteristics of the Kara Sea trough during 20–23 June. Figure 4. Circulation pattern (WAF—black arrows, m2/s−2 and PV-Θ—fill, K) on 25 June (2nd pe- riod). Here and below the red arrow show the position of cold PV-Θ advection toward Siberia, the red curve demonstrates the horizontal scheme of RWB. Era-Interim data. Figure 4. Circulation pattern (WAF—black arrows, m2/s−2 and PV-Θ—ﬁll, K) on 25 June (2nd period). Here and below the red arrow show the position of cold PV-Θ advection toward Siberia, the red curve demonstrates the horizontal scheme of RWB. Era-Interim data. igure 4. Circulation pattern (WAF—black arrows, m2/s−2 and PV-Θ—fill, K) on 25 June (2nd pe- iod). Here and below the red arrow show the position of cold PV-Θ advection toward Siberia, the ed curve demonstrates the horizontal scheme of RWB. Era-Interim data. Figure 4. Circulation pattern (WAF—black arrows, m2/s−2 and PV-Θ—ﬁll, K) on 25 June (2nd period). Here and below the red arrow show the position of cold PV-Θ advection toward Siberia, the red curve demonstrates the horizontal scheme of RWB. Era-Interim data. 10 of 25 he Fire 2023, 6, 122 riod). H red curv (a) (b) Figure 5. The total amount of precipitation and daily streamline for the first flood wave ((a) 24 June) and for the second flood wave ((b) 27 July). The red color—the region of interest. Era-Interim data, at 850 hPa. Figure 5. The total amount of precipitation and daily streamline for the ﬁrst ﬂood wave ((a) 24 Jun and for the second ﬂood wave ((b) 27 July). The red color—the region of interest. Era-Interim data, 850 hPa. (a) (a) (b) (b) mount of precipitation and daily streamline for the first flood wave ((a) 24 June) ood wave ((b) 27 July). The red color—the region of interest. 3. Results Era-Interim data, Figure 5. The total amount of precipitation and daily streamline for the ﬁrst ﬂood wave ((a) 24 June) and for the second ﬂood wave ((b) 27 July). The red color—the region of interest. Era-Interim data, at 850 hPa. So we have concluded that the CWB and precipitation were caused by the deep intrusion of polar air masses and were observed simultaneously. The extreme rainfall can be attributed to the wave breaking. We believe that two main factors contributed to the scenario of intrusion and breaking on 20–25 June 2019: an intense wave ﬂux associated with an unusual positive AEA teleconnection throughout June and the characteristics of the Kara Sea trough during 20–23 June. g g Third—Period of ﬁrst forest ﬁre large cluster formation (26 June–5 July) (Figure 9, Video S3 in Supplementary Materials). On 26 June (Video S3), the ﬁrst blocking in westerly ﬂow (GHGS > 0, Figure 1b) was detected over the longitudes of Lake Baikal. Furthermore, the intensiﬁcation of the WAF from the North Atlantic continued, and it propagated towards the east (26–30 June). As a result of wave propagation into Siberia, two warm PV-Θ waves (28 June and 29 June–2 July) maintained the blocking with a vast anticyclone in the northern part of Siberia (the area depicted schematically based on the geopotential maps). The wave-like structure is evident from 26 June to 1 July (Figure 3, Interval 3, 0 E–50 E). y g The ﬁrst blocking over Eastern Siberia (Figure 1b) was caused by the intrusion of a low PV-Θ air mass, a cutoff low, and CWB, as well as the simultaneous intensiﬁcation and propagation of WAF from the North Atlantic to Siberia. The subsequent AWB (Figure 9a) was a consequence of blocking and the propagation of WAF. The AWB also served as an additional source for the maintenance of blocking (the anticyclone in North Siberia, Video S3). The intense advection of heat from lower latitudes was associated with patterns on 28 June and 2 July (high PV-Θ). Clusters of ﬁres began to form on 3 July and persisted until 15 July (Figure 2, blue color). 3. Results Fire 2023, 6, 122 , 6, x FOR PEER REV 11 of 25 11 of 25 (a) (b) (c) Goggle maps with the position of precipitation zone (total daily for 24 (a), 25 (b) Ju c)) and PV overturning (2–7PVU) for 24 June 18 UTC (a), 25 June 12 UTC (b) and 27 Figure 6. Goggle maps with the position of precipitation zone (total daily for 2 27 July (c)) and PV overturning (2–7PVU) for 24 June 18 UTC (a), 25 June 12 UT UTC (c), letters A and B shows the schematic position of the start and end cros Era-Interim data. (a) (b) ( ) (c) Goggle maps with the position of precipitation zone (total daily for 24 (a), 25 (b) June and ) and PV overturning (2–7PVU) for 24 June 18 UTC (a), 25 June 12 UTC (b) and 27 July 12 Figure 6. Goggle maps with the position of precipitation zone (total daily for 24 (a), 25 (b) June and 27 July (c)) and PV overturning (2–7PVU) for 24 June 18 UTC (a), 25 June 12 UTC (b) and 27 July 12 UTC (c), letters A and B shows the schematic position of the start and end cross-section in Figure 7. Era-Interim data. 12 of 25 12 of 25 Fire 2023, 6, 122 UTC (c), letters A and B shows the schematic position of the start and end cross-section in Figure 7. ra-Interim data. (a) (b) 13 of 27 (c) Figure 7. Cross-sections of horizontal wind speed (m/s) (blue and red fill), PV (PVU) (black lines) and vertical speed (Pa/s) (red and blue dashed lines) for 24 June 18 UTC (а), 25 June 12 UTC (b) and 27 July 12 UTC (c), letters A and B in Figure 6. The negative value for vertical velocity—ascending motions. Red shading—westerly wind (positive); blue—easterly wind (negative). Era-Interim data. Figure 7. Cross-sections of horizontal wind speed (m/s) (blue and red ﬁll), PV (PVU) (bla and vertical speed (Pa/s) (red and blue dashed lines) for 24 June 18 UTC (a), 25 June 12 UTC 27 July 12 UTC (c), letters A and B in Figure 6. The negative value for vertical velocity—as motions. Red shading—westerly wind (positive); blue—easterly wind (negative). Era-Inter -Interim data. (a) (b) 13 of 27 (a) ( ) (b) (b) (c) (c) e 7. 3. Results Cross-sections of horizontal wind speed (m/s) (blue and red fill), PV (PVU) (black lines) ertical speed (Pa/s) (red and blue dashed lines) for 24 June 18 UTC (а), 25 June 12 UTC (b) and y 12 UTC (c), letters A and B in Figure 6. The negative value for vertical velocity—ascending ns. Red shading—westerly wind (positive); blue—easterly wind (negative). Era-Interim data. Figure 7. Cross-sections of horizontal wind speed (m/s) (blue and red ﬁll), PV (PVU) (black lines) and vertical speed (Pa/s) (red and blue dashed lines) for 24 June 18 UTC (a), 25 June 12 UTC (b) and 27 July 12 UTC (c), letters A and B in Figure 6. The negative value for vertical velocity—ascending motions. Red shading—westerly wind (positive); blue—easterly wind (negative). Era-Interim data. Fire 2023, 6, 122 13 of 25 ) and di 13 of 25 ) and di 27 July 12 UTC (c), letters A and B in Figure 6. The negative value for vertical velocity ascending motions. Red shading—westerly wind (positive); blue—easterly wind (negative). Era-Interim data. (a) (b) (c) (d) Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a,b) and 26–27 (c,d) July 2019. Era-Interim data. Third—Period of first forest fire large cluster formation (26 June–5 July) (Figure 9, Video V3 in Supplementary Materials). On 26 June (Video V3), the first blocking in westerly flow (GHGS > 0, Figure 1b) was detected over the longitudes of Lake Baikal. Furthermore, the Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a,b) and 26–27 (c,d) July 2019. Era-Interim data. re 2023, 6, x FOR PEER REVIEW 14 of 27 intensification of the WAF from the North Atlantic continued, and it propagated towards the east (26–30 June). As a result of wave propagation into Siberia, two warm PV-Θ waves (28 June and 29 June–2 July) maintained the blocking with a vast anticyclone in the north- ern part of Siberia (the area depicted schematically based on the geopotential maps). The wave-like structure is evident from 26 June to 1 July (Figure 3, Interval 3, 0 E–50 E). The first blocking over Eastern Siberia (Figure 1b) was caused by the intrusion of a low PV-Θ air mass, a cutoff low, and CWB, as well as the simultaneous intensification and propagation of WAF from the North Atlantic to Siberia. 3. Results The subsequent AWB (Figure 9a) was a consequence of blocking and the propagation of WAF. The AWB also served as an additional source for the maintenance of blocking (the anticyclone in North Siberia, Video V3). The intense advection of heat from lower latitudes was associated with patterns on 28 June and 2 July (high PV-Θ). Clusters of fires began to form on 3 July and persisted until 15 July (Figure 2, blue color). (a) (b) Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (ac- cording to maps of geopotential at 500 hPa). Era-Interim data. Fourth—Period of second and third forest fire region formation and forest fire amplification (6–21 July) (Figure 10, Video V4 in Supplementary Materials). On 5–6 July, the anticyclone Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (according to maps of geopotential at 500 hPa). Era-Interim data. (a) (b) (c) (d) Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a b) and 26–27 (c d) July 2019 Era-Interim data Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a,b) and 26–27 (c,d) July 2019. Era-Interim data. e 2023, 6, x FOR PEER REVIEW 14 of 27 intensification of the WAF from the North Atlantic continued, and it propagated towards the east (26–30 June). As a result of wave propagation into Siberia, two warm PV-Θ waves (28 June and 29 June–2 July) maintained the blocking with a vast anticyclone in the north- ern part of Siberia (the area depicted schematically based on the geopotential maps). The wave-like structure is evident from 26 June to 1 July (Figure 3, Interval 3, 0 E–50 E). The first blocking over Eastern Siberia (Figure 1b) was caused by the intrusion of a low PV-Θ air mass, a cutoff low, and CWB, as well as the simultaneous intensification and propagation of WAF from the North Atlantic to Siberia. The subsequent AWB (Figure 9a) was a consequence of blocking and the propagation of WAF. 3. Results The AWB also served as an additional source for the maintenance of blocking (the anticyclone in North Siberia, Video V3). The intense advection of heat from lower latitudes was associated with patterns on 28 June and 2 July (high PV-Θ). Clusters of fires began to form on 3 July and persisted until 15 July (Figure 2, blue color). (a) Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a,b) and 26–27 (c,d) July 2019. Era-Interim data. Figure 8. Total column water for maximal EASM jump during cyclonic wave breaking 23–25 June (a,b) and 26–27 (c,d) July 2019. Era-Interim data. ) p 28 June and 2 July (high PV-Θ). Clusters of fires began to form on 3 July and persisted until 15 July (Figure 2, blue color). Third—Period of first forest fire large cluster formation (26 June–5 July) (Figure 9, Video V3 in Supplementary Materials). On 26 June (Video V3), the first blocking in westerly flow (GHGS > 0, Figure 1b) was detected over the longitudes of Lake Baikal. Furthermore, the (a) (b) Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period Third—Period of first forest fire large cluster formation (26 June–5 July) (Figure 9, Video V3 in Supplementary Materials). On 26 June (Video V3), the first blocking in westerly flow (GHGS > 0, Figure 1b) was detected over the longitudes of Lake Baikal. Furthermore, the (a) (b) Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (ac- cording to maps of geopotential at 500 hPa). Era-Interim data. Fourth Period of second and third forest fire region formation and forest fire amplification Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (according to maps of geopotential at 500 hPa). Era-Interim data. Third—Period of first forest fire large cluster formation (26 June–5 July) (Figure 9, Video V3 in Supplementary Materials). 3. Results On 26 June (Video V3), the first blocking in westerly flow (GHGS > 0, Figure 1b) was detected over the longitudes of Lake Baikal. Furthermore, the (a) (b) Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (ac- ording to maps of geopotential at 500 hPa). Era-Interim data. Figure 9. Circulation pattern (WAF and PV-Θ, similar to Figure 4) on 2 (a) and 4 (b) July (3rd period. Here and below, a red circle with the letter “A” marks the position of the blocking anticyclone (according to maps of geopotential at 500 hPa). Era-Interim data. Fire 2023, 6, 122 14 of 25 14 of 25 Fourth—Period of second and third forest ﬁre region formation and forest ﬁre ampliﬁcation (6–21 July) (Figure 10, Video S4 in Supplementary Materials). On 5–6 July, the anticyclone in the northern part of eastern Siberia began to decay (4 July in Video S3, 6–8 July in Video S4). The transformation of the air masses associated with the previous period with the blocking anticyclone over ES can be traced (4 July in Video S3, 6–8 July in Video S4). The high PV-Θ air masses in the northern part of Eurasia have moved to North Europe. On 10 July, the trough that had been observed over North Europe since 4 July intensiﬁed. It was due to the ampliﬁcation of WAF and the advection of cold air masses in the front part of the anticyclone. The WAF in the front part of the trough increased from 10–14 July; however, the synoptic waves do not propagate well into the northern part of ES (Figure 3, 10–30 July, Intervals 4 and 5). From 45◦E to 150◦E, the weak PV-Θ latitudinal gradient was observed (and weak geopotential height gradient Figure 1). For the period from 14 July to 17 July, two PV contours breaking were detected, ﬁrst AWB (14–15 July) and second CWB (14–17 July) (Figure 10a), with a renewing cutoff low between them. Starting from 15 July, the large forest ﬁre clusters in the second and third forest ﬁre regions began to form (Figure 2, red clusters). The crucial role for forest ﬁre spreading belongs to the CWB to the east of Lake Baikal (Figure 10a). Fifth—Period of the second flood formation (22–28 July) (Figure 12). On 22–28 July, two main large-scale dynamical events can be identified (Video V5 in Supplementary Materi- als). First, it is anticyclonic breakings over Europe and increased WAF associated with them (Figure 3, Interval 5), (Figure 12, Video V5). Second, it is the transformation and eastward movement of the low PV Θ area over sector 60° E 90° E (Figure 12 Video V5) Fifth—Period of the second ﬂood formation (22–28 July) (Figure 12). On 22–28 July, two main large-scale dynamical events can be identiﬁed (Video S5 in Supplementary Materials). First, it is anticyclonic breakings over Europe and increased WAF associated with them (Figure 3, Interval 5), (Figure 12, Video S5). Second, it is the transformation and eastward movement of the low PV-Θ area over sector 60◦E–90◦E (Figure 12, Video S5), including the 3. Results Figure 11 shows the schematic maps based on the Worldview satellite image of hotspots for 16 and 17 July and potential vorticity levels. The increase in the number of hotspots and biomass emissions, as well as the northward shift of the forest ﬁre area (as shown in Figure 2), occurred concurrently with the occurrence of CWB on 16–17 July. On 16 July, the anticyclone in the northern part of Siberia ampliﬁed. On 21 July, the anticyclone decayed simultaneously with the regeneration of the cutoff low (Figure 10b). Blocking over Siberia was observed until 22 July. 15 of 27 ntours breaking were detected, first AWB (14–15 July) and second CWB re 10a), with a renewing cutoff low between them. Starting from 15 July, e clusters in the second and third forest fire regions began to form (Fig- ). The crucial role for forest fire spreading belongs to the CWB to the kal (Figure 10a). Figure 11 shows the schematic maps based on the te image of hotspots for 16 and 17 July and potential vorticity levels. The mber of hotspots and biomass emissions, as well as the northward shift ea (as shown in Figure 2), occurred concurrently with the occurrence of y. On 16 July, the anticyclone in the northern part of Siberia amplified. nticyclone decayed simultaneously with the regeneration of the cutoff Blocking over Siberia was observed until 22 July. (a) (b) Figure 10. Circulation pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b)July (4th period). Era-In- erim data. Figure 10. Circulation pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b) July (4th period). Era-Interim d (a) (b) Figure 10. Circulation pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b)July (4th period). Era-In- erim data. Figure 10. Circulation pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b) July (4th period). Era-Interim data. (b) on pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b)July (4th period). Era-In- Figure 10. Circulation pattern (WAF and PV-Θ, K) on 17 (a) and 21 (b) July (4th period). Era-Interim data Fire 2023, 6, 122 Fire 2023, 6, x FOR 15 of 25 16 of 27 (a) (b) Figure 11. Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU according to Era-Interim on 16 July 12 UTC (a), 17 July 12 UTC (b), 60–120 E, 30–70 N, Era-Interi Figure 11. Figure 11. Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU) according to Era-Interim on 16 July 12 UTC (a), 17 July 12 UTC (b), 60–120 E, 30–70 N, Era-Interim data. Figure 11. Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU) according to Era-Interim on 16 July 12 UTC (a), 17 July 12 UTC (b), 60–120 E, 30–70 N, Era-Interim data. 3. Results Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU di t E I t i 16 J l 12 UTC ( ) 17 J l 12 UTC (b) 60 120 E 30 70 N E I t i d t (a) (a) (a) (a) (b) Figure 11 Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU Fi 11 W ld id i h h h i i d h i PV b ki (2 d 7PVU (b) Figure 11. Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU) according to Era-Interim on 16 July 12 UTC (a), 17 July 12 UTC (b), 60–120 E, 30–70 N, Era-Interim data Figure 11. Worldwide maps with the hotspots position and schematic PV breaking (2 and 7PVU) according to Era-Interim on 16 July 12 UTC (a), 17 July 12 UTC (b), 60–120 E, 30–70 N, Era-Interim data. Fire 2023, 6, 122 16 of 25 reinforced cutoff low on 21 July (Video S4, 90◦E). During 21–28 July, the low PV-Θ moved to the east along the area of PV-Θ high gradient (45◦N, 90◦E–120◦E). On 23, 27, 28 July the CWB was detected over 45◦N–60◦N, 90◦E–120◦E with maximal reversing on 27–28 July (Figure 6c). The main features of CWB 26–28 July were similar to the CWB 25–26 June: the increase in vertical motions (up to 0.7 hPa/s) (Figure 7c), the sharp drop of tropopause height between the south and the north part of breaking (not illustrated here) and high column water in the atmosphere (Figure 8c,d). The high column water can be related to the increase and northward jump of the EASM in the lower troposphere between 26–27 July (Video S2, Figure 8). The maximum intensity of rainfall (Video S2) in the southern part of eastern Siberia was observed during the period of maximum gradient of overturning 27–28 July (Figure 6c). The CWB in Eastern Siberia coincided with the occurrence of the AWB over Europe (as shown in Figure 12 and Video S5) and an increase in WAF (as shown in Figure 3, Interval 5). Regarding the formation of blocking, the CWBs on 26–28 July are similar to the CWB that accompanied the ﬁrst precipitation period (25–26 June). 3. Results Although the geopotential ﬁeld at the end of July did not indicate the presence of a blocking pattern (GHGS < 0) (Figure 1), the overturning in the region helped maintain a low PV gradient over Eastern Siberia. 18 of 27 18 of 27 Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6) Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Sixth—The last forest ﬁre period, secondary ampliﬁcation forest ﬁres activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video S6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video S6). Two AWBs (2 and 7 August) at 75◦E–90◦E occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Videos S5 and S6). Figure 1b shows that the blocking had ﬂuctuations during 1–12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest ﬁre hotspots and biomass burning emissions increased (green forest ﬁre clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). After 12 August, the trough was ﬁnally established over Eastern Siberia. 3. Results The general WAF over Eurasia moved to subtropics (12 August in Video S6). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Figure 12. Circulation pattern (WAF and PV-Θ) on 28 July (5th period). Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). 3. Results Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6) Sixth—The last forest ﬁre period, secondary ampliﬁcation forest ﬁres activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video S6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video S6). Two AWBs (2 and 7 August) at 75◦E–90◦E occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Videos S5 and S6). Figure 1b shows that the blocking had ﬂuctuations during 1–12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest ﬁre hotspots and biomass burning emissions increased (green forest ﬁre clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). After 12 August, the trough was ﬁnally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video S6). Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). 3. Results Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6). Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6) Sixth—The last forest ﬁre period, secondary ampliﬁcation forest ﬁres activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video S6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video S6). Two AWBs (2 and 7 August) at 75◦E–90◦E occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Videos S5 and S6). Figure 1b shows that the blocking had ﬂuctuations during 1–12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest ﬁre hotspots and biomass burning emissions increased (green forest ﬁre clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). After 12 August, the trough was ﬁnally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video S6). 3.2. The Factors Contributing to Forest Fire Ignitions Anthropogenic sources have been identified a fires [5 72] However in 2019 a mega wildfire in S 3.2. The Factors Contributing to Forest Fire Ignitions Anthropogenic sources have been identified a fires [5 72] However in 2019 a mega wildfire in S Anthropogenic sources have been identiﬁed as the signiﬁcant cause of Russian wild- ﬁres [5,72]. However, in 2019 a mega wildﬁre in Siberian was observed in hard-to-reach areas of Siberia (in northern Krasnoyarsk Krai, the Sakha Republic, and Zabaykalsky Krai) and was reported as being caused by natural factors of ignition. According to a quote from the Krasnoyarsk Forest Fire Center by the Russian News Agency TASS, the causes of forest ﬁres were natural and were due to a 30-degree Celsius heat (86 degrees Fahrenheit), gusts of wind, and dry thunderstorms. (https://www.nasa.gov/image-feature/goddard/2019 /huge-wildﬁres-in-russias-siberian-province-continue, accessed on 11 March 2023). fires [5,72]. However, in 2019 a mega wildfire in Siberian was observed in hard to reach areas of Siberia (in northern Krasnoyarsk Krai, the Sakha Republic, and Zabaykalsky Krai) and was reported as being caused by natural factors of ignition. According to a quote from the Krasnoyarsk Forest Fire Center by the Russian News Agency TASS, the causes of for- est fires were natural and were due to a 30-degree Celsius heat (86 degrees Fahrenheit), gusts of wind, and dry thunderstorms. (https://www.nasa.gov/image-feature/god- dard/2019/huge-wildfires-in-russias-siberian-province-continue accessed on 11 March 2023). In Figure 1a the yellow bar chart shows the total amount of lightning strokes (LS) in g p In Figure 1a, the yellow bar chart shows the total amount of lightning strokes (LS) in the forest ﬁre area. The LS increase occurred between the ﬁrst day of the ﬁrst blocking (Table S1 in Supplementary Materials, 27 June) and the last day of the second blocking (19 July). Figure 14a,b shows anomalies in LS for the 1–10 and 10–20 July along forest ﬁre clusters obtained for these periods. Figure 14c–e shows the days with the highest LS. In July 2019, LS levels higher than those in the previous period of 2009–2018 were observed; furthermore, in most of Eastern Siberia, the LS were not accompanied by rain higher than 3 mm (Figure 1, blue bar chart, Figure 14c–e). In Figure 1a, the yellow bar chart shows the total amount of lightning strokes (LS) in the forest fire area. The LS increase occurred between the first day of the first blocking (Table S1 in Supplementary Materials, 27 June) and the last day of the second blocking (19 July). 3. Results Sixth—The last forest fire period, secondary amplification forest fires activity in three regions, end of the blocking period (29 July–15 August) (Figure 13, Video V6 in Supplementary Materials) The pattern is similar to 26 June–2 July (third period); after the CWB two breaking from the west, leading to the strengthening of ES-blocking high (Figure 13, Video V6). Two AWBs (2 and 7 August) at 75° E–90° Е occurred due to WAF increasing in the front part of the trough formed after breaking over Europe during 26–31 July (Video V5, Video V6). Figure 1b shows that the blocking had fluctuations during 1– 12 August, with the peak of GHGS on 3–4 and 8 August. On 3–8 August, the number of forest fire hotspots and biomass burning emissions increased (green forest fire clusters, Figure 2) however, in most parts, for hotspot number than for emissions (Figure 1a). Af- ter 12 August, the trough was finally established over Eastern Siberia. The general WAF over Eurasia moved to subtropics (12 August in Video V6). Figure 13. Circulation pattern (WAF and PV-Θ) on 7 August (6th period). Era-Interim data. Figure 13. Circulation pattern (WAF and PV-Θ) on 7 August (6th period). Era-Interim data. Figure 13. Circulation pattern (WAF and PV-Θ) on 7 August (6th period). Era-Interim data. Figure 13 Circulation pattern (WAF and PV Θ) on 7 August (6th period) Era Interim data Figure 13. Circulation pattern (WAF and PV-Θ) on 7 August (6th period). Era-Interim data. Figure 13. Circulation pattern (WAF and PV-Θ) on 7 August (6th period). Era-Interim data. 17 of 25 17 of 25 . Fire 2023, 6, 122 3.3. Factors That Affected the Magnitude of Forest Fire Intensity (Hotspot Number and Emission) We have shown that forest fire periods in 2019 are strongly associated with wave breaking and atmospheric blocking. Both breaking and blocking caused the positive surface temperature anomalies in Northern Siberia (Figure 1b). There are three main effects of surface temperature increase linked to breaking and blocking anticyclone formation. The first effect is the transport of air masses from lower latitudes [75,76], which is indicated by high PV-Θ values on the maps. The second effect is the adiabatic warming in the upper-level anticyclone, caused by the subsidence of air parcels [77]. Additionally, the downward motion anomalies associated with blocking anticyclones result in less cloud cover and increased incoming solar radiation. The latter also contributes to an increase in surface temperature [50]. Ponomarev et al. (2016) [3] demonstrated the significant relationship between forest fire characteristics in Siberia, surface temperature, and incoming solar radiation. Thus, regardless of the contribution of the three factors, we evaluated the role of wave breaking and blocking as crucial drivers of surface temperature anomalies in summer 2019. However, it is necessary to investigate the other effects of blocking in Siberia, which contribute to the maximum magnitude of CO emissions and hotspot count (Figure 1a): atmospheric moisture and wind speed [75,76]. Therefore, we analyzed the variations in the total atmospheric moisture content and wind speed during July-August (two critical factors, along with temperature and precipitation, included in ﬁre hazard calculations) for three blocking events (Table S1 in Supplementary Materials) (Figure 1b, tcw and WS). We tracked the cloud cover based on Worldview maps (https://worldview.earthdata.nasa.gov/, ac- cessed on 21 December 2021) along with moisture and wind speed. Video S3 (4 July), Video S4 (19 July), and Video S6 (4 August, 8 August) illustrate the circulation patterns that correspond to the blocking life cycle from the last RWB to blocks decay, as shown in Table S1. During these periods, there is an ampliﬁcation of WAF and intrusions of cold air masses downstream of the blocking anticyclone in North Siberia. The location of the blocking anti- cyclone is the most favorable for forest ﬁre spreading. At the same time, the forest ﬁre area is located in a region with minimal cloud cover (https://go.nasa.gov/2NHy6Ik, accessed on 21 December 2021, https://go.nasa.gov/3idmNpt, https://go.nasa.gov/2ZknNzj, ac- cessed on 21 December 2021, https://go.nasa.gov/2YLBzMp, accessed on 21 December 2021) and near the downstream area under the intrusion of cold air masses. 3.2. The Factors Contributing to Forest Fire Ignitions Anthropogenic sources have been identified a fires [5 72] However in 2019 a mega wildfire in S Figure 14a,b shows anomalies in LS for the 1–10 and 10–20 July along forest fire clusters obtained for these periods. Figure 14c-e shows the days with the highest LS. In July 2019, LS levels higher than those in the previous period of 2009–2018 were observed; furthermore, in most of Eastern Siberia, the LS were not accompanied by rain higher than 3 mm (Figure 1, blue bar chart, Figure 14c–e). (a) (b) ire 2023, 6, x FOR PEER REVIEW 19 of 27 (c) (d) (e) Figure 14. Anomaly in lightning strokes compares to periods 2009–2018 (a);1–10 July 2019 (b);10–20 of July 2019. The total amount of lightning strokes and precipitation on 2 (c), 6 (d), 17 (e) July. Thunderstorms that occur without significant rainfall on the ground are called “dry Figure 14. Anomaly in lightning strokes compares to periods 2009–2018 (a);1–10 July 2019 (b);10–20 of July 2019. The total amount of lightning strokes and precipitation on 2 (c), 6 (d), 17 (e) July. Th d h i h i iﬁ i f ll h d ll d “d 19 2023, 6, x FOR PEER REVIEW (a) (c) (b) (e) (a) (b) (d) (c) (d) (e) Figure 14. Anomaly in lightning strokes compares to periods 2009–2018 (a);1–10 July 2019 (b);10–20 of July 2019. The total amount of lightning strokes and precipitation on 2 (c), 6 (d), 17 (e) July. Figure 14. Anomaly in lightning strokes compares to periods 2009–2018 (a);1–10 July 2019 (b);10–20 of July 2019. The total amount of lightning strokes and precipitation on 2 (c), 6 (d), 17 (e) July. Thunderstorms that occur without significant rainfall on the ground are called “dry thunderstorms” (DTh) [73,74]. DTh are the most dangerous in the case of dry soil and atmospheric conditions. More importantly, dry thunderstorm events often lead to large wildfire outbreaks [73] The increase in DTh during blocking can be explained by the in- Thunderstorms that occur without signiﬁcant rainfall on the ground are called “dry thunderstorms” (DTh) [73,74]. DTh are the most dangerous in the case of dry soil and atmospheric conditions. More importantly, dry thunderstorm events often lead to large Fire 2023, 6, 122 18 of 25 wildﬁre outbreaks [73]. The increase in DTh during blocking can be explained by the increase in the vertical instability of the atmosphere. The instability is associated with the formation of cutoff ﬁlaments with a low potential temperature (or high PV) [73]. 3.2. The Factors Contributing to Forest Fire Ignitions Anthropogenic sources have been identified a fires [5 72] However in 2019 a mega wildfire in S The example on 17 July shows that the increase in DTh during blocking can be explained by the vertical instability of the atmosphere (Figure 10a) and high near-surface temperature (Figure S2b). In Figure S2a, CAPE (convective available potential energy) is shown, which characterizes the amount of energy available for convection in the atmosphere. The com- bination of instability and high surface temperatures is due to the conﬁguration of the cyclonic breaking on 17 July. The estimates of the effect of LS on ﬁres and the ratio of their contribution require additional research. In the present paper, we only draw attention to the increase in the number of dry thunderstorms during the 2019 blocking periods. Additionally, our ﬁndings are in agreement with the statement made by the Krasnoyarsk Forest Fire Center. 3.3. Factors That Affected the Magnitude of Forest Fire Intensity (Hotspot Number and Emission) a I e i ) 3.4. The Factors That Affected the Extreme Precipitation and Flood ) 3.4. The Factors That Affected the Extreme Precipitation and Flood e Factors That Affected the Extreme Precipitation and Flood s forest fires floods in 2019 are strongly associated with waves breaking and at- eric blocking (Figure 1b). The main cause of floods is precipitation [4]. Precipita- tensity depends on ascent condition (instability) and total column water of air The three main mechanisms of air instability generation are local heating, orogra- movement of air up the slopes of mountains) and large-scale potential vorticity As forest ﬁres ﬂoods in 2019 are strongly associated with waves breaking and atmo- spheric blocking (Figure 1b). The main cause of ﬂoods is precipitation [4]. Precipitation intensity depends on ascent condition (instability) and total column water of air mass. The three main mechanisms of air instability generation are local heating, orography (movement of air up the slopes of mountains) and large-scale potential vorticity dynamics (strengthening of meridionality). p p ) g p y ics (strengthening of meridionality). e have shown that in 2019, extreme precipitation over the southern part of Eastern was due to a large-scale strengthening of meridionality: increased wave activity urope (Figure 3, periods 2 and 5) and cyclonic type of Rossby wave breaking over Siberia. This large-scale mechanism of instability was strengthened by the oro- mechanism; atmospheric front, precipitation area and ascending movements calized clearly along the Eastern Sayan ridge (Figures 6 and 7). The contrast be- warm moist air from the East Asian summer monsoon area (Video V2) and cold ir from the Kara Sea (Video V1 and Video V5), which was involved in the southern Eastern Siberia during the cyclonic Rossby wave breaking, led to an increase in ngth of the vertical circulation associated with frontogenesis. This resulted in ex- We have shown that in 2019, extreme precipitation over the southern part of Eastern Siberia was due to a large-scale strengthening of meridionality: increased wave activity over Europe (Figure 3, periods 2 and 5) and cyclonic type of Rossby wave breaking over Eastern Siberia. This large-scale mechanism of instability was strengthened by the orographic mechanism; atmospheric front, precipitation area and ascending movements were localized clearly along the Eastern Sayan ridge (Figures 6 and 7). a I e i ) 3.4. The Factors That Affected the Extreme Precipitation and Flood The contrast between warm moist air from the East Asian summer monsoon area (Video S2) and cold arctic air from the Kara Sea (Videos S1 and S5), which was involved in the southern part of Eastern Siberia during the cyclonic Rossby wave breaking, led to an increase in the strength of the vertical circulation associated with frontogenesis. This resulted in extreme precipitation, contributing to the 2019 Siberian summer anomaly. 3.3. Factors That Affected the Magnitude of Forest Fire Intensity (Hotspot Number and Emission) For all periods, a decrease in total column water was observed (Figure 1a). The signiﬁcant reduction in total column water in the lower atmosphere, up to 10 kg/m2, led to ampliﬁed ﬁre danger. On June 20th, there was a cold air intrusion from the east (Figure S2c), which resulted in an increase in geopotential contrasts (Figure S2d). The zone of low moisture content in the north expanded (Figure S2e), and the near-surface wind intensiﬁed (Figure S2f). This period is considered critical and, combined with the increase in lightning activity on 17th July, may have contributed to the maximum increase in ﬁres. Fire 2023, 6, 122 19 of 25 19 of 25 The deepest low PV-Θ intrusion in the front of the anticyclone was observed in mid- July (17–19 July, Video S4). Mid-July is the period of maximum intensiﬁcation of CO emissions and the number of hotspots (Figure 1a). Additionally, Figure 1a shows that in mid-July the ﬁre area experienced a maximum decrease in total column water and a maximum increase in surface wind speed. The advection of cold PV-Θ resulted in the renewal of the cutoff cyclone in the south of Western Siberia on 21 July (Figure 10b). The direction of cold advection was extremely atypical for the summer period. Figure 15 displays the geopotential height at 500 hPa and the total amount of hotspots for 12–23 July 2019. The massive block was located over Siberia, and the cutoff cyclone with a center of 55◦N–90◦E promoted the transfer of smoke plumes from ﬁres to the west of Siberia. In mid-summer, as shown in Figure 1a, the air composition in Western Siberia, which is not subject to ﬁres, changed simultaneously with the change in ﬁre intensity. The area from the Ural Mountains to the Far East located under a blocking dome was ﬁlled with smoke from ﬁres (https://go.nasa.gov/2CBktIx, accessed on 21 December 2021). 21 of 27 15. The total number of hotspots (FIRMS) and geopotential height at 500 hPa (12–23 July Era Interim) Figure 15. The total number of hotspots (FIRMS) and geopotential height at 500 hPa (12–23 July 2019) (Era-Interim). 5. The total number of hotspots (FIRMS) and geopotential height at 500 hPa (12–23 July I t i ) Figure 15. The total number of hotspots (FIRMS) and geopotential height at 500 hPa (12–23 July 2019) (Era-Interim). precipitation, contributing to the 20 4. Discussion and Conclusions ussion and Conclusions ecord-breaking forest fires and floods were observed in Eastern Siberia (ES) be- 24 June and 12 August in 2019. We investigated these events as one phenomenon Record-breaking forest ﬁres and ﬂoods were observed in Eastern Siberia (ES) between 24 June and 12 August in 2019. We investigated these events as one phenomenon due to the synoptic (Rossby) waves breaking (RWB) and atmospheric blocking (blocks). We Fire 2023, 6, 122 20 of 25 have demonstrated periods for RWB based on overturning potential vorticity (PV) contour; blocks based on geopotential gradient; transport and transformation of air masses based on potential temperature on the dynamical tropopause (PV-Θ); forest ﬁre based on carbon monoxide emission and hotspots; ﬂood based on the total amount of precipitation. Every speciﬁc period was described by synoptic analysis. We have obtained the following key results: 1. The rainfall and forest ﬁre in Siberia in 2019 were strongly associated with wave breaking and the life-cycle of blocking-high in the northern part of Siberia. The blocks were formed and maintained by two types of RWB: cyclonic type (CWB) from the east ES (110◦E–115◦E) and anticyclonic type (AWB) to the west ES (70◦E–90◦E). The CWB that occurred in latitude belt 40◦N–60◦N did not lead to the formation of the blocking high in geopotential; nevertheless, the CWB caused a low gradient in PV around Lake Baikal (Eastern Siberia). The AWB and CWB that extended above 60◦N resulted in the blocking high on geopotential in the northern part of ES. Both types of breaking occurred in the front part of the cutoff low or trough. The main CWB and AWB were observed: 24–26 June (CWB), 28–29 June/1–2 July (AWB), 14–17 July (both AWB and CWB), 25–28 July (CWB), and 2 and 7 August (AWB). According to the geopotential gradient, the blocking over ES was observed three times: 26 June–3 July, 12–21 July, and 4–10 August (with the break 5–7 August); 2. y y g g 2. The rainfall in the southern part of ES (24–25 June, 25–28 July) was associated with the baroclinic growth of synoptic eddies accompanying CWB around Lake Baikal. Depending on the degree of PV overturning, the rainfall can be quasi-stationary for some days. The total precipitation caused by breaking depends on the initial baroclinicity (vertical velocity) and the border East Asian summer monsoon (EASM). The EASM can sharply turn to the southwest and add extra precipitable water. precipitation, contributing to the 20 4. Discussion and Conclusions CWB associated with extreme precipitation either preceded blocking or occurred after blocking decay (simultaneously with the eastward movement of the low PV-Θ part of the blocking). Additionally, CWB occurred with high precipitation and had lower amplitude (only up to 60◦N) compared to AWB (occurring above 60◦N); 3. p y p p g 3. The periods of forest fires are associated with the establishment of blocking high in the northern part of ES, mainly due to high amplitude breaking from west of ES (75◦E–90◦E). The peak of forest fires was on 4–5 July, 19–24 July, 5–6 and 8 August; occurred in periods of blocking decay. The location of blocking anticyclones can be favorable for the spread of forest fires. The forest fire area is located simultaneously in areas with minimum cloud cover and near downstream areas where cold air masses are intruding. For all forest fire peaks, a decrease in water content was observed. The most extreme peak in forest fire intensity was related to an anticyclone resulting from double breaking from the west and east (Video S4). The periods of blocking decay were characterized by the intrusion of cold air masses along the eastern part of the blocking high. In addition to the decreasing total column water and cloud cover, an increase in surface wind speed was observed. These processes drove forest fire intensification and spreading. Furthermore, from 26 June to 19 July, the formation and decay of blocking were accompanied by dry thunderstorms (DTs). DTs occurred along the periphery of low PV-Θ filaments. DTs are potentially the primary cause of for fire ignition in Siberia regions with the lowest population density (above 60◦N) [78]; 3. The periods of forest fires are associated with the establishment of blocking high in the northern part of ES, mainly due to high amplitude breaking from west of ES (75◦E–90◦E). The peak of forest fires was on 4–5 July, 19–24 July, 5–6 and 8 August; occurred in periods of blocking decay. The location of blocking anticyclones can be favorable for the spread of forest fires. The forest fire area is located simultaneously in areas with minimum cloud cover and near downstream areas where cold air masses are intruding. For all forest fire peaks, a decrease in water content was observed. The most extreme peak in forest fire intensity was related to an anticyclone resulting from double breaking from the west and east (Video S4). precipitation, contributing to the 20 4. Discussion and Conclusions The periods of blocking decay were characterized by the intrusion of cold air masses along the eastern part of the blocking high. In addition to the decreasing total column water and cloud cover, an increase in surface wind speed was observed. These processes drove forest fire intensification and spreading. Furthermore, from 26 June to 19 July, the formation and decay of blocking were accompanied by dry thunderstorms (DTs). DTs occurred along the periphery of low PV-Θ filaments. DTs are potentially the primary cause of for fire ignition in Siberia regions with the lowest population density (above 60◦N) [78]; 4. We have concluded that both types of extremes, namely forest ﬁres in Northern Siberia and ﬂoods in Southern Siberia, are closely related. We demonstrated the relationship by synoptic analysis of wave breaking and blocking formation. Both types of wave breaking have been detected by PV overturning on 350 K, indicating that they can be associated with exchange related to the subtropical tropopause. The CWB occurred in the southern part of Siberia (45◦N–60◦N), caused extreme rainfall, and maintained a low PV gradient eastward of Lake Baikal (24–25 June, 22–28 July). The low PV (PV-Θ) gradient and strengthening of the wave activity ﬂux from the Europe–Atlantic sector may be the reason why the AWB has a high amplitude westward towards the region Fire 2023, 6, 122 21 of 25 21 of 25 with low PV (28 June–2 July, 14–15 July, and 2–7 August). In turn, AWB creates the condition for the formation of blocking highs and strengthens the WAF in the northern part of Siberia, which affects temperature, cloud cover, wind speed, and moisture content. The complicated combinations of the CWB and AWB were the main drivers of the extreme forest ﬁres. In the speciﬁc case of the summer of 2019, the repeated position of the blocking anticyclone three times before its decay played a crucial role. We want to discuss the relationship between the variability of ﬁre intensity in the cluster region, blocking, and wave breaking as a debatable issue. Figure 16 presents the total emissions of CO (from biomass burning) from 2003 to 2019, with the highest emissions recorded in 2019. This conﬁrms the earlier ﬁndings of Ponomarev et al. (2016) regarding changes in wildﬁre numbers and burned areas in Siberia [3]. As stated in [3], the number of forest ﬁres and the size of burned areas have increased (1996–2015). precipitation, contributing to the 20 4. Discussion and Conclusions As highlighted by previous studies [79,80], rising temperatures and drier conditions have led to longer and more severe ﬁre seasons, resulting in a signiﬁcant increase in forest ﬁres across North America and Siberia. Our paper further emphasizes the crucial role of atmospheric blocking in driving positive temperature anomalies, and underscores the need to monitor changes in Rossby wave breaking and blocking formation over Siberia in the future. It is not only important to track changes in their frequency, but also in the characteristics of RWB and blocking formation. garding changes in wildfire numbers and burned areas in Siberia [3]. As stated in number of forest fires and the size of burned areas have increased (1996–2015). A lighted by previous studies [79,80], rising temperatures and drier conditions hav longer and more severe fire seasons, resulting in a significant increase in fore across North America and Siberia. Our paper further emphasizes the crucial rol mospheric blocking in driving positive temperature anomalies, and underscores th to monitor changes in Rossby wave breaking and blocking formation over Siberi future. It is not only important to track changes in their frequency, but also in th acteristics of RWB and blocking formation. Figure 16. Total CO emissions (kg/s) from biomass burning in the cluster area are shown in 2. Area 1st—58° N–63° N, 95° E–102° E, 2nd—56° N–61° N, 104° E–115° E, 3rd—61° N–64° E–108° E (according figure 2). Figure 16. Total CO emissions (kg/s) from biomass burning in the cluster area are shown in Figure 2. Area 1st—58◦N–63◦N, 95◦E–102◦E, 2nd—56◦N–61◦N, 104◦E–115◦E, 3rd—61◦N–64◦N, 102◦ E–108◦E (according Figure 2). Figure 16. Total CO emissions (kg/s) from biomass burning in the cluster area are shown in 2. Area 1st—58° N–63° N, 95° E–102° E, 2nd—56° N–61° N, 104° E–115° E, 3rd—61° N–64° E–108° E (according figure 2). Figure 16. Total CO emissions (kg/s) from biomass burning in the cluster area are shown in Figure 2. Area 1st—58◦N–63◦N, 95◦E–102◦E, 2nd—56◦N–61◦N, 104◦E–115◦E, 3rd—61◦N–64◦N, 102◦ E–108◦E (according Figure 2). The development of the 2019 summer scenario, in our opinion, can be attrib three key factors: the anomaly AEA index during June [50], the specific topography Kara Sea trough, and the increasing wave-train from the Atlantic during 21–26 Jun It is important to note that some researchers have suggested that there has change in the pattern of Rossby wave propagation over Eurasia since the mid [13,53]. precipitation, contributing to the 20 4. Discussion and Conclusions Li and Ruan [29] have highlighted the increasing trend of the Atlantic–Eu teleconnection index over the past three decades. The evolution and amplification Rossby wave pattern can lead to modifications in the relationship between breaki blocking. Moreover, the role of the positive feedback loop of “heat wave-soil mo may also increase [46,49,81,82]. The aridification of Eurasia in the 21st century [ The development of the 2019 summer scenario, in our opinion, can be attributed to three key factors: the anomaly AEA index during June [50], the speciﬁc topography of the Kara Sea trough, and the increasing wave-train from the Atlantic during 21–26 June. It is important to note that some researchers have suggested that there has been a change in the pattern of Rossby wave propagation over Eurasia since the mid-1990s [13,53]. Li and Ruan [29] have highlighted the increasing trend of the Atlantic–Eurasian telecon- nection index over the past three decades. The evolution and amplification of the Rossby wave pattern can lead to modifications in the relationship between breaking and block- ing. Moreover, the role of the positive feedback loop of “heat wave-soil moisture” may also increase [46,49,81,82]. The aridification of Eurasia in the 21st century [83] can further contribute to the change in the relationship between RWB, blocking, and wildfires. The development of the 2019 summer scenario, in our opinion, can be attrib three key factors: the anomaly AEA index during June [50], the specific topograph Kara Sea trough and the increasing wave train from the Atlantic during 21 26 Ju The development of the 2019 summer scenario, in our opinion, can be attributed to three key factors: the anomaly AEA index during June [50], the speciﬁc topography of the Kara Sea trough, and the increasing wave-train from the Atlantic during 21–26 June. Kara Sea trough, and the increasing wave-train from the Atlantic during 21–26 Jun It is important to note that some researchers have suggested that there has change in the pattern of Rossby wave propagation over Eurasia since the mi [13,53]. Li and Ruan [29] have highlighted the increasing trend of the Atlantic–E teleconnection index over the past three decades. The evolution and amplification Rossby wave pattern can lead to modifications in the relationship between break blocking. Moreover, the role of the positive feedback loop of “heat wave-soil mo may also increase [46,49,81,82]. precipitation, contributing to the 20 4. Discussion and Conclusions The aridification of Eurasia in the 21st century [ g g g J It is important to note that some researchers have suggested that there has been a change in the pattern of Rossby wave propagation over Eurasia since the mid-1990s [13,53]. Li and Ruan [29] have highlighted the increasing trend of the Atlantic–Eurasian telecon- nection index over the past three decades. The evolution and amplification of the Rossby wave pattern can lead to modifications in the relationship between breaking and block- ing. Moreover, the role of the positive feedback loop of “heat wave-soil moisture” may also increase [46,49,81,82]. The aridification of Eurasia in the 21st century [83] can further contribute to the change in the relationship between RWB, blocking, and wildfires. Kara Sea trough, and the increasing wave train from the Atlantic during 21 26 Jun It is important to note that some researchers have suggested that there has change in the pattern of Rossby wave propagation over Eurasia since the mid [13,53]. Li and Ruan [29] have highlighted the increasing trend of the Atlantic–E teleconnection index over the past three decades. The evolution and amplification Rossby wave pattern can lead to modifications in the relationship between breaki blocking. Moreover, the role of the positive feedback loop of “heat wave-soil mo may also increase [46,49,81,82]. The aridification of Eurasia in the 21st century [ It is important to note that some researchers have suggested that there has been a change in the pattern of Rossby wave propagation over Eurasia since the mid-1990s [13,53]. Li and Ruan [29] have highlighted the increasing trend of the Atlantic–Eurasian telecon- nection index over the past three decades. The evolution and amplification of the Rossby wave pattern can lead to modifications in the relationship between breaking and block- ing. Moreover, the role of the positive feedback loop of “heat wave-soil moisture” may also increase [46,49,81,82]. The aridification of Eurasia in the 21st century [83] can further contribute to the change in the relationship between RWB, blocking, and wildfires. precipitation, contributing to the 20 4. Discussion and Conclusions Informed Consent Statement: Not applicable. Data Availability Statement: Global Fire Assimilation System (GFAS) (CAMS Global Fire As- similation System: http://apps.ecmwf.int/datasets/data/cams-gfas/, accessed on 12 December 2021). Fire Information for Resource Management System (FIRMS) (https://modaps.modaps. eosdis.nasa.gov/services/about/products/c6-nrt/MOD14.html, accessed on 12 December 2021) Hotspots visualizer Worldview based on the satellite image (http://worldview.earthdata.nasa.gov, accessed on 12 December 2021). Fonovaya station (http://lop.iao.ru/EN/fon/diffbat/, accessed on 12 December 2021) Worldwide lightning location network, WWLLN, http://wwlln.net/, ac- cessed on 12 December 2021 Global Precipitation Climatology Centre (Deutscher Wetterdienst) (GPCC) (https://opendata.dwd.de/climate_environment/GPCC/html/gpcc_ﬁrstguess_daily_doi_ download.html, accessed on 12 December 2021) ECMWF ERA-Interim reanalysis datasets (https: //apps.ecmwf.int/datasets/data/interim-full-daily/levtype=sfc/, accessed on 12 December 2021). Acknowledgments: The authors are very grateful to three anonymous reviewers who helped improve the article. Acknowledgments: The authors are very grateful to three anonymous reviewers who helped improve the article. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 3. Ponomarev, E.I.; Kharuk, V.I.; Ranson, K.J. Wildﬁres Dynamics in Siberian Larch Forests. Forests 2016, 7, 125. [CrossRef] 4. Frolova, N.L.; Kireeva, M.B.; Magrickiy, D.V.; Bologov, M.B.; Kopylov, V.N.; Hall, J.; Semenov, V.A.; Kosolapov, A.E.; Dorozhkin, E.V.; Korobkina, E.A.; et al. Hydrological hazards in Russia: Origin, classiﬁcation, changes and risk assessment. Nat. Hazards 2016, 88, 103–131. [CrossRef] g g y g 2. Groisman, P.; Shugart, H.; Kicklighter, D.; Henebry, G.; Tchebakova, N.; Maksyutov, S.; Monier, E.; Gutman, G.; Gulev, S.; Qi, J.; et al. Northern Eurasia Future Initiative (NEFI): Facing the challenges and pathways of global change in the twenty-ﬁrst century. Prog. Earth Planet. Sci. 2017, 4, 41. [CrossRef] p y 6. Soja, A.J.; Cofer, W.R.; Shugart, H.; Sukhinin, A.I.; Stackhouse, P.W., Jr.; McRae, D.J.; Conard, S. Estimating ﬁre emissions and disparities in boreal Siberia (1998–2002). J. Geophys. Res. Atmos. 2004, 109, D004570. [CrossRef] 5. Kajii, Y. Boreal forest ﬁres in Siberia in 1998: Estimation of area burned and emissions of pollutants by advanced very high resolution radiometer satellite data. J. Geophys. Res. 2002, 107, ACH 4-1–ACH 4-8. [CrossRef] 1. Field, C.B.; Barros, V.; Stocker, T.F.; Dahe, Q.; Dokken, D.J.; Ebi, K.L.; Mastrandrea, M.D.; Mach, K.J.; Plattner, G.K.; Allen, S.K.; et al. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation: Special Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK, 2012. precipitation, contributing to the 20 4. Discussion and Conclusions Fire 2023, 6, 122 22 of 25 Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/fire6030122/s1, Figure S1: The longitude time cross-section of the GHGS blocking index for summer 2019; Figure S2: 17 July 12 UTC (a) Convective amiable potential energy (CAPE), (b) surface temperature; 20 July 12 UTC (c) PV-Θ, (d) geopotential height at 500 hPa, (e) total column water, (f) wind at 10 m; Table S1: The time table of high precipitation, forest ﬁres clusters formation, RWB and blocks; Video S1: Between 20 and 21 June, a strengthening of WAF (Figure 3, Interval 2, –50 W–0 E) and deepening of the trough were observed; On 21–22 June, a part of the trough was cut off and started to move towards the southeast; Video S2: The weak rainfall in the front part of the cutoff low started on 23 June, and on 24 June, the precipitation became extreme; Video S3: Period of ﬁrst forest ﬁre large cluster formation (26 June–5 July); Video S4: Period of second and third forest ﬁre region formation and forest ﬁre ampliﬁcation (6–21 July); Video S5: Period of the second ﬂood formation (22–28 July); Video S6: The last forest ﬁre period, secondary ampliﬁcation forest ﬁres activity in three regions, end of the blocking period (29 July–15 August). Author Contributions: Conceptualization, O.Y.A.; methodology, O.Y.A., A.V.G., P.N.A. and B.D.B.; software, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; validation, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; investigation, O.Y.A.; writing—original draft preparation, O.Y.A. and E.V.D.; writing—review and editing, O.Y.A., E.V.D. and P.N.A.; visualization, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; supervision, B.D.B.; project administration, E.V.D. All authors have read and agreed to the published version of the manuscript. Author Contributions: Conceptualization, O.Y.A.; methodology, O.Y.A., A.V.G., P.N.A. and B.D.B.; software, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; validation, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; investigation, O.Y.A.; writing—original draft preparation, O.Y.A. and E.V.D.; writing—review and editing, O.Y.A., E.V.D. and P.N.A.; visualization, L.D.T., K.N.P., P.N.A., Y.V.M. and A.V.G.; supervision, B.D.B.; project administration, E.V.D. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Russian Science Foundation №23-27-00167 “Relationship between the formation of extreme precipitation in Southern Siberia and Rossby waves breaking and atmospheric blocking” (https://rscf.ru/en/project/23-27-00167/, accessed on 10 March 2023). Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. 3. Ponomarev, E.I.; Kharuk, V.I.; Ranson, K.J. Wildﬁres Dynamics in Siberian Larch Forests. Forests 2016, 7, 125. [CrossRef] 4 Frolova, N L ; Kireeva, M B ; Magrickiy, D V; Bologov, M B ; Kopylov, VN ; Hall, J ; Semenov, VA ; Kosolapov, A E ; Dorozhkin 3. Ponomarev, E.I.; Kharuk, V.I.; Ranson, K.J. Wildﬁres Dynamics in Siberian Larch Forests. Forests 2016, 7, 125. [CrossRef] 3. Ponomarev, E.I.; Kharuk, V.I.; Ranson, K.J. Wildﬁres Dynamics in Siberian Larch Forests. Forests 2016, 7, 125. [CrossRef] 4. Frolova, N.L.; Kireeva, M.B.; Magrickiy, D.V.; Bologov, M.B.; Kopylov, V.N.; Hall, J.; Semenov, V.A.; Kosolapov, A.E.; Dorozhkin E V; Korobkina E A ; et al Hydrological hazards in Russia: Origin classiﬁcation changes and risk assessment Nat Hazards References 1. Field, C.B.; Barros, V.; Stocker, T.F.; Dahe, Q.; Dokken, D.J.; Ebi, K.L.; Mastrandrea, M.D.; Mach, K.J.; Plattner, G.K.; Allen, S.K.; et al. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation: Special Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK, 2012. 1. Field, C.B.; Barros, V.; Stocker, T.F.; Dahe, Q.; Dokken, D.J.; Ebi, K.L.; Mastrandrea, M.D.; Mach, K.J.; Plattner, G.K.; Allen, S.K.; et al. Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation: Special Report of the Intergovernmental Panel on Climate Change; Cambridge University Press: Cambridge, UK, 2012. g g y g 2. Groisman, P.; Shugart, H.; Kicklighter, D.; Henebry, G.; Tchebakova, N.; Maksyutov, S.; Monier, E.; Gutman, G.; Gulev, S.; Qi, J.; et al. Northern Eurasia Future Initiative (NEFI): Facing the challenges and pathways of global change in the twenty-ﬁrst century. Prog. Earth Planet. Sci. 2017, 4, 41. [CrossRef] 3. Ponomarev, E.I.; Kharuk, V.I.; Ranson, K.J. Wildﬁres Dynamics in Siberian Larch Forests. Forests 2016, 7, 125. [CrossRef] y 4. Frolova, N.L.; Kireeva, M.B.; Magrickiy, D.V.; Bologov, M.B.; Kopylov, V.N.; Hall, J.; Semenov, V.A.; Kosolapov, A.E.; Dorozhkin, E.V.; Korobkina, E.A.; et al. Hydrological hazards in Russia: Origin, classiﬁcation, changes and risk assessment. Nat. Hazards 2016, 88, 103–131. [CrossRef] 5. Kajii, Y. Boreal forest ﬁres in Siberia in 1998: Estimation of area burned and emissions of pollutants by advanced very high resolution radiometer satellite data. J. Geophys. Res. 2002, 107, ACH 4-1–ACH 4-8. [CrossRef] p y 6. Soja, A.J.; Cofer, W.R.; Shugart, H.; Sukhinin, A.I.; Stackhouse, P.W., Jr.; McRae, D.J.; Conard, S. Estimating ﬁre emissions and disparities in boreal Siberia (1998–2002). J. Geophys. Res. Atmos. 2004, 109, D004570. [CrossRef] 23 of 25 23 of 25 Fire 2023, 6, 122 7. Bowman, D.M.J.S.; Balch, J.K.; Artaxo, P.; Bond, W.J.; Carlson, J.M.; Cochrane, M.A.; D’Antonio, C.M.; DeFries, R.S.; Doyle, J.C.; Harrison, S.P.; et al. Fire in the Earth System. Science 2009, 324, 481–484. [CrossRef] [PubMed] y 8. Sitnov, S.A.; Mokhov, I.I. A Comparative Analysis of the Characteristics of Active Fires in the Boreal Forests of Eurasia and North America Based on Satellite Data. Izv. Atmos. Ocean. Phys. 2018, 54, 966–978. [CrossRef] y 9. Vinogradova, A.A.; Vasileva, A.V. Black carbon in air over northern regions of Russia: Sources and spatiotemporal variations. Atmos. Ocean. Opt. 2017, 30, 533–541. [CrossRef] 10. Fedorova, A.; Makarieva, O.; Nesterova, N.; Shikhov, A.; Vinogradova, T. References Observed ch stratiform precipitation in Northern Eurasia over the last ﬁve decades. Environ. Res. Lett. 2019, 14, 045 ulsky, A.; Kozlov, F.; Zolina, O.; Bulygina, O.N.; I Mokhov, I.; A Semenov, V. Observed changes in conv stratiform precipitation in Northern Eurasia over the last ﬁve decades. Environ. Res. Lett. 2019, 14, 045001. [CrossRef] 24. Pfahl, S.; Wernli, H. Quantifying the Relevance of Cyclones for Precipitation Extremes. J. Clim. 2012, 25, 6770–6780. [CrossRef] p p 24. Pfahl, S.; Wernli, H. Quantifying the Relevance of Cyclones for Precipitation Extremes. J. Clim. 2012, 25, 6770–6780. [CrossRef] 24. Pfahl, S.; Wernli, H. Quantifying the Relevance of Cyclones for Precipitation Extremes. J. Clim. 201 25. Catto, J.L.; Pfahl, S. The importance of fronts for extreme precipitation. J. Geophys. Res. Atmos. 2013, 118, 10791–10801. [CrossRef] 26. Barlow, M.; Gutowski, W.J.; Gyakum, J.R.; Katz, R.W.; Lim, Y.-K.; Schumacher, R.S.; Wehner, M.F.; Agel, L.; Bosilovich, M.; Collow, A.; et al. North American extreme precipitation events and related large-scale meteorological patterns: A review of statistical methods, dynamics, modeling, and trends. Clim. Dyn. 2019, 53, 6835–6875. [CrossRef] 27. Bosart, L.F.; Moore, B.J.; Cordeira, J.M.; Archambault, H.M. Interactions of North Paciﬁc Tropical, Midlatitude, and Polar Disturbances Resulting in Linked Extreme Weather Events over North America in October 2007. Mon. Weather Rev. 2017, 145, 1245–1273. [CrossRef] 28. Antokhina, O.Y.; Antokhin, P.N.; Devyatova, E.V.; I Mordvinov, V.; Martynova, Y.V. Precipitation in the Selenga River basin during atmospheric blocking over Europe and the Russian Far East in July. IOP Conf. Ser. Earth Environ. Sci. 2018, 211, 012054. [CrossRef] 29. Li, J.; Ruan, C. The North Atlantic–Eurasian teleconnection in summer and its effects on Eurasian climates. Environ. Res. Lett. 2017, 13, 024007. [CrossRef] 30. Xu, P.; Wang, L.; Chen, W. The British–Baikal Corridor: A Teleconnection Pattern along the Summertime Polar Front Jet over Eurasia. J. Clim. 2019, 32, 877–896. [CrossRef] 31. Basha, G.; Ratnam, M.V.; Kishore, P. Asian summer monsoon anticyclone: Trends and variability. Atm 6789–6801. [CrossRef] M.V.; Kishore, P. Asian summer monsoon anticyclone: Trends and variability. Atmos. Chem. Phys. 2020, 20 ef] 32. Bowley, K.A.; Gyakum, J.R.; Atallah, E.H. A New Perspective toward Cataloging Northern Hemisphere Rossby Wave Breaking on the Dynamic Tropopause. Mon. Weather Rev. 2019, 147, 409–431. [CrossRef] y p p 33. Jing, P.; Banerjee, S. Rossby Wave Breaking and Isentropic Stratosphere-Troposphere Exchange During 1981-2015 in the Northern Hemisphere. J. Geophys. Res. Atmos. 2018, 123, 9011–9025. [CrossRef] p p y 34. References Modelling maximum discharge of the catastrophic ﬂood at the Iya River (Irkutsk region, Russia) in 2019. E3S Web Conf. 2020, 163, 01004. [CrossRef] 11. Vilfand, R.M.; A Kulikova, I.; E Makarova, M. Weather and climate features of the Northern Hemisphere in 2019 in the context of long-period variability. IOP Conf. Ser. Earth Environ. Sci. 2020, 606, 012067. [CrossRef] , S.J. Evidence linking Arctic ampliﬁcation to extreme weather in mid-latitudes. Geophys. Res. Lett. 2012, 39 f] 12. Francis, J.A.; Vavrus, S.J. Evidence linking Arctic ampliﬁcation to extreme weather in mid-latitudes. G GL051000. [CrossRef] 13. Schubert, S.D.; Wang, H.; Koster, R.D.; Suarez, M.J.; Groisman, P.Y. Northern Eurasian Heat Waves and Droughts. J. Clim. 2014, 27, 3169–3207. [CrossRef] 14. Huang, C.S.Y.; Nakamura, N. Local Finite-Amplitude Wave Activity as a Diagnostic of Anomalous Weather Events. J. Atmos. Sci. 2015, 73, 211–229. [CrossRef] 15. Rodrigues, R.R.; Woollings, T. Impact of Atmospheric Blocking on South America in Austral Summer. J. Clim. 2017, 30, 1821–1837. [CrossRef] 16. Wirth, V.; Riemer, M.; Chang, E.K.M.; Martius, O. Rossby Wave Packets on the Midlatitude Waveguide—A Review. Mon. Weather Rev. 2018, 146, 1965–2001. [CrossRef] rth, V.; Bossmann, P.; Fink, A.H. Linking Northern Hemisphere temperature extremes to Rossby wave packets . 2018, 144, 553–566. [CrossRef] 17. Fragkoulidis, G.; Wirth, V.; Bossmann, P.; Fink, A.H. Linking Northern Hemisphere temperature extreme Q. J. R. Meteorol. Soc. 2018, 144, 553–566. [CrossRef] 18. Dacre, H.F.; Pinto, J.G. Serial clustering of extratropical cyclones: A review of where, when and why it occurs. NPJ Clim. Atmos. Sci. 2020, 3, 48. [CrossRef] 19. Ali, S.M.; Martius, O.; Röthlisberger, M. Recurrent Rossby Wave Packets Modulate the Persistence of Dry and Wet Spells Across the Globe. Geophys. Res. Lett. 2021, 48, e2020GL091452. [CrossRef] p y 20. Moore, B.J.; Keyser, D.; Bosart, L.F. Linkages between Extreme Precipitation Events in the Central and Eastern United States and Rossby Wave Breaking. Mon. Weather Rev. 2019, 147, 3327–3349. [CrossRef] y Wave Breaking. Mon. Weather Rev. 2019, 147, 3327–33 21. Liu, C. Role of Rossby Wave Breaking in the Variability of Large-Scale Atmospheric Transport and Mixing. Doctor’s Thesis, Colorado State University, Fort Collins, CO, USA, 2017. 22. Lau, W.K.M.; Kim, K.-M. The 2010 Pakistan Flood and Russian Heat Wave: Teleconnection of Hydrometeorological Extremes. J. Hydrometeorol. 2012, 13, 392–403. [CrossRef] 2. Lau, W.K.M.; Kim, K.-M. The 2010 Pakistan Flood 23. Chernokulsky, A.; Kozlov, F.; Zolina, O.; Bulygina, O.N.; I Mokhov, I.; A Semenov, V. ns, B.I. Towards a PV-θ view of the general circulation. Tellus A Dyn. Meteorol. Oceanogr. 1991, 43, 27–36. [C References Erdenebat, E.; Sato, T. Recent increase in heat wave frequency around Mongolia: Role of atmospheric forcing and possible inﬂuence of soil moisture deﬁcit. Atmos. Sci. Lett. 2015, 17, 135–140. [CrossRef] 47. Kornhuber, K.; Osprey, S.; Coumou, D.; Petri, S.; Petoukhov, V.; Rahmstorf, S.; Gray, L. Extreme weathe 2018 connected by a recurrent hemispheric wave-7 pattern. Environ. Res. Lett. 2019, 14, 054002. [CrossR 48. Hall, J.V.; Loboda, T.V. Quantifying the variability of potential black carbon transport from cropland burning in Russia driven by atmospheric blocking events. Environ. Res. Lett. 2018, 13, 055010. [CrossRef] p g , , [ ] 49. Whan, K.; Zscheischler, J.; Orth, R.; Shongwe, M.; Rahimi, M.; Asare, E.O.; Seneviratne, S.I. Impact of soil moisture on extreme maximum temperatures in Europe. Weather Clim. Extrem. 2015, 9, 57–67. [CrossRef] p p 50. Zhao, W.; Zhou, N.; Chen, S. The Record-Breaking High Temperature over Europe in June of 2019. Atmosphere 2020, 11, 524. [CrossRef] 51. Hayasaka, H.; Yamazaki, K.; Naito, D. Weather conditions and warm air masses during active ﬁre-periods in boreal forests. Polar Sci. 2019, 22, 100472. [CrossRef] 52. Reeder, M.J.; Spengler, T.; Musgrave, R. Rossby waves, extreme fronts, and wildﬁres in southeastern Australia. Geophys. Res. Lett. 2015, 42, 2015–2023. [CrossRef] Antokhin, P.N.; Martynova, Y.V.; Mordvinov, V.I. The Linkage of the Precipitation in the Selenga River Basi mospheric Blocking Atmosphere 2019 10 343 [CrossRef] 53. Antokhina, O.Y.; Antokhin, P.N.; Martynova, Y.V.; Mordvinov, V.I. The Linkage of the Precipitation Midsummer Atmospheric Blocking. Atmosphere 2019, 10, 343. [CrossRef] khina, O.Y.; Antokhin, P.N.; Martynova, Y.V.; Mordvinov, V.I. The Linkage of the Precipitation in the Selenga Antokhin, P.N.; Martynova, Y.V.; Mordvinov, V.I. The Linkage of the Precipitation in the Selenga River Basi mospheric Blocking. Atmosphere 2019, 10, 343. [CrossRef] Antokhina, O.Y.; Antokhin, P.N.; Martynova, Y.V.; Mordvinov, V.I. The Linkage of the Precipitation in the Midsummer Atmospheric Blocking. Atmosphere 2019, 10, 343. [CrossRef] y g p g ummer Atmospheric Blocking. Atmosphere 2019, 10, 343. [CrossRef] 54. de Vries, A.J. A global climatological perspective on the importance of Rossby wave breaking and intense moisture transport for extreme precipitation events. Weather Clim. Dyn. 2021, 2, 129–161. [CrossRef] 55. Kaiser, J.W.; Heil, A.; Andreae, M.O.; Benedetti, A.; Chubarova, N.; Jones, L.; Morcrette, J.-J.; Razinger, M.; Schultz, M.G.; Suttie, M.; et al. Biomass burning emissions estimated with a global ﬁre assimilation system based on observed ﬁre radiative power. Biogeosciences 2012, 9, 527–554. [CrossRef] 56. Ester, M. Density-Based Clustering. References McIntyre, M.E.; Palmer, T.N. Breaking planetary waves in the stratosphere. Nature 1983, 305, 593–6 34. McIntyre, M.E.; Palmer, T.N. Breaking planetary waves in the stratosphere. Nature 1983, 305, 593–600. [CrossRef] 35. McIntyre, M.E.; Palmer, T. A note on the general concept of wave breaking for Rossby and gravity waves. Pure Appl. Geophys. 1985, 123, 964–975. [CrossRef] 35. McIntyre, M.E.; Palmer, T. A note on the general concept of wave breaking for Rossby and gravity waves. Pure Appl. Geophys. 1985, 123, 964–975. [CrossRef] [ ] 36. Hoskins, B.I. Towards a PV-θ view of the general circulation. Tellus A Dyn. Meteorol. Oceanogr. 1991, 43, 27–36. [CrossRef] 24 of 25 24 of 25 Fire 2023, 6, 122 37. Postel, G.A.; Hitchman, M.H. A Climatology of Rossby Wave Breaking along the Subtropical Tropopause. J. Atmos. Sci. 1999, 56, 359–373. [CrossRef] 38. Masato, G.; Hoskins, B.J.; Woollings, T.J. Wave-breaking characteristics of midlatitude blocking. Q. J. R. Meteorol. Soc. 2011, 138, 1285–1296. [CrossRef] 39. Barnes, E.A.; Hartmann, D.L. Detection of Rossby wave breaking and its response to shifts of the midlatitude jet with climate change. J. Geophys. Res. Atmos. 2012, 117, D017469. [CrossRef] 40. Strong, C.; Magnusdottir, G. Tropospheric Rossby Wave Breaking and the NAO/NAM. J. Atmos. Sci. 2008, 65, 2861–2876. [CrossRef] 41. Liu, C.; Barnes, E.A. Quantifying Isentropic Mixing Linked to Rossby Wave Breaking in a Modiﬁed Lagrangian Coordinate. J. Atmos. Sci. 2018, 75, 927–942. [CrossRef] 41. Liu, C.; Barnes, E.A. Quantifying Isentropic Mixing Linked to Rossby Wave Breaking in a Modiﬁed Lagrangian Coordinate. J. Atmos. Sci. 2018, 75, 927–942. [CrossRef] 42 Shi N ; Wang X ; Zhang L ; Xu H Features of Rossby Wave Propagation Associated with the Evolution of Summertime Blocking 42. Shi, N.; Wang, X.; Zhang, L.; Xu, H. Features of Rossby Wave Propagation Associated with the Evolution Highs with Different Conﬁgurations over Northeast Asia. Mon. Weather Rev. 2016, 144, 2531–2546. [Cr 43. Chyi, D.; Xie, Z.; Shi, N.; Guo, P.; Wang, H. Wave-Breaking Features of Blocking over Central Siberia and Its Impacts on the Precipitation Trend over Southeastern Lake Baikal. Adv. Atmos. Sci. 2019, 37, 75–89. [CrossRef] p [ ] 44. Lin, Z.; Lu, R. Abrupt Northward Jump of the East Asian Upper-Tropospheric Jet Stream in Mid-Summer. J. Meteorol. Soc. Jpn. Ser. II 2008, 86, 857–866. [CrossRef] 45. Iwao, K.; Takahashi, M. Interannual change in summertime precipitation over northeast Asia. Geophys. Res. Lett. 2006, 33, GL027119. [CrossRef] 46. 63. Kharyutkina, E.; Loginov, S.; Martynova, Y.; Sudakov, I. Time Series Analysis of Atmospheric Precipitation Characteristics in Western Siberia for 1979–2018 across Different Datasets. Atmosphere 2022, 13, 189. [CrossRef] References In Data Clustering; Chapman and Hall/CRC: New York, NY, USA, 2018; pp. 111–127. [CrossRef] 57. Giglio, L.; Schroeder, W.; Hall, J.V.; Justice, C.O. MODIS Collection 6 Active Fire Product User’s Guide. Revision B; NASA: Washington, DC, USA, 2018; p. 64. 58. Giglio, L.; Schroeder, W.; Justice, C.O. The collection 6 MODIS activee ﬁre detection algorithm and ﬁre products. Remote Sens. Environ. 2016, 178, 31–41. [CrossRef] 59. Antonovich, V.V.; Antokhin, P.N.; Arshinov, M.Y.; Belan, B.D.; Balin, Y.S.; Davydov, D.K.; Ivlev, G.A.; Kozlov, A.V.; Kozlov, V.S.; Kokhanenko, G.P.; et al. Station for the comprehensive monitoring of the atmosphere at Fonovaya Observatory, West Siberia: Current status and future needs. In Proceedings of the 24th International Symposium on Atmospheric and Ocean Optics: Atmospheric Physics, Tomsk, Russia, 2–5 July 2018. [CrossRef] p y y 60. Hutchins, M.L.; Holzworth, R.H.; Rodger, C.J.; Brundell, J.B. Far-Field Power of Lightning Strokes as Wide Lightning Location Network. J. Atmos. Ocean. Technol. 2012, 29, 1102–1110. [CrossRef] 61. Kirstin, S.; Markus, Z.; Andreas, B.; Peter, F.; Anja, M.-C.; Bruno, R.; Udo, S. GPCC First Guess Daily Product at 1.0◦: Near Real-Time First Guess daily Land-Surface Precipitation from Rain-Gauges Based on SYNOP Data; Global Precipitation Climatology Centre: Offenbach, Germany, 2013. [CrossRef] y 62. Schneider, U.; Becker, A.; Finger, P.; Meyer-Christoffer, A.; Ziese, M.; Rudolf, B. GPCC’s new land surface precipitation climatology based on quality-controlled in situ data and its role in quantifying the global water cycle. Theor. Appl. Clim. 2013, 115, 15–40. [CrossRef] 63. Kharyutkina, E.; Loginov, S.; Martynova, Y.; Sudakov, I. Time Series Analysis of Atmospheric Precipitation Characteristics in Western Siberia for 1979–2018 across Different Datasets. Atmosphere 2022, 13, 189. [CrossRef] 25 of 25 Fire 2023, 6, 122 64. Dee, D.P.; Uppala, S.M.; Simmons, A.J.; Berrisford, P.; Poli, P.; Kobayashi, S.; Andrae, U.; Balmaseda, M.A.; Balsamo, G.; Bauer, P.; et al. The ERA-Interim reanalysis: Conﬁguration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 2011, 137, 553–597. [CrossRef] 65. Barriopedro, D.; García-Herrera, R.; Lupo, A.R.; Hernández, E. A Climatology of Northern Hemisphe 19, 1042–1063. [CrossRef] 66. Lejenäs, H.; Økland, H. Characteristics of northern hemisphere blocking as determined from a long time series of observational data. Tellus A Dyn. Meteorol. Oceanogr. 1983, 35A, 350–362. [CrossRef] y g 67. Tibaldi, S.; Molteni, F. On the operational predictability of blocking. Tellus A Dyn. Meteorol. Oceanogr. 1990, 42, 343. [CrossRef] y g 67. Tibaldi, S.; Molteni, F. References On the operational predictability of blocking. Tellus A Dyn. Meteorol. Oceanogr. 1990, 42, 343. [CrossRef] 68. Thorncroft, C.D.; Hoskins, B.J.; McIntyre, M.E. Two paradigms of baroclinic-wave life-cycle behaviour. Q. J. R. Meteorol. Soc. 1993, 119, 17–55. [CrossRef] 67. Tibaldi, S.; Molteni, F. On the operational predictability of blocking. Tellus A Dyn. Meteorol. Oceanogr. 1990, 42, 343. [CrossRef] 68. Thorncroft, C.D.; Hoskins, B.J.; McIntyre, M.E. Two paradigms of baroclinic-wave life-cycle behaviour. Q. J. R. Meteorol. Soc. 1993, , ; , p p y g y g , , [ ] 68. Thorncroft, C.D.; Hoskins, B.J.; McIntyre, M.E. Two paradigms of baroclinic-wave life-cycle behaviour. Q. J. R. Meteorol. Soc. 1993, 119, 17–55. [CrossRef] 69. Plumb, R.A. On the Three-Dimensional Propagation of Stationary Waves. J. Atmos. Sci. 1985, 42, 217–229. [CrossRef] 70. Xu, P.; Wang, L.; Liu, Y.; Chen, W.; Huang, P. The record-breaking heat wave of June 2019 in Central Europe. Atmos. Sci. Lett. 2020, 21, e964. [CrossRef] p g y J [ ] 70. Xu, P.; Wang, L.; Liu, Y.; Chen, W.; Huang, P. The record-breaking heat wave of June 2019 in Central Europe. Atmos. Sci. Lett. 2020, 21, e964. [CrossRef] , [ ] 71. Yihui, D.; Chan, J.C.L. The East Asian summer monsoon: An overview. Meteorol. Atmos. Phys. 2005 71. Yihui, D.; Chan, J.C.L. The East Asian summer monsoon: An overview. Meteorol. Atmos. Phys. 2005, 89, 117 142. [CrossRef] 72. Kim, J.-S.; Kug, J.-S.; Jeong, S.-J.; Park, H.; Schaepman-Strub, G. Extensive ﬁres in southeastern Siberian permafrost linked to preceding Arctic Oscillation. Sci. Adv. 2020, 6, eaax3308. [CrossRef] 72. Kim, J.-S.; Kug, J.-S.; Jeong, S.-J.; Park, H.; Schaepman-Strub, G. Extensive ﬁres in southeastern Siber preceding Arctic Oscillation. Sci. Adv. 2020, 6, eaax3308. [CrossRef] 73. Wallmann, J. A Procedure for Forecasting Dry Thunderstorms in the Great Basin Using the Dynamic Tropopause and Alternate Tools for Assessing Instability. NOAA/NWS WR Tech. Attach. 2004. Available online: www.wrh.noaa.gov/media/wrh/online_ publications/TAs/ta0408.pdf (accessed on 12 December 2022). 74. Nauslar, N.; Kaplan, M.L.; Wallmann, J.; Brown, T. A forecast procedure for dry thunderstorms. J. Oper. Meteorol. 2013, 1, 200–214. [CrossRef] 75. Hayasaka, H. Rare and Extreme Wildland Fire in Sakha in 2021. Atmosphere 2021, 12, 1572. [CrossRef] 76. Hayasaka, H. Fire Weather Conditions in Boreal and Polar Regions in 2002–2021. Atmosphere 2022, 13, 1117. Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. References [CrossRef] 77 Z h d l i P Pf hl S W li H Fi k A H A L i l i f h i i l i d i h h 76. Hayasaka, H. Fire Weather Conditions in Boreal and Polar Regions in 2002–2021. Atmosphere 2022, 13, 77. Zschenderlein, P.; Pfahl, S.; Wernli, H.; Fink, A.H. A Lagrangian analysis of upper-tropospheric anticyclones associated with heat waves in Europe. Weather Clim. Dyn. 2020, 1, 191–206. [CrossRef] 78. Yegorov, Y. Population Structures in Russia: Optimality and Dependence on Parameters of Global Evolution. Region 2016, 3, 87–104. [CrossRef] 79. Soja, A.J.; Tchebakova, N.M.; French, N.H.; Flannigan, M.D.; Shugart, H.H.; Stocks, B.J.; Sukhinin, A.I.; Parfenova, E.; Chapin, F.S.; Stackhouse, P.W. Climate-induced boreal forest change: Predictions versus current observations. Glob. Planet. Change 2007, 56, 274–296. [CrossRef] 80. Flannigan, M.D.; Krawchuk, M.A.; De Groot, W.J.; Wotton, B.M.; Gowman, L.M. Implications of changing climate for global wildland ﬁre. Int. J. Wildland Fire 2009, 18, 483–507. [CrossRef] 81. Cherenkova, E.; Kononova, N.; Muratova, N. SUMMER DROUGHT 2010 IN THE EUROPEAN RUSSIA 2013, 6, 55–66. [CrossRef] Kononova, N.; Muratova, N. SUMMER DROUGHT 2010 IN THE EUROPEAN RUSSIA. Geogr. Environ. Sus CrossRef] 82. Vogel, M.M.; Zscheischler, J.; Seneviratne, S.I. Varying soil moisture–atmosphere feedbacks explain divergent temperature extremes and precipitation projections in central Europe. Earth Syst. Dyn. 2018, 9, 1107–1125. [CrossRef] 83. Sato, T.; Nakamura, T. Intensiﬁcation of hot Eurasian summers by climate change and land–atmosphere interactions. Sci. Rep. 2019, 9, 10866. [CrossRef] 82. Vogel, M.M.; Zscheischler, J.; Seneviratne, S.I. Varying soil moisture–atmosphere feedbacks explain divergent temperature extremes and precipitation projections in central Europe. Earth Syst. Dyn. 2018, 9, 1107–1125. [CrossRef] 82. Vogel, M.M.; Zscheischler, J.; Seneviratne, S.I. Varying soil moisture–atmosphere feedbacks explain divergent temperature extremes and precipitation projections in central Europe. Earth Syst. Dyn. 2018, 9, 1107–1125. [CrossRef] 83. Sato, T.; Nakamura, T. Intensiﬁcation of hot Eurasian summers by climate change and land–atmosphere interactions. Sci. Rep. 2019, 9, 10866. [CrossRef] 83. Sato, T.; Nakamura, T. Intensiﬁcation of hot Eurasian summers by climate change and land–atmosphere interactions. Sci. Rep. 2019, 9, 10866. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
https://openalex.org/W2890210998	https://europepmc.org/articles/pmc6160213?pdf=render	English	null	The association of rainfall and Buruli ulcer in southeastern Australia	PLoS neglected tropical diseases	2,018	cc-by	6,080	RESEARCH ARTICLE Arvind YerramilliID1,2, Ee Laine Tay3, Andrew J. Stewardson4,5, Janet Fyfe6, Daniel P. O’Brien2,7, Paul D. R. Johnson1,4 Arvind YerramilliID1,2, Ee Laine Tay3, Andrew J. Stewardson4,5, Janet Fyfe6, Daniel P. O’Brien2,7, Paul D. R. Johnson1,4 1 University of Melbourne, Department of Medicine, Austin Health, Heidelberg, Victoria, Australia, 2 Department of Medicine and Infectious Diseases, The Royal Melbourne Hospital, University of Melbourne, Parkville, Victoria, Australia, 3 Health Protection Branch, Department of Health & Human Services, Melbourne, Victoria, Australia, 4 Infectious Diseases Department, Austin Health, Heidelberg, Victoria, Australia, 5 Department of Infectious Diseases, Alfred Hospital and Central Clinical School, Monash University, Melbourne, Victoria, Australia, 6 Victorian Infectious Diseases References Laboratory, Peter Doherty Institute for Infection and Immunity, Parkville, Victoria, Australia, 7 Department of Infectious Diseases, Barwon Health, Geelong, Victoria, Australia a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 * a.yerramilli15@outlook.com (AY); paul.johnson@austin.org.au (PDRJ) Methodology Editor: Gerd Pluschke, Swiss Tropical and Public Health Institute, SWITZERLAND Editor: Gerd Pluschke, Swiss Tropical and Public Health Institute, SWITZERLAND We created yearly and monthly graphs comparing rainfall with local Buruli ulcer incidence for the period 2004–2016 by endemic region and then considered a range of time lag inter- vals of 0–24 months to investigate patterns of correlation. Received: April 13, 2018 Accepted: August 16, 2018 Published: September 17, 2018 Conclusions Optimal positive correlation for the Bellarine Peninsula occurred with a 12-month prior rain- fall lag, however, no significant correlation was observed on the Mornington Peninsula for any time lag. These results provide an update in evidence to further explore transmission mechanisms which may differ between these geographically proximate endemic regions. Copyright: © 2018 Yerramilli et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Relevant rainfall and case data can be found within the paper with additional rainfall data prior to 2004 freely available from the Australian Bureau of Meteorology (http:// www.bom.gov.au/climate/data/). Background Citation: Yerramilli A, Tay EL, Stewardson AJ, Fyfe J, O’Brien DP, Johnson PDR (2018) The association of rainfall and Buruli ulcer in southeastern Australia. PLoS Negl Trop Dis 12(9): e0006757. https://doi.org/10.1371/journal. pntd.0006757 Citation: Yerramilli A, Tay EL, Stewardson AJ, Fyfe J, O’Brien DP, Johnson PDR (2018) The association of rainfall and Buruli ulcer in southeastern Australia. PLoS Negl Trop Dis 12(9): e0006757. https://doi.org/10.1371/journal. pntd.0006757 Buruli ulcer has been increasing in incidence in southeastern Australia with unclear trans- mission mechanisms. We aimed to investigate the link between rainfall and case numbers in two endemic areas of the state of Victoria; the Bellarine and Mornington Peninsulas. Introduction Buruli ulcer (BU), endemic to west and sub-Saharan Africa, Australia and several other coun- tries, is a potentially severe infection of subcutaneous tissue, caused by the environmental pathogen Mycobacterium ulcerans (MU) [1]. Exact transmission mechanisms remain unclear despite new understanding of complex ecological interactions [2, 3]. Increased disease preva- lence has previously been linked to areas of forested land cover with low level terrain and prox- imity to wetlands and swamps [2–5]. Optimal mean temperatures with reduced sunlight and ample oxygen may also play a pivotal role given the pathogen grows best in these environmen- tal conditions [2, 3]. Rainfall has also been shown to be a central driving environmental trigger with recent field studies in Africa linking rainfall with BU diagnoses and an increased detec- tion of MU in the environment [6, 7]. Established time lags between rain and subsequent recognition of incident BU cases may provide a conceptual framework for understanding the ecology and transmission of MU. The first is a long mean incubation period of 4–5 months [8, 9]. Secondly, BU tends to have a slow clinical progression resulting in diagnosis 1–2 months after a lesion is first observed [10, 11]. There is therefore an expected lag of at least 5–6 months between potential permissive environ- mental events such as increased rainfall and an increase in human disease [8, 9, 11]. Climate is well recognised to influence infectious disease outbreaks through alterations in pathogen, reservoir and vector dynamics as well as influencing human behavior [12, 13]. In southeastern Australia, native mammals such as ringtail (Pseudocheirus peregrinus) and brush- tail (Trichosurus vulpecula) possums have been implicated as reservoir/amplifiers of MU while certain species of mosquitoes have been suggested to act as mechanical vectors [14–18]. Rain may influence the number of possums infected through increased environmental pathogen abundance, subsequent digestion and excretion of contaminated vegetation, and caecotrophic behavior [5, 16]. Further, local salt marsh mosquito population increases are linked to rainfall, with previous studies noting an increase in local abundance after periods of heavy rain [19, 20]. Additional time therefore needs to be considered from an environmental trigger such as rain and the resultant growth of M. ulcerans in the environment, along with time delays related to potential modes of transmission such as uptake in reservoirs with or without spread by insect vectors. Author summary Buruli ulcer, a mycobacterial infection resulting in destructive soft tissue lesions, has been increasing in incidence in southeastern Australia over recent years. Exact transmission mechanisms and therefore preventative measures remain unclear. We aimed to investi- gate a possible link between Buruli ulcer and rainfall for two major endemic areas of the state of Victoria; the Bellarine and Mornington Peninsulas. Our results demonstrate a pos- itive correlation with rainfall on the Bellarine Peninsula with a prior rainfall lag of 12 Funding: AJS is supported by a National Health and Medical Research Council Early Career Fellowship (APP1141398). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 1 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Competing interests: The authors have declared that no competing interests exist. months but no correlation on the Mornington Peninsula. Established time-frames such as the incubation period and average delay to diagnosis (due to slow and often asymptomatic progression of the disease) make up a total of at least 5–6 months. On the Bellarine Penin- sula, there appears to be another 5–6 months of unaccounted time from rainfall to trans- mission of the pathogen. This provides a basis to explore previously proposed transmission mechanisms which, based on the discrepancy noted with rainfall correlation, may differ between endemic regions. Introduction To our knowledge, there have been no studies to date in southeastern Australia which have investigated the longitudinal temporal and seasonal relationship of rain and BU. We therefore aimed to determine whether a correlation exists between these two entities over a 13-year study period including the effect of a range of time lag adjustments to further explore local transmission mechanisms. Methods This study is a retrospective analysis of BU notifications data from 2004 to 2016 obtained from the Victorian Government’s Department of Health & Human Services (DHHS) and the 2 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Fig 1. Map of Buruli endemic zones in the southeastern Australian state of Victoria with approximate location of weather stations (numbered). Corresponding numbers for weather stations can be found in Table 3. Figure with adapted text from Trubiano et al [8] as per the Creative Commons Attribution License [https://doi.org/10.1371/journal.pntd.0002463]. Approximate location of weather stations was sourced from the Bureau of Meteorology’s Climate Data Online services [22]. Fig 1. Map of Buruli endemic zones in the southeastern Australian state of Victoria with approximate location of weather stations (numbered). Corresponding numbers for weather stations can be found in Table 3. Figure with adapted text from Trubiano et al [8] as per the Creative Commons Attribution License [https://doi.org/10.1371/journal.pntd.0002463]. Approximate location of weather stations was sourced from the Bureau of Meteorology’s Climate Data Online services [22]. https://doi.org/10.1371/journal.pntd.0006757.g001 https://doi.org/10.1371/journal.pntd.0006757.g001 Victorian Infectious Diseases Reference Laboratory (VIDRL) [21]. Buruli ulcer was made legally notifiable in the state of Victoria from January 2004 under the Public Health and Well- being Act 2008. Aggregated anonymized data used were BU cases by month and year of notifi- cation, postcode of residence at the time of notification and the most likely exposure location based on patient history (residence or travel to an endemic area or perceived acquisition in a specific geographical region). In Victoria, BU is acquired in well-defined endemic zones; par- ticularly the Bellarine and Mornington Peninsulas to the southwest and southeast of the city of Melbourne, respectively [8, 9]. We obtained rainfall data from all weather stations in these major endemic areas from the Australian Bureau of Meteorology’s (BOM) Climate Data Online (Fig 1) [22]. Due to very low case numbers since 2004 we excluded two areas from sub- sequent analyses; Phillip Island, near the endemic region of the Mornington Peninsula, and East Gippsland, located approximately 250km east of Melbourne. We compared annual monthly rainfall averages with BU incidence per 100,000 person- years in the Bellarine and Mornington Peninsulas from 2004–2016 to examine secular trends. Rainfall and Buruli ulcer in southeastern Australia Table 1. Postcode regions as defined for Buruli ulcer cases from residents in corresponding endemic regions. Bellarine Peninsula Mornington Peninsula Postcode Regions Postcode Regions 3222 Clifton Springs, Curlewis, Drysdale 3912 Somerville 3223 St Leonards, Indented Heads 3915 Hastings 3224 Leopold 3918 Bittern, Crib Point 3225 Point Lonsdale, Queenscliff 3930 Mt Eliza 3226 Ocean Grove 3931 Mornington 3227 Barwon Heads, Connewarre 3933 Moorooduc 3934 Mt Martha 3936 Dromana 3939 Rosebud, Boneo 3940 Rosebud West 3941 Rye, St Andrews Beach, Tootgarook 3942 Blairgowrie 3943 Sorrento 3944 Portsea https://doi.org/10.1371/journal.pntd.0006757.t001 Table 1. Postcode regions as defined for Buruli ulcer cases from residents in corresponding endemic regions. monthly data points (data point = amount of rainfall for a given month in a given year) in an endemic region. The result for each year was then divided by the number of data points avail- able to negate the effect of any missing data points. Finally, these results were divided by the number of weather stations with valid data for those years to produce annual monthly rainfall averages from 2004–2016. BU incidence was defined as the number of BU cases residing in an endemic area divided by the approximate population of the endemic area (defined through postcode region with at least one case of BU over the study period–Table 1) multiplied by 100,000. Population data was sourced from The Australian Bureau of Statistics (ABS) Census data 2005, 2011 and 2016 [23]. To investigate seasonal trends in the two endemic regions, we compared monthly rainfall (averaged over weather stations with valid data points) with the total number of BU cases (both residents and travelers) for each calendar month over the same 13-year period. Transmission delays were examined by comparing BU incidence with averaged total rain- fall (over corresponding weather stations) for the 12-months prior for each of 0 to 24-month time lags which we investigated for each year in the reporting period. Total 12-month prior rainfall was used to match the temporal resolution of incidence calculations which were also derived from 12-month (yearly) BU case totals as defined above. Methods This was achieved by first computing the sum of yearly rainfall across all weather stations with PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 3 / 14 https://doi.org/10.1371/journal.pntd.0006757.t001 Results The total number of BU cases notified in Victoria from 2004–2016 was 839 (median = 61, IQR = 32–80) with an overall increase observed during the study period (Table 2). The Bellar- ine Peninsula, where BU was unknown prior to 1998, had 403 diagnosed cases in the study period (residents + travelers) and 7 weather stations with publicly available records. Out of the 403 diagnosed cases, 253 cases were from residents as determined by postcode region (Tables 1–3). On the Mornington Peninsula, there were 257 cases (residents + travelers) during the study period and 15 weather stations (Tables 1–3). Out of the 257 diagnosed cases, 155 were residents (Table 2). BU was first documented on the Mornington Peninsula in 1990, however after 1998 there were few cases linked to this area until 2012, from when there has been a marked year-on-year increase in notified cases. The remaining 179 cases were located outside the endemic zones, largely distributed sporadically in various other areas of Victoria or, for a small subset, for which precise exposure locations were unknown. Comparisons of raw data (0-month lag) from 2004–2016 within each endemic zone revealed very weak correlation between the number of BU diagnoses and annual monthly rain- fall averages (Bellarine Peninsula; r = 0.07, 95%, CI = -0.50–0.60, Mornington Peninsula; r = -0.09, 95% CI = -0.61–0.48). Adjusting for total 12-month prior rainfall lags from 0–24 months, optimal with very strong positive correlation for the Bellarine Peninsula occurred with a 12-month lag (r = 0.82, 95% CI = 0.49–0.94) (Fig 2). No moderate to very strong Total number of Buruli ulcer cases and incidence by year with annual monthly rainfall averages calculated for endemic r rainfall lag). ber of Buruli ulcer cases and incidence by year with annual monthly rainfall averages calculated for endemic regions and g) idence by year with annual monthly rainfall averages calculated for endemic regions and state-wide for Victoria Table 2. Total number of Buruli ulcer cases and incidence by year with annual monthly rainfall averages calculated for endemic regions and state-wide for Victoria (no prior rainfall lag). Rainfall and Buruli ulcer in southeastern Australia (Prism 7 for Windows, Version 7.01, California, USA). Qualitative inspection from map over- lays of terrain, amount of sunlight exposure, as well as minimum and maximum temperatures were performed to determine differences between endemic regions. These map overlays were sourced from BOM Climate Data Online [22]. As the data used in this study were collected under the legislative authority of the Public Health and Wellbeing Act 2008, approval from a Human Research Ethics Committee was not required. Approval for data access for this project was given by DHHS. https://doi.org/10.1371/journal.pntd.0006757.t002 For example, for a 1-month lag, total cases for 2004 would be matched with averaged (across weather stations) 12-month prior rainfall from December 2003 to November 2004 while for a 2-month lag, BU incidence for 2004 would be matched with averaged rainfall from November 2003 to October 2004 and so on for every year from 2004–2016. Annual monthly rainfall averages for each monthly time lag (0–24 months) were then each plotted against BU incidence from 2004–2016, making a total of 25 separate plots. The correlation coefficient for each plot was then calculated and plot- ted against the degree of time lag (each of 0–24 months) for each endemic region to observe which time lag produced the greatest correlation (S1 Fig). Statistical analyses were performed with measured outcomes including Pearson correlation coefficients (r) and 95% confidence intervals (CI). The strength of the correlation was defined based on previously described cut-offs by Evans for the absolute value of r: 0.00–0.19: “very weak”; 0.20–0.39: “weak”; 0.40–0.59: “moderate”; 0.60–0.79: “strong” and 0.80–1.0: “very strong” [24]. All graphical and statistical analyses were performed using Graphpad Prism 4 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Table 3. Weather stations with recorded rainfall data from the Mornington and Bellarine Peninsulas. Consecu- tive missing data for 1 year or more noted below. Number ID corresponds to locations of weather stations in Fig 1. Number ID Bellarine Peninsula Mornington Peninsula 1 Torquay Golf Club# Bonbeach (Carrum) 2 Breakwater (Geelong Racecourse) ¶ Cranbourne South 3 Barwon Heads Golf Club Cranbourne Botanical Gardens 4 Ocean Grove Mornington 5 Clifton Springs§ Mooroduc 6 Drysdale (Brimdale) Devilbend† 7 Portarlington Hastings 8 Cerberus 9 Dromana Sussex Farm 10 Merricks Stonier’s Winery 11 Rosebud (Country Club) 12 Main Ridge‡ 13 Shoreham 14 Cape Schanck¥ 15 Flinders https://doi.org/10.1371/journal.pntd.0006757.t003 correlations were observed for the Mornington Peninsula regardless of the degree of time lag (Fig 3). Additionally, from qualitative inspection of map overlays using BOM Climate Data Online, the amount of sunlight exposure, maximum and minimum temperatures, as well as elevation above sea level did not differ greatly between endemic regions [22]. Analysis of seasonal trends revealed most rain falls in the Buruli-endemic regions over the winter months of June, July and August with an isolated spike towards the end of the year in November (Fig 4). Comparison of rainfall between endemic regions resulted in a very strong positive correlation (r = 0.94, 95% CI = 0.79–0.98), suggesting the amount and pattern of rain- fall was similar between the two regions over the study period. On the Bellarine Peninsula, high rainfall in the winter months also corresponded with the greatest number of BU diagno- ses. Statistical analysis of cases in this region compared to rainfall by calendar month (Fig 4A) revealed very strong positive correlation throughout the calendar year over the 13-year study period (r = 0.86, 95% CI = 0.56–0.96). Cases peaked on the Mornington Peninsula during the spring season months of September and October when rainfall was higher than summer but lower than in winter. Compared to the Bellarine Peninsula, only a moderate positive correla- tion was noted in this region (Fig 4B) when comparing cases with rainfall throughout the cal- endar year (r = 0.53, 95% CI = -0.06–0.84). Results Year BU cases (incidence), Victoria BU cases (incidence), BP† residents BU cases (incidence), MP‡ residents Annual monthly rainfall average, BP (mm) Annual monthly rainfall average, MP (mm) 2004 26 (0.5) 8 (18.2) 1 (0.8) 50.5 69.2 2005 41 (0.8) 19 (43.2) 2 (1.7) 50.4 67.4 2006 61 (1.2) 31 (70.5) 1 (0.8) 27.3 42.1 2007 17 (0.3) 3 (6.8) 1 (0.8) 50.0 59.5 2008 34 (0.6) 18 (41) 0 (0.0) 37.5 53.1 2009 28 (0.5) 13 (26.3) 1 (0.8) 41.4 56.6 2010 32 (0.6) 15 (30.3) 0 (0.0) 61.3 80.9 2011 80 (1.4) 41 (82.9) 5 (3.9) 63.9 77.4 2012 77 (1.4) 34 (68.8) 4 (3.1) 56.1 78.6 2013 65 (1.1) 16 (32.4) 14 (11.0) 45.2 71.8 2014 89 (1.5) 26 (43.6) 17 (12.4) 38.9 57.0 2015 107 (1.8) 17 (28.5) 45 (32.7) 35.6 51.3 2016 182 (2.9) 12 (20.1) 64 (46.6) 50.2 67.2 Table 2. Total number of Buruli ulcer cases and incidence by year with annual monthly rainfall averages calculated for endemic regions and state-wide for Victoria (no prior rainfall lag). Table 2. Total number of Buruli ulcer cases and incidence by year with annual monthly rainfall averages (no prior rainfall lag). 5 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Discussion We have shown a clear correlation between rainfall and BU incidence in one endemic region of Victoria, the Bellarine Peninsula, over a 13-year period when we applied a 12-month prior rainfall lag. The degree of time lag is supported by seasonal trends with rainfall very strongly 6 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Fig 2. Yearly BU incidence compared to annual monthly rainfall averages for the Bellarine Peninsula from 2004–2016. Fig 2A (top) shows comparisons with no prior rainfall lag (0-month lag) while Fig 2B (bottom) shows case totals compared to 12 months prior rainfall lag (12-month lag). BU = Buruli ulcer, BP = Bellarine Peninsula. https://doi.org/10.1371/journal.pntd.0006757.g002 Fig 2. Yearly BU incidence compared to annual monthly rainfall averages for the Bellarine Peninsula from 2004–2016. Fig 2A (top) shows comparisons with no prior rainfall lag (0-month lag) while Fig 2B (bottom) shows case totals compared to 12 months prior rainfall lag (12-month lag). BU = Buruli ulcer, BP = Bellarine Peninsula. https://doi.org/10.1371/journal.pntd.0006757.g002 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 7 / 14 Rainfall and Buruli ulcer in southeastern Australia Fig 3. Yearly BU incidence compared to annual monthly rainfall averages for the Mornington Peninsula from 2004–2016. Fig 3A (top) shows comparisons with no prior rainfall lag (0-month lag) while Fig 3B (bottom) shows case totals compared to 12 months prior rainfall lag (12-month lag). BU = Buruli ulcer, MP = Mornington Peninsula. https://doi.org/10.1371/journal.pntd.0006757.g003 Fig 3. Yearly BU incidence compared to annual monthly rainfall averages for the Mornington Peninsula from 2004–2016. Fig 3A (top) shows comparisons with no prior rainfall lag (0-month lag) while Fig 3B (bottom) shows case totals compared to 12 months prior rainfall lag (12-month lag). BU = Buruli ulcer, MP = Mornington Peninsula. https://doi.org/10.1371/journal.pntd.0006757.g003 https://doi.org/10.1371/journal.pntd.0006757.g003 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 8 / 14 Rainfall and Buruli ulcer in southeastern Australia Fig 4. Total Buruli ulcer cases compared to averaged monthly rainfall, for each calendar month, from 2004–2016 on the Bellarine Peninsula (Fig 4A; top) and Mornington Peninsula (Fig 4B; bottom). Numbers 1–12 on the x-axes represents months Jan-Dec respectively. BU = Buruli ulcer. https://doi.org/10.1371/journal.pntd.0006757.g004 Rainfall and Buruli ulcer in southeastern Australia Fig 4. Rainfall and Buruli ulcer in southeastern Australia correlating with the number of BU diagnoses on the Bellarine Peninsula throughout a given calendar year. That is, the amount of rainfall at any period in time in a calendar year strongly correlates with the number of cases of BU at the same point in time or by intervals of one whole calendar year. Accounting for recognised time lags of 5–6 months from transmission of MU to clinical diagnosis, there appears to be an additional time lag of 5–6 months after rain events in this region. We also provide evidence for a recent exponential rise of BU incidence in the other major endemic area of the Mornington Peninsula from 2012. However, no correla- tion with rainfall was observed on the Mornington Peninsula regardless of the degree of time lag we applied. This result may therefore suggest differences in MU ecology and transmission between endemic zones despite their geographical proximity. At present, we have no clear explanation for the discrepancy in rainfall correlation between these two major endemic regions of Victoria. However, there is increasing evidence that mos- quitoes may act as mechanical vectors in Victoria, acquiring M. ulcerans directly or indirectly from infected possums which shed MU in their faeces [14–18]. On the Bellarine peninsula, salt marsh mosquitoes (Aedes camptorhynchus) are the major pest species and their numbers increase rapidly after significant rain events or high tides and warm weather [14, 19, 20]. As a salt marsh breeding species, adult A. camptorhynchus generally lay their eggs in damp soil. After rainfall occurs there is a subsequent rise in soil water levels in turn resulting in a drop in dissolved oxygen which stimulates hatching [19, 25]. An expected lag from rainfall is thus favourable to allow for these ecological events with a 1-month lag previously documented by Barton et al as producing the greatest species abundance in the East Gippsland area [19]. This research group also found that mosquito numbers peak in November with a steady reduction every month thereafter through to April. Due to potential differences in landscape and other climate factors, however, the seasonal variability and correlation with rainfall of A. camptor- hynchus may differ on the Bellarine Peninsula, where M. ulcerans DNA was found in these spe- cies, compared to East Gippsland. Nevertheless, this 1-month lag with rainfall and increased species abundance may fit with our 12-month rainfall-to-diagnosis model. Discussion Total Buruli ulcer cases compared to averaged monthly rainfall, for each calendar month, from 2004–2016 on the Bellarine Peninsula (Fig 4A; top) and Mornington Peninsula (Fig 4B; bottom). Numbers 1–12 on the x-axes represents months Jan-Dec respectively. BU = Buruli ulcer. Fig 4. Total Buruli ulcer cases compared to averaged monthly rainfall, for each calendar month, from 2004–2016 on the Bellarine Peninsula (Fig 4A; top) and Mornington Peninsula (Fig 4B; bottom). Numbers 1–12 on the x-axes represents months Jan-Dec respectively. BU = Buruli ulcer. https://doi.org/10.1371/journal.pntd.0006757.g004 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 9 / 14 Rainfall and Buruli ulcer in southeastern Australia habits of A. notoscriptus implies less reliance on rainfall and more on constant water inunda- tion such as the watering of gardens and the use of bore water that continue all year round and are likely to increase in hot, dry weather when A. camptorhynchus die off. Other possibilities including movement of the pathogen into a new reservoir on the Mornington Peninsula that is less dependent on rain such as permanent wetlands supported by irrigation, or an increasing prevalence in possums. Furthermore, population dynamics, altered human behavior including travel predilection, the abundance of natural water bodies and different pathogen environmen- tal niche requirements may also contribute to the discrepancy noted in rainfall correlation between regions. A major limitation of our study is that rain was the only environmental variable quantified as opposed to others such as humidity, temperature, sunlight exposure and ground elevation which may all also influence the number of BU diagnoses [2]. From qualitative inspection of map overlays using the Bureau of Meteorology Climate Data Online services, average mini- mum and maximum temperatures, as well as the amount of sunlight exposure and elevation above sea level, do not appear to differ greatly between endemic areas; however, these need to be further explored[20]. Other environmental niche characteristics for MU such as humidity levels, the availability of nutrients and ample oxygen are recognizably difficult to account for. These factors may alter the eco-biological relationship between climate factors such as rain and BU incidence with a linear correlation perhaps being too simplistic. Nevertheless, demon- strating a linear correlation between the two entities reinforces the importance of rainfall in BU disease and must therefore be considered in a model with other eco-biological factors in future investigations. Additionally, occurrences of missing rainfall data for individual weather stations may have resulted in systematic errors (Table 2). Most missing data points, however, were sporadic and these were excluded from our calculations by averaging over a large set of valid data points. The impact on resultant trend analyses is therefore unlikely to be significant. The incidence calculations may have also led to systematic errors due to a limitation of 5-yearly Census data- sets collected by ABS whereas population rates are likely to have fluctuated on a yearly basis. With rainfall peaking in August, there may be an increased abundance of mosquitoes shortly after causing an increase in infection among pos- sum reservoirs such as the common ringtail. Furthermore, rainfall has been shown to signifi- cantly increase the survival of the common ringtail possum compared to periods of drought [26]. Thus, an increase in infected possums and contaminated mosquito populations would further contribute to environmental contamination and the overall increasing environmental abundance of M. ulcerans. The subsequent isolated peak in rainfall in November may then lead to another peak in mosquito abundance from late November through to December. It is during this period, the beginning of the hot Australian summer, that human hosts are most vulnerable with an increased tendency to wear less protective clothing, a well-documented risk factor for developing the disease [2]. Furthermore, in a prior study, we confirmed the distribu- tion of Buruli lesions on the human body to be highly specific, localised to areas such as the distal limb surfaces which tend to be more exposed in warmer months [27]. We also confirmed the summer months of Dec-Feb as the most probable date of infection for a large cohort of patients (n = 338) by using diagnosis dates and back calculating from the already established lag times of 5–6 months (incubation period plus delay to diagnosis) [27]. From acquisition of the pathogen in Dec-Feb, adding another 5–6 months would result in diagnosis in the winter months of Jun-Aug which is consistent with our seasonal observations in the present study. On the more densely populated Mornington peninsula, Aedes notoscriptus are likely to pre- dominate [personal communication, Peter Mee: AgriBio/DEDJTR]. Compared to A. camptor- hynchus, A. notoscriptus is a container breeder and therefore develops in natural or artificial water-holding containers. M. ulcerans DNA has been detected in both species but the breeding PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 10 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 S1 Checklist. STROBE checklist. (DOCX) S1 Fig. Correlation plot for monthly time lags. Correlation coefficient compared with time lags for each of 0–24 months of annual monthly rainfall averages vs annual Buruli incidence (2004–2016) on the Bellarine Peninsula (top) and Mornington Peninsula (bottom). BU = Buruli ulcer. Project administration: Ee Laine Tay. Project administration: Ee Laine Tay. Resources: Ee Laine Tay, Paul D. R. Johnson. Software: Arvind Yerramilli. Supervision: Ee Laine Tay, Andrew J. Stewardson, Paul D. R. Johnson. Validation: Ee Laine Tay, Janet Fyfe, Daniel P. O’Brien, Paul D. R. Johnson. Visualization: Arvind Yerramilli. Visualization: Arvind Yerramilli. Writing – original draft: Arvind Yerramilli. Writing – original draft: Arvind Yerramilli. Writing – review & editing: Arvind Yerramilli, Ee Laine Tay, Andrew J. Stewardson, Janet Fyfe, Daniel P. O’Brien, Paul D. R. Johnson. Author Contributions Conceptualization: Arvind Yerramilli, Ee Laine Tay, Paul D. R. Johnson. Conceptualization: Arvind Yerramilli, Ee Laine Tay, Paul D. R. Johnson. Data curation: Arvind Yerramilli, Ee Laine Tay, Andrew J. Stewardson, Janet Fyfe, Daniel P. O’Brien, Paul D. R. Johnson. Formal analysis: Arvind Yerramilli, Andrew J. Stewardson, Daniel P. O’Brien, Paul D. R. Johnson. Investigation: Arvind Yerramilli, Daniel P. O’Brien, Paul D. R. Johnson. Methodology: Arvind Yerramilli, Ee Laine Tay, Andrew J. Stewardson, Janet Fyfe, Paul D. R. Johnson. Project administration: Ee Laine Tay. Nevertheless, we excluded travelers who had contracted BU from our study and only used case totals of patients who reside in endemic regions with population data sourced for postcodes within these regions only. For these reasons, discrepancies in incidence are also unlikely to sig- nificantly alter our principal findings. To investigate time lags, a maximum of 24 months with intervals of 1 month was used to allow meaningful comparison between the results of previous studies [5–7]. One such study by van Ravensway et al using Victorian data has shown a period of high rainfall 19 months, along with drier conditions 5 months prior, to correlate with increased BU emergence [5]. However, this study used case data from 1981–2008 when case numbers were relatively sparse compared to the last decade and examined climate variables with case data from specific points in time rather than investigating longitudinal temporal and seasonal relationships. Another study in an endemic area of Cameroon, Africa, proposed a 5–6 month delay from the high rainy season (Aug-Oct) and peaks in BU incidence (Mar-Apr) [6]. In this instance, the discrepancy with our results may be explained by the absence of an intermediate small animal reservoir which has never been confirmed in Africa [2]. Nevertheless, human-body lesion distribution, which remains highly specific, appears to be similar across countries suggesting unifying global trans- mission mechanisms [10, 27, 28]. Evidence provided by our study will allow for more focused future investigations of BU transmission taking into consideration time frames for potential modes of transmission, in an overall effort to prevent the spread of disease. 11 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia 7. Aboagye SY, Ampah KA, Ross A, Asare P, Otchere ID, Fyfe J, et al. Seasonal Pattern of Mycobacte- rium ulcerans, the Causative Agent of Buruli Ulcer, in the Environment in Ghana. Microb Ecol. 2017; 74 (2):350–61. https://doi.org/10.1007/s00248-017-0946-6 PMID: 28238016 8. Trubiano JA, Lavender CJ, Fyfe JA, Bittmann S, Johnson PD. The incubation period of Buruli ulcer (Mycobacterium ulcerans infection). PLoS Negl Trop Dis. 2013; 7(10):e2463. https://doi.org/10.1371/ journal.pntd.0002463 PMID: 24098820 9. Loftus MJ, Trubiano JA, Tay EL, Lavender CJ, Globan M, Fyfe JAM, et al. The incubation period of Bur- uli ulcer (Mycobacterium ulcerans infection) in Victoria, Australia—Remains similar despite changing geographic distribution of disease. PLoS Negl Trop Dis. 2018; 12(3):e0006323. https://doi.org/10.1371/ journal.pntd.0006323 PMID: 29554096 10. Boyd SC, Athan E, Friedman ND, Hughes A, Walton A, Callan P, et al. Epidemiology, clinical features and diagnosis of Mycobacterium ulcerans in an Australian population. Med J Aust. 2012; 196(5):341–4. PMID: 22432674 11. Quek TY, Henry MJ, Pasco JA, O’Brien DP, Johnson PD, Hughes A, et al. Mycobacterium ulcerans infection: factors influencing diagnostic delay. Med J Aust. 2007; 187:561–3. PMID: 18021043 12. Kuhn K, Campbell-Lendrum D, Haines A, Cox J. Using climate to predict infectious disease epidemics. The World Health Organisation (WHO) 2005. [http://www.who.int/globalchange/publications/ infectdiseases/en/]; First accessed July 2016. 13. Altizer S, Dobson A, Hosseini P, Hudson P, Pascual M, Rohani P. Seasonality and the dynamics of infectious diseases. Ecol Lett. 2006; 9(4):467–84. https://doi.org/10.1111/j.1461-0248.2005.00879.x PMID: 16623732 14. Johnson PD, Azuolas J, Lavender CJ, Wishart E, Stinear TP, Hayman JA, et al. Mycobacterium ulcer- ans in mosquitoes captured during outbreak of Buruli ulcer, southeastern Australia. Emerg Infect Dis. 2007; 13(11):1653–60. https://doi.org/10.3201/eid1311.061369 PMID: 18217547 15. Johnson PD, Lavender CJ. Correlation between Buruli ulcer and vector-borne notifiable diseases, Vic- toria, Australia. Emerg Infect Dis. 2009; 15(4):614–5. https://doi.org/10.3201/eid1504.081162 PMID: 19331750 16. Fyfe JA, Lavender CJ, Handasyde KA, Legione AR, O’Brien CR, Stinear TP, et al. A major role for mam- mals in the ecology of Mycobacterium ulcerans. PLoS Negl Trop Dis. 2010; 4(8):e791. https://doi.org/ 10.1371/journal.pntd.0000791 PMID: 20706592 17. Wallace JR, Gordon MC, Hartsell L, Mosi L, Benbow ME, Merritt RW, et al. Interaction of Mycobacte- rium ulcerans with mosquito species: implications for transmission and trophic relationships. Appl Envi- ron Microbiol. 2010; 76(18):6215–22. https://doi.org/10.1128/AEM.00340-10 PMID: 20675453 18. Wallace JR, Mangas KM, Porter JL, Marcsisin R, Pidot SJ, Howden B, et al. References 1. MacCallum P, Tolhurst JC, et al. A new mycobacterial infection in man. J Pathol Bacteriol. 1948; 60 (1):93–122. 2. Merritt RW, Walker ED, Small PL, Wallace JR, Johnson PD, Benbow ME, et al. Ecology and transmis- sion of Buruli ulcer disease: a systematic review. PLoS Negl Trop Dis. 2010; 4(12):e911. https://doi.org/ 10.1371/journal.pntd.0000911 PMID: 21179505 3. Huang GK, Johnson PD. Epidemiology and management of Buruli ulcer. Expert Rev Anti Infect Ther. 2014; 12(7):855–65. https://doi.org/10.1586/14787210.2014.910113 PMID: 24918117 4. Wagner T, Benbow ME, Brenden TO, Qi J, Johnson RC. Buruli ulcer disease prevalence in Benin, West Africa: associations with land use/cover and the identification of disease clusters. Int J Health Geogr. 2008; 7:25. https://doi.org/10.1186/1476-072X-7-25 PMID: 18505567 5. van Ravensway J, Benbow ME, Tsonis AA, Pierce SJ, Campbell LP, Fyfe JA, et al. Climate and land- scape factors associated with Buruli ulcer incidence in Victoria, Australia. PLoS One. 2012; 7(12): e51074. https://doi.org/10.1371/journal.pone.0051074 PMID: 23251425 6. Landier J, Constantin de Magny G, Garchitorena A, Guegan JF, Gaudart J, Marsollier L, et al. Seasonal Patterns of Buruli Ulcer Incidence, Central Africa, 2002–2012. Emerg Infect Dis. 2015; 21(8):1414–7. https://doi.org/10.3201/eid2108.141336 PMID: 26196525 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 12 / 14 28. Bratschi MW, Bolz M, Minyem JC, Grize L, Wantong FG, Kerber S, et al. Geographic distribution, age pattern and sites of lesions in a cohort of Buruli ulcer patients from the Mape Basin of Cameroon. PLoS Negl Trop Dis. 2013; 7(6):e2252. https://doi.org/10.1371/journal.pntd.0002252 PMID: 23785529 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Mycobacterium ulcerans low infectious dose and mechanical transmission support insect bites and puncturing injuries in the spread of Buruli ulcer. PLoS Negl Trop Dis. 2017; 11(4):e0005553. https://doi.org/10.1371/journal.pntd. 0005553 PMID: 28410412 19. Barton PS, Aberton JG, Kay BH. Spatial and temporal definition of Ochlerotatus camptorhynchus (Thomson) (Diptera: Culicidae) in the Gippsland Lakes system of eastern Victoria. Australian Journal of Entomology. 2004; 43(1):16–22. 20. Dhileepan K, Peters C, Porter A. Prevalence of Aedes camptorhynchus (Thomson) (Diptera: Culicidae) and other mosquitoes in the eastern coast of Victoria. 1997:183. 21. Department of Health and Human Services. Surveillance of notifiable conditions in Victoria: Victoria State Government 2018. [http://www.health.vic.gov.au/ideas/downloads/daily_reports/rptVS_ SNIDSVictorianSummary_GR.pdf]; First accessed July 2016. 22. Bureau of Meteorology. Climate Data Online: Commonwealth of Australia 2013. [http://www.bom.gov. au/climate/data/]; First accessed July 2016. 23. Australian Bureau of Statistics. Census (2006, 2011, 2016): Commonwealth of Australia 2018. [http:// www.abs.gov.au/census]; First accessed July 2018. 24. Evans JD. Straightforward Statistics for the Behavioral Sciences. Pacific Grove, California, USA Brooks/Cole Publishing Company; 1996. 25. Dale P & Breitfuss M. Ecology and management of mosquitoes. In: (Ed.) INS, editor. Australian Salt- marsh Ecology. Australia: CSIRO Publishing 2009. p. 167–78 26. Pahl L. Survival, Age-Determination and Population Age Structure of the Common Ringtail Possum, Pseudocheirus peregrinus, in a Eucalyptus Woodland and a Leptospermum Thicket in Southern Victo- ria. Australian Journal of Zoology. 1987; 35(6):625–39. 27. Yerramilli A, Tay EL, Stewardson AJ, Kelley PG, Bishop E, Jenkin GA, et al. The location of Australian Buruli ulcer lesions-Implications for unravelling disease transmission. PLoS Negl Trop Dis. 2017; 11(8): e0005800. https://doi.org/10.1371/journal.pntd.0005800 PMID: 28821017 13 / 14 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0006757 September 17, 2018 Rainfall and Buruli ulcer in southeastern Australia Rainfall and Buruli ulcer in southeastern Australia 14 / 14
https://openalex.org/W4295048252	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0273616&type=printable	English	null	A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis	PloS one	2,022	cc-by	8,901	RESEARCH ARTICLE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis RESEARCH ARTICLE Mira OhID1☯, Hyun-Seung Park2☯, Soohyun Um1, Tae-Jin Yang2, Seung Hyun KimID1* 1 College of Pharmacy, Yonsei Institute of Pharmaceutical Sciences, Yonsei University, Incheon, Korea, 2 Department of Agriculture, Forestry and Bioresources, Research Institute of Agriculture and Life Sciences, and Plant Genomics and Breeding Institute, College of Agriculture and Life Sciences, Seoul National University, Seoul, Republic of Korea a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 ☯These authors contributed equally to this work. * kimsh11@yonsei.ac.kr OPEN ACCESS Citation: Oh M, Park H-S, Um S, Yang T-J, Kim SH (2022) A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis. PLoS ONE 17(9): e0273616. https://doi. org/10.1371/journal.pone.0273616 Editor: Clara Sousa, Universidade Cato´lica Portuguesa Escola Superior de Biotecnologia: Universidade Catolica Portuguesa Escola Superior de Biotecnologia, PORTUGAL Editor: Clara Sousa, Universidade Cato´lica Portuguesa Escola Superior de Biotecnologia: Universidade Catolica Portuguesa Escola Superior de Biotecnologia, PORTUGAL Copyright: © 2022 Oh et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Abstract The diversity of secondary metabolites of individual plants results from multiple enzymatic processes in planta and various environmental factors, such as temperature, moisture, and soil conditions. Chemical composition analysis of plants can lead to a new method to under- stand relationship among comparable plants along with biological classification such as genetic and anatomical method. In this study, the chemical diversity of nine different Laura- ceae species was investigated, and the plant samples were chemically analyzed and classi- fied. Multivariate analysis methods, such as PLS-DA, were used to select important metabolites distinguishing the nine Lauraceae species. The selected metabolites were iden- tified through preparative LC-MS or MS/MS fragment pattern analysis. In addition, the chemical dendrogram for the nine Lauraceae species was interpreted through molecular network analysis and compared with the genetic dendrogram. This approach enabled us to compare the complete chemical compositions of multiple plant samples to identify relation- ships among plants. Introduction Individual plants are chemically rich with diverse chemical compositions regardless of their genus or phenotype [1]. The chemical diversity of plant metabolites has been reported to be one of the phenotypic results of evolution [2]. We can gain insights into the evolution and tax- onomy of different plant species from the distribution of specialized metabolites [3]. Although phytochemical studies are recognized for being able to deepen the understanding of relation- ships among targeted plants, comparative chemical analyses among Lauraceae have not been performed [4, 5]. Many taxonomic analyses for Lauraceae species have been conducted on using morphological [6, 7], anatomical [8–11], and genetic [12–14] differences to distinguish species. Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. PLOS ONE PLOS ONE Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Funding: This research was supported by the National Research Foundation of Korea (NRF) grant [2020R1A2C1012632] and Basic Research Infrastructure Support Program [2018R1A6A1A03023718]. The Lauraceae family is composed of approximately 55 genera totalling over 3000 species distributed throughout Southeast Asia and tropical America [15]. Phytochemicals of the Competing interests: The authors have declared that no competing interests exist. PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 1 / 16 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Lauraceae family are used in various industrial fields and worthy of investigation. This family contains valuable secondary metabolites and is therefore has economically important the food, pharmaceutical, and perfumery industries [16, 17]. For example, Cinnamomum burmannii, Litsea pungens, and Laurus nobilis are used in cooking as spices, and the fruits of Persea ameri- cana Mill. are used as a food ingredient [5, 15]. The essential oil, a representative component of the Lauraceae family, is both used as a perfume ingredient and has a wide range of biological activities, such as cytotoxic, antimicrobial, antioxidant, and anti-inflammatory activities [18– 22]. Lauraceae family are used in various industrial fields and worthy of investigation. This family contains valuable secondary metabolites and is therefore has economically important the food, pharmaceutical, and perfumery industries [16, 17]. For example, Cinnamomum burmannii, Litsea pungens, and Laurus nobilis are used in cooking as spices, and the fruits of Persea ameri- cana Mill. are used as a food ingredient [5, 15]. The essential oil, a representative component of the Lauraceae family, is both used as a perfume ingredient and has a wide range of biological activities, such as cytotoxic, antimicrobial, antioxidant, and anti-inflammatory activities [18– 22]. Metabolomics is a useful tool for comparing complete metabolites and differentiating geno- types of individual plants in a rapid and unbiased way [23]. The scope of metabolomics has been considerably expanded as an important method of fingerprinting and profiling for pri- mary and secondary metabolites, as well as for the selection and identification of targeted metabolites [24]. Liquid chromatography–mass spectrometry (LC-MS) is a widely used tech- nique in plant metabolomics. A high sensitivity and strong compatibility with biomolecules make LC-MS suitable for profiling chemical composition from a large quantity of data [25, 26]. Metabolomics with multivariate analysis techniques is useful for nontargeted metabolic profiling [27]. Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. For large-scale datasets, multivariate analyses, such as principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA), can be used to sim- plify complex data by transforming high-dimensional raw mass data to a lower number of var- iables as points in maps [28]. This statistical method enables the chemical compositions of plant metabolites to be conveniently discriminated. Specifically, the VIP score of PLS-DA can be determined by measuring the importance of individual metabolites in samples to improve the classification accuracy of samples [29]. Fourteen species in seven genera of the Lauraceae family are found in Korea that are also native to Japan and Taiwan. In this study, nine Korean endemic Lauraceae species were com- pared and classified by their metabolite composition. The LC-MS spectral data of the nine samples were applied to multivariate analyses. Then, the variable importance in projection (VIP) scores determined using PLS-DA were used to select specific metabolites that are key factors for discriminating samples. The selected discriminatory metabolites were identified by using NMR or comparison with MS/MS fragment patterns obtained from spectral databases, such as GNPS (Global Natural Products Social Molecular Networking). The complete chemi- cal composition of the samples was visualized in a molecular networking system and compared with a chemical dendrogram. We used this approach to classify nine Lauraceae species based on their chemical composition, and the chemical dendrogram was compared with phyloge- nomic results obtained from a plastid genome sequence analysis. UPLC-QTOF-MS analysis The concentrated samples were dissolved in 50% MeOH (JT Baker, Phillipsburg, USA) to a concentration of 1 mg/mL in preparation for LC-MS analysis. The analysis was performed on a UPLC-QTOF-MS analytical system consisting of an Agilent 1290 Infinity LC (Agilent Tech- nologies, Palo Alto, USA) and an Agilent 6550 iFunnel QTOF LC-MS equipped with a dual Agilent Jet Stream (AJS) ESI source. The UPLC column was a YMC-Triart C18 column (2.0 × 150 mm, 1.9 μm; YMC KOREA Co., Seongnam, Korea) that was maintained at 25˚C during the analysis. The mobile phases were 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B), and the following gradients were used: 10–30% B (0–15 min), 30–50% B (15–17 min), 50–80% B (17–20 min), 80–100% B (20–20.1 min), 100% B (20.1–25 min), 100–10% B (25–25.1 min), and 10% B (25.1–28 min). The flow rate was 0.4 mL/min. Each sample was injected in six 10 μL replicates, and a blank (50% MeOH) was injected at the begin- ning of the sample sequence. The MS experiment was performed with a dual AJS ESI source under the following condi- tions: drying gas temperature 300˚C, drying gas flow 8 L/min, nebulizer gas pressure 35 psi, sheath gas temperature 350˚C, sheath gas flow 11 L/min, and capillary voltage +3.5 kV for the positive ionization mode. The QTOF parameters were set to an acquisition rate of five spectra/ sec for MS (mass range of 100−3200 m/z) and three spectra/sec for MS/MS (mass range of 100 −3200 m/z). The collision energy for fragmentation was set to 20, 40, and 60 eV. To obtain the exact mass, calibration was performed with an Agilent tune mix (Agilent Technologies, Palo Alto, USA) from 100 to 1600 Da. Data were acquired in centroid mode at high resolution (4 GHz). PLOS ONE PLOS ONE Table 1. Information for the nine Lauraceae species used in this experimental study. Family Genus Species Location Code Lauraceae Cinnamomum C. camphora (L.) J.Presl Jeju, Korea CC C. yabunikkei H.Ohba CY Lindera L. erythrocarpa Makino Jeju, Korea LDE Litsea L. coreana H.Le´v. Jeju, Korea LC L. japonica (Thunb.) Jussieu LJ Machilus M. japonica Siebold & Zucc. Jeju, Korea MJ M. thunbergii Siebold & Zucc. MT Neolitsea N. aciculata (Blume) Koidz. Jeju, Korea NA N. sericea (Blume) Koidz. NS h //d i /10 1371/j l 0273616 001 Table 1. Information for the nine Lauraceae species used in this experimental study. Sample preparation Leaves of nine Lauraceae samples were collected from Jeju, Korea (between August 2nd and 8th, 2020), and authenticate by Dae Yang Park of Korea Medicinal Plant Farming Corporation. A voucher specimen (YU2020CC/CY/LDE/LC/LJ/MJ/MY/NA/NS) was deposited in the Her- barium at the College of Pharmacy, Yonsei Institute of Pharmaceutical Sciences, Yonsei Uni- versity, Incheon, Korea. The samples were dried at 40˚C for 48 h and homogenized. A powder of each leaf type (50 mg) was transferred to a 5 mL glass vial, to which 1 mL of 50% MeOH was added. The samples were then sonicated for an hour at room temperature. The supernatant was filtered using a 0.2 μm pore syringe filter (Whatman, Clifton, USA). The filtrate was completely evaporated under a nitrogen flow and stored at −20˚C until analysis. The species information and codes of the nine Lauraceae samples used in this study are shown in Table 1. PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 2 / 16 A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Metabolite identification Preparative LC-MS (prep LC-MS) was carried out using an Agilent 1100 series capillary LC system (Agilent Technologies, Palo Alto, USA) coupled with a Waters micromass ZQ mass spectrometer (Waters Co., Milford, USA). The prep column was a YMC-Triart C18 semiprep column (10.0 × 150 mm, 5 μm; YMC KOREA Co., Seongnam, Korea). The mobile phases were 0.1% formic acid in water and 0.1% formic acid in acetonitrile with the same solvent gra- dients used in the QTOF-LC-MS analysis. Broad fractions containing each target peak were eluted from the samples (50% MeOH extract), followed by separating the single peaks more carefully under isocratic solvent conditions. Compound 2 was obtained from the Neolitsea ser- icea (Blume) Koidz. extract, 4 and 9 were obtained from Litsea coreana H. Le´v., 6 was obtained from Machilus japonica Siebold & Zucc., and 13 was obtained from Lindera erythrocarpa Makino (the compound names corresponding to the compound numbers are presented in Table 2). All the NMR spectra of the isolated compounds obtained using prep LC-MS were recorded on a JEOL JNM-ECZ600R spectrometer (JEOL, Tokyo, Japan) operated at 600 and 150 MHz for hydrogen and carbon, respectively. The chemical shifts are reported in parts per million from tetramethylsilane. Data processing was carried out by the MestReNova ver.12.0.1 program. Compounds 1, 3, 7, 8, 11, and 12 were putatively annotated by comparison with masses, chemical formulas, and MS/MS fragment spectra from chemical databases (Table 2). The exact mass and molecular formulas were calculated by MassHunter qualitative analysis software B.06.00 (Agilent Technologies, Palo Alto, USA) with a mass accuracy below 7.5 ppm. The accurate mass and chemical formulas were used to search the compounds in each sample in the Scifinder1 chemical database (http://www.scifinder.org) (Compounds 1, 8, and 12 were identified in the L. erythrocarpa Makino database, 3 was identified in the L. japonica (Thunb.) Jussieu database, and 7 and 11 were identified in the N. sericea (Blume) database). Finally, our MS/MS fragment data for the target compounds were compared to the GNPS spectral database or other literature. Data preprocessing For chemometric analysis, the mass features were detected from the LC-MS raw data using MZmine 2.53 under the following conditions: a retention time range of 0–20 min; a mass detection noise level of 1500 for MS1 and 20 for MS2; a minimum time span of 0.01 min, a minimum height of 5000 and an m/z tolerance of 0.001 m/z for the chromatogram builder; a baseline cut-off algorithm with a minimum peak height of 10,000, a peak duration range of 0.01–0.5 min, and a baseline level of 500 for chromatogram deconvolution; an isotopic peak- grouper algorithm with an m/z tolerance of 0.006 and a retention time tolerance of 0.15 min; a join aligner module with an m/z tolerance of 0.01, an absolute retention time tolerance of 0.3 min, an m/z weight of 70, and a retention time weight of 30. Duplicate peaks with blanks were manually removed from the aligned peak table. 3 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Chemometric data analysis Multivariate analysis was performed using MetaboAnalyst 5.0 (http://www.metaboanalyst.ca) software [30]. The preprocessed peak intensity table was uploaded, and then PCA and PLS-DA were performed using the R package’s prcomp and plsr functions, respectively. A list of important metabolites distinguishing nine Lauraceae species was generated by PLS-DA using the VIP score as a measure. A hierarchical clustering dendrogram was obtained using the mass feature matrix, functioned by hclust of the stats R package. The measure of the Euclidean distance was used in conjunction with Ward’s clustering algorithm to generate the dendrogram. To perform molecular networking, the raw mass spectral data were converted into mzML file formats using MSConvert 3.0 and uploaded to the GNPS server. The tolerances for the pre- cursor and product ions were set to 2.0 and 0.5 Da, respectively. A network was generated using the MS-Cluster algorithm, with parameter settings of minimum cosine pairs of 0.7, min- imum matched fragment ions of 6, and a minimum cluster size of 2. The molecular network- ing job on GNPS can be found at https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task= 1d5650eba32b484aa176f75adc7c02d4. LC-MS profiles LC-MS total ion current (TIC) chromatograms were acquired from the leaves of nine Laura- ceae samples under the optimized UPLC-QTOF-MS conditions (S1 Fig in S1 File). The peaks were detected more sensitively in positive ionization mode than in negative mode. Addition- ally, the chromatograms showed a higher peak abundance when 0.1% formic acid was added to the mobile phase. S1 Fig in S1 File shows very different MS patterns for the nine Lauraceae species even though the species belong to the same genus. The species information and codes of the nine Lauraceae samples used in this study are shown in Table 1. PLOS ONE PLOS ONE Table 2. List of fourteen metabolites selected as important features for sample discrimination. No. Compound name VIP score Class Formula RTa (min) m/z Adduct ion MS/MS fragment 1 neochlorogenic acid 4.8429 cinnamic acids derivatives C16H18O9 3.42 355.0999 M+H 215, 185, 163 2 afzelinb 4.3162 flavonoids C21H20O10 15.25 433.1118 M+H - 3 laurolitsine 4.028 isoquinoline alkaloids C18H19NO4 6.83 314.1372 M+H 297, 265, 237 4 catechinb 3.9379 flavonoids C15H14O6 5.72 291.0927 M+H 147, 139, 123 5 unidentified 1 3.7352 - - 7.18 563.1522 - 378, 313, 223, 123 6 chlorogenic acidb 3.2855 cinnamic acids derivatives C16H18O9 5.40 355.1026 M+H 215, 185, 163 7 coclaurine 3.1933 isoquinoline alkaloids C17H19NO3 6.66 286.1462 M+H 269, 237, 175, 107 8 dihydrokaempferol 3.0558 flavonoids C15H12O6 11.78 289.0710 M+H 215, 153, 149, 107 9 epicatechinb 2.7646 flavonoids C15H14O6 7.35 291.0845 M+H 139, 123 10 unidentified 2 2.7263 - - 2.82 188.0695 - - 11 roemerine 2.7097 isoquinoline alkaloids C18H17NO2 15.95 280.1321 M+H 249, 219 12 phenylalanine 2.695 amino acids C9H11NO2 2.22 166.0854 M+H 120, 103 13 quercitrinb 2.5482 flavonoids C21H20O11 13.08 449.1064 M+H - 14 unidentified 3 2.5469 - - 19.77 281.2104 - - a Retention time b Identified using NMR (the NMR data are shown in S3–S12 Figs in S1 File) Table 2. List of fourteen metabolites selected as important features for sample discrimination. a Retention time manufacturer’s instructions. DNA barcoding analysis was performed using universal primers to amplify each trnH-GUG and rbcL region under conditions suggested in a previous paper [31]. Amplicons were sequenced using ABI 3730xL (NICEM, Seoul, Korea). After combining the sequences from two regions, pairwise alignment was performed by multiple alignment using the fast Fourier transform, with the plastid genomes of each species collected from Gen- Bank for making consensus sequences. A neighbor-joining phylogenetic tree was drawn by MEGA X with 1000 bootstrap replicates. https://doi.org/10.1371/journal.pone.0273616.t002 DNA extraction and barcoding analysis Each leaf sample was ground with liquid nitrogen using a mortar and pestle. DNA was extracted using a GeneAll Plant SV mini kit (GeneAll, Seoul, Korea) following the 4 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Chemometric analyses A mass peak list was produced from the LC-MS spectral data of each sample by using MZmine prefiltering. A total of 447 mass features were uploaded to a MetaboAnalyst server to perform multivariate statistical analyses on the chemical differences among the tested Lauraceae spe- cies. Score plots generated from both the PCA and PLS-DA models simply visualized informa- tion-rich spectral data by reducing the dimensions. PLS-DA is a supervised method that is an alternative to the unsupervised PCA method and interprets data with intergroup variability to better represent group structures [27]. The nine Lauraceae species were clustered more clearly in the PLS-DA (Fig 1B) than in the PCA (Fig 1A). The cross-validated coefficients of the 5 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Fig 1. Multivariate statistical analysis plots of nine Lauraceae species based on LC-MS spectral data. (A) PCA score plot. (B) PLS-DA score plot. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. Fig 1. Multivariate statistical analysis plots of nine Lauraceae species based on LC-MS spectral data. (A) PCA score plot. (B) PLS-DA score plot. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g001 https://doi.org/10.1371/journal.pone.0273616.g001 PLS-DA model were Q2 = 0.985 and R2 = 0.99593, indicating good model quality (S2 Fig in S1 File). The compounds showing a significant difference among the nine Lauraceae samples were selected using the PLS-DA model (VIP score > 2.5), and a PLS-DA loading plot was gen- erated to obtain the distribution of the selected compounds for all the compounds in the sam- ples. The loading scatter plot of the PLS-DA model indicates the relationship between a characteristic variable and a categorical variable, which reflects the contribution of a consid- ered variable on the score plot. PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 Chemometric analyses The 14 important metabolites selected according to the VIP scores are indicated on the corresponding plots. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. htt //d i /10 1371/j l 0273616 002 Fig 2. VIP scores and loading plot obtained using the PLS-DA model for nine Lauraceae metabolites. (A) Important metabolites identified by PLS-DA. The colored boxes on the right indicate the relative concentrations of the corresponding metabolite in each group under study. (B) Loading plot showing PC1-PC2. The 14 important metabolites selected according to the VIP scores are indicated on the corresponding plots. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H Le´v https://doi.org/10.1371/journal.pone.0273616.g002 related to different genera, such as LC, MT, LDE, LJ, and CY, than to the same genera, such as NS. In the PCA score plots, NS was strongly separated from the other samples along the PC1 axis (28.2%), and CC and MJ were also separated along the PC2 axis (12.6%), corresponding with the HCA result (Fig 1A). Considering these results together with the HCA results indi- cates that the close chemical compositions of the nine Lauraceae species were not identical with their genomic closeness. He et al. (2014) [32] similarly observed inconsistencies between the chemical taxonomy and molecular phylogeny of four Coptis species. Wen et al. (2020) [33] also found that the chemical classification of Nardostachys jatamansi collected from different habitats was inconsistent with molecular phylogenetic analysis results. Environmental factors and developmental conditions can affect phytochemical synthesis in planta and the accumula- tion of metabolic constituents [34]. In nature, plant secondary metabolism pathways elicit an array of plant defensive compounds called secondary metabolites. A secondary metabolite is synthesized in organs or tissues in response to various environmental stimuli, such as light, temperature, soil conditions, and microbiota [35]. Accordingly, the corresponding genes to each plant secondary metabolite are regulated at the transcriptional level by multiple transcrip- tion factors and not only by the genetic structure [36]. Chemometric analyses The fourteen selected metabolites (Fig 2A) (representative chemical structures of the selected metabolites are shown in Fig 3) are displayed on the loading plot (Fig 2B), indicating that chemical compounds, such as cinnamic acid derivatives, flavo- noids, and isoquinoline alkaloids, were the major features distinguishing the nine Lauraceae species. Isoquinoline alkaloids, such as coclaurine (7) and roemerine (11), were selected as marker metabolites to separate Neolitsea sericea (Blume) Koidz (NS) from the other species on the PLS-DA score plot because of the relatively high concentrations of 7 and 11 in NS. The bar plot data presented in Fig 4 show that 7 was also found in small quantities in LJ, whereas 11 was only detected in NS and could therefore be a more definitive marker than 7. Flavonoids, such as catechin (4), dihydrokaempferol (8), and epicatechin (9), can similarly be used as dif- ferential markers to distinguish Litsea coreana H. Le´v. (LC). Lindera erythrocarpa Makino (LDE) contained an abundance of neochlorogenic acid (1), phenylalanine (12), and quercitrin (13) and Machilus japonica Siebold & Zucc. (MJ) contained an abundance of chlorogenic acid (6). The fourteen metabolites are ranked by their VIP scores in Table 2, and the relative peak height of the metabolites are represented by using box and whisker plots in Fig 4. Hierarchical clustering analysis (HCA) was performed to investigate the grouping patterns of nine Lauraceae species according to the corresponding phytochemicals (Fig 5). The distinc- tion between samples was confirmed by unsupervised HCA of the mass feature matrix. The dendrogram consisted of one separated branch (NS) and two clusters, of which one comprised CC and MJ and the other comprised NA, LC, MT, LDE, LJ, and CY. Although CC was clus- tered with MJ in the same cluster at a short phytochemical distance, CC and CY showed rela- tively weak correlations despite belonging to the same genus. Similarly, NA was more closely PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 6 / 16 A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis PLOS ONE Fig 2. VIP scores and loading plot obtained using the PLS-DA model for nine Lauraceae metabolites. (A) Important metabolites identified by PLS-DA. The colored boxes on the right indicate the relative concentrations of the corresponding metabolite in each group under study. (B) Loading plot showing PC1-PC2. https://doi.org/10.1371/journal.pone.0273616.g002 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 Chemometric analyses Therefore, the difference between our chemical and gene taxonomic results can be explained by environmental factors affecting the biosynthesis of metabolic compounds in plants. A molecular network was constructed to investigate the complete chemical composition of the nine Lauraceae species. A molecular network is a spectral analysis tool for grouping various compounds by their fragmentation patterns. In GNPS, each MS/MS spectrum is aligned in a dataset, and structurally related molecules are clustered by using the MS-Cluster algorithm [37]. A molecular network of nine Lauraceae species was generated using GNPS and visualized through Cytoscape 3.8.0, an open-source software for visualizing complex networks (Fig 6). A total of 1414 nodes and 73 clusters from the nine samples were detected by the GNPS analysis 7 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis PLOS ONE Fig 3. Representative chemical structures of eleven metabolites contributing to sample discrimination. (1) neochlorogenic acid; (2) afzelin; (3) laurolitsine; (4) catechin; (6) chlorogenic acid; (7) coclaurine; (8) dihydrokaempferol; (9) epicatechin; (11) roemerine; (12) phenylalanine; and (13) quercitrin. https://doi.org/10.1371/journal.pone.0273616.g003 Fig 3. Representative chemical structures of eleven metabolites contributing to sample discrimination. (1) neochlorogenic acid; (2) afzelin; (3) laurolitsine; (4) catechin; (6) chlorogenic acid; (7) coclaurine; (8) dihydrokaempferol; (9) epicatechin; (11) roemerine; (12) phenylalanine; and (13) quercitrin. https://doi.org/10.1371/journal.pone.0273616.g003 https://doi.org/10.1371/journal.pone.0273616.g003 (Fig 6): four isoquinoline alkaloid clusters (91 nodes in total), six flavonoid clusters (61 nodes in total), and two lignan clusters (15 nodes in total). The annotated isoquinoline alkaloid and flavonoid clusters (shown as blue and red squares in Fig 6) were composed of the nodes from eight (LJ, NS, MT, CC, CY, NA, MJ, and LC) and nine (LDE, LJ, NS, MT, CC, CY, NA, MJ, and LC) samples, respectively. However, the nodes in the lignan clusters were only detected from MJ, MT, and CC. For the detected isoquinoline alkaloid clusters, the node composition ratio was the highest in NS (32.4%), followed by CC (23.9%) and NA (13.3%), whereas MJ accounted for the lowest ratio at 0.5%. Under our analysis conditions, no isoquinoline alkaloid nodes were detected in the LDE sample (Fig 7A and 7B). Flavonoid nodes were detected for all nine Lauraceae species. MT had the highest ratio of flavonoid nodes at 19.3%, followed by LC (16.4%) and LJ (13.5%). Chemometric analyses LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. Fig 4. Bar plots of fourteen metabolites selected as important features according to the VIP scores. The bar plots show the intensity of the corresponding ions in the nine analyzed Lauraceae samples. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi org/10 1371/journal pone 0273616 g004 https://doi.org/10.1371/journal.pone.0273616.g004 samples (Fig 6). In addition, CC and MJ, which were grouped into one cluster in the chemical dendrogram (Fig 5), contained most of the lignan compounds detected in the network analysis (Fig 6). samples (Fig 6). In addition, CC and MJ, which were grouped into one cluster in the chemical dendrogram (Fig 5), contained most of the lignan compounds detected in the network analysis (Fig 6). Chemometric analyses CY accounted for the lowest ratio at 2.9% (Fig 7A and 7B). According to the molecular network results, NS, which was the only sample separated from the others in the chemical dendrogram (Fig 5), has the highest isoquinoline alkaloid contents among the (Fig 6): four isoquinoline alkaloid clusters (91 nodes in total), six flavonoid clusters (61 nodes in total), and two lignan clusters (15 nodes in total). The annotated isoquinoline alkaloid and flavonoid clusters (shown as blue and red squares in Fig 6) were composed of the nodes from eight (LJ, NS, MT, CC, CY, NA, MJ, and LC) and nine (LDE, LJ, NS, MT, CC, CY, NA, MJ, and LC) samples, respectively. However, the nodes in the lignan clusters were only detected from MJ, MT, and CC. For the detected isoquinoline alkaloid clusters, the node composition ratio was the highest in NS (32.4%), followed by CC (23.9%) and NA (13.3%), whereas MJ accounted for the lowest ratio at 0.5%. Under our analysis conditions, no isoquinoline alkaloid nodes were detected in the LDE sample (Fig 7A and 7B). Flavonoid nodes were detected for all nine Lauraceae species. MT had the highest ratio of flavonoid nodes at 19.3%, followed by LC (16.4%) and LJ (13.5%). CY accounted for the lowest ratio at 2.9% (Fig 7A and 7B). According to the molecular network results, NS, which was the only sample separated from the others in the chemical dendrogram (Fig 5), has the highest isoquinoline alkaloid contents among the PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 8 / 16 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Fig 4. Bar plots of fourteen metabolites selected as important features according to the VIP scores. The bar plots show the intensity of the corresponding ions in the nine analyzed Lauraceae samples. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi org/10 1371/journal pone 0273616 g004 Fig 4. Bar plots of fourteen metabolites selected as important features according to the VIP scores. The bar plots show the intensity of the corresponding ions in the nine analyzed Lauraceae samples. Metabolite identification In this study, metabolites were identified by using extensive spectroscopic methods or verified through a literature search based on various spectral databases. In Table 2, the five compounds annotated with ‘b’ (Compounds 2, 4, 6, 9, and 13) were isolated by using prep LC-MS, and the chemical structures of the purified compounds were confirmed through NMR. Prep LC-MS is an effective and highly efficient approach to purify small molecules [38]. Unlike UV-based sep- aration with HPLC, this technique enables selective isolation of specific compounds with exact masses, thereby eliminating the need for additional purification analysis to determine the mass of the isolated compounds [39]. The purified compounds were compared with reported 1H PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 9 / 16 E A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Fig 5. Chemical dendrogram of nine Lauraceae species. A dendrogram was generated based on the chemical components of the samples using the Euclidean distance and Ward’s clustering algorithm. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g005 Fig 5. Chemical dendrogram of nine Lauraceae species. A dendrogram was generated based on the chemical components of the samples using the Euclidean distance and Ward’s clustering algorithm. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g005 and 13C NMR and MS data and identified as afzelin (2) [40], catechin (4) [41], chlorogenic acid (6) [42], epicatechin (9) [41], and quercitrin (13) [42] (the NMR data are shown in S3– S12 Figs in S1 File). Compounds 1, 3, 7, 8, 11, and 12 were putatively identified as neochlorogenic acid (1), laur- olitsine (3), coclaurine (7), dihydrokaempferol (8), roemerine (11), and phenylalanine (12), respectively, based on a literature search and comparison of MS/MS patterns [43–46]. Metabolite identification The exact mass, chemical formula, retention time, and MS/MS fragment information of the com- pounds were obtained from the UPLC-QTOF-MS data using MassHunter software. Neo- chlorogenic acid (1) was expected to be structurally similar to chlorogenic acid (6) because the mass and MS/MS spectrum were the same as those of 6, except for the retention time (Table 2). According to the data from Kurita et al. (2016) [47], the retention times of neo- chlorogenic acid (8.11 min) and chlorogenic acid (16.67 min) were clearly different under the respective HPLC analysis conditions, despite having the same mass and chemical formula. According to our analysis, the two compounds were detected at different retention times: Compound 1 (3.42 min) was detected earlier than Compound 6 (5.40 min) (S13 Fig in S1 File). Considering the results together, Compound 1 was putatively identified as neochloro- genic acid (1). Compounds 5, 10, and 14 were purified by using prep LC-MS at the beginning; however, their yields were too low to obtain NMR data. In addition, MS/MS pattern comparison was not possible because the MS/MS fragment information of the compounds could not be obtained through QTOF-LC-MS analysis and there was no match in a chemical database with 5, 10, and 14. DNA barcoding and genetic relationship of Lauraceae Genomic information of each species was obtained from two universal barcoding regions in the plastid genome, trnH-GUG and rbcL. In the phylogenetic tree, most of our sequences were 10 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis PLOS ONE Fig 6. The networking analysis results of the nine Lauraceae species. A network was generated using MS/MS spectra through classical molecular networking on the GNPS server and visualized with nodes and edges through Cytoscape 3.8.0. The nodes consist of pie charts based on the peak intensity proportion for each metabolite. The thickness of the edges was determined by the similarity between two connected nodes with edge widths ranging from 6.0 to 16.0. The blue, red, and green boxes indicate isoquinoline alkaloids, flavonoids, and lignin clusters, respectively. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g006 Fig 6. The networking analysis results of the nine Lauraceae species. A network was generated using MS/MS spectra through classical molecular networking on the GNPS server and visualized with nodes and edges through Cytoscape 3.8.0. The nodes consist of pie charts based on the peak intensity proportion for each metabolite. The thickness of the edges was determined by the similarity between two connected nodes with edge widths ranging from 6.0 to 16.0. The blue, red, and green boxes indicate isoquinoline alkaloids, flavonoids, and lignin clusters, respectively. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g006 Fig 6. The networking analysis results of the nine Lauraceae species. A network was generated using MS/MS spectra through classical molecular networking on the GNPS server and visualized with nodes and edges through Cytoscape 3.8.0. The nodes consist of pie charts based on the peak intensity proportion for each metabolite. The thickness of the edges was determined by the similarity between two connected nodes with edge widths ranging from 6.0 to 16.0. PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 DNA barcoding and genetic relationship of Lauraceae The blue, red, and green boxes indicate isoquinoline alkaloids, flavonoids, and lignin clusters, respectively. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g006 properly clustered with publicly available plastid genome sequences (Fig 8). Intraspecific diver- sity due to habitat isolation could have affected the escape of our CY sequence from the CY and CC groups because the published CY sequence was collected in southern China and our CY sequence was collected in Korea. The genetic phylogenetic relationship showed a different topology from that of the chemical-based phylogenetic relationship. A phylogenetic tree based on plastid genome information is widely used for constructing evolutionary relationships properly clustered with publicly available plastid genome sequences (Fig 8). Intraspecific diver- sity due to habitat isolation could have affected the escape of our CY sequence from the CY and CC groups because the published CY sequence was collected in southern China and our CY sequence was collected in Korea. The genetic phylogenetic relationship showed a different topology from that of the chemical-based phylogenetic relationship. A phylogenetic tree based on plastid genome information is widely used for constructing evolutionary relationships PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 11 / 16 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Fig 7. The chemical composition of Lauraceae samples based on the network analysis. (A) Clusters annotated as isoquinoline alkaloids or flavonoids through classical molecular networking. The thickness of the edges was determined by the similarity between two connected nodes with edge widths ranging from 6.0 to 16.0. (B) The node composition ratio of isoquinoline alkaloids or flavonoids in each sample. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. htt //d i /10 1371/j l 0273616 007 Fig 7. The chemical composition of Lauraceae samples based on the network analysis. (A) Clusters annotated as isoquinoline alkaloids or flavonoids through classical molecular networking. DNA barcoding and genetic relationship of Lauraceae The thickness of the edges was determined by the similarity between two connected nodes with edge widths ranging from 6.0 to 16.0. (B) The node composition ratio of isoquinoline alkaloids or flavonoids in each sample. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. between species [48, 49] because of the good converseness of this genetic material across the plant lineage and a high resolution for interspecific diversity. However, secondary metabolites in plants are produced from complex biological pathways that involve multiple genes and are influenced by both genetic and environmental factors. For example, at least twelve catalyzing steps are needed for the biosynthesis of the ginsenoside backbone, and more enzymes (mainly UDP glycosyltransferase) are required in the branching step to produce approximately 20 kinds of different ginsenosides [50, 51]. Moreover, these genes are usually involved in multiple steps and controlled under precise and systematic regulation [52]. Therefore, metabolite-based relationships can differ even in genetically close species because plant metabolites are not made simply by the translation of genetic information but through communication with the environment via complex and diverse processes. https://doi.org/10.1371/journal.pone.0273616.g007 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 Conclusions In the present study, the primary and secondary metabolites of nine Lauraceae species collected in Korea were compared and classified using chemometric multivariate analysis and molecular networking. Multivariate analyses were performed on LC-MS spectral data of the samples to generate PCA and PLS-DA score plots, between which the nine samples were clustered more clearly on the PLS-DA plot. Fourteen important metabolites were selected based on the VIP scores of the PLS-DA model (VIP score > 2.5). The fourteen selected metabolites were also scat- tered on the PLS-DA loading plot, indicating that these metabolites contributed to discriminat- ing the nine Lauraceae samples. Eleven of the fourteen metabolites were annotated as cinnamic acid derivatives, flavonoids, and isoquinoline alkaloids by using prep-MS or MS/MS fragment pattern analysis. In addition, the chemical diversity of the nine samples was analyzed through molecular network analysis, whereby isoquinoline alkaloids, flavonoids, and lignan clusters were assigned as major clusters. Molecular network analysis facilitated interpretation of the grouping patterns in chemical dendrograms, indicating that the chemical differences between NS and the other samples derive from a high isoquinoline alkaloid content, whereas those between the CC and MJ groups derive from a high lignan content. This approach enabled us to PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 12 / 16 PLOS ONE LOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis Fig 8. Phylogenetic relationship of Lauraceae samples. Combined sequences of two universal barcoding regions, trnH-GUG and rbcL, from each sample were used to draw a neighbor-joining tree with 1000 bootstrap replicates. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. https://doi.org/10.1371/journal.pone.0273616.g008 Fig 8. Phylogenetic relationship of Lauraceae samples. Combined sequences of two universal barcoding regions, trnH-GUG and rbcL, from each sample were used to draw a neighbor-joining tree with 1000 bootstrap replicates. LDE, Lindera erythrocarpa Makino; LJ, Litsea japonica (Thunb.) Jussieu; NS, Neolitsea sericea (Blume) Koidz.; MT, Machilus thunbergii Siebold & Zucc.; CC, Cinnamomum camphora (L.) J. Presl; CY, C. yabunikkei H. Ohba; NA, N. aciculata (Blume) Koidz.; MJ, M. japonica Siebold & Zucc.; and LC, L. coreana H.Le´v. Conclusions https://doi.org/10.1371/journal.pone.0273616.g008 compare the complete chemical composition of several Lauraceae samples simultaneously and interpret the clustering pattern of the chemical dendrogram at the metabolite level. compare the complete chemical composition of several Lauraceae samples simultaneously and interpret the clustering pattern of the chemical dendrogram at the metabolite level. Author Contributions Conceptualization: Mira Oh, Seung Hyun Kim. Conceptualization: Mira Oh, Seung Hyun Kim. Data curation: Mira Oh, Hyun-Seung Park. Formal analysis: Mira Oh, Hyun-Seung Park. Funding acquisition: Seung Hyun Kim. Investigation: Mira Oh, Hyun-Seung Park, Soohyun Um, Tae-Jin Yang, Seung Hyun Kim. Resources: Mira Oh. Investigation: Mira Oh, Hyun-Seung Park, Soohyun Um, Tae-Jin Yang, Seung Hyun Kim. Resources: Mira Oh. Investigation: Mira Oh, Hyun-Seung Park, Soohyun Um, Tae-Jin Yang, Seung Hyun Kim. Resources: Mira Oh. Supervision: Tae-Jin Yang, Seung Hyun Kim. Supervision: Tae-Jin Yang, Seung Hyun Kim. Validation: Hyun-Seung Park, Soohyun Um, Tae-Jin Yang. Writing – original draft: Mira Oh, Hyun-Seung Park. Acknowledgments We would like to thank Dae Yang Park of Korea Medicinal Plant Farming Corporation for providing plant samples. Supporting information PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 13 / 16 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis S1 Data. (CSV) References Conflicting phylogenetic signals in plastomes of the tribe Laureae (Lauraceae). PeerJ. 2020; 8:e10155. https://doi.org/10.7717/peerj.10155 PMID: 33088627 15. Chau DTM, Chung NT, Huong LT, Hung NH, Ogunwande IA, Dai DN, et al. Chemical compositions, mosquito larvicidal and antimicrobial activities of leaf essential oils of eleven species of Lauraceae from Vietnam. Plants. 2020; 9(5):606. https://doi.org/10.3390/plants9050606 PMID: 32397613 16. Damasceno CSB, Higaki NTF, Dias JdFG, Miguel MD, Miguel OG. Chemical composition and biologi- cal activities of essential oils in the family Lauraceae: A systematic review of the literature. Planta Med. 2019; 85(13):1054–72. https://doi.org/10.1055/a-0943-1908 17. Silva JRdA, do Carmo DF, Reis E´ M, Machado G, Leon LL, Silva BOD, et al. Chemical and biological evaluation of essential oils with economic value from Lauraceae species. J Braz Chem Soc. 2009; 20:1071–6. https://doi.org/10.1590/S0103-50532009000600011 18. Salleh WMNHW, Ahmad F. Antioxidant and Anti-inflammatory Activities of Essential Oils of Actino- daphne macrophylla and A. pruinosa (Lauraceae). Nat Prod Commun. 2016b; 11 (6):1934578X1601100639. https://doi.org/10.1177/1934578X1601100639 19. Loizzo MR, Tundis R, Menichini F, Saab AM, Statti GA, Menichini F. Cytotoxic activity of essential oils from Labiatae and Lauraceae families against in vitro human tumor models. Anticancer Res. 2007; 27 (5A):3293–9. PMID: 17970073 20. Unlu M, Ergene E, Unlu GV, Zeytinoglu HS, Vural N. Composition, antimicrobial activity and in vitro cytotoxicity of essential oil from Cinnamomum zeylanicum Blume (Lauraceae). Food Chem Toxicol. 2010; 48(11):3274–80. https://doi.org/10.1016/j.fct.2010.09.001 21. Joshi SC, Verma AR, Mathela CS. Antioxidant and antibacterial activities of the leaf essential oils of Himalayan Lauraceae species. Food Chem Toxicol. 2010; 48(1):37–40. https://doi.org/10.1016/j.fct. 2009.09.011 PMID: 19766161 22. Ramos C, Teixeira B, Batista I, Matos O, Serrano C, Neng N, et al. Antioxidant and antibacterial activity of essential oil and extracts of bay laurel Laurus nobilis Linnaeus (Lauraceae) from Portugal. Nat Prod Res. 2012; 26(6):518–29. https://doi.org/10.1080/14786419.2010.531478 23. Oh M, Park S, Kim H, Choi GJ, Kim SH. Application of UPLC-QTOF-MS Based Untargeted Metabolo- mics in Identification of Metabolites Induced in Pathogen-Infected Rice. Plants. 2021; 10(2):213. https://doi.org/10.3390/plants10020213 PMID: 33499273 24. Okada T, Mochamad Afendi F, Altaf-Ul-Amin M, Takahashi H, Nakamura K, Kanaya S. Metabolomics of medicinal plants: the importance of multivariate analysis of analytical chemistry data. Curr Comput Aided Drug Des. 2010; 6(3):179–96. https://doi.org/10.2174/157340910791760055 PMID: 20550511 25. Arbona V, Iglesias DJ, Talo´n M, Go´mez-Cadenas A. Plant phenotype demarcation using nontargeted LC-MS and GC-MS metabolite profiling. J Agric Food Chem. 2009; 57(16):7338–47. https://doi.org/10. 1021/jf9009137 PMID: 19639992 26. References 1. Fellows L, Scofield A. Chemical diversity in plants. Intellectual Property Rights and Biodiversity Conser- vation—An Interdisciplinary Analysis of the Values of Medicinal Plants. 1995; 19–44. 2. Moghe GD, Last RL. Something old, something new: conserved enzymes and the evolution of novelty in plant specialized metabolism. Plant Physiol. 2015; 169(3):1512–23. https://doi.org/10.1104/pp.15. 00994 PMID: 26276843 3. Wink M. Introduction: biochemistry, physiology and ecological functions of secondary metabolites. Annu Plant Rev. 2010; 40:1–19. https://doi.org/10.1002/9781444320503.ch1/ 4. Salleh WMNH, Ahmad F, Yen KH, Zulkifli RM. Essential oil compositions of Malaysian Lauraceae: A mini review. Pharm Sci. 2016a; 22(1):60–7. https://doi.org/10.15171/PS.2016.11 5. Wang Y, Zhang L-T, Feng Y-X, Zhang D, Guo S-S, Pang X, et al. Comparative evaluation of the chemi- cal composition and bioactivities of essential oils from four spice plants (Lauraceae) against stored- product insects. Ind Crops Prod. 2019; 140:111640. https://doi.org/10.1016/j.indcrop.2019.111640 6. Ceolin GB, Rosito JM, Canto-Dorow TSd. Leaf surface characters applied to Lauraceae taxonomy in a seasonal forest of southern Brazil. Braz. Arch. Biol. Technol. 2009; 52:1453–60. https://doi.org/10. 1590/S1516-89132009000600017 7. Sastri R. Studies in the Lauraceae: V. Comparative Morphology of The Flower. Ann Bot. 1965; 29 (1):39–44. https://doi.org/10.1093/oxfordjournals.aob.a083935 8. Sajo MG, Moraes PL, Assis LC, Rudall PJ. Comparative floral anatomy and development in neotropical Lauraceae. Int J Plant Sci. 2016; 177(7):579–89. https://doi.org/10.1086/687281 9. Gao Z, Sun J, Wu H, Tu D, Guan L, Li B. Comparative study on wood anatomical structure of 14 species belonging to 10 genera of Lauraceae. For Res. 2009; 22(3):413–7. 10. Gomes-Bezerra KM, Gomes SM, SILVEIRA CEDS, Soares-Silva LH. Leaf epidermal descriptors applied to the taxonomy of Lauraceae, including new anatomical characters. Phytotaxa. 2018; 358 (1):49–66. https://doi.org/10.11646/phytotaxa.358.1.3 11. Sun J, Wu J, Wang X, Gu J, Gao Z. Comparative wood anatomy of 56 species of Lauraceae from Yun- nan, China. Braz J Bot. 2015; 38(3):645–56. https://doi.org/10.1007/s40415-015-0172-8 12. Zhao M-L, Song Y, Ni J, Yao X, Tan Y-H, Xu Z-F. Comparative chloroplast genomics and phylogenetics of nine Lindera species (Lauraceae). Sci Rep. 2018; 8(1):1–11. https://doi.org/10.1038/s41598-018- 27090-0 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 14 / 16 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis 13. Song Y, Yao X, Tan Y, Gan Y, Corlett RT. Complete chloroplast genome sequence of the avocado: gene organization, comparative analysis, and phylogenetic relationships with other Lauraceae. Can J For Res. 2016; 46(11):1293–301. https://doi.org/10.1139/cjfr-2016-0199 14. Xiao T-W, Xu Y, Jin L, Liu T-J, Yan H-F, Ge X-J. References Combinatorial chemistry and mass spectrometry in the 21st century drug discovery labora- tory. Chem Rev. 2001; 101(2):255–68. https://doi.org/10.1021/cr990085q 40. Eldahshan OA. Isolation and structure elucidation of phenolic compounds of carob leaves grown in Egypt. Curr Res J Biol Sci. 2011; 3(1):52–5. 41. Cren-Olive´ C, Wieruszeski J-M, Maes E, Rolando C. Catechin and epicatechin deprotonation followed by 13C NMR. Tetrahedron Lett. 2002; 43(25):4545–9. https://doi.org/10.1016/S0040-4039(02)00745-1 42. Jung HA, Park JC, Chung HY, Kim J, Choi JS. Antioxidant flavonoids and chlorogenic acid from the leaves of Eriobotrya japonica. Arch Pharm Res. 1999; 22(2):213–8. https://doi.org/10.1007/ BF02976549 PMID: 10230515 43. Qing Z, Xu Y, Yu L, Liu J, Huang X, Tang Z, et al. Investigation of fragmentation behaviours of isoquino- line alkaloids by mass spectrometry combined with computational chemistry. Sci Rep. 2020; 10(1):1– 13. https://doi.org/10.1038/s41598-019-57406-7 44. Le Gall G, DuPont MS, Mellon FA, Davis AL, Collins GJ, Verhoeyen ME, et al. Characterization and content of flavonoid glycosides in genetically modified tomato (Lycopersicon esculentum) fruits. J Agric Food Chem. 2003; 51(9):2438–46. https://doi.org/10.1021/jf025995e 45. Nikolić D, Go¨decke T, Chen S-N, White J, Lankin DC, Pauli GF, et al. Mass spectrometric dereplication of nitrogen-containing constituents of black cohosh (Cimicifuga racemosa L.). Fitoterapia. 2012; 83 (3):441–60. https://doi.org/10.1016/j.fitote.2011.12.006 46. Hu H, Wang Z, Hua W, You Y, Zou L. Effect of chemical profiling change of processed Magnolia offici- nalis on the pharmacokinetic profiling of honokiol and magnolol in rats. J Chromatogr Sci. 2016; 54 (7):1201–12. https://doi.org/10.1093/chromsci/bmw052 PMID: 27107095 47. Kurita S, Kashiwagi T, Ebisu T, Shimamura T, Ukeda H. Identification of neochlorogenic acid as the pre- dominant antioxidant in Polygonum cuspidatum leaves. Ital J Food Sci. 2016; 28(1):25–31. https://doi. org/10.14674/1120-1770/ijfs.v453 48. Song Y, Yu W-B, Tan Y, Liu B, Yao X, Jin J, et al. Evolutionary comparisons of the chloroplast genome in Lauraceae and insights into loss events in the Magnoliids. GBE. 2017; 9(9):2354–64. https://doi.org/ 10.1093/gbe/evx180 PMID: 28957463 49. Park H-S, Jayakodi M, Lee SH, Jeon J-H, Lee H-O, Park JY, et al. Mitochondrial plastid DNA can cause DNA barcoding paradox in plants. Sci Rep. 2020; 10(1):1–12. https://doi.org/10.1038/s41598-020- 63233-y 50. Lee YS, Park H-S, Lee D-K, Jayakodi M, Kim N-H, Koo HJ, et al. Integrated transcriptomic and metabo- lomic analysis of five Panax ginseng cultivars reveals the dynamics of ginsenoside biosynthesis. Front Plant Sci. 2017; 8:1048. https://doi.org/10.3389/fpls.2017.01048 51. Kim NH, Jayakodi M, Lee SC, Choi BS, Jang W, Lee J, et al. References Lee J, da Silva RR, Jang HS, Kim HW, Kwon YS, Kim J-H, et al. In silico annotation of discriminative markers of three Zanthoxylum species using molecular network derived annotation propagation. Food Chem. 2019; 295:368–76. https://doi.org/10.1016/j.foodchem.2019.05.099 27. Worley B, Halouska S, Powers R. Utilities for quantifying separation in PCA/PLS-DA scores plots. Anal Biochem. 2013; 433(2):102–4. https://doi.org/10.1016/j.ab.2012.10.011 PMID: 23079505 28. Abdi H, Williams LJ. Principal component analysis. Wiley interdisciplinary reviews: computational statis- tics. 2010; 2(4):433–59. https://doi.org/10.1002/wics.101 29. Peerbhay KY, Mutanga O, Ismail R. Commercial tree species discrimination using airborne AISA Eagle hyperspectral imagery and partial least squares discriminant analysis (PLS-DA) in KwaZulu–Natal, South Africa. ISPRS J Photogramm Remote Sens. 2013; 79:19–28. https://doi.org/10.1016/j.isprsjprs. 2013.01.013 30. Chong J, Wishart DS, Xia J. Using MetaboAnalyst 4.0 for comprehensive and integrative metabolomics data analysis. Curr. Protoc. Bioinform. 2019; 68(1):e86. https://doi.org/10.1002/cpbi.86 PMID: 31756036 31. CBOL Plant Working Group. A DNA barcode for land plants. PNAS. 2009; 12794–12797. http://www. jstor.org/stable/40484604 32. He Y, Hou P, Fan G, Arain S, Peng C. Comprehensive analyses of molecular phylogeny and main alka- loids for Coptis (Ranunculaceae) species identification. Biochem Syst Ecol. 2014; 56:88–94. https:// doi.org/10.1016/j.bse.2014.05.002 15 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022 PLOS ONE A comparative phytochemical study of nine Lauraceae species by using chemometric data analysis 33. Wen J, Wu J, Yu H, Fan G, He Y. Correlation analysis between genetic and chemical differences of Nar- dostachys jatamansi from different habitats in Ganzi Tibetan Autonomous Prefecture, Sichuan Prov- ince, China. Biochem Syst Ecol. 2020; 92:104133. https://doi.org/10.1016/j.bse.2020.104133 34. Kamath S, Skeels M, Pai A. Significant differences in alkaloid content of Coptis chinensis (Huanglian), from its related American species. Chin Med. 2009; 4(1):1–4. https://doi.org/10.1186/1749-8546-4-17 35. Yang L, Wen K-S, Ruan X, Zhao Y-X, Wei F, Wang Q. Response of plant secondary metabolites to environmental factors. Molecules. 2018; 23(4):762. https://doi.org/10.3390/molecules23040762 PMID: 29584636 36. Yang CQ, Fang X, Wu XM, Mao YB, Wang LJ, Chen XY. Transcriptional regulation of plant secondary metabolism F. J Integr Plant Biol. 2012; 54(10):703–12. https://doi.org/10.1111/j.1744-7909.2012. 01161.x 37. Aron AT, Gentry EC, McPhail KL, Nothias L-F, Nothias-Esposito M, Bouslimani A, et al. Reproducible molecular networking of untargeted mass spectrometry data using GNPS. Nat Protoc. 2020; 15 (6):1954–91. https://doi.org/10.1038/s41596-020-0317-5 PMID: 32405051 38. Blom KF, Glass B, Sparks R, Combs AP. Preparative LC−MS Purification: Improved Compound-Spe- cific Method Optimization. J Comb Chem. 2004; 6(6):874–83. https://doi.org/10.1021/cc049890v PMID: 15530113 39. Kassel D. References Genome and evolution of the shade-requir- ing medicinal herb Panax ginseng. Plant Biotechnol J. 2018; 16(11):1904–17. https://doi.org/10.1111/ pbi.12926 PMID: 29604169 52. Kang KB, Jayakodi M, Lee YS, Park H-S, Koo HJ, Choi IY, et al. Identification of candidate UDP-glyco- syltransferases involved in protopanaxadiol-type ginsenoside biosynthesis in Panax ginseng. Sci Rep. 2018; 8(1):1–10. https://doi.org/10.1038/s41598-018-30262-7 16 / 16 PLOS ONE \| https://doi.org/10.1371/journal.pone.0273616 September 9, 2022
https://openalex.org/W2187034499	https://etj.uotechnology.edu.iq/article_25916_6bf3ca7c2c53e9e61d8c57803b7e1819.pdf	English	null	Color Texture Classification Using Adaptive Discrete Multiwavelets Transform	Maǧallaẗ al-handasaẗ wa-al-tiknūlūǧiyā	2,012	cc-by	4,846	Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 https://doi.org/10.30684/etj.30.4.8 2412-0758/University of Technology-Iraq, Baghdad, Iraq This is an open access article under the CC BY 4.0 license http://creativecommons.org/licenses/by/4.0 Received on: 3/10 /2011 & Accepted on: 5/1/2012 Received on: 3/10 /2011 & Accepted on: 5/1/2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Dr. Matheel Emad El-Deen Abdulmunim Computer Sciences Department, University of Technology/Baghdad E-mail: matheel_74@yahoo.com ABSTRACT The classification of textures images has attracted the attention of many researchers. The multiscale techniques for gray level texture analysis have been intensively studied. In this paper, we aim on extending texture classification of color images by using the multiwavelets transform, a new notion addition to wavelet. The recognition of textures deals with both feature extraction and classification phases. In the classification phase the evolutionary computation techniques (genetic programming) was used for more speed recognition result evaluation. In our experiment results the proposed method has achieved 99.6% test accuracy on an average. In addition, the experimental results also show that classification rules generated by this approach are robust to some noises on textures. Keywords: classification, multiwavelets, texture, and genetic programming. Keywords: classification, multiwavelets, texture, and genetic programming. ﺗﺼﻨﯿﻒ اﻟﻨﺴﯿﺞ اﻟﻤﻠﻮن ﺑﺄﺳﺘﺨﺪام اﻟﺘﺤﻮﯾﻞ ﻣﺘﻌﺪد اﻟﻤﻮﯾﺠﺎت اﻟﻤﻜﯿﻒ ﺍﻟﺨﻼ ﻪﺼ ﺍﻥ ﺘﺼﻨﻴﻑ ﺼﻭﺭ ﺍﻟﻨﺴﻴﺞ ﻗﺩ ﺠﺫﺏ ﺍﻨﺘﺒﺎﻩ ﺍﻟﻌﺩﻴﺩ ﻤﻥ ﺍﻟﺒﺎﺤﺜﻴﻥ . ﺍﻟﺘﻘﻨﻴﺎﺕ ﻤﺘﻌﺩﺩﺓﺍﻟﻤﺠﺎﻻﺕ ﻟﺘﺤﻠﻴل ﺍﻟﻨﺴﻴﺞ ﺒﺎﻟﻤﺴﺘﻭﻯ ﺍﻟﺭﻤﺎﺩﻱ ﺩﺭﺴﺕ ﺒﺸﻜل ﻤﻜﺜﻑ . ﻓﻲ ﻫﺫﺍ ﺍﻟﺒﺤﺙ، ﻫﺩﻓﻨﺎ ﺍﻟﻰ ﺘﻭﺴﻴﻊ ﺍﻟﻌﻤل ﺒﺘﺼﻨﻴﻑ ﺍﻷﻨﺴﺠﺔ ﻓﻲ ﺍﻟﺼﻭﺭ ﺍﻟﻤﻠﻭﻨﺔ ﺒﺎﺴﺘﺨﺩﺍﻡ ﺘﺤﻭﻴل ﻤﺘﻌﺩﺩ ﺍﻟﻤﻭﻴﺠﺎﺕ ) ﺍﻀﺎﻓﺔ ﺠﺩﻴﺩﺓ ﻟﻠﺘﺤﻭﻴل ﺍﻟﻤﻭﻴﺠﻲ (. ﺍﻥ ﺘﻤﻴﻴﺯ ﺍﻷﻨﺴﺠﺔ ﻴﻌﻨﻰ ﺒﻁﻭﺭﻴﻥ، ﺍﺴﺘﺨﺭﺍﺝ ﺍﻟﺨﺼﺎﺌﺹ ﻭﻤﻥ ﺜﻡ ﺘﺼﻨﻴﻔﻬﺎ . ﻓﻲ ﻁﻭﺭ ﺍﻟﺘﺼﻨﻴﻑ ﺍﺴﺘﺨﺩﻤﺕ ﺘﻘﻨﻴﺎﺕ ﺤﺴﺎﺒﺎﺕ ﺍﻟﺘﻁﻭﺭ ) ﺍﻟﺒﺭﻤﺠﺔ ﺍﻟﺠﻴﻨﻴﺔ ( ﻟﻠﺤﺼﻭل ﻋﻠﻰ ﺴﺭﻋﺔ ﻋﺎﻟﻴﺔ ﺒﺎﺴﺘﺨﺭﺍﺝ ﻨﺘﺎﺌﺞ ﺍﻟﺘﺼﻨﻴﻑ . ﻤﻥ ﻨﺘﺎﺌﺞ ﺍﻟﺒﺤﺙ، ﺍﻟﻁﺭﻴﻘﺔ ﺍﻟﻤﻘﺘﺭﺤﺔ ﺤﻘﻘﺕ ﻨﺴﺒﺔ ﻓﺤﺹ 99.6 %ﻜﻤﻌﺩل . ﺍﻀﺎﻓﺔ ﻟﺫﻟﻙ، ﺍﻟ ﻨﺘﺎﺌﺞ ﻭﻀﺤﺕ ﺍﻥ ﺍﻟﻘﻭﺍﻨﻴﻥ ﺍﻟﻤﻭﻟﺩﺓ ﺒﻬﺫﺍ ﺍﻷﺴﻠﻭﺏ ﺘﻌﻨﻰ ﺒﻭﺠﻭﺩ ﺍﻟﻀﻭﻀﺎﺀ ﻓﻲ ﺍﻷﻨﺴﺠﺔ . BACKGROUND In this section, the theoretical foundations for the expert system used in the present study are given in the following subsections. INTRODUCTION T exture analysis plays an important role in many image processing tasks, ranging from remote sensing to medical image processing, computer vision applications, and natural scenes. A number of texture analysis methods have been proposed in the past decades (e.g., [1]) but most of them use gray scale images, which represent the amount of visible light at the pixel’s position, while ignoring the color information. The performance of such methods can be improved by adding the color information because, besides texture, color is the most important property, especially when dealing with real world images [2]. T In this paper, a color texture image classification scheme is proposed which uses multiwavelets transform and ensembles of evolutionary computation technique (genetic programming). Feature extraction in the multiwavelets domain and 615 Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform classification with ensembles of genetic programming are proposed. Multiwavelets entropies and multiwavelets energies of each color plane at different scales are used for forming the feature vectors. The experimental studies show the efficiency of the proposed system. Moreover, a comparison of the proposed schema with the wavelet energy correlation signatures is conducted. The organization of this paper is as follows: in Section 2, we summarize the theory for multiwavelets transform and genetic programming; several brief definitions are given in this section; in Section 3, the methodology and the implementation of the proposed process is given; in Section 4, an experimental study is introduced and the classification results are shown; in Section 5, we finally conclude the study. DISCRETE MULTIWAVELETS TRANSFORM As in the scalar wavelet case, the theory of multiwavelets is based on the idea of multiresolution analysis (MRA) [3]. The difference is that multiwavelets have several scaling functions. The standard multiresolution has one scaling function Ø(t) [4]. Fig. (1) Shows the multiwavelets framework for image decomposition. The prefilter is first applied to all the rows of the image, before the first level decomposition is applied to each of the resultant rows. The first half of each row of the decomposition results contains coefficients corresponding to the first scaling function and the second half contains coefficients corresponding to the second scaling function. Then the prefilter and decomposition operations are repeated to the columns, such that the first half of each column contains coefficients corresponding to the first scaling function and the second half of each column corresponding to the second scaling function. At the end of the first of 2-D multiwavelets decomposition, we have a 16-subband intermediate image. In practice multiscaling and wavelet functions often have multiplicity r=2. An important example was constructed by Geronimo, Hardin and Massopust [4], which we shall refer to as the GHM system. For the GHM multiscaling functions there are two scaling functions ∅1(t), ∅2(t) and the two wavelets w1(t), w2(t) shown in Fig. 2. There are four remarkable properties of the GHM scaling functions, as follows [4]: • They each have short support (the intervals [0, 1] and [0, 2]). • Both scaling functions are symmetric, and the wavelets form a symmetric/asymmetric pair. y y p • All integers translates of the scaling functions are orthogonal. • All integers translates of the scaling functions are o • The system has second order of approximation. The system has second order of approximation. Before the operation of decomposition is applied to the input data, the preprocessing operation must be done. The aim of preprocessing is to associate the given scalar input signal of length N to a sequence of length-2 vectors {v0,k} in order to start the analysis algorithm. Here N is assumed to be a power of 2, and so is of even length. After the wavelet reconstruction (synthesis) step a postfilter is applied. Clearly, prefiltering, wavelet transform, inverse transform, and postfiltering should 616 PDF created with pdfFactory Pro trial version www.pdffactory.com GENETIC PROGRAMMING Genetic programming (GP) is a methodology for obtaining computer programs to solve a particular problem by a process of simulated evolution. An initial population of programs is constructed. Each program is executed on the problem at hand and its success on the task, its fitness, is measured. A new population of programs is then constructed by selecting the fitter programs as parents and generating children by recombining selected parts of the parents (crossover) and/or making random changes to the parents (mutation). This process continues until the problem is solved or until some preset number of generations has been completed. If the process is working well, the programs will gradually become fitter and fitter through the generations until the problem is solved [6]. GP continues the trend of dealing with the problem of representation in genetic algorithms (GAs) by increasing the complexity of [7]: 1. Structures undergoing adaptation. In GP the individuals in the population are programs, these programs are represented as a tree structure. 1. Structures undergoing adaptation. In GP the individuals in the population are programs, these programs are represented as a tree structure. 2. Introducing the initial structure for the tree representation. The generation of each individual in the initial population is done by randomly generating a rooted, point-labeled tree with ordered branches representing the string. p p g g 3. Fitness measure evaluation. p p g g 3. Fitness measure evaluation. 3. Fitness measure evaluation. 4. The operations that modify the structures. 4. The operations that modify the structures. GP has been used in a wide range of applications and have succeeded in solving many problems such as prediction and classification, image and signal processing, optimization, financial trading, robots and autonomous agents, artificial life, and neural networks [8]. In tree based GP, programs are represented as tree structures. An example tree and the corresponding code are shown in Fig. (3). The internal nodes are functions and the terminals are inputs to the program. The tree is evaluated in a bottom up fashion and the value of the root node is the output of the program. In specifying the configuration of a GP run it is necessary to give the functions, the terminals, a method of evaluating fitness and a number of parameters for the evolutionary parameters. PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform recover the input signal exactly if nothing else has been done. Fig. (2) illustrate the idea of the multiwavelet transform [5]. recover the input signal exactly if nothing else has been done. Fig. (2) illustrate the idea of the multiwavelet transform [5]. GENETIC PROGRAMMING These include the population size, the maximum number of generations to compute if a solution is not found, the elitism rate (the percentage of best individuals in the current generation copied without change to the next generation), the crossover rate (percentage of individuals in the new population that are created by crossover) mutation rate (the percentage of individuals in the new population created by mutation) and the maximum permitted tree depth [6]. TEXTURE ANALYSIS The goal in image analysis is to extract information useful for solving application- based problem. This is done by intelligently reducing the amount of image data with the tools have explored [8]. Different measures were used for analyzing the image and texture features like entropy and energy. 617 ENERGY Energy is commonly used for texture analysis. In this study, we use the averaged l2-norm, which is defined as follows: … (3) … (3) PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform ENTROPY Entropy is a quantity that is widely used in information theory and is based on probability theory [9]. Entropy is a common concept in many fields, mainly in mechanics, image processing, and signal processing. The general form of the entropy is given by: … (1) … (1) where X is a random variable which can be one of the values X1 ,X2 ,..., Xn with probability p1 ,p2 ,..., pn. In this paper, we use the norm entropy. The norm entropy H is defined as: where X is a random variable which can be one of the values X1 ,X2 ,..., Xn with probability p1 ,p2 ,..., pn. In this paper, we use the norm entropy. The norm entropy H is defined as: …… (2) Where , , , is the wavelet coefficient at (i ,j) location at l scale in B ( B €{LH,HL,HH } ) sub band and Xm is the color space (m = 1,2,3) [10]. Where , , , is the wavelet coefficient at (i ,j) location at l scale in B ( B €{LH,HL,HH } ) sub band and Xm is the color space (m = 1,2,3) [10]. METHODOLOGY The proposed color texture classification algorithm is illustrated in Fig.4. The steps involved in color texture classification are as follows: 1: The input to the classifier (GP) based color texture classification system is the color textures of size 512x512. We make sub images of size 128x128 by randomly choosing from the original input texture. Thus, color texture sub images may overlap. For gray scale texture analysis, the color information is discharged by RGB to gray scale trans-formation. Then the red, green, and blue components are decomposed from the color texture images and saved for subsequent processing. 2: This step involves both feature extraction and classification. The feature extraction is composed of two layers. These are the multiwavelets decomposition layer and texture analysis layer. Layer 1. Multiwavelets decomposition layer: For multiwavelets decomposition of each of the red, green, and blue components of color textures, the pyramid multiwavelets structure is used. We obtain one-level multiwavelets decomposition, and save only the three detail images HH, LH, and HL where H and L stand for the high pass and low pass band in each of the horizontal and vertical orientations for the subsequent calculation of entropy and energy quantities. Layer 2. Texture analysis layer: This layer is responsible for calculating the entropy and energy quantities of each LH, HL, and HH of the red, green, and blue 618 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform components of the color texture images. Thus, entropy and the energy quantities are the features that characterize the color texture images. components of the color texture images. Thus, entropy and the energy quantities are the features that characterize the color texture images. Most of the research in texture classification is focused on the feature extraction stage and a large number of different ways of getting useful, highly discriminatory features have been investigated. The conventional approach has three main drawbacks. Firstly, there is no universal set of optimal texture features. Secondly, some of the approaches generate an enormous number of features, perhaps more than there are pixels in the image. This necessitates complex dimensionality reduction in feature space. Thirdly, most of the texture feature extraction algorithms are computationally expensive. They require the generation of Fourier-type transforms or other complex intermediate data structures and then additional computation on these structures. In this paper we show that the use of the GP techniques can overcome some of these drawbacks [6]. There has been prior work on evolving GP classifiers [12] [13]. To obtain the classifier the available data is split into a training and test set, following the machine learning methodology for learning from examples. To get the fitness of an evolved program, it is applied to each example in the training data and the number of classification errors is counted. The fewer errors are the fitter program. Once a program achieves a fitness of zero, the evolutionary run can be terminated. In this case GP terminate the set of texture features wanted, so it overcome the first drawback. We have used a refinement of this basic method which is described in [12][14]. In this refinement, called dynamic range selection, the real line is split into a variable number of ranges, not just two as in the straightforward approach. So the second drawback was overcome because the feature space was restricted with image boundary. The range boundaries are evolved along with the classification program. If the output of the program is less than fittest value then the example is class 2, if the output of the program is between ranges of values it is class 1. PDF created with pdfFactory Pro trial version www.pdffactory.com These classifiers are more accurate and are evolved in fewer generations than the ones from the straight forward approach [12]. So, the third drawback was overcome because the computations were less expensive. Also, they can be easily extended to more than two classes. In classification problem, the classification accuracy, which is ratio of the number of true predication and total number of samples, is usually used as the fitness function. A comprehensive method is proposed to calculate the fitness, which uses the area under the convex hull of the Receiver Operating Characteristic (ROC) curve, as a fitness measure. The ROC shows tradeoff between missing positive cases and raising false alarms. In this paper, for the speed and simplicity, only one single threshold point on the curve was used, which leads to the fitness function: = ( + ) … (4) … (4) … (4) Where, Ntp, Ntn, Np, Nn are the number of true positive prediction, the number of true negative predication, the number positive samples, and the number of negative samples respectively. It is clear that the rules which classify all the samples correctly 619 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 achieve a fitness value of 1 and the ones which misclassify all the samples obtain a fitness value of 0. GP CLASSIFIER In this approach conventional texture feature extraction programs are used to generate a feature vector. The feature vectors used are Variance, Sum Variance, Entropy, Sum Entropy, Difference Variance, and Difference Entropy. The functions used in GP runs are (+, -, ×, ÷, AND, OR, XOR, =, ≥, ≤, <, >, IF). The descriptions of these functions are as follows: (Arithmetic addition, arithmetic subtraction, arithmetic multiplication, arithmetic division, Boolean expression AND, Boolean expression OR, Boolean expression XOR, equal, greater than or equal, less than or equal, less than, greater than, and If arg1 is true return arg2) respectively. The parameters used in the propose GP classifier are: Population size, 300; max generations, 70; elitism rate, 0.2, crossover rate, 0.5; mutation rate, 0.04; maximum tree depth, 25. EXPERIMENT AND DISCUSSIONS In this paper, 12 real world RGB color images of size 512 x 512 from different natural scenes are used in the experimental studies. Fig. 5 shows the color texture images. A data base of 1000 color image regions of 16 texture classes of size 128 x128 was constructed randomly by subdividing each color texture image. 300 of the color texture sub images were used for GP classifier. For comparison purposes, four different randomly chosen multiwavelets filters are used. These filters are GHM, CL, Sa4, D4. The following feature vectors were constructed: g (1) Intensity (gray scale) images were obtained from the RGB form, thereby discarding the color information. The features are extracted according to the step 2 of Section 3. Thus, 6 features are obtained for gray scale textures (1 gray scale image x 3 multiwavelets detail images x 2 features (norm entropy and l2 norm energy)). (2) Each R, G, and B component was multiwavelets transformed using one-level decomposition and the proposed feature extraction scheme was employed. This process constructed a feature vector of size 18 x 3 color channels x 3 multiwavelets detail images x 2 feature values (norm entropy and l2 norm energy)). The experiments are conducted based on the methodology which is illustrated in Fig. (4) and the experimental results are presented at Table (1). The correct classification rates are indicated for all color texture types and the related multiwavelets filters. Table (1) indicates the experimental results for gray scale texture images. Flowers1, Clouds, and Misc texture images are correctly classified for all multiwavelets filter types. The correct classification rate is 100 % for these texture images. This high correct classification rate is obtained for these color texture types because of their homogeneity. On the other hand, the high correct classification rates are not obtained for the rest of the gray scale texture images. Another important property which can be extracted from the results is that the CL multiwavelets type is produced much more accurately than the other multiwavelets filter types. One observation from experimental results is that features and the subsequent classification performance are significantly improved when color information is added to the texture property. The correct classification rates for color texture images are also given in Table (1). The correct classification rates are 99.6\|% for all 620 CONCLUSIONS In this paper, the effect of the color and multiwavelets domain features on the texture classification problem was discussed. The main conclusions and suggestions for the study are: 1. Combining the color and texture information to improve the classification of the texture images. We proposed a system which uses the multiwavelets domain entropy and energy quantities of the red, green, and blue component of the RGB texture images. 2. Experimental studies and subsequent results using a set of real world colored texture images show the usefulness of the multiwavelets entropy and energy for color texture analysis. The results show that color is an important component for improving the classification results for the texture analysis problem. p 3. In this study, several important parameters such as multiwavelet decomposition level and multiwavelets filter type are constant. Selecting the best decomposition level is an important issue. Furthermore, selecting the best multiwavelets filter type will be studied in the future. 4. Another important point is the chosen color space. Several color spaces, such as K-L color space, I1I2I3 color space, and UVW color space, will be added in future works. 5. Also using GP classifier adds some speed to the classification scheme because GP terminate the set of texture features wanted, so it overcomes the drawback in other methods. This result suggests that the GP classifiers could be used in real time situations where speed is more important than accuracy. A drawback of the GP classifiers is that the programs are hard to comprehend. An exciting outcome of this work is the accuracy and speed of the one step classifiers. 6. A set of rules, which classify all the samples correctly are obtained by the GP. Using these rules, most parts of the noisy images generated from the training samples can be classified accurately. All these results give us the confidence that the proposed method is suitable for the texture classification problem. p 7. GP can achieve high accuracy in classifying texture features. The further study in texture classification shows the feasibility of the single-step approach. That is, texture classifiers can be evolved directly based on raw- pixels, without the conventional feature extraction phrase. So by using GP, a new paradigm of texture classification can be established. PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform multiwavelets filter types. Correct classification rate is obtained for Flowers1, Clouds, Misc, Fabric2, Fabric3, and Water texture images. REFERENCES [1]Sklansky J., “Image Segmentation and Feature Extraction”, IEEE Trans. System Man Cybernat., 8, pp. 237–247, 1978. [2]Wouwer G. Van de. Scheunders P. . Livens S. .Van Dyck D., “Wavelet Correlation Signatures for Color Texture Characterization”, Pattern Recognition, 32(3), pp. 443–451, 1999. 621 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 [3]Strela V. . Walden A., "Orthogonal and Biorthogonal Multiwavelets for Signal Denoising and Image Compression", Dep. Of Math., Dartmouth College, 6188 Bradley Hall, Hanover, NH 03755, U.S.A, 1999. [4]Strela V.. Heller P.N., . Strang G., Topiwala P.. Heil C, "The Application of Multiwavelet Filter Banks to Image Processing", Technical report, Massachusets Institute of Technology, USA, Submitted to IEEE transform on Image Processing, 1995. [5]Shutao LI. Yaonan WANG. Boris Lohmann, “Texture Classification Using Discrete Multiwavelet Transform”, Institute of Automatic Control, TUM publications, 2001. [6]Ciesielski Vic. Song Andy. Lam Brian, “Visual Texture Classification and Segmentation by Genetic Programming”, in cagnoni, Luttonand Olague, Genetic and evolutionary Image processing and analysis Hindawi Publishing Corporation, 2007. [7]Koza J. R., “Genetic Programming on the Programming of Computers by Means of Natural Selection”, Massachusetts Institute of Technology, 1993. [8]Umbangh, S. E., "COMPUTER VISION AND IMAGE PRO-CESSING", a practical approach using CVIP tools, Prentice Hall, Inc., Upper Saddle River, NJ 07458 (U.S.A.), 1998. [9]Breiman L., “Bagging Predictors”, Machine Learning, 24, pp. 123–140, 1996. Coifman R. R. Wickerhauser M.V., “Entropy-Based Algorithms for Best Basis Selection”, IEEE Transactions on Information Theory, 38(2), pp. 713–718, 1992. [9]Breiman L., “Bagging Predictors”, Machine Learning, 24, pp. 123–140, 1996. Coifman R. R. Wickerhauser M.V., “Entropy-Based Algorithms for Best Basis Selection”, IEEE Transactions on Information Theory, 38(2), pp. 713–718, 1992. 5 [10]University of Oulu Texture Database, http://www.outex.oulu.fi/outex.php, 2005. [11] Thomas Loveard and Vic Ciesielski. “Representing Classification Problems in Genetic Programming”, In Jong-Hwan Kim, editor, Proceedings of the 2001 Congress on Evolutionary Computation, pages 1070–1077, Seoul, South Korea, May 2001. [12]Kishore J. K. Patnaik L.M. Mani V. Agrawal V. K., “Application of Genetic Programming for Multicategory Pattern Classification”, IEEE Transactions on Evolutionary Computation, 4(3):242–258, Sep. 2000. y p p [13]Andy Song. Thomas Loveard, .Vic Ciesielski., “Towards Genetic Programming for Texture Classification”., In Markus Sumpter, Dan Corbett, and Mike Brooks, editors, Proceedings of the 14th Australian Joint Conference on Artificial Intelligence, LNAI 2256, pages 461–472, Berlin, Dec. 2001. 622 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Table (1) Results of Texture Classification Using GP with Different Multiwavelets Filters (for 100 Image Regions) Images Correct classification % Multiwavelets filter type GHM CL Sa4 D4 Feature Vector 1 (FV1) Grass Flower1 Flower2 Bark1 Clouds Fabric7 Leaves Metal Misc Tile Bark2 Fabric2 Fabric3 Food1 Water Food2 100 93 88 99 100 100 82 91 90 100 84 88 98 97 100 100 100 100 98 91 100 100 99 100 100 100 94 97 100 99 100 100 99 87 85 91 92 100 96 88 100 100 95 87 88 100 93 99 100 100 90 98 96 100 85 88 95 93 100 92 89 87 100 100 Feature Vector 2 (FV2) Grass Flower1 Flower2 Bark1 Clouds Fabric7 Leaves Metal Misc Tile Bark2 Fabric2 Fabric3 Food1 Water Food2 100 90 81 88 85 100 100 96 89 100 92 91 98 100 97 93 100 100 100 97 98 100 100 100 100 99 100 95 98 95 99 100 99 89 96 94 100 100 94 89 86 84 100 99 95 92 90 100 100 99 96 89 95 94 90 90 100 98 91 88 85 100 94 93 Feature Vector 3 (FV3) Grass Flower1 Flower2 Bark1 Clouds Fabric7 Leaves Metal Misc Tile 100 100 90 98 95 99 100 87 84 98 100 99 96 100 100 100 100 100 98 97 99 98 97 100 86 82 97 94 94 90 90 100 100 88 86 85 98 95 93 100 Table (1) Results of Texture Classification Using GP with Different Multiwavelets Filters (for 100 Image Regions) 623 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Discrete Multiwavelets Transform Bark2 Fabric2 Fabric3 Food1 Water Food2 98 99 100 99 94 89 100 100 99 99 94 99 90 98 87 96 100 85 98 95 87 98 93 100 Overall results Total FV1 FV2 FV3 94.3 93.7 95.6 98.6 98.8 98.8 93.7 94.1 93.3 94.5 93.8 94.1 Figure (1): The two scaling and wavelet GHM functions. Figure (2): The multiwavelets transform decompositions. Numeric output Bark2 Fabric2 Fabric3 Food1 Water Food2 98 99 100 99 94 89 100 100 99 99 94 99 90 98 87 96 100 85 98 95 87 98 93 100 Overall results Total FV1 FV2 FV3 94.3 93.7 95.6 98.6 98.8 98.8 93.7 94.1 93.3 94.5 93.8 94.1 Figure (1): The two scaling and wavelet GHM functions. Figure (1): The two scaling and wavelet GHM functions. g ( ) g Figure (2): The multiwavelets transform decompositions. Numeric output Figure (2): The multiwavelets transform decompositions. Numeric output 624 PDF created with pdfFactory Pro trial version www.pdffactory.com Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Discrete Multiwavelets Transform Tree Form (* A (IF (<=0.2 B) [then] -0.5 (Prefix notation) [else] (+B C) Figure (3): A program in Tree Based Genetic Programming * A IF ≤ -0.5 + 0.2 B B C * IF + ≤ -0.5 B B Figure (3): A program in Tree Based Genetic Programming Figure (3): A program in Tree Based Genetic Programming 625 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Figure (4): The proposed color texture classification scheme Apply classifier Feature extraction Multiwavelet energy Feature extraction Multiwavelet energy Feature extraction Multiwavelets energy GP Decision Figure (4): The proposed color texture classification scheme 626 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com Color Texture Classification Using Adaptive Discrete Multiwavelets Transform Eng.& Tech. Journal ,Vol. 30 , No.4 , 2012 Discrete Multiwavelets Transform Figure (5) Color texture images from left to right and top to bottom: Grass, flowers1, Flowers2, Bark1, Clouds, Fabric7, Leaves, Metal, Misc, Tile, Bark2, Fabric2, Fabric3, Food1, Water, and Food2. Figure (5) Color texture images from left to right and top to bottom: Grass, flowers1, Flowers2, Bark1, Clouds, Fabric7, Leaves, Metal, Misc, Tile, Bark2, Fabric2, Fabric3, Food1, Water, and Food2. 627 PDF created with pdfFactory Pro trial version www.pdffactory.com PDF created with pdfFactory Pro trial version www.pdffactory.com
https://openalex.org/W4388710044	https://journal-nam.com.ua/index.php/journal/article/download/652/609	Ukrainian	null	КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО	Ûridična nauka	2,020	cc-by	3,635	УДК: 343.8 doi: 10.32844/2222-5374-2020-104-2-2.04 Дмитращук О. С., здобувач Науково-дослідного інституту публічного права КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 25 КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО З одного боку, чинники злочинності пов’язані з об’єктивними соціальними явищами, які негативно впливають на формування особи і проявляються в егоїстичних, антисуспільних поглядах і звичках, що лежать в основі кримінальної поведінки, а з іншого – з обставинами, що сприяють реалізації таких поглядів і звичок у конкретному злочині. Зазначені явища і обставини не залежать від людини, є зовнішніми стосовно неї, але свого криміногенного значення вони набувають, формуючи особу злочинця і таким чином трансформуючись у внутрішні детермінанти злочину [1, с. 71]. Як зазначає І.М. Даньшин, особа злочинця – це складне інтегруюче поняття, що включає в себе і біологічні, і психологічні, і соціальні сторони людини [2, с. 34]. Однак, переконані, що, в першу чергу, особу злочинця слід розглядати як єдність соціологічного і юридичного аспектів цього явища правової дійсності. Не виключенням є й дослідження характерних рис особи, що вчиняє домашнє насильство, з'ясування яких в умовах сучасності залишається актуальним питанням, що потребує здійснення подальших наукових розвідок. Актуальність теми. Однією зі складових предмета кримінології є вчення про особу злочинця. Без нього важко дослідити інші кримінологічні проблеми, зокрема, детермінанти злочинів та засоби боротьби з ними. З одного боку, чинники злочинності пов’язані з об’єктивними соціальними явищами, які негативно впливають на формування особи і проявляються в егоїстичних, антисуспільних поглядах і звичках, що лежать в основі кримінальної поведінки, а з іншого – з обставинами, що сприяють реалізації таких поглядів і звичок у конкретному злочині. Зазначені явища і обставини не залежать від людини, є зовнішніми стосовно неї, але свого криміногенного значення вони набувають, формуючи особу злочинця і таким чином трансформуючись у внутрішні детермінанти злочину [1, с. 71]. р ф р у у у р д р у [ , ] Як зазначає І.М. Даньшин, особа злочинця – це складне інтегруюче поняття, що включає в себе і біологічні, і психологічні, і соціальні сторони людини [2, с. 34]. Однак, переконані, що, в першу чергу, особу злочинця слід розглядати як єдність соціологічного і юридичного аспектів цього явища правової дійсності. Не виключенням є й дослідження характерних рис особи, що вчиняє домашнє насильство, з'ясування яких в умовах сучасності залишається актуальним питанням, що потребує здійснення подальших наукових розвідок. Стан дослідження. Запобіганню домашньому насильству були присвячені наукові праці таких вітчизняних та зарубіжних вчених: О. Бандурки, В. Бондаровської, Т. Бугаєць, О. Ковальової, Л. Козуб, О. Кочемировської, Г. Лактіонової, К. Левченко, Т. Малиновської, О. Суслової, E. Salisbury, G. Donavin та інших. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО Актуальність статті полягає в тому, що однією зі складових предмета кримінології є вчення про особу злочинця. Без нього важко дослідити інші кримінологічні проблеми, зокрема, детермінанти злочинів та засоби боротьби з ними. З одного боку, чинники злочинності пов’язані з об’єктивними соціальними явищами, які негативно впливають на формування особи і проявляються в егоїстичних, антисуспільних поглядах і звичках, що лежать в основі кримінальної поведінки, а з іншого – з обставинами, що сприяють реалізації таких поглядів і звичок у конкретному злочині. Зазначені явища і обставини не залежать від людини, є зовнішніми стосовно неї, але свого криміногенного значення вони набувають, формуючи особу злочинця і таким чином трансформуючись у внутрішні детермінанти злочину. Стаття присвячена кримінологічній характеристиці особи, яка вчиняє злочинні дії в сімейно-побутовому осередку. З’ясовано існуючі доктринальні підходи до розуміння категорій «особа-кривдник» та «особа злочинець», на основі чого обґрунтовано, що кривдником є особа, що вчиняє систематичне фізичне, психологічне або економічне насильство відносно подружжя чи колишнього подружжя або іншої особи, з якою особа перебуває (перебував) у сімейних або близьких відносинах. Визначено соціально-демографічні властивості особи- кривдника, а саме: а) стать (переважна більшість осіб чоловічої статі); б) рівень освіти; в) вік; г) стан алкогольного сп’яніння; д) трудова зайнятість; є) наявність судимості. Встановлено кримінологічний портрет особи, яка вчиняє домашнє насильство (громадянин України чоловічої статі, 30-50 років, що має базову середню освіту, працездатний, але не працюючий, не судимий, схильний до вживання алкоголю). Резюмується, що особу, яка вчиняє домашнє насильство, слід віднести до насильницького типу особи злочинця, а за глибиною антисуспільної спрямованості до ситуаційно- нестійких злочинців. Наголошено, що структура особи злочинця передбачає наявність трьох груп ознак (властивостей, якостей): 1) соціально- демографічні (стать, вік, соціальне становище, рівень освіти, сімейний стан, рід занять тощо); 2) кримінально-правові (зміст і мотивація, одноосібний чи груповий характер злочинної діяльності, тривалість та інтенсивність злочинної діяльності, наявність чи відсутність судимості тощо); 3) соціально-психологічні властивості та якості осіб, які вчинили злочин (світогляд, інтереси та життєві установки). Ключові слова: домашнє насильство, особа злочинця, особа-кривдник, соціально-демографічні властивості, кримінологічний портрет особи, насильницький тип особи злочинця. ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 © Дмитращук О. С., 2020 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 26 Актуальність теми. Однією зі складових предмета кримінології є вчення про особу злочинця. Без нього важко дослідити інші кримінологічні проблеми, зокрема, детермінанти злочинів та засоби боротьби з ними. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО Однак науковцями недостатньо уваги було приділено питанням кримінологічної характеристики особи, яка вчиняє домашнє насильство, що зумовлює необхідність проведення додаткових наукових пошуків у цій царині. Метою статті є сформувати кримінологічний портрет особи, яка вчиняє домашнє насильство, шляхом визначення його характерних рис. Метою статті є сформувати кримінологічний портрет особи, яка вчиняє домашнє насильство, шляхом визначення його характерних рис. Виклад основного матеріалу. Принагідно зазначимо, що в кримінологічній науці розуміння категорії «особа-кривдник» характеризується полісемічністю. Так, наприклад, з погляду Л.В. Самарай, особа кривдника – це наслідок вчинення протиправної діяльності, а не тип людини, який їй передує. Логічно, що в цьому випадку проблематичним є визначення норм, відповідно до яких позначають злочинність чи не злочинність діяння. Адже норма фіксує безумовний мотив. Є люди, які народжуються з певними генетичними й психологічними вадами, що за несприятливих умов середовища може стати фактором, який провокує агресивну поведінку, наслідком якої може стати (а може й не стати) вчинення насильницьких дій. Але це відбувається лише в тому разі, якщо є сукупність соціальних детермінант, які відіграють роль збудників такої поведінки [3, с. 191]. На переконання І.А. Ботнаренко, особу злочинця, який вдається до насильства в сім’ї, слід розуміти як особу, що вчиняє умисні насильницькі дії фізичного, сексуального, психологічного чи економічного характеру ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 27 відносно члена (членів) своєї сім’ї. Відповідно, кримінологічна характеристика такої особи – це кримінологічний опис комплексу соціально значущих ознак особи, наділеної такими якостями з огляду на її взаємодію із зовнішніми умовами й обставинами, що зрештою призводить до вчинення нею актів насильства відносно члена (членів) сім’ї [4, с. 84]. Проте з визначенням І.А. Ботнаренко слід не погодитись, адже, воно не відповідає нормативним положенням ст. 126-1 КК України. Тому вважаємо, що кривдника більш коректно розуміти як особу, що вчиняє систематичне фізичне, психологічне або економічне насильство відносно подружжя чи колишнього подружжя або іншої особи, з якою особа перебуває (перебував) у сімейних або близьких відносинах. Задля повного та об’єктивного дослідження особи-кривдника, що вчиняє домашнє насильство, обов’язковим є з’ясування його структурних елементів. З цього приводу загальноприйнятим є підхід, що структура особи злочинця передбачає наявність трьох груп ознак (властивостей, якостей): 1) соціально-демографічні (стать, вік, соціальне становище, рівень освіти, сімейний стан, рід занять тощо); 2) кримінально-правові (зміст і мотивація, одноосібний чи груповий характер злочинної діяльності, тривалість та інтенсивність злочинної діяльності, наявність чи відсутність судимості тощо); 3) соціально-психологічні властивості та якості осіб, які вчинили злочин (світогляд, інтереси та життєві установки) [5, с. 97]. Проте А.Б. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО Благою, з точки зору методології структурного аналізу, запропоновано дещо інший підхід, що передбачає такі елементи: соціально- демографічні, кримінально-правові, соціально-рольові та морально- психологічні [6, с. 92]. Соціально-демографічні властивості особи злочинця пов’язані з: умовами формування особи та її життєдіяльності, потребами і мотивацією, соціальними ролями і статусом особи, тому вони є істотним елементом поняття особи злочинця і мають важливе значення для розробки і здійснення заходів профілактики [7, с. 34]. Першим показником соціально-демографічних властивостей особи- кривдника домашнього насильства уважаємо стать. Відповідно до даних Державної судової адміністрації України за 2019 рік домашнє насильство вчинили 159 осіб, з них 151 особа (95 %) чоловічої статі та 8 (5 %) осіб жіночої статі. За січень-листопад 2020 року домашнє насильство вчинили 149 осіб, з яких 134 (90 %) особи чоловічої статі та 15 (10 %) осіб жіночої статі. Отже, як бачимо, переважна більшість кривдників особи чоловічої статі. Окремої уваги заслуговує і жіноча злочинність в побутовій сфері. Адже, незважаючи на те, що даний вид злочинності в жодному суспільстві немає домінуючого характеру, його рівень і структура є показниками морального здоров’я суспільства, індикаторами його духовності, ставлення до основних людських цінностей тощо [8, с. 189]. Так, Г.М. Філоненко, аналізуючи злочини, що вчиняються жінками у сімейно-побутовій сфері, дійшла до висновку, що найбільшу групу складають злочини проти життя та здоров’я особи, спрямовані проти чоловіків або співмешканців, що вчиняються активними діями з умисною формою вини [9, ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 28 с. 156]. З результатами дослідженням Т.М. Продана, питома вага зазначених злочинів становить (60,8 %) [10, с. 311], а загальний відсоток жіночої злочинності у суспільстві коливається в межах 10-12 % від усієї злочинності [11, с. 193]. Жертвами злочинів, що вчиняються жінками можуть стати: чоловіки, співмешканці, коханці, колишні чоловіки; діти; батьки, родичі; друзі і сусіди [12, с. 73]. у у Однією з соціально-демографічних властивостей особи-кривдника уважаємо її рівень освіти. Подібних висновків доходять багато кримінологів, якими зауважується, що найбільше випадків домашнього насильства вчиняють особи, які мають повну та неповно загальну середню освіту (56 %), середню спеціальну (36 %), і початкову вищу (3 %). Невелика кількість нападників мають повну вищу освіту (5 %) або взагалі не мають жодної [13, с. 390]. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО Із звіту Державної судової адміністрації за 2019 рік встановлено, що 81 особа (51 %) вчинила домашнє насильство в стані алкогольного сп’яніння. б [1 ] Із звіту Державної судової адміністрації за 2019 рік встановлено, що 81 особа (51 %) вчинила домашнє насильство в стані алкогольного сп’яніння. Що становить більше половини загальної кількості [17]. Відповідно звіту про осіб, які вчинили кримінальні правопорушення за січень-листопад 2020 року встановлено, що 59 осіб (39,6 %) вчиняли домашнє насильство в стані алкогольного сп’яніння, 6 осіб (4 %) в стані алкогольного сп’яніння повторно [18]. Вказані дані дають підстави стверджувати, що алкогольне сп’яніння сприяє вчиненню домашнього насильства, як і будь-якому іншому злочину. Також встановлено, що за січень-листопад 2020 року, із загальної кількості осіб-кривдників 39 осіб раніше вчиняли кримінальні правопорушення, що становить 26,2 %, майже одна четверта частина. Судимість не знята та не погашена була у 32 осіб (21,5 %). Дві особи (1,3 %) раніше вчиняли домашнє насильство, судимість за яке не знята і не погашена. Наголосимо, що всі випадки домашнього насильства за звітний період були вчинені одноосібно [18]. Розглядаючи соціально-психологічні властивості та якості особи- кривдника, який вчиняє домашнє насильство необхідно звернути увагу на соціальну площину. Соціальна площина агресії відзначається тим, що в суспільстві вона не виникає без певних мотивів і настанов, і тому це явище більш складне, аніж агресія в тваринному середовищі. В суспільстві агресія постає як насильство і пояснюється тими соціальними відносинами, які склались в суспільстві, а високий або низький рівень агресивності особистості виступає результатом її соціалізації [16, с. 189]. Аналізуючи звіт Державної судової адміністрації за 2019 рік можемо побачити, що переважна більшість, а це 143 особи (89,9 %) не працювали і не навчались. Робітники становили – 8 осіб (5 %), пенсіонери – 4 особи (2,5 %), безробітні – 2 (1,2 %), інші види занять також – 2 особи (1,2 %). У 2020 році простежується така ж сама тенденції безробіття. Встановлено, що 133 особи (89,3 %), які вчинили домашнє насильство були працездатні, але не працюють і не навчаються [18]. Отже, надмірна кількість вільного часу, відсутність зайнятості, небажання до суспільно-корисної праці та самовдосконалення, тільки сприяє вчиненню домашнього насильства. Погоджуємось із твердження А.Б. Благи, що кардинальні зміни соціально-економічної ситуації в країні торкнулися також сфери зайнятості та організації освіти (в нашому випадку, маємо на увазі запровадження карантину з метою запобігання поширенню на території України гострої респіраторної хвороби COVID-19, спричиненої коронавірусом SARS-CoV-2), що, у свою чергу, вплинуло і на розподіл злочинців досліджуваної категорії за родом занять. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО Аналізуючи звіт Державної судової адміністрації за 2019 рік можемо побачити, що у 5 осіб (3,1 %) була повна вища освіта, у 1 особи (0,6 %) базова вища, у 37 осіб (23,3 %) професійна (професійно технічна), у 86 осіб (54 %) повна загальна середня, у 29 осіб (18,2 %) базова загальна середня, без освіти лише 1 особа (0,6 %) [17]. Відповідно до звіту про осіб, які вчинили кримінальні правопорушення за січень-листопад 2020 року встановлено, що у 4 (2,7 %) осіб була вища або фахова передвища освіта, у 32 (21,5 %) осіб професійна (професійно технічна), у 112 осіб (75,1 %) базова середня та профільна середня та лише у 1 особи (0,7 %) була початкова освіта. Вказана статистика свідчить про те, що рівень освіти безпосередньо впливає на кількість випадків домашнього насильства, адже лише 4 (2,7 %) особи із 149 осіб мали вищу або фахову передвищу освіту. Наступним показником соціально-демографічних властивостей особи- кривдника, що вчиняє домашнє насильства є вік. Вікова характеристика злочинців дає змогу дійти висновку про ступінь інтенсивності прояву криміногенної активності та особливості злочинної поведінки представників різних вікових груп [14, с. 34]. Як зазначає А.Б. Блага пік кримінальної активності осіб, які вчинили злочини, пов’язані з насильством в сім’ї, припадає на дві вікові категорії – 21‒29 роки, що становить 29 % від загальної маси, та 30-39 років, що складає 26,9 % [15, с. 96]. Згідно звіту Державної судової адміністрації за 2019 рік віковий ценз особи-кривдника був наступний: 18‒25 років – 3 особи (1,9 %), 25‒30 років – 15 осіб (9,4 %), 30-50 років – 108 осіб (67,9 %), 50‒65 років – 32 особи (20,1 %), 65 років і старше – 1 особа (0,6 %) [17]. Відповідно звіту про осіб, які вчинили кримінальні правопорушення за січень-листопад 2020 року встановлено, що серед загальної кількості осіб, які вчинили домашнє насильство 18‒28 років становило 11 осіб (7,4 %), 29‒39 років – 52 особи (34,9 %), 40‒54 років – 71 особа (47, 6 %), 55‒59 – 8 осіб (5,4 %), 60 років і більше – 7 осіб (4,7 %) [18]. ( ) [ ] Ураховуючи, що у зазначених звітах віковий ценз різниться, вважаємо за необхідне вивести середнє арифметичне. Таким чином, зазначена вище ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 29 статистика дає змогу стверджувати, що більша кількість осіб, що вчиняють домашнє насильство припадає на 30‒50 років. статистика дає змогу стверджувати, що більша кількість осіб, що вчиняють домашнє насильство припадає на 30‒50 років. КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 30 Висновки. Отже, враховуючи вищевикладене, маємо змогу сформувати кримінологічний портрет особи, яка вчиняє домашнє насильство: це громадянин України чоловічої статі, 30‒50 років, що має базову середню освіту, працездатний, але не працюючий, не судимий, схильний до вживання алкоголю. Кримінологічна характеристика особи-кривдника, який вчиняє домашнє насильство, вимагає дослідження й питання його типологізації. Поділяємо позицію більшості кримінологів, які вважають, що підставами (критеріями) для класифікації особи-злочинця є: характер антисуспільної спрямованості особи злочинців; ступінь глибини цієї антисуспільної спрямованості. Перший тип включає: насильницький, корисливий, злісний, а також необережний тип злочинців. Другий тип включає: випадкових, ситуаційних, нестійких, злісних та особливо злісних злочинців [2, с. 40‒41]. у [ ] Зважаючи на вказане, особу, яка вчиняє домашнє насильство, необхідно віднести до насильницького типу особи злочинця, для якого характерне зневажливе ставлення до іншої особи. За глибиною антисуспільної спрямованості особу-кривдника, який вчиняє домашнє насильство, необхідно віднести до ситуаційно-нестійких злочинців. СПИСОК ВИКОРИСТАНИХ ДЖЕРЕЛ Д 1. Кримінологія : підручник для студентів вищих навч. закладів / за заг. ред. О.М. Джужи. Київ : Юрінком Інтер, 2002. 416 с. 1. Кримінологія : підручник для студентів вищих навч. закладів / за заг. ред. О.М. Джужи. Київ : Юрінком Інтер, 2002. 416 с. 2. Кримінологія: Загальна та Особлива частина : підручник / за заг. ред. В.В. Голіни. Харків : Право, 2009. 288 с. 2. Кримінологія: Загальна та Особлива частина : підручник / за заг. ред. В.В. Голіни. Харків : Право, 2009. 288 с. 3. Самарай Л.В. Кримінологічна характеристика особи кривдника, який вчинив сімейне насильство під впливом агресії. Науковий вісник Національної академії внутрішніх справ. 2013. № 1. С. 190‒197. 3. Самарай Л.В. Кримінологічна характеристика особи кривдника, який вчинив сімейне насильство під впливом агресії. Науковий вісник Національної академії внутрішніх справ. 2013. № 1. С. 190‒197. 4. Ботнаренко І.А. Кримінологічна характеристика осіб, які вчиняють насильство в сім’ї. Науковий вісник Національної академії внутрішніх справ. 2017. № 2 (103). С. 83‒96. 4. Ботнаренко І.А. Кримінологічна характеристика осіб, які вчиняють насильство в сім’ї. Науковий вісник Національної академії внутрішніх справ. 2017. № 2 (103). С. 83‒96. 5. Біленко С.П. Кримінологічна характеристика кривдника. Вісник Луганського державного університету внутрішніх справ ім. Е.О. Дідоренко. 2019. Вип. 2 (86). С. 95‒105. 5. Біленко С.П. Кримінологічна характеристика кривдника. Вісник Луганського державного університету внутрішніх справ ім. Е.О. Дідоренко. 2019. Вип. 2 (86). С. 95‒105. 6. Блага А.Б. Насильство в сім’ї (кримінологічний аналіз і запобігання) : монографія. Харків : ФО-П Макаренко, 2014. 360 с. 6. Блага А.Б. Насильство в сім’ї (кримінологічний аналіз і запобігання) : монографія. Харків : ФО-П Макаренко, 2014. 360 с. 7. Кримінологія : навч. посіб. / за заг. ред. О.М. Джужи. Київ : Прецедент, 2006. 198 с. 7. Кримінологія : навч. посіб. / за заг. ред. О.М. Джужи. Київ : Прецедент, 2006. 198 с. 8. Гаргат-Українчук О.М. Кримінологічна характеристика жіночої злочинності. Науковий вісник Ужгородського національного університету. Серія Право. 2014. Вип. 25. С. 189‒191. 8. Гаргат-Українчук О.М. Кримінологічна характеристика жіночої злочинності. Науковий вісник Ужгородського національного університету. Серія Право. 2014. Вип. 25. С. 189‒191. 9. Філоненко Г.М. Класифікація злочинів, що вчиняються жінками у сімейно-побутовій сфері. Наше право. 2014. № 8. С. 150‒156. 9. Філоненко Г.М. Класифікація злочинів, що вчиняються жінками у сімейно-побутовій сфері. Наше право. 2014. № 8. С. 150‒156. 10. Продан Т.В. Стан жіночої насильницької злочинності в України. Право і суспільство. 2014. № 5‒2. С. 303‒311. 10. Продан Т.В. Стан жіночої насильницької злочинності в України. Право і суспільство. 2014. КРИМІНОЛОГІЧНА ХАРАКТЕРИСТИКА ОСОБИ, ЯКА ВЧИНЯЄ ДОМАШНЄ НАСИЛЬСТВО На момент вчинення злочину переважна більшість сімейних кривдників не працювали і не навчались, перебивались тимчасовими заробітками. Домінування цих категорій може свідчити про значимість трудової дезадаптації як фактору злочинної поведінки [15, с. 102]. ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО СПИСОК ВИКОРИСТАНИХ ДЖЕРЕЛ № 5‒2. С. 303‒311. ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 31 11. Філіпенко В.О. Кримінологічна характеристика злочинності жінок у побутовій сфері. Юридичний науковий електронний журнал. 2018. № 1. С. 193‒196. 12. Філоненко Г.М. Жертви злочинів, що вчиняються жінками у сімейно-побутовій сфері. Наше право. 2014. № 9. С. 72‒77. 13. Лук’янова Ю.О. Кримінологічна характеристика осіб, що вчиняють домашнє насильство, та шляхи запобігання вчиненню цього злочину. 2018. № 4. С. 388‒391. 14. Кримінологія : навчальний посібник / за заг. ред. О.М. Джужи. Київ : Атіка, 2010. 312 с. 15. Блага А.Б. Насильство в сім’ї (кримінологічний аналіз і запобігання) : монографія. Харків : ФО-П Макаренко, 2014. 360 с. 16. Качмар О.В. Вплив соціального макро- і мікросередовища на прояви агресивності особистості. Гуманітарний вісник ЗДІА. 2016. № 66. С. 183‒191. 16. Качмар О.В. Вплив соціального макро- і мікросередовища на прояви агресивності особистості. Гуманітарний вісник ЗДІА. 2016. № 66. С. 183‒191. 17. Склад засуджених осіб за судовими рішеннями, що набрали законної сили за 2019 рік. Судова влада України. URL: https://court.gov.ua/inshe/ sudova_statystyka/nasilstvo_19 17. Склад засуджених осіб за судовими рішеннями, що набрали законної сили за 2019 рік. Судова влада України. URL: https://court.gov.ua/inshe/ sudova_statystyka/nasilstvo_19 18. Склад засуджених за судовими рішеннями, що набрали законної сили за 2020 рік. Судова влада України. URL: https://court.gov.ua/inshe/ sudova_statystyka/rik_2020 18. Склад засуджених за судовими рішеннями, що набрали законної сили за 2020 рік. Судова влада України. URL: https://court.gov.ua/inshe/ sudova_statystyka/rik_2020 CRIMINOLOGICAL CHARACTERISTICS OF A PERSON WHICH DOES DOMESTIC VIOLENCE The relevance of the article is that one of the components of the subject of criminology is the doctrine of the identity of the offender. Without it, it is difficult to investigate other criminological problems, in particular, the determinants of crimes and means of combating them. On the one hand, the factors of crime are related to objective social phenomena that negatively affect the formation of personality and are manifested in selfish, anti-social views and habits that underlie criminal behavior, and on the other - with the circumstances that contribute to the implementation of such views and habits in a particular crime. These phenomena and circumstances do not depend on the person, are external to him, but they acquire their criminogenic significance, forming the identity of the offender and thus transforming into internal determinants of the crime. The article is devoted to the criminological characteristics of a person who commits criminal acts in a family home. The existing doctrinal approaches to understanding the categories «offender» and «offender» are clarified, on the basis of which it is substantiated that the offender is a person who commits systematic physical, psychological or economic violence against a spouse or ex-spouse or another person with whom the person is (was) in a family or close relationship. The socio-demographic characteristics of the offender are determined, namely: a) sex (the vast majority of people are men); b) level of education; c) age; d) state of alcohol intoxication; e) employment; g) the presence of a criminal record. Has been ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2 КРИМІНАЛЬНЕ ПРАВО ТА КРИМІНОЛОГІЯ; КРИМІНАЛЬНО-ВИКОНАВЧЕ ПРАВО 32 established a criminological portrait of a person who commits domestic violence (a male citizen of Ukraine, 30‒50 years old, with a basic secondary education, able to work, but not working, not convicted, prone to alcohol consumption). It is concluded that a person who commits domestic violence should be classified as a violent type of criminal, and by the depth of anti-social orientation to situational unstable criminals. CRIMINOLOGICAL CHARACTERISTICS OF A PERSON WHICH DOES DOMESTIC VIOLENCE It is emphasized that the structure of the offender's personality presupposes the presence of three groups of characteristics (properties, qualities): 1) socio- demographic (sex, age, social status, level of education, marital status, occupation, etc.); 2) criminal law (content and motivation, individual or group nature of criminal activity, duration and intensity of criminal activity, presence or absence of a criminal record, etc.); 3) socio-psychological properties and qualities of persons who have committed a crime (worldview, interests and attitudes). Keywords: domestic violence, person of the criminal, person-offender, socio- demographic characteristics, criminological portrait of the person, violent type of person of the criminal. ЮРИДИЧНА НАУКА № 2(104)/2020. ТОМ 2
https://openalex.org/W4361023397	https://animalproduction.net/index.php/JAP/article/download/536/475	English	null	Farmer Children’s Willingness for Dairy Farming succession in Banyumas Regency	Animal production/Animal Production	2,016	cc-by-sa	3,910	Faculty of Animal Science, Jenderal Soedirman University, Purwokerto 53123, Indonesia Corresponding author email: krismiwati@gmail.com Faculty of Animal Science, Jenderal Soedirman University, Purwokerto 53123, Indonesia Corresponding author email: krismiwati@gmail.com Abstract. The objectives of this research were to 1) observe the age, education, farm parents’ cattle ownership, parents’ income, children’s perception and willingness for farming succession, and 2) investigate the relationship between age, education, farm parents’ cattle ownership, parents’ income, children’s perception of and willingness for farming succession. Purposive sampling was used to determine the regional sample in three districts in Banyumas regency with the most dairy cow population and included in the supervised area subject to “Pesat” Milk Cooperative, namely Pekuncen, Cilongok and Baturaden districts. Ninety respondents from 10-30 years old were determined. Result showed that farmers’ children had better education than the parents with average age of 21,67 years old. Average cattle ownership was 1-15 heads per household with Rp 564,167 – Rp 4,523,333 monthly income. Farmers’ children shaped positive perception on dairy farming and had willingness for farming succession. Age was positively correlated with farming perception and willingness of the farmers’ children, while education was negatively correlated. Key words: farmer children’s perception, willingness to farm, education, age, income Key words: farmer children’s perception, willingness to farm, education, age, income Abstrak. Penelitian bertujuan: 1) Mengkaji umur, pendidikan, jumlah ternak sapi perah milik orangtua, pendapatan orangtua, persepsi anak dan kemauan anak untuk melanjutkan usaha sapi perah orangtuanya dan 2) Mengetahui hubungan antara umur, pendidikan, jumlah ternak sapi perah milik orangtua, pendapatan orangtua dengan persepsi anak dan kemauan anak untuk melanjutkan usaha sapi perah orangtuanya. Metode penetapan sampel wilayah yang digunakan yaitu purposive sampling, dengan memilih 3 kecamatan di Kabupaten Banyumas yang memiliki populasi peternak sapi perah terbanyak dan merupakan daerah binaan Koperasi susu “Pesat” yaitu Kecamatan Pekuncen, Cilongok, dan Baturraden. Penetapan sampel responden usia 10-30 tahun) sebanyak 90 orang. Hasil penelitian menunjukkan bahwa anak peternak memiliki pendidikan yang lebih baik dari orangtuanya dan berumur rata-rata 21,67 tahun. Kepemilikan ternak berkisar antara 1 hingga 15 ekor per rumah tangga peternak, sedangkan pendapatan yang diperoleh sebesar Rp 564.167 – Rp 4.523.333/bulan. Anak peternak memiliki persepsi baik terhadap usaha sapi perah juga memiliki kemauan untuk melanjutkan usaha orangtuanya. Umur berkorelasi positif dengan persepsi dan kemauan beternak anak peternak tetapi pendidikan memiliki korelasi negatif terhadap kemauan anak peternak untuk melanjutkan usaha orangtuanya. Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Faculty of Animal Science, Jenderal Soedirman University, Purwokerto 53123, Indonesia Corresponding author email: krismiwati@gmail.com Kata kunci: Persepsi anak peternak, kemauan beternak, pendidikan, umur, pendapatan Kata kunci: Persepsi anak peternak, kemauan beternak, pendidikan, umur, pendapatan aspect in farming with strategic role to fulfill food demand and increase national revenue (Simamora, 2015). Farmer Children’s Willingness for Dairy Farming Succession in Banyumas Regency Krismiwati Muatip and Moch Sugiarto Operational definitions 1. Age is the age of farmers’ children in year unit, round off to the nearest year. 2. Education is the length of formal education taken by farmers’ children in year unit. 3. Income is farmers’ revenue from dairy farm (milk and waste/manure sell, and cattle added value) in rupiah unit within a year. 4. Dairy cow ownership is the number of cattle owned by farmers in head unit. Purposes of this research were 1) to observe the age, education, parents’ cattle ownership, parents’ income, children’s perception and willingness of farming succession. 2) to investigate the relationship between ages, education, parents’ cattle ownership, parents’ income, children’s perception and willingness of farming succession. 5. Perception is the view of farmers’ children on the sustainability of dairy farm measured from the children’s point of view, the benefit for society and economic value of dairy farm. Perception was measured in 1-4 likert scale. 6. Willingness to farm is the desire of farmers’ children for farming succession. Introduction Dairy farming in Banyumas has long been established by the society precisely in 1987 through funding by European Economic Community (EEC). To date, however, dairy farming is relatively stagnant. Muatip and Aunurohman (2004) reported that dairy cow ownership in 2003 in Banyumas was 3-5 head/household. In 2015, the number remained within the range. It indicated that dairy farming has not shown a sustainable farming development system, whereas milk is one In fact, when diligently managed, dairy farming should be able to improve farmers’ welfare because it produces milk every day to support farmers’ household. Milk as dairy farming product is an important food product for health because of the nutrient content; therefore, dairy farming serves as the most important fundamental to establish food security. Besides milk, manure and urine are 118 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Pekuncen, Cilongok and Sumbang districts were the monitored area of “Pesat” Milk Cooperative. Sample respondents were 50% of the total 209 members of the cooperative (Annual Meeting Report, 2015), thus 105 farmers. Then farmers children were subject to census (10-30 years old), making up 90 respondents. dairy farming by-products with economic value to fertilize plants or to compose renewable energy to cut down farmers’household expense. Heifer are also dairy farming product to increase cattle ownership. Regeneration is very essential for a sustainable dairy farming. Regeneration means how dairy cow business run by the parents can be preserved or developed by the children. Whether or not farmers’ children want to continue the dairy cow business has been noted since the age of 10. At this stage, children start to understand and comprehend life events, including the character that sparks perception of a profession. Environment and circumstances around the farmers’ children are visible and sensible and therefore affect perception and decision to determine their future. A good perception of dairy farmingis expected to encourage farmers’ children to follow their path. Regeneration of dairy farmingis desired to make it more developed and sustainable to support milk self-sufficiency Indonesia. The observed variables were the perception of farmers’ children of dairy farming(Y1), farmers’ children willingness to farming (Y2), age of farmers’ children umur anak peternak (X1), education of farmers’ children (X2),parents’ income (X3), and parents’ cattle ownership (X4). Materials and Methods Descriptive analysis was performed to provide information on age, education, farm parents’ cattle ownership, parents’ income, children’s perception of and willingness for farming succession. Then, Spearman’s rank correlation was used to analyze the relationship between on age, education, farm parents’ cattle ownership, parents’ income, children’s perception of and willingness for Sampling and variables Purposive sampling was used to determine the regional sample for three districts with the highest dairy cow population. According to Sugiyono (2013), purposive sampling was setting the sample based on certain consideration. Sample items were specifically chosen in line with research purpose to obtain the more representative data. In this research, 119 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 farming succession. Systematic Spearman’s rank correlationwas as follows: Based on the Annual Report of “Pesat” Cooperative, milk production of the members was 5,097 liter/day. Based on the Annual Report of “Pesat” Cooperative, milk production of the members was 5,097 liter/day. farming succession. Systematic Spearman’s rank correlationwas as follows: Farmer children’s age was 12 to 30 years, 21.67 years on average. Age is correlated with ability to work and adoption level. Dairy cattle farming takes strong physic, high consistency and discipline. Young age has flexibility to apply innovation (Zanu et al., 2012). They are also highly adaptable and tolerant to innovation development. Where rs = Spearman’s rank coefficient correlation n = number of sample di = the i gap Where rs = Spearman’s rank coefficient correlation n = number of sample di = the i gap Coefficient Correlation: Coefficient Correlation: 0,80 – 1,00 : very strong or perfect correlation 0,60 – 0,79 : strong correlation 0,40 – 0,59 : moderate correlation 0,20 – 0,39 : low correlation 0,00 – 0,19 : no or weak correlation Research on dairy farmers in Ireland reported that farmer’s age significantly affected motivation to develop dairy farming (McDonald et al., 2013). Young farmers had less asset or savings than the aged farmers, but showed a proficient ability of business planning. Therefore, young farmers need capital support to develop their farm. 0,00 – 0,19 : no or weak correlation Results and Discussion Dairy farming in Banyumas has long been established by the society. Precisely in 1987, the government through Developing Baturraden Dairy Cow project funded by European Economic Community (EEC) spread 756 dairy cows. To date, the population of dairy cows keeps growing. According to Banyumas Statistics Bureau in 2015, 1567 dairy cows were spread in 8 districts in Banyumas. Dairy farming was strengthened by the Livestock Breeding Center for Excellence and Forage (BPTU-HMT) in Baturraden. Farmer’s children had adequate education although some were only primary school graduates. The average education was 11.08 years or equal to the second year in high school. Farmer children’s education was relatively better than the parents’. Education is a long- term investment for life. Education gives someone broad insight and good analytical ability prior to decision-making. McDonald et al. (2013) reported that dairy farmers in Ireland that followed farming development program had certificate of at least 180-hour dairy farming course. Education quality gives positive effect on financial performance of dairy farming. Education also significantly affects entrepreneurial behavior that enables business development (Patel et al., 2014).According to Fita et al. (2012), farmers’ education, training and exposure to media are the contributing factors of farming ability. Furthermore, farming ability can serve as a reference to improve farming knowledge or skill. Parents or relatives are also the source of farming knowledge (McDonald et al, 2013). Observed from the initial ownership, cattle ownership was from legacy (40.95%), self- purchase (21.91%),government support with credit system (through cooperatives) (28.57%) and entrusted cattle (8.57%). Based on the data, it was obvious that the majority of dairy cow farmers in Banyumas succeeded parents’ farm (legacy). In line with Sentosa (2013), dairy farming knowledge was obtained from farm parents. All dairy farmers in Banyumas sold milk product to “Pesat” Milk Cooperative. Although many farmers were not the cooperative members, they entrusted their milk product to other farmer member to sell in the cooperative. Dairy farming in Banyumas are small-holder farming with 1 to 15 cattle ownership per 120 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 household. The ownership was relatively stagnant year to year because dairy farming was merely a side business. Santoso et al. (2012) stated that a business belongs to side job when the venture is run in simple manner and contributes less than 30% of household income. Farmer children’s perception on dairy farming Farmer children’s perception on dairy farming Analysis result showed that most farmer children or 79 respondents had positive perception of dairy farming. Good perception might rise since the children witnessed their parents’ ability to provide for family and send children to school through dairy farming although the venture was traditionally managed or only a side job. Moreover, the children felt the benefit of dairy farming. It was in line with Kunasekaran et al. (2012) that good perception rises when someone has seen and felt the benefit of an innovation (venture). The perception will get better when farmers own knowledge of entrepreneurial cow farming. Dairy farming annual income was Rp 6,770,000-Rp 54,280,000 from selling milk and manure and cattle price raise. Calculated monthly, the income was Rp 564,167–Rp 4,523,333. The price of milk was determined by the cooperative based on milk quality namely fat content and milk hygiene. The best quality was Rp 5,800–Rp 6,000/liter, medium quality was Rp 5,600–Rp 5,700/liter, and low quality was Rp 5,400–Rp 5,500/liter. Spearman’s rank analysis indicated that coefficient correlation between farmer children’s age and their perception was 0.016 or positive correlation with very weak coefficient correlation. Howley et al. (2012) stated that aged farmers were relatively conservative and difficult to change standpoint. Farmers’ income could be improved if the farmers consider cattle balance (lactating and non-lactating). To date, farmers did not pay attention to cattle balance, even eight farmers in the study did not have lactating cow. Income was obtained from the rise of cattle price. It caused detrimental effect to farmers. Improving farmers’ income was also by improving feed quality. “Pesat” cooperative provided concentrate in various types and process, but most farmers chose the cheapest concentrate (RP 3,200/kg). It detrimentally affected the quality and price of milk. Income from milk was merely enough to buy concentrate. As seen from income and milk selling, farmers experienced loss. Wolf et al. (2014) in the US reported that the government determined milk price that was adequate to pay feed and labor. Correlation between farmer children’s education and dairy farming was -0.085 or negative correlation with very weak coefficient correlation. Education of farmers’ children was relatively good but family environment was less supportive to build a good perception on dairy farming. Farmers had less time to communicate with the children and gave less guidance for their future. Results and Discussion The main job was agricultural farmer, civil servant or army and merchant. The other contributing factors to the less-developed dairy farming were limited field and capital. Dairy farming needs field to grow forage to ensure forage availability. However, farming business that is integrated with agriculture makes farmers divide the field for the two venture, and they tend to prioritize the agriculture. Integrated farming systemis a venture to improve dairy farmers’ income. However, it requires farmers to own knowledge, creativity, motivation and resources. Integrated farming system can improve profit, environment conservation and farmers’ community. Despite the many benefits, this system complexity can obstruct adoption (Russelle et al., 2007). Farmer children’s perception on dairy farming It might due to farmers’ low level of education, limited communication skill and less knowledge on job empowerment. Many farmers did not want their children to continue farming business because dairy farming did not have fixed income and future. The children 121 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 All research respondents were male children. Total of 82 children (91.11%) were willing for dairy farming succession, while 8 children (8.89%) were doubtful. Farmer children’s willingness to continue dairy farming was likely due to interest in the business and perception that it has a promising potential. As reported by Kiama (2016), women are more careful to make decision. Children’s interest to dairy farming needs support by a good communication between them and parents. Children should be involved in breeding because it served as non-formal education process. were encouraged to be an employee with stable monthly income. Higher level of education should be linear to the amount of knowledge obtained by the farmers’ children. Higher formal education provides the farmers’ children greater job opportunity than farming business. In fact, education level affects farmers’ access to information and ability to understand the technical aspects of innovation that help accelerate adopting innovation (Zanu et al., 2012). Farm parents’ cattle ownership was positively correlated (0.054) with the children’s perception of dairy farming. The more cattle owned, the more positive the perception. However, the very weak correlation indicated that parents’ cattle ownership was not correlated to children’s perception on dairy farming; instead, there was tendency to positive correlation. It was due to relatively equal cattle ownership among farmers. Farmers were reluctant to add more cattle due to some reasons: 1) farmers had limited forage area, 2) farmers had limited workers or time because dairy farming was integrated with agriculture and 3) farmers had limited resources that discouraged them to adopt innovation and to build collaboration with other parties. Correlation between farmer children’s age and willingness for succession was a positive 0.027. Small correlation value is disregarded, thereby age is not correlated with the willingness to farm in farmers’ children, but it tends to have positive correlation. A good perception of dairy farming makes children willing to continue their parents’ business, although they did not instruct the children to follow their path. The older the children, the more willing for succession. Farmer children’s perception on dairy farming It was probably because the majority of dairy cows were obtained by legacy. A credence in Banyumas was that legacy is not for sale. Education of farmer children negatively correlated (-0,007) with willingness to farm. The very low correlation indicated that education level was not correlated but tended to be negatively correlated. Higher education enables someone to make predictions on a business. Copeland and Dharmasena (2016) stated that milk consumption in America has declined for the last 25 years, might due to the increasing awareness of health. Alternative drink (soya milk, almond milk) is generally considered a more healthy choice because of less calorie and minus growth hormone content as in dairy milk. Higher education enables someone to make a more objective assessment because they have more The more cattle ownership, the more positive children’s perception on dairy farming. Positive perception of a profession affects the interest to toil in the profession. Perception is related to the subject (individual) interest on the observed object (Abdullah, 2015). Parents’ income was positively correlated with farmer children’s perception with 0,279 coefficient correlation, indicated a weak correlation but very significant relationship. Parents’ income from dairy farming has shaped positive perception to children of dairy farming. Willingness to farm of dairy farmers’ children 122 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 information from a higher ability to access information. Conclusions Children of dairy cow farmers are in productive age with considerably high education. The age of farmers’ children positively correlated with children’s perception of dairy farm and willingness to continue parents’ business, while children’s education negatively correlated with the perception of dairy farm and willingness to continue parents’ business. Parents’ income and cattle ownership positively correlated with children’s perception of dairy farm and the willingness to continue parents’ business, moreover the better children’s perception the more willingness to continue dairy cow business. Farm parents’ cattle ownership was positively correlated (0.044) with the children’s willingness for farming succession. Small correlation value showed that parents’ cattle ownership was not correlated to children’s willingness to farm but it tended to have positive value. If the more cattle owned was balanced with the correct composition of cow’s age, it could sustain farmers’ income. Farmers’ income was positively correlated (0.149) with children’s willingness for farming succession. The higher the income, the more expectation to improve farmers welfare. A study by Wahyudi (2014) of “SAE” cooperative members in Malang found that cow ownership encouraged members’ productivity. Besides, the more cows owned, the higher milk production. The increasing output was linear to farmers’ income and therefore stimulated the farmer children to continue parents’ farming. References Abdullah A. 2015. Farmers’ perception on the work of extension agent in improving technology of hay processing and beef cattle waste. Jurnal Ilmu and Teknologi Peternakan Tropis 1(1) Perception and willingness for farming succession had moderate coefficient correlation (0.586). Daniar et al. (2014) stated that the youth perception affects their interest on certain object. According to Arifianto and Sukanti (2014) willingness can derive from themselves and make them do efforts to achieve the goal. Willingness or desire from themselves is usually influenced by some factors, one of which is upbringing pattern. Family upbringing can affect children’s interest to be dairy farmer. Previous studies reported that a mother is the dominant nurture, so mother holds an important role to bring up the children to be a dairy farmer. Willingness to farm in this study was in high category because not all mother who brought up the children also helped their husband to take care of cows in the farm. Accordingly, knowledge and experience in dairy farming was limited, so only a little information on dairy farming was passed on to the children. Koperasi Pesat. 2015. Report of Annual Member Meeting 2015. Purwokerto Arifianto F and S Sukanti. 2014. Effect of self- motivation and perception on public accountant profession on the interest to be public accountant in the students of Accounting Program, Economics Faculty, Yogyakarta State University. Jurnal Nominal 3(2). National Statistics Bureau. 2015. Banyumas In Numbers. Purwokerto. Copeland A and S Dharmasena. 2016. Impact of increasing demand for dairy alternative beverages on dairy farmer welfare in the United States. Southern Agricultural Economics Association’s 2016 Annual Meeting, San Antonio, Texas. Daniar GR , BA Nugroho and E Nugroho. 2014. Perception and interest of young people in Madura cattle agribusiness (Study in Waru Districts, Pamekasan Region). Jurnal Ilmu-Ilmu Peternakan 24(3):69–78. Fita L, MM Trivedi and B Tassew. 2012. Adoption of improved dairy husbandry practices and its relationship with the socio-economic characteristics of dairy farmers in Ada’a Distric of Oromia State, Ethiopia. Jurnal Agricultural Extension and Rural Development 4(14):392-395. Howley P, CO Donoghue, K Heanue. 2012 Factor affecting farmers’ adoption of agricultural 123 Krismiwati Muatip and Moch Sugiarto/Animal Production. 18(2):118-124, May 2016 Accredited by DGHE No. 81/DIKTI/Kep./2011. ISSN 1411-2027 Indian Research Journal of Extension Education 14(2):46-49. innovation: A panel data analysis of the use of artificial insemination among dairy farmers in Ireland. Jurnal of Agricultural Science. 4(6):171- 179. Russelle MP, Entz MH and Franzluebbers AJ. 2007. Reconsidering integrated crop–livestock systems in North America. Agronomy Journal 99(2):325- 334. Kiama TN, Lindahl JF, Sirma AJ, Senerwa DM, Waithanji EM, Ochungo PA, Poole EJ, Kang’ethe EK and Grace D. 2016. Kenya dairy farmer perception of moulds and mycotoxins and implications for exposure to aflatoxins: a gendered analysis. African Journal of Food, Agriculture, Nutrition and Development. 16(3):11106-11125. Santoso K, S Warsito, and A Andoko. 2012. Feedlot business. Agromedia Pustaka. Jakarta. Santosa SI. 2013. Analysis of potential development of dairy cattle business under agribusiness paradigm in Musuk district Boyolali Region. Jurnal Buletin Peternakan. 37(2):125-135. Kunasekaran P, S Ramachandran, Z Samdin, KW Awang. 2012. Factors affecting farmers’ agro tourism involvement in Cameron Highlands, Pahang. OIDA International Journal of Sustainable Development. 4(1):83-90. Sugiyono. 2013. Non-parametric Statistics for Research.Alfabeta. Bandung. Simamora T, AM Fuah, A Atabany and Burhanuddin. 2015. Evaluation on technical aspect of smallholder dairy farming in Karo Region North Sumatra. Jurnal Ilmu Produksi and Teknologi Hasil Peternakan. ISSN 2303-2227. 03(1):52-58. McDonald R, Pierce K, Fealy R and Horan B. 2013. Characteristics, intentions and expectations of new entrant dairy farmers entering the Irish dairy industry through the new entrant scheme. International Journal of Agricultural Management. 2(4):189-198. Wahyudi A. 2014. Analysis of Contributing factors to the income of Dairy Cattle cooperative members (A case study in SAE Cooperative member Pujon District, Malang Region). Jurnal Ilmiah. Universitas Brawijaya. Malang. Muatip K, M Sugiarto and I Haryoko. 2015. Study on cow milk marketing in Central Java to support milk sustainability. Kajian Pemasaran Susu Sapi Perah di Jawa Tengah untuk Mendukung Swasembada Susu. In: Proceeding Unsoed. Purwokerto. Wolf CA and Widmar NJO. 2014. Adoption of milk and feed forward pricing methods by dairy farmers. Journal of Agricultural and Applied Economics 46(4):527. Zanu HK, Antwiwaa A And Agyemang CT. 2012. Factors influencing technology adoption among pig farmers in Ashanti Region of Ghana. Jurnal of Agricultural Technology 8(1):81-92. Patel P, Patel MM, Badodia SK and Sharma P. 2014. Entrepreneurial behaviour of dairy farmers. 124
https://openalex.org/W3179449664	https://www.e3s-conferences.org/articles/e3sconf/pdf/2021/59/e3sconf_iccaue2021_02002.pdf	English	null	Research on Key Technical Indexes of Coastal Reclamation Planning and Design	E3S web of conferences	2,021	cc-by	3,335	1 Introduction rainstorms resulted in insufficient drainage capacity of the inland and outer rivers in many coastal cities in Zhejiang, the water level rose, and serious floods were caused. The direct economic loss reached 4.737 billion yuan [8]. It can be seen that these unscientific reclamation projects and the consequences have restricted the sustainable economic development of coastal areas in southeast China. Due to the growing demand for the resources, coastal reclamation has become a global land use practice. Since the middle of the 20th century, China has reclaimed a total of 13.4 million ha of tidal flats.[1] Zhejiang is located in the southeast coastal area of China, with tortuous coasts, numerous islands and rich tidal flat resources. A total of 75134.3 ha of coastal wetland in Hangzhou bay were reclaimed in recent 4 decades.[2] Although the reclamation of coastal wetlands brings substantial economic benefits, it can lead to a series of adverse ecological impacts[2] due to the lack of scientific theoretical guidance and effective quantitative control measures. The natural and complex coastal landscape has undergone permanent changes [3], and the service function of the ecosystem has also been severely weakened [4]. According to the 2017 National Marine Environmental Status Announcement, most of China's typical bays, estuaries, coastal wetlands, and other ecosystems are in a sub-healthy or unhealthy state. Among them, Hangzhou Bay is a heavily polluted sea area [5]. What's more serious is that the spatial shape of the estuary coast and the complicated boundary changes, resulting in an increase in tidal range, rising tide level [6], and rising water level in the estuary [7]. This has increased the probability of flooding in coastal towns. For example, after the landfall of Typhoon Fitow, the 23th typhoon in2013 in China, the high tide level and Zhoushan Archipelago is China's first new state-level area formed by archipelago and a key development area in the Hangzhou Bay Economic Belt. In the high-speed urban development, the reclamation of tidal flats in Zhoushan area mostly adopts 3 methods: harbor reclamation around a single island, reclamation connecting a small island and a big island, or connecting two adjacent small islands. This leads to a substantial change in the landscape structure along the coast, especially the natural and complex morphological characteristics of the coastline. Undoubtedly, it will have a more serious impact on the surrounding ecological environment. Research on Key Technical Indexes of Coastal Reclamation Planning and Design Zhang Shixia1, , Zhan Zhenyu2, and Cao Guangtian1 1Zhejiang Tongji Vocational College of Science and Technology, Hangzhou 311231, China 2Huadong Engineering Corporation Limited, Hangzhou 310000, China Abstract. Large-scale coastal reclamation has become a significant land use issue worldwide for urban construction and economic development. The reclamation of coastal wetlands brings substantial economic benefits, however, the structure and function of coastal ecosystem are affected by drivers of human-caused landscape change. This research takes Hangzhou Bay and Zhoushan Islands as the case study to investigate the correlation between the coastal geomorphic complexity and the tidal range reduction rate, and to explore the control technical indexes of the design in reclamation area by a multidisciplinary approach that integrates the basic theories and quantitative methods of fractal geometry with the hydrodynamic mechanism of ocean dynamics. The results show that the coastal tidal range reduction rate is closely related to the coastline fractal dimension and patch shape index (D, S), and reveals the influence of the complexity of the coastal landscape on the tidal energy loss. In addition, based on model predictions, it can be found that the large-scale reclamation in Zhoushan will cause a serious decline in the complexity of the coastal landscape and the reduction of tidal energy, which is extremely detrimental to coastal disaster prevention. In the end, the scientific design theory and quantitative control indexes of reclamation are put forward to provide theoretical basis and design reference for future coastal reclamation and disaster prevention. © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). Corresponding author: zhsx@zju.edu.cn © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). Corresponding author: zhsx@zju.edu.cn https://doi.org/10.1051/e3sconf/202128302002 https://doi.org/10.1051/e3sconf/202128302002 E3S Web of Conferences 283, 02002 (2021) ICCAUE 2021 3 Results & Discussion (2) 1 Introduction At present, there is still a lack of theoretical research on the influence of the complexity of coastal landforms on tidal energy dissipation, as well as its mechanism of tidal prevention and disaster reduction. Especially in the planning and design of reclamation area, there is a lack of quantitative control indexes and methods for the complexity of tidal flat landforms. Therefore, with Hangzhou Bay and Zhoushan Islands as a case study (Figure 1), the relationship between the E3S Web of Conferences 283, 02002 (2021) ICCAUE 2021 https://doi.org/10.1051/e3sconf/202128302002 complexity of coastal geomorphology and tidal energy loss is studied, and quantifiable technical indicators of scientific reclamation are established, and ecological reclamation models that are conducive to coastal disaster prevention and mitigation are explored. It can provide a theoretical basis and design reference for the rational use of tidal flat wetland resources and the maintenance of the sustainable development of coastal areas. Fig. 1. Concurrent lines and isotidal lines of M2 tide peak in Hangzhou Bay Fig. 1. Concurrent lines and isotidal lines of M2 tide peak in Hangzhou Bay study, the number of boxes as a function of the grid size was calculated, and fractal dimensions were determined by Fractalize software. Normally, the more complex the shoreline shape is, the higher D is. 2.2 Shape index of patch (S) The shape index (S) is an important indicator for quantitative analysis of landscape spatial pattern and landscape heterogeneity. It shows the regional ecological environment from a macro perspective [11]. The shape index of patch (S), assessing the complexity of coastal landscape patch by calculating the degree of deviation between the shape of a patch and the circle or square of the same area, was measured by the following equation[15]. Fractal dimensions can quantitatively describe the complexity of irregular shapes, and reflect the naturalness and complexity of landscape composition and spatial pattern [11]. Fractal geometry and models have been successfully and widely used in quantitatively describing the fractal features of irregular objects such as coastlines and river networks [12-14]. A fractal model between the length of the observation object and the observation scale will be established in the calculation, and the grid method will be used[12-13] to calculate the fractal dimension of the coastline. The calculation formula is as follows: (3) reference as square a In 0.25 reference as round a In 2        ，， A P A P S  (3) where P is the perimeter of the patch, and A is the area of the patch. The value of shape index can reflect the size and the complexity of the landscape patch. The larger S is, the more complex the shape of patch is. Based on general the shape of study objects, this study measured the complexity of coastal islands and patches by using square objects of reference. ( ) D N     (1) lg N() DlgC (2) ( ) D N     (1) lg N() DlgC (2) (1) 2.1 Fractal Dimension and Calculation Method Fractal geometry theory takes irregular fragments as objects, such as the complex, chaotic, and rich forms of clouds, trees, coastlines in nature[9], and its fragments are self-organized and self-similar. Fractal theory plays an important role in observing and analysing the relationship between various complex characteristic forms and their internal ecological functions in nature [10]. Fractal model is a research method based on this theory to explain the spatial shape and structure of objects. Fractal dimensions can quantitatively describe the complexity of irregular shapes, and reflect the naturalness and complexity of landscape composition and spatial pattern [11]. Fractal geometry and models have been successfully and widely used in quantitatively describing the fractal features of irregular objects such as coastlines and river networks [12-14]. A fractal model between the length of the observation object and the observation scale will be established in the calculation, and the grid method will be used[12-13] to calculate the fractal dimension of the coastline. The calculation formula is as follows: 3.1The Correlation between Coastal Geomorphic Complexity and Tidal Energy Loss 63     S D A (4) The above formula shows that the tidal range reduction rate is closely related to the irregularity of the island and shoreline, the curvature of the boundary and the complexity of the spatial shape. The more tortuous the boundary of the coastline, the more complex the spatial shape, the larger the value of D and S, the greater the tidal range drop per unit length, and the greater the resistance to tidal current. The tidal energy is proportional to the square of the tidal range, so the more tidal energy is lost. It can also be seen from this model that the D value of the coastline plays a decisive role in tidal energy dissipation. Therefore, it is recommended to use D and S values as the technical indicators to control the coastal complexity in the planning and design of reclamation. The model also provides a new method for evaluation and simulation of reclamation planning. p In order to study the impact of the landscape complexity of the tidal path boundary (shoreline and islands) on the tidal current, Hangzhou Bay and Zhoushan Islands will be the main research objects. Taking Shupu at the mouth of the Qiantang River as a benchmark, the study area is divided into two areas in the north and south according to the central axis of Hangzhou Bay (Figure 2). Using fractal dimension model and shape index calculation method (formula 3), using fractalyse software, calculate the coastline fractal dimension D and patch shape index S of the two regions respectively. The calculation result is shown in Figure 3. It can be seen from Figure 3 that the D and S values in the south are larger than those in the north, and the shape index of the south is 2.14 times that of the north. The reason is that the southern part is the Zhoushan archipelago, which has islands of different shapes. The shape of the shoreline is much more complicated than that of the northern part. The landscape and landform are more complicated, and the resistance to tidal current is greater, and the tide energy is reduced more. It can be preliminarily shown that the complex geomorphic structure can reduce tidal energy and play a role in flood prevention and disaster reduction. 3.1The Correlation between Coastal Geomorphic Complexity and Tidal Energy Loss Where D is the fractal dimension of the coastline, and it is determined by computing the slope of the plotted relationship. An irregular curve will have a fractal dimension higher than 1 but smaller than 2. And C is a proportionality constant which allows for the relationship between N and ε mentioned above. In this According to statistics, from 1951 to 2005, the total area of reclaimed tidal flats in Zhoushan has reached 199,200 mu, which has fundamentally changed the landscape structure along the coast, especially the natural and 2 E3S Web of Conferences 283, 02002 (2021) ICCAUE 2021 https://doi.org/10.1051/e3sconf/202128302002 Fig. 3. Coastal Landscape Characteristic Indexes (S, D) complex morphological characteristics of the coastline. Therefore, taking Hangzhou Bay and the Zhoushan Islands as the main research objects (Figure 1), the impact of the landscape complexity of the tidal flow boundary (shoreline and islands) on the tidal energy loss is studied. The tidal resonant along the coast of Zhejiang are formed by Pacific tidal waves and celestial tidal forces. Tides in Hangzhou Bay are dominated by semidiurnal M2 tide which is a progressive wave in East China Sea. Figure 1 shows the concurrent lines and isotidal lines of M2 tide peak in Hangzhou Bay. It can be seen from the concurrent lines that the progressive wave in the East China Sea first arrived at Zhujiajian in the Zhoushan Islands in the southern part of Hangzhou Bay. Due to the scattered distribution of Zhoushan Islands in the southern part of Hangzhou Bay, the propagation speed of tidal waves slowed down after being blocked by the islands. However, there are few islands in the northern part, and the tidal wave propagates fast. When it reaches the mouth of the bay, the concurrent lines turn close to north-south. The tidal range in Zhoushan Islands is much lower than other places. Fig. 3. Coastal Landscape Characteristic Indexes (S, D) According to ocean dynamics theory, tidal energy is proportional to the square of tidal range, that is, for the same water surface area, the smaller the average tidal range, the smaller the tidal energy. So, it is reasonable to use the tidal range reduction rate, that is, the change in tidal range per unit distance, to indicate the degree of tidal energy loss. On the basis of my previous related research, the model was obtained (formula (4)) [16]. 437 . 71 644 . 1 585 . 3.2 Impact of Zhoushan Coastal Reclamation on Tidal Energy Loss and Countermeasures According to the "Zhoushan Tidal Flat Reclamation Planning Report", the current land reclamation plan for Zhoushan Island still uses the three traditional methods described in Chapter 1, which greatly reduces the tortuosity and complexity of the shoreline. Based on the 1:100,000 hardcopy topographic map of Zhoushan City in the early 1970s and the digital topographic map of 2004 (after large-scale reclamation), this study used the model (formula 4) to compare and analyze the changes in the landscape index and tidal range reduction rate before and after the implement of reclamation plan. As shown in Table 1. Fig. 2. Hangzhou Bay, Zhoushan Islands, and research boundaries Fig. 2. Hangzhou Bay, Zhoushan Islands, and research boundaries 3 https://doi.org/10.1051/e3sconf/202128302002 E3S Web of Conferences 283, 02002 (2021) ICCAUE 2021 Table 1. Changes in the landscape index and tidal range reduction rate before and after the implementation of the reclamation plan. D S N A  /（mm/km） Before 1.144 3.66 98 7.32 After 1.068 20.67 83 0.86 Table 1. Changes in the landscape index and tidal range reduction rate before and after the implementation of the reclamation plan. the complex topography of the shoreline. Both the fractal dimension D and the shape index S are greatly reduced, and the tidal range reduction rate is reduced by more than 88%. It is extremely unfavorable for disaster prevention in coastal and estuaries and bay areas. D S N A  /（mm/km） Before 1.144 3.66 98 7.32 After 1.068 20.67 83 0.86 p y Therefore, when planning and designing tidal flat reclamation in a specific area, it is necessary to conduct detailed investigation on the landscape features that can withstand disasters, and protect the natural and complex spatial pattern as much as possible. It is recommended to adopt the quantitative indicators and methods of scientific reclamation based on "D\S" to improve and control the complexity of the landform, and to ensure that the complexity of the coastal landscape is not lower than the original value. The research results not only provide theoretical basis and practical guidance for the future large-scale tidal flat development of Hangzhou Bay and Zhoushan Islands, but also can enrich the basic theory and quantitative control indexes of tidal flat reclamation engineering design, which has very important scientific significance. It can be seen from Table 1 that due to traditional reclamation methods, the coastline shape of Zhoushan and nearby areas has undergone tremendous changes. 3.2 Impact of Zhoushan Coastal Reclamation on Tidal Energy Loss and Countermeasures The original tortuous shape of the shoreline became straight after reclamation. At the same time, the reduction of islands (the number of patches N) also reduces the complexity and richness of the spatial form, and reduces the length of the land and water boundary. Both the fractal dimension D and the shape index S are greatly reduced, which will greatly weaken the coastal hydrodynamic effect and reduce the tidal wave energy loss. The calculation results show that the tidal range reduction rate has been reduced by 88.3%, which has adversely affected the disaster prevention work in the coastal and downstream estuaries and bay areas. Acknowledgements: This research is sponsored by the Scientific and Technical Project of Water Conservancy of Zhejiang Province (grant number: R14842, RC2008). Therefore, determining the mode of coastal reclamation should start from determining the shape of the shoreline, and taking the fractal dimension as the main control index. It is necessary to follow the self- similarity of the natural coastline in Zhoushan area and try to imitate the natural morphological characteristics of the coastline. Meanwhile, on the basis of enriching the shape of the shoreline, referring to the auxiliary indicator of the shape index, the spatial layout of multiple islands should be retained. In this way, the fractal dimension and landscape shape index after the reclamation of the tidal flat are greater than or equal to the value before the reclamation to ensure that the complexity of the landscape after the project is not reduced. This can extend the length and space complexity of the land and water boundary, increase the resistance of tidal currents, reduce tidal energy and alleviate disasters. References 1. Fu, Y.B., Cao, K., Wang, F.(2010) Preliminary study of the methods used to evaluate the potential impacts of coastal reclamation.Ocean Development and Management., 27 (1):27-30. 2. Qiu,L.F.,Zhang,M.,Zhou,B.B.,Cui,Y.Z.,Yu,Z.L.,Liu ,T.,Wu,S.H.(2021)Economic and ecological trades- offs of coastal reclamation in Hangzhou Bay, China. Ecological Indicators.,125:1:12. 3. Peng, B.R., Lin, C.C., Jin, D.(2013)Modeling the total allowable area for coastal reclamation: A case study of Xiamen, China.Ocean & Coastal Management.,76:38-44. 4. Rao,H.H.,Lin,C.C.,Kong,H.(2014)Ecological damage compensation for coastal sea area uses. Ecological Indicators.,(38):149-158. 4 Conclusions This paper uses fractal dimension model, related calculation software (Fractalyse) and landscape quantitative indicators to calculate the coastline fractal dimension, shape index of island patch and tidal range reduction rate in typical areas, and study the correlation between each index and tidal range reduction rate. The results show that the tidal range reduction rate is closely related to the complexity of coastal morphology. The more tortuous the coastline and the more complex the spatial shape, the greater the value of D and S, the greater the resistance to the tidal current, the greater the tidal energy loss. 5. State Oceanic Administration (SOA).(2018)Announcement on the State of Marine Ecology and Environment of China in 2017.,www.soa.gov.cn/zwgk/hy 6. You,A.J.,Hang,Z.C.,He,R.Y.(2010)Characteristics and effecting factors of the tidal level in the Qiantangjiang River Estuary under changing environment. Journal of Marine Sciences., 28(1):18- 25. 7. Chen,X.W.,Liu,X.,Zhang,W.(2011)Shore reclamation in Pearl River Esturay and its impact analysis . Journal of Hohai University (Natural Sciences).,39(1):39-43. In addition, the model was used to evaluate the reclamation plan in the Zhoushan area, and the results showed that the traditional reclamation method changed 4 E3S Web of Conferences 283, 02002 (2021) ICCAUE 2021 https://doi.org/10.1051/e3sconf/202128302002 8. http://news.xinmin.cn/domestic/bjtj/2013/10/08/221 66339.html. 8. http://news.xinmin.cn/domestic/bjtj/2013/10/08/221 66339.html. 8. http://news.xinmin.cn/domestic/bjtj/2013/10/08/221 66339.html. 9. Wazawa,F., Kohiko,K., Kenaki,I.(2000)Landscape Environment Theory. Earth Society Press,Tokyo. 10. Griggs,J.A.(2018)Using fractal geometry to examine failed implants and prostheses. Dental Materials., 34(12): 1748–1755. 11. LI, B. L.(2000)Fractal geometry applications in description and analysis of patch patterns and patch dynamics. Ecological Modelling., 132(1/2): 33–50. 12. Zhu,X.H.,Cai,Y.L.(2004)Study on fractal dimension of Chinese Coastline and character.Advance in Marine Science.,22(2):156-162. 13. Zhang,S.X., Guo,Y.K.,Wang, Z.W. (2015)Correlation between flood frequency and geomorphologic complexity of rivers network–a case study of Hangzhou China. Journal of Hydrology.,527: 113–118. 14. Many,G., Durrieu de Madron,X.,Verney, R.,Bourrin, F.,Renosha,P.R.,Jourdin,F.,Gangloff,A.(2019)Geom etry, fractal dimension and settling velocity of flocs during flooding conditions in the Rhône ROFI. Estuarine, Coastal and Shelf Science., 219: 1–13. 15. Zhang S. X.,Wang, Z W.,Wu, S.N.(2020)Investigation of correlation between complexity of coastal landscape and water ecological environment.Marine Environment Science.,39:880-886. 16. Zhang S. X.,Wang, Z. W.,Wu, S.N.(2012) Landscape mechanism quantitative study on impact of tidal flat reclamation on the function of coastal disaster prevention：illustrated with Hangzhou Bay and Qiantang estuary. Journal of Zhejiang University(Engineering Science).,46:1281-1288. 5 5
https://openalex.org/W2954642868	https://ejurnal.stmik-budidarma.ac.id/index.php/jurikom/article/download/44/24	Indonesian	null	SISTEM PAKAR DIAGNOSA PENYAKIT BATU GINJAL MENGGUNAKAN TEOREMA BAYES	JURIKOM (Jurnal Riset Komputer)/Jurikom	2,016	cc-by	2,836	ISSN : 2407-389X ISSN : 2407-389X Jurnal Riset Komputer (JURIKOM), Volume : 3, Nomor: 1, Februari 2016 SISTEM PAKAR DIAGNOSA PENYAKIT BATU GINJAL MENGGUNAKAN TEOREMA BAYES Intan Russari Mahasiswa Program Studi Teknik Informatika STMIK Budi Darma Medan Jl. Sisingamangaraja No. 338 Simpang Limun Medan http : //www.stmik-budidarma.co.id // Email :intanrussari2@gmail.com Kata kunci : Batu Ginjal, Sistem Pakar, Metode Bayes Kata kunci : Batu Ginjal, Sistem Pakar, Metode Bayes 1.1 Latar Belakang Masalah Perumusan masalah yang akan dibahas dalam skripsi ini adalah : Batu ginjal (renal lithiasis) adalah penyakit yang berasal dari gumpalan kecil dan keras yang terbentuk di dalam ginjal. Batu ginjal dapat disebabkan oleh berbagai hal. Pada skenario yang umum, batu ginjal terbentuk ketika urin berkonsentrasi, mineral mengkristal dan menggumpal. Sakit batu ginjal biasanya dimulai pada sisi tubuh atau punggung, dibawah pinggul serta bergerak ke perut bagian bawah dan pangkal paha. Rasa nyeri sering berubah seiring pergerakan batu ginjal pada saluran urin. Batu ginjal dapat ataupun tidak menyebabkan tanda dan gejala sampai batu tersebut bergerak didalam ureter pipa yang menghubungkan ginjal dan kandung kemih. Batu ginjal terbentuk ketika komponen urin cairan dan berbagai mineral dan asam hilang keseimbangan. Ketika hal ini terjadi, urin terdapat lebih banyak zat yang mengkristal, seperti kalsium, oxalate dan uric acid, daripada cairan. 1. Bagaimana mengumpulkan data mengenai penyakit batu ginjal? 2. Bagaimana menerapkan metode Bayes dalam menentukan hasil diagnosa penyakit batu ginjal ? 3. Bagaimana merancang aplikasi sistem pakar diagnosa penyakit batu ginjal ? g p y 1.3 Batasan Masalah Sistem Pakar Diagnosa Penyakit Batu Ginjal Menggunakan Teorema Bayes. Oleh : Intan Russari ABSTRAK Batu ginjal (renal lithiasis) adalah penyakit yang berasal dari gumpalan kecil dan keras yang terbentuk di dalam ginjal. Batu ginjal dapat disebabkan oleh berbagai hal. Pada skenario yang umum, batu ginjal terbentuk ketika urin berkonsentrasi, mineral mengkristal dan menggumpal. Sakit batu ginjal biasanya dimulai pada sisi tubuh atau punggung, dibawah pinggul serta bergerak ke perut bagian bawah dan pangkal paha. Rasa nyeri sering berubah seiring pergerakan batu ginjal pada saluran urin. Sistem Pakar adalah suatu program komputer yang dirancang untuk mengambil keputusan seperti keputusan yang diambil oleh seseorang atau beberapa orang pakar. Menurut Marimin (1992), Sistem Pakar adalah sistem perangkat lunak komputer yang menggunakan ilmu, fakta dan teknik berpikir dalam pengambilan keputusan untuk menyelesaikan masalah- masalah yang biasanya hanya dapat diselesaikan oleh tenaga ahli dalam bidang yang bersangkutan.Adapun tujuan yang akan dicapai adalah untuk membuat aplikasi sistem pakar yang berguna sebagai alat bantu untuk mendapatkan informasi dan dugaan awal dalam mendiagnosa penyakit batu ginjal. Hasil dalam penelitian ini adalah sistem pakar untuk mendiagnosa penyakit batu ginjal dengan menggunakan metode bayes dapat menyelesaikan masalah diagnosis penyakit batu ginjal, karena dapat memberikan hasil diagnosis dengan nilai probabilitas kemunculan setiap jenis penyakit. 1.3 Batasan Masalah Adapun yang menjadi batasan masalah dalam penulisan skripsi ini adalah: Adapun yang menjadi batasan masalah dalam penulisan skripsi ini adalah: 1. Bahasa pemrograman yang digunakan adalah Microsoft visual basic.Net 2008 dan Database yang digunakan adalah Mysql. 2. Sistem pakar yang dibuat hanya mendiagnosa penyakit batu ginjal tidak membahas cara pengobatan dan pencegahannya. 3. Hasil keluaran sistem berupa persentase user terdiagnosa penyakit batu ginjal. Sistem pakar (expert system) adalah sistem yang berusaha mengadopsi pengetahuan manusia ke komputer, agar komputer dapat menyelesaikan masalah seperti yang biasa dilakukan oleh para ahli. Sistem pakar yang baik dirancang agar dapat menyelesaikan suatu permasalahan tertentu dengan meniru kerja dari para ahli. Dengan sistem pakar ini, orang awampun dapat menyelesaikan masalah yang cukup rumit yang sebenarnya hanya dapat diselesaikan dengan bantuan para ahli. Didalam aplikasi tersebut diperlukan adanya metode yang dapat melakukan diagnosa penyakit batu ginjal tersebut dan mengambil suatu keputusan dari pengetahuan awal yg diperlukan 4. Pengambilan Data gejala penyakit batu ginjal diperoleh melalui interview dengan dokter spesialis ginjal yaitu Dr. Harnold P. Sihaloho, Sp.U, media internet. y j 2.4. Metode Bayes Metode Find-S tidak dapat digunakan untuk data yang tidak konsisten dan data yang bias, sehingga untuk bentuk data semacam ini salah satu metode sederhana yang dapat digunakan adalah metode bayes. Metode bayes merupakan metode yang baik di dalam mesin pembelajaran berdasarkan data training, dengan menggunakan probabilitas bersyarat sebagai dasarnya, Machin Learning team (2006, 2). Probabilitas bayes merupakan salah satu cara untuk mengatasi ketidakpastian data dengan cara menggunakan formula bayes yang dinyatakan dengan:, Sri Winiarti (2003, 191). 2. Menurut Ignizio : Sistem pakar adalah suatu model dan prosedur yang berkaitan, dalam suatu domain tertentu, yang mana tingkat keahliannya dapat dibandingkan dengan keahlian seorang pakar. 3. Menurut Giarratano dan Riley : Sistem pakar adalah suatu sistem komputer yang bisa menyamai atau meniru kemampuan seorang pakar. 2.2.1. Perbandingan sistem konvensional dengan sistem pakar 2.1. Kecerdasan Buatan 2.1. Kecerdasan Buatan Kecerdasan buatan berasal dari bahasa Inggris “Artificial Intelligence” atau disingkat AI, yaitu Intelligence adalah kata sifat yang berarti cerdas, sedangkan Artificial artinya buatan. Kecerdasan buatan yang dimaksud adalah di sini merujuk pada mesin yang mampu berpikir, menimbang tindakan yang akan diambil, dan mampu mengambil keputusan seperti yang dilakukan oleh manusia. Berikut adalah beberapa defenisi kecerdasan buatan yang telah didefenisikan oleh beberapa ahli. T.Sutojo, Edy Mulyono, V.Suhartono (2011). Batu, terutama yang kecil, bisa tidak menimbulkan gejala. Batu di dalam kandung kemih bisa menyebabkan nyeri di perut bagian bawah. Batu yang menyumbat ureter, pelvis renalis maupun tubulus renalis bisa menyebabkan nyeri punggung atau kolik renalis (nyeri kolik yang hebat). Kolik renalis ditandai dengan nyeri hebat yang hilang- timbul, biasanya di daerah antara tulang rusuk dan tulang pinggang, yang menjalar ke perut, daerah kemaluan dan paha sebelah dalam. Gejala lainnya adalah mual dan muntah, perut menggelembung, demam, menggigil dan darah di dalam air kemih. Penderita mungkin menjadi sering berkemih, terutama ketika batu melewati ureter. Batu bisa menyebabkan infeksi saluran kemih. Jika batu menyumbat aliran kemih, bakteri akan terperangkap di dalam air kemih yang terkumpul diatas penyumbatan, sehingga terjadilah infeksi. Jika penyumbatan ini berlangsung lama, air kemih akan mengalir balik ke saluran di dalam ginjal, menyebabkan penekanan yang akan menggelembungkan ginjal (hidronefrosis) dan pada akhirnya bisa terjadi kerusakan ginjal. 2. LANDASAN TEORI Batu Ginjal di dalam saluran kemih (kalkulus uriner) adalah massa keras seperti batu yang terbentuk di sepanjang saluran kemih dan bisa menyebabkan nyeri, pendarahan, penyumbatan aliran kemih atau infeksi. Batu ini bisa terbentuk di dalam ginjal (batu ginjal) maupun di dalam kandung kemih (batu kandung kemih). Proses pembentukan batu ini disebut urolitiasis (litiasis renalis, nefrolitiasis). 2.2 Sistem Pakar sistem pakar (expert system) adalah sistem yang berusaha mengadopsi pengetahuan manusia ke komputer, agar komputer dapat menyelesaikan masalah seperti yang biasa dilakukan oleh para ahli. Sistem pakar yang baik dirancang agar dapat menyelelasikan suatu permasalahan tertentu dengan meniru kerja dari para ahli. Dengan sistem pakar ini, orang awampun dapat menyelesaikan masalah yang cukup rumit yang sebenarnya hanya dapat diselesaikan dengan bantuan para ahli. Bagi para ahli, sistem pakar ini juga akan membantu aktivitasnya sebagai asisten yang sangat berpengalaman. Ada beberapa definisi tentang sistem pakar, antara lain : Ada beberapa definisi tentang sistem pakar, antara lain : 1. Menurut Durkin Sistem pakar adalah suatu program komputer yang dirancang untuk memodelkan kemampuan penyelesaian masalah yang dilakukan seorang pakar. Sistem Pakar Diagnosa Penyakit Batu Ginjal Menggunakan Teorema Bayes. Oleh : Intan Russari 1.4 Tujuan dan Manfaat Penelitian 1.4 Tujuan dan Manfaat Penelitian j Berdasarkan perumusan masalah sebagai tujuan dan manfaat penelitian adalah sebagai berikut: Berdasarkan perumusan masalah sebagai tujuan dan manfaat penelitian adalah sebagai berikut: 1.4.1 Tujuan Penelitian Adapun tujuan dari penelitian ini adalah: Adapun tujuan dari penelitian ini adalah: 1. Untuk mengetahui gejala-gejala tentang penyakit batu ginjal. 2. Untuk mengimplementasikan metode Bayes dalam mendiagnosa penyakit batu ginjal. 2. Untuk mengimplementasikan metode Bayes dalam mendiagnosa penyakit batu ginjal. g p y g j 3. Untuk merancang sebuah aplikasi sistem pakar yang dapat mendiagnosa penyakit batu ginjal. 18 Jurnal Riset Komputer (JURIKOM), Volume : 3, Nomor: 1, Februari 2016 ISSN : 2407-389X 1.4.2 Manfaat Adapun manfaat dari penelitian ini adalah: beroperasi jika sistem tersebut sudah lengkap. Eksekusi dilakukan langkah demi langkah menggunakan data tujuan utamanya adalah efisiensi. 2. Sistem Pakar 1. Agar dapat digunakan untuk para penderita yang tidak mampu untuk mengkonsultasikan penyakitnya dengan para ahli dikarenakan biaya yang mahal Basis pengetahuan merupakan bagian dari mekanisme inferensi penjelasan adalah bagian terpenting dari sistem pakar. Pengubahan aturan dapat dilaksanakan dengan mudah, sistem dapat beroperasi hanya dengan beberapa aturan.Eksekusi dilakukan pada keseluruhan basis pengetahuan menggunakan pengetahuan tujuan utamanya adalah efektivitas. 2. Agar penderita mengetahui penyakit yang dialami dengan adanya gejala yang dirasakan 3. Agar penderita mengetahui tingkat keparahan penyakit batu ginjal dengan adanya perhitungan dari keseluruhan gejala 3.2 Penerapan Teorema Bayes Pada Diagnosa Batu Ginjal p ( ) p evidence E, jika hipotesis Hi benar Sistem pakar adalah sebuah sistem yang menggunakan pengetahuan manusia di mana pengetahuan tersebut dimasukkan ke dalam sebuah komputer dan kemudian digunakan untuk menyelesaikan masalah-masalah yang biasanya membutuhkan kepakaran atau keahlian manusia. (menurut hasil sebelumnya) tanpa memandang evidence apapun. mungkin. mungkin. Output: P(Hi\E) Proses: 𝑃 𝐻1 = 𝑃 𝐻1 1 𝑘=1 P 𝐻𝑖 = 𝑃 𝐸\𝐻𝑖 ∗P(Hi) n k=1 P E\Hi ∗P(Hi) l t i Output: P(Hi\E) Proses: 𝑃 𝐻1 = 𝑃 𝐻1 1 𝑘=1 P 𝐻𝑖 = 𝑃 𝐸\𝐻𝑖 ∗P(Hi) n k=1 P E\Hi ∗P(Hi) l t i Setiap rule terdiri dari dua bagian, yaitu bagian IF disebut evidence (fakta-fakta) dan bagian THEN disebut Hipotesis atau kesimpulan. Metode Find-S tidak dapat digunakan untuk data yang tidak konsisten dan data yang bias, sehingga untuk bentuk data semacam ini salah satu metode sederhana yang dapat digunakan adalah metode bayes. Metode bayes merupakan metode yang baik di dalam mesin pembelajaran berdasarkan data training, dengan 3.1 Analisa Diagnosa Penyakit Batu Ginjal 3.1 Analisa Diagnosa Penyakit Batu Ginjal p p P(E) = Probabilitas evidence E p p P(E) = Probabilitas evidence E Gejala akibat batu ginjal baru bisa terasa jika batu berukuran lebih besar dari diameter saluran ureter. Batu yang besar akan bergesekan dengan lapisan dinding uterer sehingga menyebabkan iritasi dan bahkan luka. Oleh sebab itu, urin kadang bisa mengandung darah. Selain mengiritasi ureter, batu ginjal juga bisa tersangkut di dalam ureter sehingga urin tidak bisa keluar dan menumpuk di ginjal. ( ) Tabel 3 Terminologi Kepastian Bayes Terminologi A/B Tidak 0 Tidak tahu 0,2 Mungkin 0,4 Kemungkinan Besar 0,6 Hampir Pasti 0,8 Pasti 1,0 Adapun gejala dari batu ginjal adalah: Tabel 2 : Gejala Penyakit Batu Ginjal Kode Gejala G1 Nyeri yang parah pada sisi tubuh atau punggung, di bawah pinggul G2 Urin berwarna pink, merah atau coklat G3 Nyeri pada saat buang air kecil G4 Nyeri yang menyebar ke bagian bawah tubuh dan pangkal paha G5 Mual dan muntah G6 Demam dan menggigil saat infeksi terjadi G7 Mengalami kesulitan saat hendak buang air kecil G8 Sering buang air kecil G9 Air seni bercampur dengan darah ketika buang air kecil 1. Sistem Konvensional Informasi dan pemrosesannya biasanya jadi satu dengan program biasanya tidak bisa menjelaskan mengapa suatu input data itu dibutuhkan, atau bagaimana output itu diperoleh. Pengubahan program cukup sulit & membosankan sistem hanya akan P(H \| E) = Probabilitas hipotesis H benar jika diberikan evidence E P(E \| H) = Probabilitas munculnya evidence E, jika diketahui hipotesis H benar. 19 ISSN : 2407-389X Jurnal Riset Komputer (JURIKOM), Volume : 3, Nomor: 1, Februari 2016 P(H) = Probabilitas hipotesis H (menurut hasil sebelumnya) tanpa memandang evidence apapun. P(H) = Probabilitas hipotesis H (menurut hasil sebelumnya) tanpa memandang evidence apapun. menggunakan probabilitas bersyarat sebagai dasarnya, Machin Learning team (2006, 2). Probabilitas bayes merupakan salah satu cara untuk mengatasi ketidakpastian data dengan cara menggunakan formula bayes yang dinyatakan dengan:, Sri Winiarti (2003, 191). menggunakan probabilitas bersyarat sebagai dasarnya, Machin Learning team (2006, 2). Probabilitas bayes merupakan salah satu cara untuk mengatasi ketidakpastian data dengan cara menggunakan formula bayes yang dinyatakan dengan:, Sri Winiarti (2003, 191). P(E) = Probabilitas evidence E Tabel 1 Keterangan nilai Terminologi kepastian pada gejala penyakit Tabel 1 Keterangan nilai Terminologi kepastian pada gejala penyakit ( ) P (H \ E) = 𝑃 𝐸\𝐻 𝑃 𝐻 ∗𝑃 𝐻 𝑃(𝐸) …… Bayes Terminologi A/B Tidak 0 Tidak ada 0,2 Mungkin 0,4 Kemungkinan besar 0,6 Hampir pasti 0,8 Pasti 1,0 ……( 1) Dimana: P(H \| E) = Probabilitas hipotesis H benar jika diberikan evidence E P(E \| H) = Probabilitas munculnya evidence E, jika diketahui hipotesis H benar. P(H) = Probabilitas hipotesis H (menurut hasil sebelumnya) tanpa memandang evidence apapun. 4.1 Algoritma Algoritma adalah kumpulan perintah- perintah yang menentukan operasi-operasi tertentu yang diperlukan untuk menyelesaikan suatu masalah atau mengerjakan suatu tugas tertentu. Perintah- perintah ini dapat diterjemahkan secara bertahap dari awal hingga akhir. Algoritma dibuat sebelum melakukan perancangan perangkat lunak, hal itu dilakukan untuk memudahkan dalam merancang suatu perangkat lunak, karena algoritma dibuat dalam bahasa yang dimengerti oleh manusia yang nantinya akan diimplementasi pada perangkat keras. p p p g Algorima teorema bayes pada sistem pakar diagnosa penyakit batu ginjal dapat lihat dibawah ini: g j p w Algoritma Nilai P(Hi\E) = probabilitas hipotesis Hi, benar jika diberikan evindance E. Input : P(E/Hi) = probabilitas munculnya evidence E, jika hipotesis Hi benar P(Hi) = probabilitas hipotesis H (menurut hasil sebelumnya) tanpa memandang evidence apapun. N = nilai hipotesis yang mungkin. Output: P(Hi\E) Proses: 𝑃 𝐻1 = 𝑃 𝐻1 1 𝑘=1 P 𝐻𝑖 = 𝑃 𝐸\𝐻𝑖 ∗P(Hi) n k=1 P E\Hi ∗P(Hi) plementasi g j p Algoritma Nilai P(Hi\E) = probabilitas hipotesis Hi, benar jika diberikan evindance E. Input : P(E/Hi) = probabilitas munculnya evidence E, jika hipotesis Hi benar P(Hi) = probabilitas hipotesis H (menurut hasil sebelumnya) tanpa memandang evidence apapun. N = nilai hipotesis yang mungkin. Output: P(Hi\E) Proses: 𝑃 𝐻1 = 𝑃 𝐻1 1 𝑘=1 P 𝐻𝑖 = 𝑃 𝐸\𝐻𝑖 ∗P(Hi) n k=1 P E\Hi ∗P(Hi) plementasi Algoritma Nilai P(Hi\E) = probabilitas hipotesis Hi, benar jika diberikan evindance E. Input : P(E/Hi) = probabilitas munculnya evidence E, jika hipotesis Hi benar P(Hi) = probabilitas hipotesis H Sistem Pakar Diagnosa Penyakit Batu Ginjal Menggunakan Teorema Bayes. Oleh : Intan Russari 3. ANALISA DAN PERANCANGAN 3. ANALISA DAN PERANCANGAN 4.2. Implementasi p Tahap selanjutnya setelah perancangan adalah tahap implementasi program. Pada tahap implementasi, rancangan form yang telah dibuat 20 ISSN : 2407-389X Jurnal Riset Komputer (JURIKOM), Volume : 3, Nomor: 1, Februari 2016 kemudian diaplikasikan dengan menggunakan Microsoft Visual Basic 2008. Implementasi Sistem dari perangkat lunak ini mencakup spesifikasi kebutuhan perangkat keras (hardware) dan perangkat lunak (software) serta pengujian perangkat lunak. Gambar 3 Tampilan Form Data Pasien Form Konsultasi 4. Form Konsultasi Form konsultasi digunakan untuk melakukan konsultasi diagnosa penyakit batu ginjal pada pasien. Pada saat pertama sekali form ditampilkan akan tampil gejala yang dialami. Tekan tombol “Ya” jika gejala jika gejala sama dengan yang dialami, dan tombol “Tidak” jika gejala tidak sama dengan yang dialami. Gambar 1 Tampilan Form Login Gambar 4. Tampilan Form Konsultasi 2. Form Menu Utama Menu utama sebagai user interface yang muncul setelah menu login pada saat aplikasi dijalankan. Gambar 2 Tampilan Form Menu Utama Gambar 4. Tampilan Form Konsultasi 4.2.2 Tampilan Output Perangkat Lunak Tampilan output dari perangkat lunak ini adalah sebagai berikut: 1. Form Login Form Login digunakan untuk mengakses program sehingga user yang berhak saja yang dapat mengakses ini. Gambar 1 Tampilan Form Login Gambar 3 Tampilan Form Data Pasien 5.2 Saran Berdasarkan kesimpulan diatas, maka dapat dikemukakan saran untuk mengembangkan sekripsi ini lebih lanjut yaitu sebagai berikut: 1. Diharapkan sistem ini dapat dikembangkan lagi, tidak hanya mendiagnosa satu jenis batu ginjal saja ataupun tidak hanya mendiagnosa bisa juga dengan cara pengobatan dan pencegahan. 2. Sistem ini dapat dikembangkan lagi oleh pembaca dengan mendiagnosa penyakit batu ginjal menggunakan sistem yang lebih lengkap lagi untuk dimasa yang akan datang. 3. Program sistem pakar ini perlu dilakukan evaluasi tidak hanya dengan visual basic Net 2008 saja, misalnya dengan berbasis web ataupun berbasis android karena ilmu pengetahuan dan teknologi yang terus berkambang sehingga dapat perlu adanya perbaikan atau penyempurnaan kembali. 5. Form Hasil Form ini digunakan untuk menampilkan hasil dari konsultasi diagnosa penyakit batu ginjal. Gambar 5 Tampilan Form Hasil 3. Form Data Pasien Form ini digunakan untuk menginput data pasien ataupun pengguna aplikasi untuk dapat melakukan proses diagnosa. Berdasarkan dilakukan selama Berdasarkan pengembangan yang telah dilakukan selama proses parancangan hingga Sistem Pakar Diagnosa Penyakit Batu Ginjal Menggunakan Teorema Bayes. Oleh : Intan Russari 21 ISSN : 2407-389X Jurnal Riset Komputer (JURIKOM), Volume : 3, Nomor: 1, Februari 2016 implementasi sistem pakar untuk diagnosa penyakit batu ginjal dengan menggunakan Teorema Bayes, maka dapat diambil kesimpulan sebagai berikut : 1. Diagnosa penyakit batu ginjal dengan sistem pakar melihat dari gejala-gejala yang dirasakan oleh pasien proses mengetahui penyakit batu ginjal dapat dilakukan dengan cepat dan akurat.. 2. Dengan menerapkan teorema bayes dalam mendiagnosa batu ginjal dapat menghasilkan perhitungan valid yang sama dengan perhitungan manual sehingga proses prediksi dapat dilakukan dengan cepat dan akurat. 3. Dalam perancangan aplikasi ini menggunakan bahasa pemrograman Visual Basic 2008 dan dengan MySQL sebagai pengolah data. Sistem Pakar Diagnosa Penyakit Batu Ginjal Menggunakan Teorema Bayes. Oleh : Intan Russari DAFTAR PUSTAKA 1. T. Sutojo, Edy Mulyanto, Dr. Vincent Suhartono., “Kecerdasan Buatan”, Penerbit Andy, Yogyakarta, 2011 2. Http//Komunitas_elearningilmukomputer /system_pakar_dasar.com, diakses 05 Maret 2015 3. Http://id.wikipedia.org/wiki/batu_ginjal, diakses 05 Maret 2015 4. Sri Winiarti., “Pemanfaatan Teorema Bayes Dalam Penentuan Penyakit Tht”, jurnal, 2008 y j 5. Jogiyanto HM., “Analisa dan desain”, Andi, Yogyakarta, 2005, 6. Http://id.wikipedia.org/wiki/MySQL, diakses 05 Maret 2015 7. Primananda Arif Aditya, S.Si, M.M, 2013 22
W3006335412.txt	https://ieeexplore.ieee.org/ielx7/6287639/8948470/08998290.pdf	en		Multiple Black Hole Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing	IEEE access	2,020	cc-by	8,421	Received February 1, 2020, accepted February 9, 2020, date of publication February 13, 2020, date of current version February 26, 2020. Digital Object Identifier 10.1109/ACCESS.2020.2973696 Multiple Black Hole Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing HAMSA NAJI NSAIF AL-SAMMARRAIE AND DAYANG N. A. JAWAWI Faculty of Engineering, School of Computing, Universiti Teknologi Malaysia (UTM), Johor Bahru 81310, Malaysia Corresponding author: Hamsa Naji Nsaif Al-Sammarraie (omeehams@gmail.com) This work was supported in part by the Ministry of Higher Education Malaysia (MOHE) for Fundamental Research Grant Scheme (FRGS) under Grant 5F117, and in part by the Universiti Teknologi Malaysia (UTM) for UTM-Transdiciplinary Research Grant (TDR) under Grant 06G23. ABSTRACT Combinatorial searching-based software testing (CSST) is a challenging optimization procedure. The achievement of optimal solutions involves a careful formulation of the optimization problem and the selection of an appropriate approach. Meta-heuristic searching procedures have proven to be effective for solving CSST issues. Black hole (BH) optimization is among the more recently developed meta-heuristic searching algorithms. While this approach has been observed to be an effective alternative to particle swarm optimization, its operation is based on only one swarm. To date, no efforts have been made to modify this approach to accommodate multiple swarms. This study proposes a new variant of BH that involves a combination of multiple swarms. The BH optimizer is modified from continuous searching to binary searching and subsequently applied for solving CSST. The evaluation is based on a modified-benchmarking mathematical function and well-known CSST problems. This modified BH method is superior to the original BH and the established particle swarm optimization (PSO) approach. In terms of CSST problems, binary multiple black hole (BMBH) optimizations generate reduction rates between 50% and more than 60% for t = 4 according to the problem. INDEX TERMS Black hole optimization, combinatorial searching-based software testing, meta-heuristic searching, multiple black hole optimization, swarm meta-heuristic. I. INTRODUCTION Currently, the emergence of a variety of complicated systems in software products is increasing. Consequently, the development of an effective process for assessing the quality of these systems has proven to be an arduous task. A typical way of testing any system is to consider the possible values of input interactions to generate test suites that are independent of each other and can be associated with different faults. According to our observations, the software developed over the last two decades has had an emphasis on customizability to user needs. This makes the configuration of such software an important issue [1]. Software testing entails the coverage of all possible values of input or configurations and their interactions. However, given the current speed of hardware, the coverage of all possible variable interactions is not practical. In a system with 9 input variables, if these The associate editor coordinating the review of this manuscript and approving it for publication was Mostafa M. Fouda 33406 . inputs are indicated as parameters and each parameter has 5 values, the coverage of all the possible combinations of input variables in the traditional way requires testing of 59 = 1, 953, 125 combinations overall. This would require an exceedingly high level of hardware power [2]. Combinatorial testing (CT) is a subtopic in software engineering for testing purposes. The goal of CT is the conversion of an original space of software variables into a reduced space. This reduced space is considered the input for the testing process to detect faults. This method facilitates the coverage of variable interactions to generate possible faults while excluding superfluous operations [3]. In the context of software generation, automatic generation of a reduced space is crucial in regard to cost-effectiveness, time savings, and quality. In a review [4], CT based on searching approaches was termed search-based software testing (SBST). In addressing the various challenges associated with SBST, it was orated that ‘‘There exists a structured parallel approach for test data generation, but an idea of using search together with parallel This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing islands has not been explored with branch selection’’ [4]. Existing search algorithms need to be upgraded before these methods can be applied to parallel-based searching. The relevant literature documented numerous approaches related to meta-heuristic searching for optimization. Each approach is based on a certain concept or metaphor. Some approaches were inspired by genetic evolution and their suitability with regard to the environment, as in [5] and [6], while others were inspired by social aspects [7]. The literature related to heuristic searching reported a wide range of applicable models. The mechanism and searching process vary from one approach to another, which is reflected in the outcome. In the research of artificial intelligence (AI), opinions vary on the superiority of one searching model over another. While some models generate superior results in certain applications, these methods perform less effectively in others [8]. It is essential that the evaluation of meta-heuristic searching be application-dependent. Some models are applicable for evolutionary searching, while others are more suitable for parallel searching. Researchers have recommended several models with a parallel search forte to overcome the problems associated with SBST. This study emphasized the upgrade of an existing meta-heuristic searching algorithm termed black hole (BH)-based optimization into a parallel-based searching algorithm. In this study, the upgraded meta-heuristic algorithm is named the multiple black hole (MBH) optimization method. The current article is separated into the following sections: Section II provides a theoretical background. Section III presents the t-way test suite generation problem. Next, section IV presents a literature survey. Then, section V gives the background and terms. Next, section VI presents the proposed algorithm and contributions followed by the proposed approach in section VII. Section VIII and section IX present the evaluation and experimental results and the corresponding analysis, respectively. Finally, section X presents the conclusion and future work. B. DEFINITION 2 The test case TC is an n-tuple (v1 , v2 , . . . vn ) where v1 ∈ V1 , v2 ∈ V2 , . . . vn ∈ Vn . If a fault is generated when testing the system based on TC, then the test case has served in the t-testing of the system. The examples provided in section III, such as the pizza ordering system, which includes vegetarian cheese, extra thick, ground beef, large, and take away, are regarded TCs. C. DEFINITION 3 All the possible test cases are denoted TC all , TC all = {(v1 , v2 , . . . vn )\|v1 ∈ V1 , v2 ∈ V2 , . . . vn ∈ Vn }. In the examples provided in section III, the encoded possible values with integer levels = 0, 1, 2. . . are portrayed in Table 2. Table 3 displays TC all . D. DEFINITION 4 The covering array (CA) is adopted as a mathematical object for the description of the generated t-way test suit. Generally, an SUT is composed of multiple parameters that cross-interact with their associated values. In this research, n denotes the number of parameters. The associated levels are denoted by p, while t is the interaction strength. When all the parameters n have the same number of values (v), the CA is represented as CA (N; t,vn ). Otherwise, when the numbers of values are not similar for all the parameters, CA can be nj n2 n3 represented as MCA (N, t, vn1 1 , v2 , v3 . . . , vj ). E. DEFINITION 4 An orthogonal array (OA) with strength t is defined as OA(N ; t; n; (a1 , a2 , . . . an )), where the size of the array is N × n and the array fulfills the two following conditions: • Each column of OA with an index i contains all possible elements from the set Vi , where ai = \|Vi \|. • Any subset of t columns covers all possible t-tuples N . exactly a1 ×a2 ×,...×a n II. THEORETICAL BACKGROUND III. T-WAY TEST SUITE GENERATION PROBLEM Several important preliminaries are necessary to explain the data and theory presented in this article, with the following definitions: A. DEFINITION 1 There are three examples from different types of applications: the first example is the service application known as the pizza ordering system case study, the second example is the smart mobile system case study, and the third example is a heart disease case study. A software under test (SUT) is a software or application that is tested based on the possible values of its parameters by assuming that the SUT has n parameters ci (i = 1, 2 . . . n); the parameters can represent a possible input, certain event (internal or external), or configuration variable, as illustrated in Example 1 in section III. The SUT has n variables V1 , V2 , . . . Vn that take any possible value such that v1 ∈ V1 , v2 ∈ V2 . . . vn ∈ Vn . Based on this approach, the application of combinatorial searching-based software testing (CSST) is appropriate for solving the unlimited array of problems associated with the software engineering industry. This example is found in article [9] as well as in other relevant studies and is summarized in Table 1. The pizza ordering system comes with five variables: pizza type, crust, toppings, size, and delivery. Three variables take one of two values (pizza type, crust, and delivery), while two variables take one of three values (toppings and size). Therefore, the covering array combined with the number of cases is equal to 32 × 23 . Table 2 shows encoding for the covering array for the pizza ordering system case study. Additionally, Table 3 shows the dataset for the pizza ordering system case study. VOLUME 8, 2020 A. EXAMPLE 1 33407 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 1. The pizza ordering system case study. TABLE 2. Encoding of possible values of input to the pizza ordering system case study. FIGURE 1. Smart mobile system case study. TABLE 4. Heart disease predictions case study. TABLE 3. The dataset extracted from the Pizza ordering system case study. From these parameters, the possible configurations that need to be tested are 210 × 18 = 1024. C. EXAMPLE 3 The heart disease case study [11] for predictions comprises seven variables: gender (2 levels: male/female), age (four levels: younger than 20, between 20 and 40, between 40 and 60, and older than 60), morning, evening, and sleeping heartbeat rates (three levels: lower than 60, between 60 and 100, and more than 100), and blood test (three levels: lower than 10, between 10 and 15, and more than 15). The full extent of the problem is 22 × 4 × 34 = 1296. The states can be represented as in Table 4. B. EXAMPLE 2 The smart mobile system, this system, is also utilized as a case study where the selection is based on its capacity for offering a wide range of factors and levels [10], as explained in Fig. 1. This renders the smart mobile system favorable for software testing. The factors are correlated such that every factor has a relationship with several other factors. The smart mobile system consists of 18 features that are divided into two categories: one-valued and two-valued parameters. There are ten features in the two-valued parameter category (video call, voice messages, video messages, basic colors, high resolution, camera, video player, music player, radio and voice recorder) and eight features in the one-valued parameter category (smart mobile system, calls, messages, GPS, screen, media, voice call, and text messages). 33408 IV. LITERATURE SURVEY The literature survey is separated into two subsections, the literature of CSST and a review of meta-heuristic-based approaches focusing on the parallel nature of some methods. A. CSST TESTING The literature on CSST is relatively extensive, embracing numerous approaches. However, all these approaches share a common factor: these approaches exploit the power of random combinatorial searching when they are integrated with heuristics for determining the t-strength covering array. The formulation of the optimization problem calls for an emphasis on two aspects: the devising of the objective function and the selection of the approach employing a pure-based VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing approach [12] or hybrid-based approach [13]. In terms of the objective function formulation, some models use a minimization formulation, while others use a maximization formulation. The fitness function is the total number of different pairs covered from all the test cases by an individual. If an individual covers a greater number of different pairs than others, that individual is superior to the others [14]. Thus, the formulation of the problem is a maximization problem, where an individual becomes a solution when it covers all pairs. In other work, the number of covered d-tuples was used as a fitness value for the candidate solution. On the other hand, the objective function was investigated using the number of non-covered interaction tuples covered by the candidate solution [15]. In another situation, the number of uncovered tuples was taken as the cost of the candidate solution, which required minimization. The cardinality of the set and the objective value were represented by the number of non-covered interaction tuples covered in [16]. In terms of the type of meta-heuristic algorithm applied, the study in [17] proposed a new variant of the teaching learningbased optimization (TLBO) algorithm. Their meta-heuristic algorithm comes with a Mamdani fuzzy inference system, which not only provides a solution to trapping in local optima but also provides greater diversity. In the undertaking of [18], the firefly meta-heuristic algorithm was applied as a test suite generator. The computational comparison between the firefly meta-heuristic algorithm, genetic algorithm, and ant colony optimization revealed that the firefly algorithm has a shorter optimization time. Other models developed for solving combinatorial software testing using meta-heuristic searching include the memetic algorithm (MA) in [19], the particle swarm optimization (PSO) in [20], the artificial bee colony (ABC) and corresponding developed variants in [13], [21]–[26], the harmony search algorithm (HSA) in [27], the bat algorithm (BA) in [12], [28]–[30], and the flower pollination algorithm (FPA) in [31] and [32]. All these models are considered pure models because they utilize a common pure meta-heuristic approach. Each of the metaheuristic searching-based approaches uses a representation related to the metaphor of the approach. For example, in the work of [21], ABC was used for t-way generation. The representation these researchers used presents the number of food sources by a test case and the fitness value by the coverage. The food source is optimized by the employed bees and presented to the onlooker bees. The food sources are selected based on the probability calculated using their fitness, while the onlooker bees investigate the solutions and select the best. On the other hand, analyzing the literature of meta-heuristic-based optimization reveals that there are differences in the approaches with respect to the parameters. This aspect has been discussed in the work of [25], where various approach parameters were presented. For example, PSO uses the maximum iteration, swarm size, learning factors and inertia factor. ABC uses the maximum iteration, the number of colony sizes, the number of food source limits, and the number of cycles. VOLUME 8, 2020 In terms of hybrid-based approaches, several researchers integrated meta-heuristic algorithms with other models. For example, [33] integrated the particle swarm optimization approach with fuzzy logic for the tuning of w,c1 , and c2 . The authors built three fuzzy inference systems to monitor the PSO performance and to adjust w, c1 , and c2 for optimization improvement efficiency. The role of PSO is as an exploitation agent for ABC. ABC takes the information from PSO via the weight factor. This algorithm is an extended variant of [21]. B. PARALLEL META-HEURISTIC SEARCHING Parallel meta-heuristic searching refers to methods that come with a parallel searching quality. Researchers have made efforts to improve the exploration qualities of existing meta-heuristic searching algorithms by converting them into parallel searching algorithms. The study in [34] developed the parallel comprehensive learning particle swarm optimizer (PCLPSO). This optimizer, which comes with multiple swarms based on the master-slave paradigm, operates both cooperatively and concurrently. The simulations in [34] fashioned a multi-swarm optimizer for multi-objective optimization. The authors used a hybrid strategy decomposition and dominance (MSMO/2D) to improve convergence and diversity by splitting the primary swarm into several sub-swarms. Similarly, [29] devised a multi-swarm algorithm, called the multi-swarm bat algorithm (MBA), for global optimization. This algorithm recommends the exchange of information between different swarms. Rule hiding is included in several applications using the MBA [35]. The study in [13] states that the strength t was increased from t = 2 to t = 6 through the development of a hybrid artificial bee colony (HABC) strategy by hybridizing an ABC algorithm and a PSO algorithm. The observed performance of this family of parallel searching approaches and the good performance of black hole optimization motivate us to develop a multi-variant of this approach for solving the problem of t-way testing. V. BACKGROUND AND TERMS Scientists define a black hole (BH) as a part of the universe where a great gravitational force attracts and subsequently swallows anything (including light). The attraction of anything moving around the black hole occurs in a region with a certain radius. This region is known as the event horizon (EH). Light that crosses the EH inevitably disappears into the black hole for good. This has inspired researchers, such as in [36], to apply this concept as a metaphor for a point in the solution space that has a better objective value than its surroundings. This serves to remove all nearby solutions within a given radius. Researchers associate this point in the solution space with a BH and its surrounding region with the EH. The pseudocode of the BH algorithm is provided in Fig. 2. The sequence for this algorithm is as follows: (a) random initialization of the solutions in the solution space, (b) evaluation of each solution using the objective function, (c) selection of the solution corresponding to the minimum cost value to 33409 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 5. The symbols of the MBH algorithm and their definitions. convergence of the fitness value). The algorithm is finished when the stopping criterion is met. The terms, notations, and solutions used in the methodology are provided in Table 5. The solution space is defined as dimension n. Any solution is presented as star xs , while the black hole is presented as xBHi . Prior to operating the algorithm, it is essential that NBH , which represents the number of black holes, be defined. The radius of the EH for each BH is denoted as R (i). VI. PROPOSED ALGORITHM AND CONTRIBUTION FIGURE 2. The pseudocode of the black hole algorithm. represent the BH, (d) shifting of all the solutions towards the BH, (e) a fresh evaluation of the population to update the BH with the new solutions corresponding to the minimum cost value, (f) calculation of the radius of the BH and replacement of the solutions that cross the radius with new solutions, (g) a fresh evaluation to update the BH with the new solution corresponding to the minimum cost value, and last, (h) a check of the stopping criterion (the number of iterations or the 33410 Among the recently developed meta-heuristic searching algorithms is the black hole (BH) algorithm [37]. This algorithm is based on the black hole phenomenon and the behavior of stars during their interaction with the BH. A star that gets too close to the black hole is swallowed by the BH. In the context of the BH algorithm, a new star is randomly generated to represent a new solution. This new star is included in the search. This study in [36] compared the black hole algorithm to other meta-heuristic optimization algorithms. Among them, the performance of PSO was observed to be exceptional. As presented in section III, our literature survey identified CSST as one of the computational problems that can be addressed through meta-heuristic searching. The computational difficulties associated with CSST call for the application of a powerful meta-heuristic optimization. Considering the substantial power of the multi-swarm-based optimization algorithm, the modification of the black hole optimization into a multi-black hole with multiple swarms of stars delivers improved optimization results. Combinatorial t-way testing is one of the most well-known optimization problems solvable through meta-heuristic optimization. A multi-swarm variant of the BH searching optimizer is VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing successfully developed, called the multiple black hole (MBH) optimization, and is applied to the problem of CSST. Based on the problem statement stated, the following contributions are presented. • A novel variant of the black hole optimization approach based on the multi-swarm concept is proposed that can be characterized as a variant multiple black hole or MBH optimization. This is supported by introducing the concept of the black hole energy to facilitate the elimination of certain black hole swarms and to generate fresh swarms. • A binary variant of the MBH optimization algorithm is presented and given the name of the binary multiple black hole (BMBH) optimization algorithm. The purpose of this variant is to solve the combinatory problems of CSST. • A benchmarking mathematical function is proposed to provide a comprehensive comparison between PSO, the classical BH, and the BH variant MBH. • A well-known benchmarking problem of CSST documented in the relevant literature is applied to evaluate BMBH in conjunction with the binary original variant BH and PSO. VII. THE PROPOSED APPROACH The purpose of this section is to describe the methodology employed to realize the objectives of this investigation, which is a new variant of the BH algorithm named the multiple black hole (MBH) algorithm. Another algorithm, a binary variant of the MBH algorithm named the binary multiple black hole (BMBH) algorithm, is also employed. The performance of the BMBH algorithm in solving CSST is demonstrated. A. MBH ALGORITHM In this section, the focus is on an explanation with regard to the multiple black hole optimization algorithm. This algorithm begins with the creation of NBH populations and the selection of a BH for each population. The black hole xBHi of population Pi is the solution that achieves the minimum objective value within the stars of the population. The movement of the stars of each BH towards the BH leads to an update of the BH by the star that achieves the lowest value for the objective function. Each update of the BH corresponds to an increase in the energy of the BH by one. The higher the energy of a particular BH is, the greater the potential for the future generation of stars by that BH. In this study, the radius of the BH and the elimination of the solutions that crossed that radius are developed. The eliminated stars or solutions are then replaced at a probability according to the existing energy within the black hole. A BH with less energy has fewer stars. Such a BH meets the omitting conditions or is removed. 1) PSEUDOCODE OF MBH The pseudocode for MBH is illustrated in Fig. 3. As portrayed, the difference between BH and MBH has to do with VOLUME 8, 2020 FIGURE 3. The pseudocode of the MBH algorithm. the latter’s acceptance of a predefined number of BHs, rather than a solitary BH. The algorithm performs the search in parallel for NBH equal to the number of black holes. Another concept introduced is the energy of the BH at the beginning of the search, or at the birth of the BH, when the BH has the maximum energy. With each occurrence of non-improvement in the best solution in the black hole, the energy is decreased by one until the energy of the BH is negative. At this point, the BH is replaced by a fresh BH. This concept facilitates simultaneous searching in multiple BHs to enhance the probability of attaining better solutions. The last difference is that 33411 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing MBH enables smarter and more dynamic searching than BH, which is dependent on the initialization of its parameters because of the static nature due to the lack of elimination and regeneration of swarms. 2) COMPARISON BETWEEN MBH AND BH There are several differences between the BH optimization and the newly developed MBH optimization. While the BH optimization involves the use of a single BH during its search in the solution space, the MBH optimization involves the use of the NBH black hole for the same purpose. Additionally, the MBH optimization algorithm includes the concept of the energy of the BH. This concept serves to eliminate non-active BHs, as well as BHs that do not show improvement with time. Improvement means the capacity of the algorithm to search for better fitness values for the solution, representing the BH. In comparison to the BH optimization algorithm, the MBH optimization algorithm is more effective for searches in the solution space. Additionally, the MBH optimization algorithm can be relied on to achieve better results with fewer iterations than required by the BH optimization algorithm. In both the MBH and BH optimizations, the movement of stars beyond the permitted region of search (violation of the constraint) results in their clipping. This renders their searching prowess less effective. The MBH optimization algorithm overcomes this dilemma through the addition of random components after being subjected to clipping. Assuming Ns is the number of solutions, It is the number of iterations, and NmaxBH is the maximum number of black holes, then the complexity of BMBH for the worst case scenario is O(It.Ns NmaxBH ), while the complexity of the BBH optimization is O(It.Ns ). B. BINARY VARIANT OF BOTH THE BH AND MBH ALGORITHMS The equation defining the movement of stars towards the black hole, rand, is a random number between 0 and 1, and the movement equation for the stars is as follows: where j indicates the index of the black hole and (1) The setback accompanying (1) is its ineffectiveness with regard to binary values because of the addition of a second term to a binary value, causing a breaching of the binary constraint, which is 0 or 1. To overcome this problem, a parameter named the pulling rate pr is introduced. This approach generates a random number rd between 0 and 1, where d denotes the index of the dimension. The value of rd is compared to that of pr, and the value of xi (t + 1)d is defined in compliance with (2) below: BH (t)d if r < pr d xi (t + 1) = (2) xi (t)d otherwise A small value for the pulling rate pr is preferred for a diligent search in the solution space. The following example 33412 provides an explanation of the performance of (2). Assuming one star has the value xt = [1 0 0 1], while the black hole BHt = [0 0 0 0] and pr = 0.2, we generate one random number rd for each element of xt . Assuming that r = [r1 r2 r3 r4 ] = [0.1 0.2 0.04 0.6], then based on (2), the value of xi (t + 1) = between the star and the [0 0 0 1]. While the initial distance √ black hole is observed to be ( 2), the movement of the star in compliance with (2) alters this distance to 1. The similarity of the BMBH and BBH concepts allows for the use of the same equation to move the stars. The addition of another subscript to the equation serves to indicate the index of the BH. Subsequently, this generates a random number rd between 0 and 1, where d denotes the index of the dimension. Then, d d j j we identify its value and alter the value of xi (t) to xi (t + 1) in compliance with (3) below: ( BH j (t) if rd < pr d j xi (t + 1) = j d (3) xi (t) otherwise j = 1, 2, . . . NBH xi (t + 1) = xi (t) + rand × (xBH − xi (t)) , i = 1, 2 . . . N FIGURE 4. The t-way optimization framework using the black hole algorithm. C. CSST BASED ON BMBH To verify the covering array of strength t, with the entire number of test cases, a function designed for this purpose is employed. This function, which verifies the validity of the covering array, returns a Boolean variable with a value of 0 if the covering condition for a certain strength is not met and with a value of 1 if the covering condition for a certain strength is met. Known as the cover CheckGR, the integration of this function with the BBH and BMBH algorithms is portrayed in Fig. 4. As can be observed, the searching algorithm generates a random candidate solution and delivers the solution to the objective function. The objective function employs cover CheckGR, which determines if the array has a covering property with strength t. Then, if the covering property strength is met, the number of rows of the covering array VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 6. The ranges of the parameters and the effect of each parameter. represents a reflection of the fitness values. The objective of this exercise is the minimization of the number of rows. In a circumstance where the covering property strength is not met, the fitness value is given as infinity. This implies that the solution is no longer included in iterations for the generation of new solutions. D. PARAMETER TUNING The goal of parameter tuning is to determine the best values of parameters for the algorithm. To determine the best values for the parameters, it is necessary to provide the effect of each of the parameters. This information is given in Table 6. The values that are used for the problem of t-way testing are selected based on a tuning process that balances exploration and execution time. The algorithm selects an average value of 3 for NBH and 250 for IterationsMax for adequate results. VIII. EVALUATION AND EXPERIMENTAL RESULTS To evaluate our proposed MBH and BMBH, we use MathWorks R MATLAB 2019b and code the PSO, BPSO, BH, MBH, BBH, and BMBH approaches with the evaluating objective functions. The experiments are conducted on a 1.4 GH Intel i5 processor. The evaluation process is separated into two sections. First, the MBH optimization is assessed, and its performance is compared with that of the original BH optimization and PSO. Benchmarking mathematical functions are employed for this first stage. In the second section, the BMBH optimization is assessed, and its performance is compared with that of BBH and BPSO. Several CSST problems reported in the relevant literature are referred to for this second stage. A. BENCHMARKING MATHEMATICAL FUNCTIONS To gauge the performance of the MBH algorithm, the algorithm is compared with the BH and PSO algorithms with respect to the capacity for determining the optimal point of the provided benchmark optimization functions. VOLUME 8, 2020 The formulas of the mathematical functions, which cover the true optimal, dimension, searching range, and setting of the algorithms in terms of the number of solutions and number of iterations, are presented in Table 7. The MBH optimization converges closer to the optimal value than either BH or PSO. Furthermore, the BH optimization is better than PSO for most functions and similar for some. This comparison emphasizes the superiority of the MBH optimization over both the BH optimization and PSO. B. BENCHMARKING CASE STUDIES BASED ON CSST Three case studies (i.e., the pizza ordering system, heart disease system, and smart mobile system) are used to evaluate the performance of the proposed BMBH and BBH and compare them with that of the BPSO. Each case study has different features and aspects of industry. The pizza ordering system considers the user as a client of restaurants. The heart disease system considers the health care center to be the user. The smart mobile system offers a wide range of factors and levels, which makes this system suitable for software testing. From a t-way testing point of view, all the case studies have different sizes of covering array, numbers of variables and levels. The results for the three case studies are generated based on CSST. Table 8 shows the features of each case study. The results of comparisons between the BMBH optimization, BBH optimization and BPSO for the three literature case studies in Examples 1, 2, and 3 are presented. The algorithms are tested with 250 solutions and iterations. The same values that resulted from the parameter tuning phase are used. The resulting numbers are used for expediting experimental operations, as well as for increasing the number of solutions and iterations to generate better results. For the pizza ordering system problem, the full size of the problem is CA (23 × 32 = 72) cases. In a comparison experiment involving BPSO, BBH optimization, and BMBH optimization, for the three values of t = 2, t = 3, and t = 4, both BBH and BMBH are superior to BPSO in this 33413 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 7. The benchmark mathematical optimization functions and their optimal values. TABLE 8. Features of case studies. area. BMBH achieves a lower cost value for t = 2 with a reduction in the full size from 72 to 22 and for t = 3 with a reduction rate in the full size from 72 to 35. Furthermore, as illustrated in Table 9, the greatest reduction rate of 68% for t = 2 is recorded by BMBH because this algorithm reduced the original size of 72 to 22. In Fig. 5, for the pizza ordering system problem, a boxplot diagram based on 20 experiments performed for t = 3 shows that the BMBH optimization is the only method that 33414 can effectively reduce the full size of the case study. This is portrayed in Fig. 5. Based on the figure, the BBH and BMBH results have similar minimum cost tabulations, but the BMBH results are better in terms of the means. However, cross-referencing Fig. 5 with Table 9 shows that the results and the standard deviations of both algorithms are similar. This is due to the multi-swarm nature of BMBH. Another factor is the BMBH performance; the energy variable of the swarm enables the algorithm to replace BHs that do not produce improvement with other black holes. For smart mobile systems, the full size of the problem is CA(210 × 18 = 1024) cases. The BMBH optimization reduces the full problem size from 1024 to 400 for t = 2 and t = 3, while the BBH optimization reduces the full problem size from 1024 to 485 for t = 2 and t = 3. However, although BPSO reduces the full problem size from 1024 to 1002, this is only for t = 2 and t = 3. Table 10 displays the reduction rate of each algorithm with respect to the full size of the problem. This table makes clear that the highest reduction rate is achieved by BMBH optimization, which is 61% for t = 2 and t = 3. Additionally, according to the boxplot diagram for the smart mobile system displayed in Fig. 6, the BMBH optimization is better than the rest with regard to reducing the problem size. This optimization exhibited a lower standard deviation, as well as fewer rows, compared to BPSO and the VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 9. Pizza ordering system problem, BMBH vs. BBH vs. BPSO. FIGURE 6. Smart mobile system problem boxplot diagram for t = 3. FIGURE 5. Pizza ordering system problem boxplot diagram for t = 3. BBH optimization. Based on the figure, the BBH and BMBH results have similar minimum cost tabulations, but the BMBH results have better mean and standard deviation results. Additionally, cross-referencing Fig. 6 with Table 10 shows that BMBH has superior results. Again, this is due to the multi-swarm nature of BMBH. Another factor is the BMBH performance; the energy variable of the swarm provides high exploration power in the searching space. For the heart disease prediction problem, the full size of the problem is CA (22 ×4×34 = 1296). The BMBH optimization reduces the full problem size from 1296 to 587 for t = 2 and t = 3, while the BBH optimization reduces the full problem size from 1296 to 610 for t = 2 and t = 3. However, while BPSO reduces the full problem size from 1296 to 883, this is only for t = 2 and t = 3. Table 10 displays the reduction VOLUME 8, 2020 rate of each algorithm with respect to the full size of the problem. This table clearly shows that the highest reduction is associated with the BMBH optimization, with 54% for t = 2, t = 3 and t = 4. Additionally, the boxplot diagram for the heart disease problem shown in Fig. 7 reveals the BMBH optimization to be superior to the rest in terms of problem size reduction. This optimization exhibited a lower standard deviation, as well as fewer rows, in comparison to BPSO and the BBH optimization. Additionally, cross-referencing Fig. 7 with Table 11 shows that BMBH has a better mean and standard deviation. Again, the superiority of this algorithm comes from the multi-swarm nature of BMBH. Another factor is BMBH performance; the energy variable of the swarm provides powerful searching. IX. RESULT ANALYSIS The results reveal that BMBH is able to provide more optimal and consistent performance than either BBH or BPSO. The combinatorial nature of the problem and the large number 33415 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing TABLE 10. Smart mobile system problem - BMBH vs. BBH vs. BPSO. TABLE 11. Heart disease problem - BMBH vs. BBH vs. BPSO. FIGURE 7. Heart disease system problem boxplot diagram for t = 3. of possible solutions implies a very large solution space. This makes searching difficult. In both BPSO and BBH, the candidate solutions are selected randomly and change their position in the space according to the best solutions, represented by the leader in BPSO and the black hole in BBH. However, it is very possible that the quality of the whole 33416 swarm in both approaches is poor, including the leader in BPSO and the black hole in BBH. Hence, the search will not provide good results at the end. However, the incorporation of multi-swarm searching under multi-black holes in BMBH with the concept of the energy of the black holes provides the algorithm with the capability of replacing poor quality black holes with good ones. The indicator of good quality is the improvement in the black hole value over the progress of searching. Another observation is the consistency in the experimental results of BMBH compared with those of both BBH and BPSO. The boxplot standard deviation is lower in BMBH, which means that this algorithm has more consistency. This is facilitated by the capability of the algorithm to determine the quality of the black holes based on the concept of energy compared with BBH and BPSO, which suffer from sensitivity to the seed that provides the solution. The differences between the performances of BMBH, BBH and BPSO occur when the t value increases, which indicates a wider solution space and more complex optimization. X. CONCLUSION AND FUTURE WORK CSST represents a challenging optimization problem. This challenge was treated with the proposal of the application VOLUME 8, 2020 H. N. N. Al-Sammarraie, D. N. A. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing of the multiple swarm concept to existing swarm methods. To achieve this, the existing meta-heuristic searching optimization method known as the BH optimization as upgraded to a multiple swarm-based BH optimization method named the MBH optimization. In this paper, the application of MBH to CSST by the conversion of the MBH optimization into a binary variant called the BMBH optimization was employed. BMBH was able to generate more optimal solutions to CSST problems compared with BBH and BPSO. This was exhibited by the exploration power on CSST; BMBH is able to generate swarm distributions in various regions in the solution space, while BBH is not. Another aspect of the powerful searching performance of BMBH is the energy variable, which increases or decreases according to the success of the black hole in finding more optimal solutions and updating its value. This factor does not exist in classical BH or the binary variants of BH. The approach suffers from various limitations. First, the current variant of BH and BMBH has no mutation operations, which makes the algorithm subject to local minima by some percentage. Incorporating mutation operations might increase the performance of the algorithm and make it immune to such local optima. Second, the algorithms BBH and BMBH have a randomly blind initialization of the first population. Developing an application-oriented initialization might improve the performance of these algorithms. Finally, the process of pulling the stars towards the black hole currently does not focus on the application part. Moving the stars in an application-dependent operation will generate better performance. In the future, the optimization will be extended to accept various constraints on the variables, which will be more complicated than having only boundary constraints. Another future direction is to incorporate reinforcement learning, which makes the searching subject to training and, hence, more fruitful. ACKNOWLEDGMENT The authors would like to express their profound gratitude to the lab members of the Software Engineering Research Group, Universiti Teknologi Malaysia (UTM), for their continuous support and constructive feedback towards the completion of this paper. REFERENCES [1] B. J. Garvin, M. B. Cohen, and M. B. Dwyer, ‘‘Evaluating improvements to a meta-heuristic search for constrained interaction testing,’’ Empirical Softw. Eng., vol. 16, no. 1, pp. 61–102, Feb. 2011. [2] J. Lin, C. Luo, S. Cai, K. Su, D. Hao, and L. Zhang, ‘‘TCA: An efficient two-mode meta-heuristic algorithm for combinatorial test generation (T),’’ in Proc. 30th IEEE/ACM Int. Conf. Automat. Softw. Eng. (ASE), Nov. 2015, pp. 494–505. [3] L. Ma, F. Zhang, M. Xue, B. Li, Y. Liu, J. Zhao, and Y. Wang, ‘‘Combinatorial testing for deep learning systems,’’ 2018, arXiv:1806.07723. [Online]. Available: http://arxiv.org/abs/1806.07723 [4] M. Khari and P. Kumar, ‘‘An extensive evaluation of search-based software testing: A review,’’ Soft Comput., vol. 23, no. 6, pp. 1933–1946, Mar. 2019. [5] R.-Z. Qi, Z.-J. Wang, and S.-Y. Li, ‘‘A parallel genetic algorithm based on spark for pairwise test suite generation,’’ J. Comput. Sci. Technol., vol. 31, no. 2, pp. 417–427, Mar. 2016. VOLUME 8, 2020 [6] X. Li, W. E. Wong, R. Gao, L. Hu, and S. Hosono, ‘‘Genetic algorithmbased test generation for software product line with the integration of fault localization techniques,’’ Empirical Softw. Eng., vol. 23, no. 1, pp. 1–51, Feb. 2018. [7] D. V. C. Prakash, S. Tatale, V. Kondhalkar, and L. Bewoor, ‘‘A critical review on automated test case generation for conducting combinatorial testing using particle swarm optimization,’’ Int. J. Eng. Technol., vol. 7, no. 3.8, p. 22, Jul. 2018. [8] R. R. Sahoo and M. Ray, ‘‘Metaheuristic techniques for test case generation: A review,’’ J. Inf. Technol. Res., vol. 11, no. 1, pp. 158–171, Jan. 2018. [9] A. R. A. Alsewari and K. Z. Zamli, ‘‘Design and implementation of a harmony-search-based variable-strength t-way testing strategy with constraints support,’’ Inf. Softw. Technol., vol. 54, no. 6, pp. 553–568, Jun. 2012. [10] Y. A. Alsariera, M. A. Majid, and K. Z. Zamli, ‘‘SPLBA: An interaction strategy for testing software product lines using the Bat-inspired algorithm,’’ in Proc. 4th Int. Conf. Softw. Eng. Comput. Syst. (ICSECS), Aug. 2015, pp. 148–153. [11] C. S. Dangare and S. S. Apte, ‘‘Improved study of heart disease prediction system using data mining classification techniques,’’ Int. J. Control Automat., vol. 47, no. 10, pp. 44–48, Jul. 2012. [12] Y. A. Alsariera and K. Z. Zamli, ‘‘A bat-inspired strategy for t-way interaction testing,’’ Adv. Sci. Lett., vol. 21, no. 7, pp. 2281–2284, Jul. 2015. [13] A. K. Alazzawi, H. M. Rais, and S. Basri, ‘‘HABC: Hybrid artificial bee colony for generating variable t-way test sets,’’ Interaction, vol. 7, no. 12, p. 13, 2019. [14] B. S. Ahmed, ‘‘Test case minimization approach using fault detection and combinatorial optimization techniques for configuration-aware structural testing,’’ Eng. Sci. Technol., Int. J., vol. 19, no. 2, pp. 737–753, Jun. 2016. [15] K. Z. Zamli, B. Y. Alkazemi, and G. Kendall, ‘‘A tabu search hyperheuristic strategy for t-way test suite generation,’’ Appl. Soft Comput., vol. 44, pp. 57–74, Jul. 2016. [16] K. Z. Zamli, F. Din, B. S. Ahmed, and M. Bures, ‘‘A hybrid Q-learning sinecosine-based strategy for addressing the combinatorial test suite minimization problem,’’ PLoS ONE, vol. 13, no. 5, May 2018, Art. no. e0195675. [17] K. Z. Zamli, F. Din, S. Baharom, and B. S. Ahmed, ‘‘Fuzzy adaptive teaching learning-based optimization strategy for the problem of generating mixed strength t-way test suites,’’ Eng. Appl. Artif. Intell., vol. 59, pp. 35–50, Mar. 2017. [18] A. A. Alsewari, L. M. Xuan, and K. Z. Zamli, ‘‘Firefly combinatorial testing strategy,’’ in Proc. Sci. Inf. Conf. Cham, Switzerland: Springer, 2018, pp. 936–944. [19] G. Fraser, A. Arcuri, and P. Mcminn, ‘‘A memetic algorithm for whole test suite generation,’’ J. Syst. Softw., vol. 103, pp. 311–327, May 2015. [20] B. S. Ahmed and K. Z. Zamli, ‘‘A variable strength interaction test suites generation strategy using particle swarm optimization,’’ J. Syst. Softw., vol. 84, no. 12, pp. 2171–2185, Dec. 2011. [21] A. K. Alazzawi, H. M. Rais, and S. Basri, ‘‘ABCVS: An artificial bee colony for generating variable t-way test sets,’’ Int. J. Adv. Comput. Sci. Appl., vol. 10, no. 4, pp. 259–274, 2019. [22] A. K. Alazzawi, A. A. B. Homaid, A. A. Alomoush, and A. A. Alsewari, ‘‘Artificial bee colony algorithm for pairwise test generation,’’ J. Telecommun., Electron. Comput. Eng., vol. 9, nos. 1–2, pp. 103–108, 2017. [23] A. A. Alsewari, A. K. Alazzawi, T. H. Rassem, M. N. Kabir, A. A. B. Homaid, Y. A. Alsariera, and N. M. Tairan, ‘‘ABC algorithm for combinatorial testing problem,’’ J. Telecommun., Electron. Comput. Eng., vol. 9, nos. 3–3, pp. 85–88, 2017. [24] A. K. Alazzawi, H. M. Rais, and S. Basri, ‘‘Artificial bee colony algorithm for t-way test suite generation,’’ in Proc. 4th Int. Conf. Comput. Inf. Sci. (ICCOINS), Aug. 2018, pp. 1–6. [25] A. K. Alazzawi, H. M. Rais, and S. Basri, ‘‘Parameters tuning of hybrid artificial bee colony search based strategy for t-way testing,’’ Int. J. Innov. Technol. Exploring Eng., vol. 8, no. 5S, pp. 204–212, 2019. [26] A. K. Alazzawi, H. M. Rais, S. Basri, and Y. A. Alsariera, ‘‘PhABC: A hybrid artificial bee colony strategy for pairwise test suite generation with constraints support,’’ in Proc. IEEE Student Conf. Res. Develop. (SCOReD), Oct. 2019, pp. 106–111. [27] A. R. A. Alsewari and K. Z. Zamli, ‘‘A harmony search based pairwise sampling strategy for combinatorial testing,’’ Int. J. Phys. Sci., vol. 7, no. 7, pp. 1062–1072, 2012. [28] Y. A. Alsariera, M. A. Majid, and K. Z. Zamli, ‘‘A bat-inspired strategy for pairwise testing,’’ ARPN J. Eng. Appl. Sci., vol. 10, pp. 8500–8506, Oct. 2015. 33417 H. N. N. Al-Sammarraie, D. N. A.. Jawawi: MBH Inspired Meta-Heuristic Searching Optimization for Combinatorial Testing [29] G.-G. Wang, B. Chang, and Z. Zhang, ‘‘A multi-swarm bat algorithm for global optimization,’’ in Proc. IEEE Congr. Evol. Comput. (CEC), May 2015, pp. 480–485. [30] Y. A. Alsariera, M. A. Majid, and K. Z. Zamli, ‘‘Adopting the bat-inspired algorithm for interaction testing,’’ in Proc. 8th Ed. Annu. Conf. Softw. Test., 2015, p. 14. [31] A. B. Nasser, Y. A. Sariera, A. A. Alsewari, and K. Z. Zamli, ‘‘Assessing optimization based strategies for t-way test suite generation: The case for flower-based strategy,’’ in Proc. IEEE Int. Conf. Control Syst., Comput. Eng. (ICCSCE), Nov. 2015, pp. 150–155. [32] A. B. Nasser, K. Z. Zamli, and B. S. Ahmed, ‘‘Dynamic solution probability acceptance within the flower pollination algorithm for combinatorial t-way test suite generation,’’ in Intelligent and Interactive Computing. Singapore: Springer, 2019, pp. 3–11. [33] T. Mahmoud and B. S. Ahmed, ‘‘An efficient strategy for covering array construction with fuzzy logic-based adaptive swarm optimization for software testing use,’’ Expert Syst. Appl., vol. 42, no. 22, pp. 8753–8765, Dec. 2015. [34] Ş. Gülcü and H. Kodaz, ‘‘A novel parallel multi-swarm algorithm based on comprehensive learning particle swarm optimization,’’ Eng. Appl. Artif. Intell., vol. 45, pp. 33–45, Oct. 2015. [35] K. E. Heraguemi, N. Kamel, and H. Drias, ‘‘Multi-swarm bat algorithm for association rule mining using multiple cooperative strategies,’’ Appl. Intell., vol. 45, no. 4, pp. 1021–1033, Dec. 2016. [36] A. Hatamlou, ‘‘Black hole: A new heuristic optimization approach for data clustering,’’ Inf. Sci., vol. 222, pp. 175–184, Feb. 2013. [37] M. Farahmandian and A. Hatamlou, ‘‘Solving optimization problems using black hole algorithm,’’ J. Adv. Comput. Sci. Technol., vol. 4, no. 1, p. 68, Feb. 2015. 33418 HAMSA NAJI NSAIF AL-SAMMARRAIE was born in Baghdad, Iraq, in 1982. She received the B.S. degree in computer science from the University of Baghdad, Baghdad, in 2004. She is currently pursuing the master’s degree in computer science from Universiti Teknologi Malaysia (UTM), Johor Bahru, Malaysia. Her main research interests include software engineering, search-based software testing, soft computing, and artificial intelligence. DAYANG N. A. JAWAWI received the bachelor’s degree in software engineering from Sheffield Hallam University, U.K., and the master’s degree in computer science and the Ph.D. degree in software engineering from Universiti Teknologi Malaysia (UTM), Malaysia. She is currently an Associate Professor with the Faculty of Engineering, School of Computing, UTM. Her main research interests include software engineering, software reuse, software quality, software testing, requirement engineering, and computing education. A major part of her research projects focuses on rehabilitation and mobile robotics, real-time embedded systems, and precision farming applications. VOLUME 8, 2020
https://openalex.org/W1570296396	https://europepmc.org/articles/pmc4402699?pdf=render	English	null	Using epigenomics data to predict gene expression in lung cancer	BMC bioinformatics	2,015	cc-by	8,340	© 2015 Li et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http:// creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Jeffery Li1†, Travers Ching2,3†, Sijia Huang2,3, Lana X Garmire2,3* Jeffery Li1†, Travers Ching2,3†, Sijia Huang2,3, Lana X Garmire2,3* Jeffery Li1†, Travers Ching2,3†, Sijia Huang2,3, Lana X Garmire2,3* From 10th International Symposium on Bioinformatics Research and Applications (ISBRA-14) Zhangjiajie, China. 28-30 June 2014 PROCEEDINGS Open Access Abstract Background: Epigenetic alterations are known to correlate with changes in gene expression among various diseases including cancers. However, quantitative models that accurately predict the up or down regulation of gene expression are currently lacking. Methods: A new machine learning-based method of gene expression prediction is developed in the context of lung cancer. This method uses the Illumina Infinium HumanMethylation450K Beadchip CpG methylation array data from paired lung cancer and adjacent normal tissues in The Cancer Genome Atlas (TCGA) and histone modification marker CHIP-Seq data from the ENCODE project, to predict the differential expression of RNA-Seq data in TCGA lung cancers. It considers a comprehensive list of 1424 features spanning the four categories of CpG methylation, histone H3 methylation modification, nucleotide composition, and conservation. Various feature selection and classification methods are compared to select the best model over 10-fold cross-validation in the training data set. Results: A best model comprising 67 features is chosen by ReliefF based feature selection and random forest classification method, with AUC = 0.864 from the 10-fold cross-validation of the training set and AUC = 0.836 from the testing set. The selected features cover all four data types, with histone H3 methylation modification (32 features) and CpG methylation (15 features) being most abundant. Among the dropping-off tests of individual data-type based features, removal of CpG methylation feature leads to the most reduction in model performance. In the best model, 19 selected features are from the promoter regions (TSS200 and TSS1500), highest among all locations relative to transcripts. Sequential dropping-off of CpG methylation features relative to different regions on the protein coding transcripts shows that promoter regions contribute most significantly to the accurate prediction of gene expression. Conclusions: By considering a comprehensive list of epigenomic and genomic features, we have constructed an accurate model to predict transcriptomic differential expression, exemplified in lung cancer. Conclusions: By considering a comprehensive list of epigenomic and genomic features, we have constructed an accurate model to predict transcriptomic differential expression, exemplified in lung cancer. Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 * Correspondence: lgarmire@cc.hawaii.edu † Contributed equally 2Molecular Biosciences and Bioengineering Graduate Program, University of Hawaii at Manoa, Honolulu, HI 96822, USA Full list of author information is available at the end of the article Data sets Several types of high throughput data were used to extract features or classification responses. These include the CpG methylation array data from 50 paired cancer and adjacent normal tissues, three types of histone marker CHIP-Seq data from cancer and normal cell lines, genomic nucleo- tide sequence and conservation data, and RNA-Seq expression data from samples that have coupled methyla- tion data. Background such as methylation arrays, CHIP-Sequencing, gene expres- sion microarray and RNA-Sequencing have enabled researchers to better understand the relationship between epigenetic modification and gene expression at the genome scale. Coupling with the progress in experimental metho- dology, we have witnessed a wealthy growth of bioinfor- matics tools to analyze the epigenetics patterns [2-4]. Epigenetics is a rapidly expanding biological field recently. Aberrant epigenetic modifications are associated with many different diseases including cancers and neurodeve- lopmental disorders [1]. Much work has demonstrated that epigenetic regulation plays an important role in gene expression, among other mechanisms such as transcription factor regulation. Advances in high throughput methods DNA methylation and histone modification are two major mechanisms of epigenetic regulation. The most widely researched type of DNA methylation in human is the cytosine methylation of CpG islands and their asso- ciated regions, such as CpG shores [5]. CpG methylation occurs genome-wide in regions related to protein coding * Correspondence: lgarmire@cc.hawaii.edu † Contributed equally 2Molecular Biosciences and Bioengineering Graduate Program, University of Hawaii at Manoa, Honolulu, HI 96822, USA Full list of author information is available at the end of the article Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 2 of 12 transcriptome by taking advantage of the publicly available data from the TCGA Project (lung cancer) and the ENCODE project. genes (promoters, exons, UTRs, etc.) as well as in certain intergenic regions. It has been shown that CpG methyla- tion tends to occur in promoters located upstream of the transcription starting site [6], and increased methylation (hypermethylation) in the promoter is negatively asso- ciated with the gene expression level[1]. On the other hand, CpG methylation in gene bodies seems to be posi- tively associated with gene expression [1]. In cancers cells, massive global loss of DNA methylation (hypomethyla- tion) has been observed, and such hypomethylation in pro- moters can activate aberrant expression of oncogenes [7]. Much new information has been gained through the recently developed methods, such as Illumina Infinium HumanMethylation450 arrays that enable the detection of CpG methylation throughout the different locations asso- ciated with over 99% of protein coding genes. Data processing Methylation data h C G Histone modification is another type of important epi- genetic modification [1]. Histones are the core of nucleosomes that DNA sequences wrap around. All his- tones are subject to some level of methylation or acety- lation, which would either open up or close the local chromatin structures to enable or repress gene expres- sion. Among them, Histones 3 (H3) has various kinds of methylation and they serve as well-studied markers for gene expression status. For example, Histone 3 Lysine 4 tri-methylation (H3K4Me3) in the promoter region is an indicator of active gene transcription, and Histone 3 Lysine 36 tri-methylation (H3k36me3) is associated with transcription elongation. Oppositely, Histone 3 Lysine 27 tri-methylation (H3k27me3) may repress gene expression. Even more complicated, histone modifica- tion markers interact with DNA methylation, and the consequent patterns of gene expression are the com- bined effects of their interactions. The genomic assays such as CHIP-sequencing have enabled the generation of large amount of histone modification data. The Cancer Genome Atlas (TCGA) Methylation data from Illumina’s Infinium HumanMethylation450 Beadchip (Illu- mina 450k) were used to extract CpG methylation related features, according to their annotation file. The genomic coordinates of CpG, their exons and coding regions were obtained from the Illumina annotation file. Since the anno- tation file only provided information of transcripts, exons, and coding DNA sequences (CDS), we re-annotated the protein coding genes using the Illumina iGenomes hg19 Refseq annotation in order to extract more comprehensive information from other regions of the transcripts: all introns (with special categories for the first and last intron), as well as first and last exons, untranslated regions in the 5’ and 3’ direction (5’ UTR and 3’ UTR, respectively), and a “single exon” or “single intron” designation for transcripts that only had a single exon or single intron. Histone data Human genome data Special effort was paid to compute distances of CpG probes to exon-exon junctions. Given that one or more CpG sites may exist on the individual exon segments of a transcript (including the first and last exons), transcript- level maximum, minimum and average distances of any hyper/hypo-methylated probe to the nearest 5’ or 3’ exon-exon junction were computed (maxHypoTo5, min- HypoTo5, avgHypoTo5, maxHypoTo3, minHypoTo3, avgHypoTo3, maxHyperTo5, minHyperTo5, avgHy- perTo5, maxHyperTo3, minHyperTo3, and avgHyperTo3). Histone marker modification features Nucleotide composition data were extracted from hg19 genome FASTA files downloaded from the UCSC gen- ome browser. Conservation scores across three classes of species: vertebrates, primates, and placental animals, were also considered. PhastCons46Way scores were downloaded from the UCSC genome browser [11,15]. Conservation scores were then intersected with the rele- vant segments of the transcripts using a custom Perl script, in order to extract conservation features. After the alignment of raw histone marker data (see Data Processing), the aligned histone marker reads were intersected with the segments of each transcript using the multicov function from the BEDTools package [19]. The histone reads were then normalized per 1000 bp length of each segment per 1 million aligned read library. Similar to the CpG methylation features, the his- tone marker modification features were extracted on a segment-by-segment basis. Initials are used to represent the individual cell lines where the features come from: A for the A549 cell line and S for the SAEC cell line. Following the initial is a number representing the speci- fic histone H3 methylation marker: 4 for H3k4me3, 27 for H3k27me3, and 36 for H3k36me3. As a result, fea- tures are named as segment_cell type and histone modi- fication type (e.g. first_exon_A4). In order to compare histone modification between the cancer and non-cancer cell types, the differences of the reads between them were divided by the average of the two (e.g. a feature named first_exon_A4_minus_S4_divavg). RNA-Seq gene expression data from lung cancer samples with coupled CpG methylation data were downloaded from TCGA Research Network: http://cancergenome.nih. gov. Lung adenocarcinoma and lung squamous cell carci- noma data were combined for this project, as they are two subtypes of non-small cell lung cancer. Differential expres- sion analysis was done with the DESeq2 package in R [16]. In cases where multiple transcripts are mapped to the same Refseq ID, the geometric mean of the differential expression results was used to represent the gene level expression. Histone data Three sets of histone marker CHIP-Seq data, H3k4me3, H3k27me3, and H3k36me3, were considered from two cell lines: A549 cell line (0.2% EtOH treatment) from the lung carcinoma tissue, and SAEC normal lung epithelial cell line (no treatment). Raw CHIP-Seq data were downloaded from the Broad Institute/Bernstein Lab at the Massachusetts General Hospital/Harvard Medical School and the University of Washington in collaboration with the ENCODE project via the UCSC genome browser at http://genome.ucsc.edu. [10,11]. The raw reads were processed in-house to ensure consis- tency of all normalization procedures. Raw data were first aligned to hg19 using bowtie2 [12], followed by removal of duplicated reads using the Samtools toolkit (specifically, the “rmdup” tool) [13]. The aligned reads were intersected with the relevant segments of the tran- script as annotated in the previous section, using the Bedtools toolkit (specifically, the “multicov” tool) [14]. A custom R script was used to normalize the data over total number of reads after removing PCR duplicates. Although integrative analyses on gene expression and epigenetics regulation abound throughout the literature [8,9], it is our observation that quantitative models which use epigenetic information to accurately predict the up or down regulation of gene expression are currently lacking. A frequent question that a biologist would ask, when methylation data are available but the gene expression data are missing, is how the epigenetic changes of a gene may affect the expression of this gene to be either up or down regulated. This report is aimed to fill in this gap, and provide the users with a model that allows them to estimate the consequence of epigenetic modification on gene expression, when the data for the latter are not avail- able. Towards this goal, we have built a classification pre- dictor for gene expression using the machine learning approach. This model examines a large set of CpG methy- lation data, histone modification data and genome data, and accurately predicts differential expression of RNA-Seq Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 3 of 12 Feature extraction In each segment of the transcript, four different types of nucleotide features were extracted: single nucleotide composition, dinucleotide composition, trinucleotide composition, and the length of each segment. Nucleo- tide sequences of Hg19 reference genome were pro- cessed using the Biostrings library in R [20]. The extracted features are categorized into four major sub-groups. All features were considered on a segment- wise basis (see Data Processing), unless noted otherwise. CpG Methylation features Differential expression of the methylated CpG sites was processed using the limma library in R. Specifically, the function toptable was used to determine the log fold change (logFC) between the cancer and normal tissues as well as the average methylation (avgMval) of each CpG site across the two types of tissue [18]. A positive logFC indicates hypermethylation whereas a negative logFC indicates hypomethylation. Additional segment-based features were also considered. These include the number of hypermethylated (numHyper) and hypomethylated probes (numHypo) on a segment of a given transcript. For example, first_exon_numHyper refers to the number of hypermethylated probes on the first exon. Two other types of features are the average of logFC and avgMval of all CpG probes on a segment of the transcript, e.g. the average logFC of all probes on the first exon of a given transcript (first_exon_avglogFC). Conservation features Conservation score per segment was calculated as the arithmetic mean of the conservation score per nucleo- tide in that segment. Three separate sets of conservation scores with different comparative species were extracted from UCSC genome browser - vertebrate, primate, or placental. Thus, features such as first_exon_vertebrate emerge from this set. Human genome data In the case that any of these read counts was zero, the counts from all transcripts were artificially increased by one in order to calculate the geometric mean, followed by final subtraction of one. The expression of a gene was then classified as binary outcomes: either up-re- gulated or down-regulated, once it passed two thresholds: 1) having an adjusted p value < .05 after Holm’s multiple hypothesis test [17] and 2) having an absolute value of log2 fold change greater than 1. As a result, 2874 genes were selected as “differentially expressed” genes. Summary of input data and features Four types of input data were used to extract the fea- tures including the Illumina 450K CpG methylation array data from cancer and normal tissues, three types of histone H3 marker CHIP-Seq data from cancer and normal cell lines, genomic nucleotide sequence and con- servation data, and RNA-Seq gene expression data from samples with coupled CpG methylation data. In total, we calculated 1424 features and summarized the fea- tures by column. These features can be divided into two categories (Table 1): (1) data type based features, includ- ing average CpG methylation, average methylation log fold change, number of hyper/hypo-methylated probes, mono-nucleotide, di-nucleotide and tri-nucleotide com- position, histone H3 methylation CHIP-Seq reads, and Phastcon conservation scores; (2) segment based CpG methylation features from Illumina 450K BeadChip annotations: upstream of the transcription start site (TSS) 1500, TSS200, 5’ and 3’ UTRs, exon/intron body, first and last exon/intron, single exon/intron and full transcript (Figure 1 and Table 1). ReliefF is an improved feature selection method from Relief. Relief uses the Manhattan distance of its nearest hit and miss from a random instance to continuously update a weight vector, which is then used to calculate a relevance score. Features above a certain relevance threshold are considered “selected” [24]. ReliefF improves on Relief in several ways, including two improvements particularly important for this report. First, ReliefF extends Relief to be able to handle incomplete or partial data. Second, ReliefF searches for k near-hits and near- misses and takes their averages, as opposed to one near- est hit or miss from Relief. k =10 was sufficient to obtain satisfactory results [23]. CFS is the only method that has a built-in system for selecting the number of features. Gain Ratio and ReliefF both work as ranker systems, meaning every input has a matching respective ranked output. In order to ensure fairness between feature selection methods, we matched the numbers of selected features from Gain Ratio and ReliefF to be the same as determined by CFS. Model selection and evaluation The model uses 2298 gene data points in the training set, with an additional 576 genes kept in the testing set. Three different feature selection methods were evaluated in com- bination with five classification methods, using 10-fold cross-validation on the training data set (Figure 2). The three feature selection methods are: correlation-based fea- ture selection (CFS), ReliefF, and Gain Ratio. In most cases with combined classification methods, except for Gaussian SVM, ReliefF gives the best AUCs among the three feature selection methods. Among the five classifica- tion methods that we considered, namely Gaussian SVM, linear SVM, Logistic Regression, Naïve Bayes and Random Forest, the two non-linear methods (Gaussian SVM and Random Forest) show superior performances to the other linear classifiers (Logistic Regression, linear SVM, and Naïve Bayes). This indicates that interactions exist among the selected features. However, the differences are not very big, suggesting that the decision boundary is close to lin- ear. Given that the model based on ReliefF feature selec- tion and Random Forest classification gives the best AUC of 0.864, it is selected as the best model for the rest of the project. Similarly, a ReliefF and Random Forest based Feature selection Three feature selection methods were considered: Cor- relation Feature Selection (CFS) [21], Gain Ratio [22] and ReliefF [23]. CFS is based on mutual information, a non-linear measure of correlation. CFS selects an approximately optimal set of features to maximize the relevance and Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 4 of 12 Software minimize redundancy. Relevance is the correlation of a feature to the class (up-regulated or down-regulated gene expression) measured by mutual information, whereas redundancy is the correlation between two fea- tures. Redundancy between selected features is mini- mized to keep the number of selected feature small. Weka 3 data mining software [25] was used for feature selection, classifier training and evaluation. Various R packages were used, including Corrplot for generation of the correlation matrix [26], and ROCR for ROC curves [27]. The classification model is available at: https://github.com/lanagarmire/epiPredictor The Gain Ratio is an improved method of Information Gain (IG). Both feature selection methods employ a decision tree in their respective algorithms. The Gain Ratio, by name, is a ratio of IG, but it overcomes the bias of IG which favors features with more data. Model evaluation From 5’ to 3’ end of the protein coding genes, listed are transcription starting sites (TSS) upstream up to 1500 bp (TSS 1500) and 200 bp (TSS 200), first exon which may include 5’ UTR, first intron, exon body, last intron, and last exon which may include 3’ UTR. A full transcript region is determined as the UTRs and coding region together. Figure 2 Performance comparison of models with various feature selection and classification methods. The Areas Under the Curve (AUC) of ROC are used as the metric to compare the performance of models with different combinations of feature selection (CFS, Gain Ratios and ReliefF) and classification (Gaussian SVM, Linear SVM, Logistic regression, Naïve Bayes and Random Forest), on the training data with 10 fold cross-validation. The model with ReliefF based feature selection and Random Forest classification is selected as the best model. Figure 1 Segments associated with protein coding genes. Features considered to predict differential gene expression are depicted on a segment-by-segment basis. Segments are determined based on the annotations of Illumina Infinium Human Methylation 450K Beadchip Array, with augmentations on segments located in gene bodies. From 5’ to 3’ end of the protein coding genes, listed are transcription starting sites (TSS) upstream up to 1500 bp (TSS 1500) and 200 bp (TSS 200), first exon which may include 5’ UTR, first intron, exon body, last intron, and last exon which may include 3’ UTR. A full transcript region is determined as the UTRs and coding region together. Figure 1 Segments associated with protein coding genes. Features considered to predict differential gene expression are depicted on a segment-by-segment basis. Segments are determined based on the annotations of Illumina Infinium Human Methylation 450K Beadchip Array, with augmentations on segments located in gene bodies. From 5’ to 3’ end of the protein coding genes, listed are transcription starting sites (TSS) upstream up to 1500 bp (TSS 1500) and 200 bp (TSS 200), first exon which may include 5’ UTR, first intron, exon body, last intron, and last exon which may include 3’ UTR. A full transcript region is determined as the UTRs and coding region together. Figure 1 Segments associated with protein coding genes. Features considered to predict differential gene expression are depicted on a segment-by-segment basis. Segments are determined based on the annotations of Illumina Infinium Human Methylation 450K Beadchip Array, with augmentations on segments located in gene bodies. Model evaluation From 5’ to 3’ end of the protein coding genes, listed are transcription starting sites (TSS) upstream up to 1500 bp (TSS 1500) and 200 bp (TSS 200), first exon which may include 5’ UTR, first intron, exon body, last intron, and last exon which may include 3’ UTR. A full transcript region is determined as the UTRs and coding region together. Figure 2 Performance comparison of models with various feature selection and classification methods. The Areas Under the Curve (AUC) of ROC are used as the metric to compare the performance of models with different combinations of feature selection (CFS, Gain Ratios and ReliefF) and classification (Gaussian SVM, Linear SVM, Logistic regression, Naïve Bayes and Random Forest), on the training data with 10 fold cross-validation. The model with ReliefF based feature selection and Random Forest classification is selected as the best model. Figure 2 Performance comparison of models with various feature selection and classification methods. The Areas Under the Curve (AUC) of ROC are used as the metric to compare the performance of models with different combinations of feature selection (CFS, Gain Ratios and ReliefF) and classification (Gaussian SVM, Linear SVM, Logistic regression, Naïve Bayes and Random Forest), on the training data with 10 fold cross-validation. The model with ReliefF based feature selection and Random Forest classification is selected as the best model. model has the best predictive performance on the 20% holdout data set, with an AUC of 0.836. We first explored the relationship among the selected features. Using hierarchical clustering on absolute corre- lation values between features (Figure 3A), we found that the selected features tend to cluster by the data type, as expected. For example, the conservation features in the coding regions (CDS) are grouped together, and so are Model evaluation The data were split into training and testing sets. The training set constituted 80% of the up-regulated and down-regulated genes, and the testing set constituted the remaining 20% genes. The training data set underwent 10-fold cross validation on various combinations of fea- ture selection and classification methods, in order to obtain the best model. After determining the best model, two sets of drop-off tests were conducted. The first set of tests considered the effect of data types, including nucleotide composition, histone markers and methylation data, on the perfor- mances of sub-models. The second set of drop-off tests considered the effects of different segments on tran- scripts, including gene body, exons, introns, UTRs, TSS1500 and TSS200, on the methylation CpG methyla- tion data based sub-models. For each drop-off test, a set of features was removed from the original input features prior to the feature selection and classification. Subse- quently the same ReliefF feature selection and RF classifi- cation for the drop-off tests were performed as described in the previous Feature Selection section. le eotide position Dinucleotide composition Trinucleotide composition Length of segment Histone reads Histone read comparisons (difference of reads/average of reads) Conservation scores (PHASTCONS) 500_A TSS1500_AA TSS1500_AAA – TSS1500_S27 TSS1500_A27_ minus_S27_divavg TSS1500_vertebrate 00_A TSS200_AA TSS200_AAA – TSS200_S27 TSS200_A27_ minus_S27_divavg TSS200_vertebrate _A UTR5_AA UTR5_AAA UTR5_length UTR5_S27 UTR5_A27_ minus_S27_divavg UTR5_vertebrate exon_A first_exon_AA first_exon_ AAA first_exon_length first_exon_S27 first_exon_A27_ minus_S27_divavg first_exon_vertebrate ntron_A first_intron_AA first_intron_AAA first_intron_length first_intron_S27 first_intron_A27_ minus_S27_divavg first_intron_vertebrate _A exon_AA exon_AAA exon_length exon_S27 exon_A27_ minus_S27_divavg exon_vertebrate n_A intron_AA intron_AAA intron_length intron_S27 intron_A27_ minus_S27_divavg intron_vertebrate A CDS_AA CDS_AAA CDS_length CDS_S27 CDS_A27_ minus_S27_divavg CDS_vertebrate ntron_A last_intron_AA last_intron_AAA last_intron_ length last_intron_S27 last_intron_A27_ minus_S27_divavg last_intron_vertebrate exon_A last_exon_AA last_exon_AAA last_exon_ length last_exon_S27 last_exon_A27_ minus_S27_divavg last_exon_vertebrate _A UTR3_AA UTR3_AAA UTR3_length UTR3_S27 UTR3_A27_ minus_S27_divavg UTR3_vertebrate anscript_A fullTranscript_AA fullTranscript_ AAA fullTranscript_ length fullTranscript_S27 fullTranscript_A27_ minus_S27_divavg fullTranscript_vertebrate e_exon_A single_exon_AA single_exon_ AAA single_exon_ length single_exon_S27 single_exon_A27_ minus_S27_divavg single_exon_vertebrate e_intron_A single_intron_AA single_intron_ AAA single_intron_ length single_intron_S27 single_intron_A27_ minus_S27_divavg single_intron_vertebrate 224 896 12 84 42 42 1412 ge nce end Average Distance to 3’ end yperTo5 avgHyperTo3 ypoTo5 avgHypoTo3 2 12 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 6 of 12 Figure 1 Segments associated with protein coding genes. Features considered to predict differential gene expression are depicted on a segment-by-segment basis. Segments are determined based on the annotations of Illumina Infinium Human Methylation 450K Beadchip Array, with augmentations on segments located in gene bodies. Analysis of selected features A total of 67 features are selected by the best model, spanning all four types of genomic and epigenomic data. Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 7 of 12 Figure 3 Top fifteen features from the best model. (a) The clustering results on the absolute values of Pearson’s correlation coefficients from 67 selected features by the best model. The names of different type of features are labeled by different colors. Note: the length of a segment is listed out separately. (b) List of top fifteen features selected by ReliefF feature selection and sorted by their correlation to the classification of differential gene expression. Figure 3 Top fifteen features from the best model. (a) The clustering results on the absolute values of Pearson’s correlation coefficients from 67 selected features by the best model. The names of different type of features are labeled by different colors. Note: the length of a segment is listed out separately. (b) List of top fifteen features selected by ReliefF feature selection and sorted by their correlation to the classification of differential gene expression. most methylation features. As expected, the CpG islands within the promoter are very important for the prediction of gene expression, as demonstrated by the three selected and highly correlated features CG composition features, TSS200_GC, TSS200_CG and TSS200_CGG. We also calculated the correlation of each feature to gene expression and plotted the top 15 features most relevant to gene expression prediction (Figure 3B). None of the features have correlations higher than 0.45, sug- gesting that no single feature is a dominant predictor for gene expression. These features are either histone modification (11 features) or methylation features (4 fea- tures), consistent with the previous observation on the significance of these two types of features. The largest group of selected features is the histone modification features (32 features), followed by the methy- lation features (15 features) (Additional File 1). The selected features underscore the importance of histone epigenetic modification in the regulation of gene expres- sion. Likewise, the importance of methylation features is evident, especially for the featured arising from TSS, 5’ UTR and first exons. Interestingly, several methylation fea- tures (TSS1500_avgMval, first_exon_avgMval and UTR5_avgMval) are clustered with histone modification features, suggesting collinearity between these two types of features, as shown by others [28,29]. Evaluation of features by data type To determine the contribution of different types of fea- tures to gene expression, we tested the performance of models when a subset of features from the same data type were dropped. We present the results of four mea- sures of model performance: AUC, accuracy, F-measure and Matthew’s correlation coefficient (MCC) (Figure 4). Dropping any individual feature set of nucleotide compo- sition, histone modification or CpG methylation, did not seem to have a large effect on the model performance, indicating that there is redundancy between feature sets. The sub-model performance for the dropping-off of a single feature set from the full model is in the following order: nucleotide composition removal > histone modifi- cation removal > CpG methylation removal. Thus drop- ping methylation features had the largest effect among individual feature set, as the AUC decreases from 0.864 On the other hand, when features are categorized by location relative to the transcripts (Additional File 1), the TSS200 has the most number of features (13 features), and TSS1500 has 6 features selected for this region. Together, the promoter comprises 28% of all the selected features. This confirms the previously well-known impor- tance of the promoter region for the epigenetic regulation of gene expression [30,31]. Additionally, CDS has the sec- ond highest number of features being selected, highlight- ing its significance in regulating gene expression [30]. Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 8 of 12 Figure 4 Evaluation of features generated from various data types. (a-b) Effects of feature set drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of feature set drop-off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. Figure 4 Evaluation of features generated from various data types. (a-b) Effects of feature set drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of feature set drop-off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. Figure 4 Evaluation of features generated from various data types. (a-b) Effects of feature set drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of feature set drop-off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. both training and testing sets. Evaluation of features by data type However, it leads to higher MCC than removing just histones does in the testing set. This suggests that there might be some overfitting with regards to the nucleotide feature set, which accounts for the majority (83%) of features prior to feature selection. in the full model to 0.832 in the training set, as well as from 0.836 to 0.810 in the testing set. Likewise, MCC, upon single feature set drop-off, shows the largest pro- portional change among the four performance measures, and decreases from 0.56 to 0.49 on the training set and 0.51 to 0.45 on the testing set. Evaluation of CpG methylation features by locations relative to transcripts We also compared the effect of removing both nucleo- tide and histone features on model performance, as com- pared to removing either of them alone. As expected, removing both nucleotide and histone features gives the lowest AUCs, lowest accuracies and lowest F-measures in Given that removing methylation features causes the most reduction of model performance among the single feature set drop-off (Figure 5), we next asked the Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 9 of 12 question of the relative importance of each methylation feature categorized by genomic location We performed the TSS), then from gene bodies including exons, introns, and UTRs and lastly from TSS1500 region such that only Figure 5 Evaluation of methylation features by segment. (a-b) Effects of segment-based methylation feature set sequential drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of segment-based methylation feature set sequential drop- off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. Figure 5 Evaluation of methylation features by segment. (a-b) Effects of segment-based methylation feature set sequential drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of segment-based methylation feature set sequential drop- off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. Figure 5 Evaluation of methylation features by segment. (a-b) Effects of segment-based methylation feature set sequential drop-off on ROC curves from the 10-fold cross-validation training set (a) and testing set (b). (c) Effects of segment-based methylation feature set sequential drop- off on other four metrics: AUC, Accuracy, F-measure and MCC, in the training set and testing set. question of the relative importance of each methylation feature categorized by genomic location. We performed drop-off tests by sequentially removing features in each genomic location category. We first removed the features from the first exons and first introns (as they are close to the TSS), then from gene bodies including exons, introns, and UTRs, and lastly from TSS1500 region such that only TSS200 features were kept. At each step, we re-performed feature selection and model construction, using the remaining methylation features. Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Li et al. The need to build predictive models of gene expression from epigenomics data Although currently integrative analyses between gene expression and epigenetic modification exist, we have found that quantitative models using epigenetic informa- tion to accurately predict the up or down regulation of gene expression are currently lacking. There are often cases where researchers can only obtain reliable epige- netics data, but not gene expression data. For example, if the samples are archived and processed by FFPE (Forma- lin-fixed, paraffin-embedded), one can still perform epige- nomics measurements, but not the gene expression experiments due to the degradation of mRNA in the sam- ples. More importantly, a predictive method such as ours can efficiently facilitate the bench scientists to narrow down the candidate lists and conduct gene expression vali- dation, especially when the epigenetics information is the only data handy. Worth noticing, a lot of features that are extracted on methylation and histone modification are naturally based on the annotations from Illumina 450K array plat- form for DNA methylation. There may be bias on the number of features that are hand coded in the model. To address potential issue, we changed TSS200_GC to TSS150_GC in our model and obtained an AUC = .861 (compared to 0.864) for cross fold validation on the training set and an AUC = .834 (compared to 0.836) for the testing set. Therefore, we think the bias due to rely- ing on the nomenclatures from Illumina’s annotation is small. Selected features and their relevance to gene expression All four types of data (CpG methylation, histone H3 modification, nucleotide sequence and conservation) exist in the 67 features that are selected by the best model, indicating that all of them contribute to the accu- rate prediction of gene expression. Moreover, selected features of the same data type tend to cluster together on the correlation matrix among the features, suggesting that the relationship within the same data type is closer than the relationship between different data types. As expected, histone modification and CpG methylation fea- tures are the largest two groups among the four types of data, signifying their importance to predict gene expres- sion. Since nonlinear classification methods perform slightly better than linear classification methods, it sug- gests that interactions do exist between different types of data. This is supported by numerous literatures that Evaluation of CpG methylation features by locations relative to transcripts BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 10 of 12 Page 10 of 12 enzymes responsible for CpG methylation also interact with histone modification events [32,33]. All categories of genomic locations provide relevant useful information that contributes to better prediction of gene expression, as each of the sequential feature set drop-off decreased the performance of the model in both training and testing sets. Compared to the removal of first exon and intron regions, and removal of the UTRs and the rest of the gene body, removal of TSS1500 leads to the largest reduction in all four metrics, confirming the importance of the promoter region in regulating gene expression. Even when only TSS200 features are consid- ered, an AUC of 0.638 and 0.636 are obtained in the test- ing and training sets respectively, suggesting that CpG methylation status in TSS200 is still somewhat predictive of gene expression. However, a more accurate prediction using methylation features should arise from all locations associated with the transcripts. Besides the value of predicting gene expression, our models also provide insights into the relative importance of different epigenomics/genome data, as well as the genomic locations. We found that CpG methylation fea- tures have more predictive values for differential gene expression, compared to the three types of histone H3 modification data. Although other kinds of histone mod- ification data can also be obtained to increase the pre- dictive values of histone modification data, it is much more costly to obtain them relative to the CpG methyla- tion data (the cost of CHIP-Seq on each of the histone modification marker is similar to an entire CpG methy- lation array). Therefore, practically speaking, when the budget is a constraining factor, we suggest that assays on CpG methylation should be considered with priority in predicting differential gene expression. Moreover, the results of our models demonstrate that all genomic loca- tions relative to each transcript, including promoters, exons and gene bodies, provide useful information to predict gene expression alternation. Although the CpG methylation signals from the promoters region are more important, the methylation signals from other regions, such as exons, introns and UTRs are indicative of changes in the gene expression as well. The authors declare that they have no competing interests 16. Love MI, Huber W, Anders S: Moderated estimation of fold change and dispersion for RNA-Seq data with DESeq2. bioRxiv 2014. Authors’ details 1 1Department of Biomedical Engineering, Johns Hopkins University, Baltimore MD 21218, USA. 2Molecular Biosciences and Bioengineering Graduate Program, University of Hawaii at Manoa, Honolulu, HI 96822, USA. 3Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI 96813, USA. 1Department of Biomedical Engineering, Johns Hopkins University, Baltimore MD 21218, USA. 2Molecular Biosciences and Bioengineering Graduate Program University of Hawaii at Manoa Honolulu HI 96822 USA 3Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI 3Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI 96813, USA. Conclusions A new model based on epigenomics data is proposed to predict transcriptome-level differential gene expression in lung cancers. Dropping-off feature sets by data type shows that CpG methylation features are most impor- tant for the prediction. Furthermore, methylation fea- tures on all genomic regions relative to protein coding genes contribute to the differential gene expression, within which promoter regions are most important. 9. Rhee JK, Kim K, Chae H, Evans J, Yan P, Zhang BT, Gray J, Spellman P, Huang TH, Nephew KP, et al: Integrated analysis of genome-wide DNA methylation and gene expression profiles in molecular subtypes of breast cancer. Nucleic acids research 2013, 41(18):8464-8474. 10. An integrated encyclopedia of DNA elements in the human genome. Nature 2012, 489(7414):57-74. 10. An integrated encyclopedia of DNA elements in the human genome. Nature 2012, 489(7414):57-74. 11. Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, Kent WJ: The UCSC Table Browser data retrieval tool. Nucleic acids research 2004, , 32 Database: D493-496. 11. Karolchik D, Hinrichs AS, Furey TS, Roskin KM, Sugnet CW, Haussler D, Kent WJ: The UCSC Table Browser data retrieval tool. Nucleic acids research 2004, , 32 Database: D493-496. 12. Langmead B, Salzberg SL: Fast gapped-read alignment with Bowtie 2. Nature methods 2012, 9(4):357-359. Authors’ contributions LXG envisioned the project and supervised the work. TC initiated the project. JL and TC designed and implemented the project. SJ assisted the project. JL, TC and LXG wrote the manuscript. All authors have read, revised and approved the final manuscript. 17. Holm S: A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics 1979, 65-70. 17. Holm S: A simple sequentially rejective multiple test proced Scandinavian journal of statistics 1979, 65-70. 18. Smyth GK: Limma: linear models for microarray data. Bioinformatics and computational biology solutions using R and Bioconductor Springer; 2005, 397-420. 18. Smyth GK: Limma: linear models for microarray data. Bioinformatics and computational biology solutions using R and Bioconductor Springer; 2005, 397-420. References 1 P l 1. Portela A, Esteller M: Epigenetic modifications and human disease. Nature biotechnology 2010, 28(10):1057-1068. 1. Portela A, Esteller M: Epigenetic modifications and human disease. Nature biotechnology 2010, 28(10):1057-1068. 2. Bock C, Lengauer T: Computational epigenetics. Bioinformatics 2008, 24(1):1-10. 2. Bock C, Lengauer T: Computational epigenetics. Bioinformatics 2008, 24(1):1-10. 3. Laird PW: Principles and challenges of genomewide DNA methylation analysis. Nature reviews Genetics 2010, 11(3):191-203. 3. Laird PW: Principles and challenges of genomewide DNA methylation analysis. Nature reviews Genetics 2010, 11(3):191-203. 4. Lim SJ, Tan TW, Tong JC: Computational Epigenetics: the new scientific paradigm. Bioinformation 2010, 4(7):331-337. 5. Gardiner-Garden M, Frommer M: CpG islands in vertebrate genomes. Journal of molecular biology 1987, 196(2):261-282. 4. Lim SJ, Tan TW, Tong JC: Computational Epigenetics: the new scientific paradigm. Bioinformation 2010, 4(7):331-337. 5. Gardiner-Garden M, Frommer M: CpG islands in vertebrate genomes. Journal of molecular biology 1987, 196(2):261-282. 6. Daura-Oller E, Cabre M, Montero MA, Paternain JL, Romeu A: Specific gene hypomethylation and cancer: New insights into coding region feature trends. Bioinformation 2009, 3(8):340. 6. Daura-Oller E, Cabre M, Montero MA, Paternain JL, Romeu A: Specific gene hypomethylation and cancer: New insights into coding region feature trends. Bioinformation 2009, 3(8):340. 7. Wild L, Flanagan JM: Genome-wide hypomethylation in cancer may be a passive consequence of transformation. Biochimica et biophysica acta 2010, 1806(1):50-57. 7. Wild L, Flanagan JM: Genome-wide hypomethylation in cancer may be a passive consequence of transformation. Biochimica et biophysica acta 2010, 1806(1):50-57. 8. Figueroa ME, Chen SC, Andersson AK, Phillips LA, Li Y, Sotzen J, Kundu M, Downing JR, Melnick A, Mullighan CG: Integrated genetic and epigenetic analysis of childhood acute lymphoblastic leukemia. The Journal of clinical investigation 2013, 123(7):3099-3111. 8. Figueroa ME, Chen SC, Andersson AK, Phillips LA, Li Y, Sotzen J, Kundu M, Downing JR, Melnick A, Mullighan CG: Integrated genetic and epigenetic analysis of childhood acute lymphoblastic leukemia. The Journal of clinical investigation 2013, 123(7):3099-3111. Limitations and future directions The fact that AUCs hover between 0.80-0.90 ranges could be well due to the fact that features from these other mechanisms are not considered in the current epigenetics model. To increase the accuracy, a more complex model that takes into account of all these events should be constructed. in the classification model. However, we split the dataset into training and independent testing subsets and show the model performs comparably well on the holdout testing subset. We believe that the model can be repli- cated if we can identify paired RNA-Seq and methyla- tion data. In fact, we had originally built this model on a private data set, which also achieved an AUC of more than 0.80. Additionally, Figure 4 and 5 both indicate that our approach does not suffer a significant over-fit- ting problem using the TCGA data, and show the domi- nant efforts of histone modification and CpG methylation, which yield an up/down gene expression prediction with an AUC>0.80. Currently the model uses lung cancer data, and it will be interesting find out more general epigenetic predictors for differential gene expression in other cancers as well. Lastly, we should point out that regulation of gene expression is complex, including other mechanisms mediated by transcription factors, microRNA, non-coding RNAs etc. The fact that AUCs hover between 0.80-0.90 ranges could be well due to the fact that features from these other mechanisms are not considered in the current epigenetics model. To increase the accuracy, a more complex model that takes into account of all these events should be constructed. Acknowledgements 19. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841-842. 19. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841-842. The authors would like to thank Dr. Jayson Masaki for reviewing the manuscript. g 20. Pages H, Aboyoun P, Gentleman R, DebRoy S: String objects representing biological sequences, and matching algorithms. R package version 2009, 2(2). Pages H, Aboyoun P, Gentleman R, DebRoy S: String objects represent 20. Pages H, Aboyoun P, Gentleman R, DebRoy S: String objects representing biological sequences, and matching algorithms. R package version 2009, 2(2). Additional material 13. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R, Genome Project Data Processing S: The Sequence Alignment/Map format and SAMtools. Bioinformatics 2009, 25(16):2078-2079. Additional file 1: Table S1. Selected 67 features in the best model sorted by category and their frequency Additional file 1: Table S1. Selected 67 features in the best model sorted by category and their frequency 14. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841-842. 14. Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841-842. 15. Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, Rosenbloom K, Clawson H, Spieth J, Hillier LW, Richards S: Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome research 2005, 15(8):1034-1050. 15. Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, Rosenbloom K, Clawson H, Spieth J, Hillier LW, Richards S: Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome research 2005, 15(8):1034-1050. Limitations and future directions We should point out that our current model does not include all histone modification data, but only three widely used methylation markers on histone H3 (H3K4Me3, H3K27Me3 and H3K36Me3). Moreover, the histone H3 data are drawn from ENCODE cell lines, since the TCGA samples do not have such data. The heterogeneity of the sample resources could affect the accuracy of the model. When more histone marker data coupled with DNA methylation and RNA-Seq data become publicly available for lung cancer, we can include them to achieve a better model. In the ideal set- ting, we would like to build a predictive model that has multiple types of epigenomics data obtained from the same samples. Another potential concern is overfitting Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 Page 11 of 12 Page 11 of 12 This article has been published as part of BMC Bioinformatics Volume 16 Supplement 5, 2015: Selected articles from the 10th International Symposium on Bioinformatics Research and Applications (ISBRA-14): Bioinformatics. The full contents of the supplement are available online at http://www.biomedcentral.com/bmcbioinformatics/supplements/16/S5. This article has been published as part of BMC Bioinformatics Volume 16 Supplement 5, 2015: Selected articles from the 10th International Symposium on Bioinformatics Research and Applications (ISBRA-14): Bioinformatics. The full contents of the supplement are available online at http://www.biomedcentral.com/bmcbioinformatics/supplements/16/S5. in the classification model. However, we split the dataset into training and independent testing subsets and show the model performs comparably well on the holdout testing subset. We believe that the model can be repli- cated if we can identify paired RNA-Seq and methyla- tion data. In fact, we had originally built this model on a private data set, which also achieved an AUC of more than 0.80. Additionally, Figure 4 and 5 both indicate that our approach does not suffer a significant over-fit- ting problem using the TCGA data, and show the domi- nant efforts of histone modification and CpG methylation, which yield an up/down gene expression prediction with an AUC>0.80. Currently the model uses lung cancer data, and it will be interesting find out more general epigenetic predictors for differential gene expression in other cancers as well. Lastly, we should point out that regulation of gene expression is complex, including other mechanisms mediated by transcription factors, microRNA, non-coding RNAs etc. Declarations 21. Hall MA, Smith LA: Feature Selection for Machine Learning: Comparing a Correlation-Based Filter Approach to the Wrapper. FLAIRS Conference: 1999 1999, 235-239. Publication charges for this article were funded by NIH/NIGMS P20 COBRE GM103457, NIH/NIEHS K01 ES025434-01 and Hawaii Community Foundation. Page 12 of 12 Page 12 of 12 Li et al. BMC Bioinformatics 2015, 16(Suppl 5):S10 http://www.biomedcentral.com/1471-2105/16/S5/S10 22. Quinlan JR: Induction of decision trees. Machine learning 1986, 1(1):81-106 23. Kononenko I, Šimec E, Robnik-Šikonja M: Overcoming the Myopia of Inductive Learning Algorithms with RELIEFF. Applied Intelligence 1997, 7(1):39-55. 24. Kira K, Rendell LA: The feature selection problem: Traditional methods and a new algorithm. AAAI: 1992 1992, 129-134. 25. Hall M, Frank E, Holmes G, Pfahringer B, Reutemann P, Witten IH: The WEKA data mining software: an update. ACM SIGKDD Explorations Newsletter 2009, 11(1):10-18. 26. Wei T: Corrplot: visualization of a correlation matrix. R package version 02- 0 Available on Comprehensive R Archive Network website 2010 [http://CRANR- project org/package=corrplot], (accessed 2010). 27. Sing T, Sander O, Beerenwinkel N, Lengauer T: ROCR: visualizing classifier performance in R. Bioinformatics 2005, 21(20):3940-3941. 27. Sing T, Sander O, Beerenwinkel N, Lengauer T: ROCR: visualizing classifier performance in R. Bioinformatics 2005, 21(20):3940-3941. 27. Sing T, Sander O, Beerenwinkel N, Lengauer T: ROCR: visualiz performance in R. Bioinformatics 2005, 21(20):3940-3941. 28. Fuks F, Hurd PJ, Wolf D, Nan X, Bird AP, Kouzarides T: The methyl-CpG- binding protein MeCP2 links DNA methylation to histone methylation. Journal of Biological Chemistry 2003, 278(6):4035-4040. 29. Cedar H, Bergman Y: Linking DNA methylation and histone modification: patterns and paradigms. Nat Rev Genet 2009, 10(5):295-304. 30. Jones PA: Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 2012, 13(7):484-492. 31. Maunakea AK, Nagarajan RP, Bilenky M, Ballinger TJ, D’Souza C, Fouse SD, Johnson BE, Hong C, Nielsen C, Zhao Y: Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature 2010, 466(7303):253-257. 32. Bowen NJ, Palmer MB, Wade PA: Chromosomal regulation by MeCP2: structural and enzymatic considerations. Cellular and molecular life sciences : CMLS 2004, 61(17):2163-2167. 33. Ooi SK, Qiu C, Bernstein E, Li K, Jia D, Yang Z, Erdjument-Bromage H, Tempst P, Lin SP, Allis CD, et al: DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 2007, 448(7154):714-717. doi:10.1186/1471-2105-16-S5-S10 Cite this article as: Li et al.: Using epigenomics data to predict gene expression in lung cancer. Declarations BMC Bioinformatics 2015 16(Suppl 5):S10. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit
https://openalex.org/W4293173520	https://www.pure.ed.ac.uk/ws/files/289155276/Taylor_et_al._Frontiers_Manuscript_R1_Submission_Clean.pdf	English	null	Single cell transcriptomics of Atlantic salmon (Salmo salar L.) liver reveals cellular heterogeneity and immunological responses to challenge by Aeromonas salmonicida	Frontiers in immunology	2,022	cc-by	14,728	General rights C i h f h General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights. Edinburgh Research Explorer Single cell transcriptomics of Atlantic salmon (Salmo salar L.) liver reveals cellular heterogeneity and immunological responses to challenge by Aeromonas salmonicida Citation for published version: Taylor, R, Ruiz Daniels, R, Dobie, R, Naseer, S, Clark, TC, Henderson, NC, Boudinot, P, Martin, SAM & Macqueen, D 2022, 'Single cell transcriptomics of Atlantic salmon (Salmo salar L.) liver reveals cellular heterogeneity and immunological responses to challenge by Aeromonas salmonicida', Frontiers in Immunology, pp. 1-17. https://doi.org/10.3389/fimmu.2022.984799 Link: Link to publication record in Edinburgh Research Explorer Document Version: Peer reviewed version Published In: Frontiers in Immunology Document Version: Peer reviewed version Take down policy Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact openaccess@ed.ac.uk providing details, and we will remove access to the work immediately and investigate your claim. Download date: 24. Oct. 2024 Single cell transcriptomics of Atlantic salmon (Salmo salar L.) liver reveals cellula heterogeneity and immunological responses to challenge by Aeromonas salmonicid Richard S. Taylor 1 †, Rose Ruiz Daniels 1 †, Ross Dobie 2,3, Shahmir Naseer 4, Thomas C. Clark 5, Neil C. Henderson 2,3, Pierre Boudinot 5, Samuel A.M. Martin 4, Daniel J. Macqueen 1* † These authors have contributed equally to this work and share first authorship. 1 The Roslin Institute and Royal (Dick) School of Veterinary Studies, The University of Edinburgh, Midlothian, UK 2 Centre for Inflammation Research, The Queen's Medical Research Institute, Edinburgh BioQuarter, University of Edinburgh, Edinburgh, UK 3 MRC Human Genetics Unit, Institute of Genetics and Cancer, University of Edinburgh, Edinburgh, UK 4 School of Biological Sciences, University of Aberdeen, Tillydrone Avenue, Aberdeen, UK 5 Université Paris-Saclay, INRAE, UVSQ, VIM, 78350, Jouy-en-Josas, France * Correspondence: Daniel J. Macqueen daniel.macqueen@roslin.ed.ac.uk Keywords: Liver, Single cell transcriptomics, Bacterial infection, Salmonid fish, Immune-Metabolism cross talk. Single cell transcriptomics of Atlantic salmon (Salmo salar L.) liver reveals cellular heterogeneity and immunological responses to challenge by Aeromonas salmonicida These authors have contributed equally to this work and share first authorship. 4 School of Biological Sciences, University of Aberdeen, Tillydrone Avenue, Aberdeen, UK Keywords: Liver, Single cell transcriptomics, Bacterial infection, Salmonid fish, Immune-Metabolism cross talk. Introduction The vertebrate liver is a multitasking organ with diverse physiological functions, including nutrient metabolism, transport and storage, growth signaling, endocrine regulation, and immunity (1). In mammals, these roles are performed by the cooperative actions of several distinct cell types including hepatocytes, cholangiocytes (epithelial cells of the bile duct), stellate cells, Kupffer cells (resident liver macrophages) and lymphocytes (1). Recent advances in single cell transcriptomics have revealed functional heterogeneity within the major hepatic cell types of mammals (2-3), providing insights into liver spatial organization (4-5) while revealing cellular and molecular drivers of disease and malignancy states (6-7). The adult liver of all vertebrates contains both immune and non-immune cells with important immunological functions (2, 7, 8-9) that support immune homeostasis and tolerance (10) and the generation of inflammatory responses upon pathogen challenge, leading to secretion of acute phase proteins (APPs) into circulation by hepatocytes (7, 11). The liver is the major site of haematopoiesis in the mammalian fetus and hence an important organ for early immune cell development (12), though this feature is not conserved in fishes (13). The multifaceted functions of the liver must demand tight coordination of different cell types to achieve appropriate responses to prevailing physiological and environmental conditions, inclusive of immune system status following pathogen challenge. Immunological functions may need to be prioritized at the cost of investment into metabolic functions in such scenarios (14). In this regard, the role of the liver in co-regulating metabolism and immunity makes it an interesting organ to understand the coordinated responses of different cell types following pathogenic challenge. Bulk RNA-Seq and proteomics have been widely applied to understand liver functions in commercially important fishes. For example, in Atlantic salmon (Salmo salar L.), which is among the most important aquaculture species globally (15), such work has uncovered the role of this organ in innate immune defense and the acute phase response (11,16). In the biomedical model zebrafish (Danio rerio), single cell transcriptomics has recently been used to reveal hepatic cellular heterogeneity, including within the hepatocyte, myeloid and lymphoid lineages (17-18). However, zebrafish are distantly related to salmonids, leaving a gap in knowledge on the role of liver cellular heterogeneity in this key group of fishes, where such information can be applied to understand and manipulate health and immune traits relevant to sustainable aquaculture and food production. Abstract The liver is a multitasking organ with essential functions for vertebrate health spanning metabolism and immunity. In contrast to mammals, our understanding of liver cellular heterogeneity and its role in regulating immunological status remains poorly defined in fishes. Addressing this knowledge gap, we generated a transcriptomic atlas of 47,432 nuclei isolated from the liver of Atlantic salmon (Salmo salar L.) contrasting control fish with those challenged with a pathogenic strain of Aeromonas salmonicida, a problematic bacterial pathogen in global aquaculture. We identified the major liver cell types and their sub-populations, revealing poor conservation of many hepatic cell marker genes utilized in mammals, while identifying novel heterogeneity within the hepatocyte, lymphoid, and myeloid lineages. This included polyploid hepatocytes, multiple T cell populations including γδ T cells, and candidate populations of monocytes/macrophages and dendritic cells. A dominant hepatocyte population radically remodeled its transcriptome following infection to activate the acute phase response and other defense functions, while repressing routine functions such as metabolism. These defense-specialized hepatocytes showed strong activation of genes controlling protein synthesis and secretion, presumably to support the release of acute phase proteins into circulation. The infection response further involved up-regulation of numerous genes in an immune-cell specific manner, reflecting functions in pathogen recognition and killing, antigen presentation, phagocytosis, regulation of inflammation, B cell differentiation and T cell activation. Overall, this study greatly enhances our understanding of the multifaceted role played by liver immune and non-immune cells in host defense and metabolic remodeling following infection and provides many novel cell-specific marker genes to empower future studies of this organ in fishes. 2 2 Introduction The aim of this study was to reveal the major cell lineages within the liver of Atlantic salmon and uncover the role of hepatic cellular heterogeneity in the host response to bacterial infection. Using single-nuclei RNA-Seq (snRNA-Seq), we report a comprehensive single cell transcriptomic atlas of the Atlantic salmon 3 liver, identifying novel heterogeneity across multiple hepatic cell types. By comparing cell-specific responses in control animals to those challenged with Aeromonas salmonicida, we uncover a dramatic transcriptomic remodeling of hepatocytes that underpins the acute phase response, alongside major changes in gene expression specific to distinct immune cell populations. Materials and Methods Disease challenge and sampling Animal work was carried out in compliance with the Animals (Scientific Procedures) Act 1986 under Home Office license PFF8CC5BE and was approved by the ethics committee of the University of Aberdeen. Atlantic salmon were kept in 250 L freshwater tanks at aquarium facilities of the University of Aberdeen. Water temperature was maintained at 14°C and fish were fed a commercial pellet diet at 2% body weight per day. After two weeks, twenty fish were anaesthetized using 2-phenoxyethanol (2.5 mL in 10L water/ 0.0025% v/v) and given an intraperitoneal injection of either PBS (0.5 mL) (n=10), or the pathogenic Hooke strain of A. salmonicida (2 × 105 colony-forming units / mL in PBS; 0.5 mL/fish) (n=10). Sampling was performed 24 h post-injection (after ref 11). All twenty fish were killed using a Schedule 1 method following anesthesia overdose using 2-phenoxyethanol (0.1% v/v) and destruction of the brain. Fish were immediately sampled, with liver samples (approx. 100mg) flash frozen on dry ice before storage at -80°C prior to snRNA-Seq library construction. Separate liver samples were placed in 1.5ml of Tri Reagent (Sigma-Aldrich) and used for quantitative PCR (qPCR) to validate the expected response in infected fish. Validation of immune response by qPCR qPCR validation was performed on the twenty liver samples (comparing n=10 control vs. n=10 infected animals) to confirm an inflammatory response in the infected fish. RNA was extracted from 100 mg liver tissue using 1 mL Trizol reagent. The tissue homogenization was performed using two Tungsten Carbide Beads (Qiagen) (3 mm) on a Tissuelyser II (Qiagen) at a frequency of 30.0 I/sec for 2.5 min. 200 µl of chloroform (Sigma-Aldrich) was added and the mixture centrifuged at 12,000 g for 20 minutes at 4 °C to separate the aqueous phase, which was retained. RNA precipitation was performed using 700 µl of isopropanol (Sigma-Aldrich) and centrifuged at 12,000 g for 20 min at 4 °C and washed 3 times with 80% ethanol. The concentration and purity of total RNA was estimated using a NanoDrop 1000 Spectrophotometer (Thermo Scientific). A QuantiTect Reverse Transcription kit (Qiagen) was used to synthesize first-strand cDNA from 1 µg total RNA per sample, with a genomic DNA removal step included, in a total volume of 20 µl. The resulting cDNA was diluted 20-fold (working stock) with RNase/DNase free water. qPCR was performed using primers targeting the APP encoding genes saa and hamp (11). For normalization of gene expression, primers targeting rps13 and rps29 were used (19). qPCR was performed with 2x SYBR Green I (Invitrogen) master mix on a Mx3005P System (Agilent 4 Technologies). Each assay was run in triplicate using 15 µl of reaction mix containing 7.5 µl Brilliant III Ultra-Fast SYBR Green (Agilent Technologies), 500 nM forward and reverse primers and 5 µl of cDNA (2.5 ng of total reverse-transcribed RNA). Assays were run with 1 cycle of 95 °C for 3 min, followed by 40 cycles of 95 °C for 20 s and 64 °C for 20 s. A melting curve (thermal gradient 55 °C - 95 °C) was used to confirm single qPCR products. Each qPCR plate included two no-template (i.e. water) controls. LinRegPCR (20) was used to establish the efficiency of each assay. Gene expression data were analyzed using GeneX 5.4.3 (MultiD Analysis), correcting for differences in efficiency across genes, before target gene expression was normalized to rps13 and rps29 and placed on a scale of relative expression. Separate Kruskal-Wallis tests were performed in Minitab® 17.1.0 to compare gene expression values for saa and hamp between the control (n=10) and infected (n=10) liver samples. snRNA-Seq library construction Based on the results of qPCR, n=4 samples were taken forward for snRNA-Seq library construction, representing n=2 control fish and n=2 infected fish (see Results; Supplementary Figure 1). A protocol adapted from ref (21) was used for nuclear extraction, employing a tween with salt and tris (TST) buffer. Approximately 45 mg of each frozen liver sample was placed in a 6-well tissue culture plate (Stem Cell Technologies) with 1 mL TST (2 mL of 2X ST buffer + 120 µL of 1% Tween-20 + 20 µL of 2% BSA brought up to 4 mL with nuclease-free water). The tissue was minced using Noyes spring scissors for 10 min on ice. The resulting homogenate was filtered through a 40 µm Falcon cell strainer, and a further 1 mL of TST was added to wash the well and filter. The volume was brought up to 5 mL using 3 mL of 1X ST buffer (diluted from 2xST buffer [292 µl of 146 mM NaCl, 100 µl of 10 mM Tris-HCl pH 7.5, 10 µl of 1 mM CaCl2, 210 µl of 21 mM MgCl2, brought up to 10 mL with nuclease-free water]). The sample was centrifuged at 4°C for 5 min at 500g before the resulting pellet was re-suspended in 1 mL 1X ST buffer and the recovered nuclei were filtered through a 40 µm Falcon cell strainer, Hoechst stained, visually inspected under a fluorescent microscope, and counted using a Bio-Rad TC20. Liver nuclei were processed through the 10X ChromiumTM Single Cell Platform using the ChromiumTM Single Cell 30 Library and Gel Bead Kit v3.1 and ChromiumTM Single Cell A Chip Kit (both 10X Genomics) as per the manufacturer’s protocol. For each sample, the nuclei were loaded into a channel of a Chromium 3’ Chip and partitioned into droplets using the Chromium controller before the captured RNA for each nucleus was barcoded and reverse transcribed. The resulting cDNA was PCR amplified for 14 cycles, fragmented, and size selected before Illumina sequencing adaptor and sample indexes were attached. Libraries were sequenced on a NovaSeq 6000 by Novogene UK Ltd (2x150bp paired end reads). Nuclei filtering and quality control Nuclei filtering was performed manually on each of the four samples using Seurat v3.1 (24). Ranked barcode plots of UMIs and gene counts were used to identify the lower ‘elbow point’ and cell barcodes with UMI count or gene counts below the elbow point removed as empty droplets. Further steps were used to identify additional empty droplets: the “SCTransform” function (25) was used to normalise the data, prior to centering, scaling, principal component analysis (PCA) and high-resolution clustering (using 30 PCs and resolution of 2). Wilcoxon rank sum differential gene expression tests were used to identify up-regulated genes in each cluster (see Methods Section ‘Differential gene expression tests’). Cell clusters that both lacked distinguishing markers and had a low median UMI or gene counts (typically 2 median absolute deviations lower than the median across all nuclei) were removed as likely empty droplets or poor-quality nuclei. This process was repeated iteratively on each sample until all such low-quality populations were removed. Likely doublets were identified and removed later in the analysis, after cell identity was established. Generation of snRNA-Seq count matrix Raw sequencing data were aligned to the unmasked ICSASG_v2 reference assembly (Ensembl release 104) of the Atlantic salmon genome (22). The analysis was restricted to protein coding genes. Mapping of reads to the genome, assignment of reads to cellular barcodes, and collapsing of unique molecular 5 identifiers (UMIs) was performed with StarSolo v2.7.7a (23). The genome index was generated with standard settings and “sjdbOverhang” set to 149. The reads were then mapped with the “STAR” command and following settings: “soloType=CB_UMI_Simple”, “outFilterMultimapNmax = 20”, “outMultimapperOrder=random”, “soloUMIdedup=1MM_Directional”, “soloFeatures = GeneFull”, “soloBarcodeReadLength = 0”, “outFilterMatchNminOverLread = 0”, “soloCellFilter = TopCells 100000”. The top 100,000 cell barcodes ranked by UMI number were retained to ensure the capture of transcriptionally quiet nuclei, lost when using the automated StarSolo filtering algorithm. Mapping statistics for each snRNA-Seq sample are provided in Supplementary Table 1. Differential gene expression tests ff g p Throughout this study, differentially expressed genes were defined using the Seurat function “FindAllMarkers”, applying the Wilcoxon rank sum test with default cut-offs (multi-test adjusted p-value < 0.05, log2-fold change > 0.25, with expression of the gene in at least 20% of nuclei in the cluster tested). For the global analyses of major cell types, differential gene expression tests were performed for each defined cluster in turn versus all other nuclei in the dataset as the background. For the hepatocyte-specific analysis we compared each hepatocyte cluster to all other hepatocyte nuclei. For the immune cell nuclei, the comparisons made were between each T cell sub-cluster and all other T cell nuclei, each myeloid sub- cluster and all other myeloid nuclei, and between each of the other immune cell clusters (i.e. NK-like, Neutrophil and B cell clusters) and all other immune cell nuclei. To assess which genes were up-regulated in infection in the defined immune sub-clusters, a differential gene expression test was performed within each cluster to compare nuclei that originated from control vs. infected fish (Wilcoxon rank sum test, p- value < 0.05, log2-fold change > 0.25, with expression of the gene in at least 20% of nuclei from either the infected or control nuclei). Assignment of cellular identity Each cell cluster was assigned to one of the major liver cell lineages using a priori marker genes (Supplementary Table 2). Populations identified as erythrocytes and thrombocytes were not included in downstream analyses due to their likely origin from contaminating blood. Populations identified as hepatocytes, cholangiocytes, mesenchymal cells, endothelial cells, and immune cells were merged into five separate Seurat objects (using the Seurat “merge” function) for separate analyses of each cell lineage. Batch effects across samples were removed in each merged object with Harmony (26). Gene annotations were taken from the Ensembl annotation for Atlantic salmon, and in cases where the gene was not assigned a name, the name of the nearest annotated putative orthologue in rainbow trout (Oncorhynchus mykiss), zebrafish or mouse (Mus musculus) was used, or in some cases informed by BLASTp searches against the NCBI non-redundant database. 6 For each of the five major liver cell lineages retained, data was log normalised, then scaled and centered, before PCA was performed and used as the input to graph-based clustering, using the established Seurat 6 pipeline. The appropriate number of PCs to use for clustering of each sample to minimise technical noise, was determined through visualisations generated through the “ElbowPlot” and “DimHeatmap” functions. The resolution parameter in the “FindClusters” function was tuned to return biologically meaningful heterogeneity with each cell type. At this stage, a further quality control step was performed to remove doublets from each lineage. Specifically, differential gene expression tests were performed and clusters that exhibited canonical markers (Supplementary Table 2) or lineage distinguishing markers (Supplementary Table 3) from two distinct cell lineages yet lacked any unique distinguishing marker genes of their own, were removed as likely doublets, before the data was re-clustered. This process was repeated until no doublet populations remained. For the immune cells, the Seurat object was further split into T cell, Myeloid, NK-like, Neutrophil and B cell objects (see Results) and the same process was repeated (using markers in Supplementary Table 3) and cell sub-clusters identified, before each object was merged back into a single “immune cell” Seurat object. Finally, all five major lineages were merged into a final global liver cell atlas object, with cell identities retained from the cell lineage specific analyses. We used this strategy as opposed to global clustering, as markedly more biologically meaningful heterogeneity could be established in the cell lineage specific analyses. Differential gene expression tests Results Single-nuclei RNA-Seq atlas of the Atlantic salmon liver Single-nuclei RNA-Seq atlas of the Atlantic salmon liver Single-nuclei RNA-Seq atlas of the Atlantic salmon liver 7 Hepatocytes, the major epithelial cell type within the liver, have been underrepresented in mammalian scRNA-Seq analyses (2), which may be caused by damage occurring during the dissociation step. As Atlantic salmon are ectotherms, we were also concerned that enzymatic dissociation (typically done 7 at >30oC) would activate cell stress and heat shock responses. We consequently decided to generate a snRNA-Seq atlas of salmon liver, using nuclei isolated from freshly flash frozen samples, an approach expected to provide an accurate representation of cell diversity (e.g. 27-28). The profiled liver nuclei were from control fish (n=2) and animals infected by a pathogenic strain of Aeromonas salmonicida (n=2), the bacterial agent of furunculosis, a long-standing problem disease in salmonid aquaculture (29). The infected group were sampled 24 hours post Aeromonas challenge, previously shown to capture the inflammatory and acute phase response (11). The fish used for sequencing were further selected based on gene expression data using marker genes for the acute phase response (11), which were robustly and significantly up-regulated in the infected group (Supplementary Fig. 1). Across all samples, we generated 47,432 nuclei transcriptomes with median UMI and gene counts per nucleus of 2,105 and 1,065, respectively. This was split across control fish as 11,679 and 11,433 nuclei, and Aeromonas-challenged fish as 19,148 and 5,172 nuclei, respectively (Supplementary Table 1). The major cell lineages were identified using a guided graph-based clustering strategy (24), with cell identity assigned using a priori defined marker genes (Fig. 1a, b; markers in Supplementary Table 2). Hepatocytes comprised most of the profiled nuclei (88.1%), followed by cholangiocytes (4.3%), immune cells (3.5%), mesenchymal cells (2.6%) and endothelial cells (1.5%) (Supplementary Fig. 2; Supplementary Table 1). Clustering and identification of these major cell lineages was repeatable across individual samples, with each sample contributing a large proportion of the nuclei (Supplementary Fig. 2). Differential gene expression analysis revealed markers for each major liver cell type (Fig. 1c, d; Supplementary Table 3). Marker genes for candidate cell types and sub-populations are hereafter reported according primarily to annotations provided by Ensembl, or in cases where no Ensembl annotation was available, using supplementary BLAST homology searches against the NCBI database to support our inferences. While Ensembl annotation utilizes phylogenetic information to inform homology relationships, it may nonetheless fail to correctly capture orthology of Atlantic salmon genes to mammalian species, particularly for fast evolving and complex gene families. Single-nuclei RNA-Seq atlas of the Atlantic salmon liver We observed a split of hepatocytes into sub-populations explained largely by infection status, with 75.9% of ‘control-associated’ hepatocytes deriving from control fish and 73.3% of ‘infection-associated’ hepatocytes deriving from Aeromonas-challenged fish (expanded in next section). Significant markers for the latter were dominated by genes encoding APPs, many of which were observed to be expressed across all cell types (Fig. 2b). This is likely a consequence of the numerically dominant hepatocytes (Fig. 1a) leaking mRNA from highly expressed genes into the ambient RNA. Similar examples of this type of leakage can be observed in many liver scRNA-Seq datasets with, for example, hepatocyte expressed genes Alb and Hp being widely ‘expressed’ across non-hepatocyte cells in past mammalian studies (e.g. 3, 32). Significant markers for the latter were dominated by genes encoding APPs, many of which were observed to be expressed across all cell types (Fig. 2b). This is likely a consequence of the numerically dominant hepatocytes (Fig. 1a) leaking mRNA from highly expressed genes into the ambient RNA. Similar examples of this type of leakage can be observed in many liver scRNA-Seq datasets with, for example, hepatocyte expressed genes Alb and Hp being widely ‘expressed’ across non-hepatocyte cells in past mammalian studies (e.g. 3, 32). The immune compartment contained identifiable T and B cells, along with candidate populations of neutrophils, myeloid cells, and NK-like cells (Fig. 2b). The transcriptome of cholangiocyte nuclei was homogeneous, while limited heterogeneity was identified in the endothelial and mesenchymal cells (Fig. 1a; Supplementary Fig. 5 and 6). The endothelia sub-clusters had a clear biological interpretation, with Atlantic salmon orthologues to marker genes from mammals distinguishing arterial and venous derived endothelial cells (33) (Supplementary Fig. 5, marker genes in Supplementary Table 5). However, the mesenchymal sub-clusters were not readily biologically interpretable (Supplementary Fig. 6; Supplementary Table 6). Hepatocyte remodelling dominates the liver response to bacterial infection Single-nuclei RNA-Seq atlas of the Atlantic salmon liver Adding to this challenge, genetic orthology is not a prerequisite for conservation of gene function or expression, which limits our ability to transfer knowledge about cell marker genes from mammalian studies to Atlantic salmon. While the reader must be aware of these caveats, they represent a general issue in functional genomics studies using non-model taxa like salmonids. While it was possible to identify the five major liver cell lineages of Atlantic salmon using orthologues to marker genes defined in mammalian liver scRNA-Seq datasets (e.g. 6), many markers were notably absent or expressed at very low levels in our snRNA-Seq dataset (Supplementary Fig. 3). For example, Atlantic salmon orthologues of the widely used epithelial marker Epcam did not show expression in the cholangiocyte cluster (Supplementary Fig. 3). Likewise, salmon orthologues of Pecam1 and Pdgfrb, which are excellent markers of mammalian endothelial and mesenchymal cells, were not detected at 8 significant levels in these cell types in our dataset (Supplementary Fig. 3). These results may be explained by differences in transcriptome composition between snRNA-Seq and scRNA-Seq datasets from liver, as identified in humans (30). In support of this idea, epcam and pecam1 were respective markers of cholangiocytes and endothelial cells in a recently published zebrafish scRNA-Seq dataset (31), making it less likely that the lack of expression of these markers in our Atlantic salmon dataset represents a true evolutionary difference between mammals and teleosts. Differences in expression were also observed between predicted Atlantic salmon orthologues of mammalian marker genes for the major hepatic cell types, with, for example, only one of two cdh5 co-orthologues marking the endothelial population, and only one of two hnf4a co-orthologues marking hepatocytes (Supplementary Fig. 3). Higher resolution clustering captured varying degrees of transcriptomic heterogeneity for each of the five major liver cell types (Fig. 2a, b; Supplementary Table 4), which was consistent across the four samples (Supplementary Fig. 4). We observed a split of hepatocytes into sub-populations explained largely by infection status, with 75.9% of ‘control-associated’ hepatocytes deriving from control fish and 73.3% of ‘infection-associated’ hepatocytes deriving from Aeromonas-challenged fish (expanded in next section). Higher resolution clustering captured varying degrees of transcriptomic heterogeneity for each of the five major liver cell types (Fig. 2a, b; Supplementary Table 4), which was consistent across the four samples (Supplementary Fig. 4). Hepatocyte remodelling dominates the liver response to bacterial infection 9 To explore how hepatocyte heterogeneity contributes to the response to Aeromonas infection, we analysed 41,792 available hepatocyte nuclei transcriptomes. Clustering using the most variable genes in this compartment identified nine sub-populations (H1-H9) (Fig. 3a; marker genes in Supplementary Table 7), with several showing marked differences in abundance between control and infected fish (Fig. 3b). H1- H4 comprised 90.2% of hepatocyte nuclei, with H1 and H2 deriving mainly from control fish and showing 9 highly correlated transcriptomes (Fig. 3b, c). H3 and H4 comprised 70.0% of nuclei from infected fish and showed closely related transcriptome profiles (Fig. 3b, c). Hepatocyte nuclei derived from infected fish increased from H1 (16.9%), to H2 (28.2%), to H3 (64.3%) to H4 (80.8%), with 2,842 genes differentially expressed on this gradient (Fig. 3d, e; Supplementary Tables 8 and 9). 379 genes were up-regulated in infection-dominated H4 vs. control-dominated H1, showing overrepresented functions linked to host defense and the acute phase response (‘complement activation’, ‘defense response to other organism’, and ‘cellular iron homeostasis’), in addition to translational processes (e.g. ‘translational elongation’) (Fig. 3f, Supplementary Table 10). This response was dominated by genes encoding APPs including hepcidin, haptoglobins, ferritins, transferrin, fibrinogens, ceruloplasmin, angiotensinogen, serum albumins, apolipoproteins, and c-reactive protein (Supplementary Table 8). One of the top up-regulated genes (3.7-fold up-regulated; ENSSSAG00000046715) encodes mechanistic target of rapamycin kinase (mTOR), a master regulator of translation (34). mTOR biasedly promotes translation of ribosomal protein genes (34), which is notable as 28 such genes, encoding many proteins comprising the large and small ribosomal subunits of salmonid fishes (35), were up-regulated in H4 vs. H1, with only one downregulated (Supplementary Tables 8 and 9). We further observed up-regulation of copb1 (ENSSSAG00000068484), grp78 (ENSSSAG00000054661), sec61a1 (ENSSSAG00000039008), and srprb (ENSSSAG00000046456), encoding major components of the COPI and translocon complexes, representing key protein secretion pathways (36). H5 nuclei were mainly from infected fish (Fig. 3b) and showed highly correlated transcriptomes to H3/H4 (Fig. 3c), sharing many of the same key markers up-regulated in H4 vs. H1, but also specific markers (Supplementary Table 7) associated with NF-κB signaling, including relb (ENSSSAG00000052551), encoding a component of the NF-κB transcription factor complex. This is notable, as H5 also showed the highest expression among all hepatocyte sub-populations of stat3 (ENSSSAG00000003657), encoding signal transducer and activator of transcription 3, which acts downstream of NF-κB in mammalian hepatocytes to activate inflammation driving the acute phase response (37). Hepatocyte remodelling dominates the liver response to bacterial infection 2,278 genes were downregulated in H4 vs. H1 (Supplementary Table 9), enriched in functions related to signaling (e.g. ‘intracellular receptor signaling pathway’), metabolism (e.g. ‘gluconeogenesis’), cell differentiation (e.g. ‘stem cell differentiation’) and transcription (e.g. ‘regulation of transcription, DNA- templated’) (Fig. 3f) (Supplementary Table 11). The top downregulated gene was prkag2a (4.0-fold down-regulated; ENSSSAG00000079550), encoding a subunit of the AMPK complex - a master regulator of metabolism including gluconeogenesis (38). Also downregulated were master hepatic transcription factors for lipid and glucose metabolism pathways that interact with AMPK (38), including genes encoding hepatocyte nuclear factor 1 (HNF1) (ENSSSAG00000006158) and HNF4 10 (ENSSSAG00000047055) (39), carbohydrate-responsive element-binding protein (ENSSSAG00000039257) (40), and forkhead box proteins O1 / O3 (ENSSSAG00000054669 / ENSSSAG00000055241) (41) (Supplementary Table 9). Further evidence for repression of anabolism included downregulation of genes encoding growth hormone receptor (GHR) (ENSSSAG00000065355 and ENSSSAG00000081526) and Stat5a/b (ENSSSAG00000010616, ENSSSAG00000003584 and ENSSSAG00000048873), which act downstream of GHR to activate growth and cell proliferation genes (42). In addition, a gene was downregulated encoding glucocorticoid receptor (GR) (ENSSSAG00000062169), which promotes expression of gluconeogenesis genes, while its interaction with Stat5 is required for transcriptional activation of growth genes via GHR signalling (42). (39), carbohydrate-responsive Additional hepatocyte heterogeneity includes polyploid cells The remaining hepatocyte subclusters were not linked to infection status (Fig. 3b). H7 likely represents hepatocytes that have undergone polyploidization, which occurs progressively during aging in mammals, such that 4n-16n cells make up a large fraction of liver cells by adulthood (43). Polyploidy in H7 is indicated by a striking concordance of genes representing highly specific markers for H7 and those shown elsewhere to be up-regulated in 4n vs. 2n mammalian cells, encoding DNA primase subunit 2 (ENSSSAG00000001875), replication protein A 70 kDa DNA-binding subunit (ENSSSAG00000050927, ENSSSAG00000050209) and DNA polymerase (ENSSSAG00000078390), among others (44). The close relationship of H7 to H1 (Fig. 3a, c) implies that polyploid hepatocytes derive from those supporting routine metabolism. H9 showed many specific markers encoding mitosis proteins, e.g. cytoskeleton-associated protein 5 (ENSSSAG00000066206), centromere protein E (ENSSSAG00000073454), abnormal spindle-like microcephaly-associated protein homolog (ENSSSAG00000053226) and kinesin family member 23 (ENSSSAG00000044381), indicating these are cycling hepatocytes. H6 expressed a small number of highly specific markers, including genes encoding ligand of numb-protein X1 (lnx1) (ENSSSAG00000072728 and ENSSSAG00000070275), a E3 ubiquitin ligase that targets a wide range of proteins, including CD8 expressed on T-cells in mammals (45). Hepatocyte remodelling dominates the liver response to bacterial infection H8 was biasedly represented by control fish and most correlated with H1/H2 in transcriptome profile (Fig. 3c), expressing marker genes described as ‘novel’ in the Ensembl annotation (Supplementary Table 7). Immune cell heterogeneity in the Atlantic salmon liver T1 and T2 therefore appear to be constituted mainly of resting T cells. T3 appears to be an activated T cell population based on specific expression of slamf1 (ENSSSAG00000043093) encoding CD150 (54). T3 expressed many highly specific markers, including the integrin coding gene itgal (ENSSSAG00000046537; encoding CD11a) and itgb2 (ENSSSAG00000022772; encoding CD18), whose products form lymphocyte function-associated antigen 1 (LFA-1), a molecule with key roles in T cell activation and migration, in addition to cytotoxic and memory responses (55). In mammals, LFA-1 is an established marker for the migration of liver T resident memory (TRM) cells into liver, which is also the case for CD103 (also known as ITGAE) (56), which was expressed more highly in T3 (ENSSSAG00000076346) than other T sub-clusters. The human orthologue of a highly expressed T3 specific marker, cacna2d2a (ENSSSAG00000071299), encoding a calcium voltage-gated channel, was not detected in any immune cell in the human protein atlas, implying a teleost-specific T cell marker. T4 expressed several activation markers including pou2f2a (aka oct2) (ENSSSAG00000071136) (57), CD226 (58), along with two distinct ctram genes (encoding cytotoxic and regulatory T-cell molecule), Five T cell sub-clusters were compared using differential expression tests (Fig. 4a; Supplementary Table 12). Only T1 and T2 expressed cd4 genes to low levels, specifically ENSSSAG00000076631, encoding CD4-1, which is also expressed by macrophages, and ENSSSAG00000076595, encoding CD4-2; shown elsewhere to be expressed by all CD4+ T cells (48). We did not identify any CD8 expressing T cells, likely due to low expression levels. T1 was the largest sub-cluster but showed few specific markers. T2 expressed genes encoding receptors involved in T cell activation. This included cd28 (ENSSSAG00000060163), encoding the main co-stimulatory T cell receptor (49) and cd44 (ENSSSAG00000076128), a widely used T cell adhesion, co-stimulation and activation marker (50). However, T2 cells did not specifically express any genes annotated as ctla-4, an IgSF member induced during T cell activation that regulates CD28 activity. Furthermore, T1 and T2 both expressed tcf7 (ENSSSAG00000006857) at a much higher level than T3-T5, encoding the master Wnt pathway transcription factor, which is most highly expressed on naïve mammalian T cells (51) and was a specific marker for resting CD4+ T cells in humans (52). T1 and T2 also expressed foxp1b (ENSSSAG00000077820) at a higher level than T3-5; a gene essential for quiescent naive T cells in mammals (53). T1 and T2 therefore appear to be constituted mainly of resting T cells. Immune cell heterogeneity in the Atlantic salmon liver 3.4% (n=1,620) of the liver nuclei were derived from immune cells (Fig. 4a, b). Combinations of canonical marker expression was used to classify T cells (cd3e, tox2 and tcf7), B cells (ighm, cd37, cd79a), NK-like cells (prf1.3 and runx3 with absence of cd3e) and myeloid cells (mpeg1, cd63, csf1r, lyz2) (Fig. 4b; marker genes provided in Supplementary Table 12). We also identified a candidate population of neutrophils based on marker genes that showed highest specificity or expression for 11 neutrophils among different immune cells in the human protein atlas, namely itgax (ENSSSAG00000049715), ncf1 (ENSSSAG00000079828), and mmp9 (ENSSSAG00000069874) (Fig. 4b). Mmp9 was an effective marker for neutrophils in other teleost species, and has a role in driving neutrophil migration in mammals (46, 47). The NK-like cells represent a tentative annotation owing to a lack of certain markers for NK cells, i.e. genes encoding granzymes. Each immune cluster had substantial contributions from all samples, except the candidate neutrophils, which derived mainly from one Aeromonas-challenged fish (Supplementary Fig. 4). Sub-clustering revealed heterogeneity in the T and myeloid cells, but not the NK-like cells, B cells (with no evidence of plasma cells) or candidate neutrophils. Five T cell sub-clusters were compared using differential expression tests (Fig. 4a; Supplementary Table 12). Only T1 and T2 expressed cd4 genes to low levels, specifically ENSSSAG00000076631, encoding CD4-1, which is also expressed by macrophages, and ENSSSAG00000076595, encoding CD4-2; shown elsewhere to be expressed by all CD4+ T cells (48). We did not identify any CD8 expressing T cells, likely due to low expression levels. T1 was the largest sub-cluster but showed few specific markers. T2 expressed genes encoding receptors involved in T cell activation. This included cd28 (ENSSSAG00000060163), encoding the main co-stimulatory T cell receptor (49) and cd44 (ENSSSAG00000076128), a widely used T cell adhesion, co-stimulation and activation marker (50). However, T2 cells did not specifically express any genes annotated as ctla-4, an IgSF member induced during T cell activation that regulates CD28 activity. Furthermore, T1 and T2 both expressed tcf7 (ENSSSAG00000006857) at a much higher level than T3-T5, encoding the master Wnt pathway transcription factor, which is most highly expressed on naïve mammalian T cells (51) and was a specific marker for resting CD4+ T cells in humans (52). T1 and T2 also expressed foxp1b (ENSSSAG00000077820) at a higher level than T3-5; a gene essential for quiescent naive T cells in mammals (53). Immune cell heterogeneity in the Atlantic salmon liver T3 appears to be an activated T cell population based on specific expression of slamf1 (ENSSSAG00000043093) encoding CD150 (54). T3 expressed many highly specific markers, including the integrin coding gene itgal (ENSSSAG00000046537; encoding CD11a) and itgb2 (ENSSSAG00000022772; encoding CD18), whose products form lymphocyte function-associated antigen 1 (LFA-1), a molecule with key roles in T cell activation and migration, in addition to cytotoxic and memory responses (55). In mammals, LFA-1 is an established marker for the migration of liver T resident memory (TRM) cells into liver, which is also the case for CD103 (also known as ITGAE) (56), which was expressed more highly in T3 (ENSSSAG00000076346) than other T sub-clusters. The human orthologue of a highly expressed T3 specific marker, cacna2d2a (ENSSSAG00000071299), encoding a calcium voltage-gated channel, was not detected in any immune cell in the human protein atlas, implying a teleost-specific T cell marker. T4 expressed several activation markers including pou2f2a (aka oct2) (ENSSSAG00000071136) (57), CD226 (58), along with two distinct ctram genes (encoding cytotoxic and regulatory T-cell molecule), 12 previously shown in mammals to be required for differentiation of cytotoxic CD4+ T cells (59). T5 specifically expressed two paralogues of sox13 (ENSSSAG00000077869, ENSSSAG00000058488), the defining vertebrate transcription factor for the γδ T lineage (60). My1 and My2 markers are associated with monocytes and macrophages (Supplementary Table 12). My1 specifically expressed cd4-1 (ENSSSAG00000076631), at a level higher than any T sub-cluster, likely representing a phagocytic CD4+ macrophage characterized in rainbow trout (48). My2 expressed specific monocyte marker genes including csf3r (ENSSSAG00000041566) and timp2 (ENSSSAG00000042353 and ENSSSAG00000064056). Two csf1r copies were identified with reciprocal higher expression in My1 (ENSSSAG00000004088) and My2 (ENSSSAG00000061479). High flt3 (ENSSSAG00000009390) expression in My4 supports an annotation as dendritic cells (DCs) (61), with specific up-regulation of cd9 (ENSSSAG00000059637) and lamp2 (encoding CD107b) (ENSSSAG00000074801) genes, consistent with monocyte-derived DCs in mammals (62, 63). My3 expressed the second highest level of flt3, while the top My3 marker gene, ptprsa (ENSSSAG00000051752), is a specific marker for plasmacytoid DCs (pDCs) in mice and human (64). The relationship of My3 to pDCs is also supported by specific expression of sbf1 (ENSSSAG00000071635) (65). However, another My4-specific marker gene, tcf4 (ENSSSAG00000071044), encodes a transcription factor required for pDC development (66). Immune cell-specific responses to Aeromonas insult In mammals, prosaposin encodes the precursor to all saposin lysosomal proteins, which are known to have antibacterial activity and play a key role in presenting lipid antigens to CD1-restricted T cells (76). Among the T cell populations, T3, T4, and the γδ T cells (T5), showed the strongest responses to Aeromonas infection. The top up-regulated genes in T3 included csl1 (4.3-fold up-regulated; ENSSSAG00000004327), encoding a L-rhamnose-binding lectin that binds bacteria and enhances phagocytosis in salmonids (71), and c8b (4.1-fold up-regulated; ENSSSAG00000073702), encoding a core component of the complement membrane attack complex. In T4, up-regulated genes included tln1, encoding Talin-1 (4.8-fold up-regulated; ENSSSAG00000063331), which is known to regulate the integrin LFA-1 complex (defining T3; see last section), and is required for sustained interactions between APCs and T cells, as well as T cell proliferation (72). T4 also up-regulated itgb1 (encoding CD29, also called β1-integrin) (4.8-fold up-regulated; ENSSSAG00000007621), a signature marker for cytotoxic T cells in humans (73). In T5, among the top-up-regulated genes was prex1 (4.8-fold up-regulated; ENSSSAG00000044871), a signaling molecule that promotes expression of key interleukin cytokines in activated human T cells, including IL-2 (74). T5 also up-regulated catl1, encoding cathepsin L (8.0-fold up-regulated; ENSSSAG00000077309), which regulates T cell cytotoxicity (75) and an unannotated gene encoding a protein with saposin-like domains, which is annotated as prosaposin-like in NCBI (7.4-fold up-regulated; ENSSSAG00000009411). In mammals, prosaposin encodes the precursor to all saposin lysosomal proteins, which are known to have antibacterial activity and play a key role in presenting lipid antigens to CD1-restricted T cells (76). My1 specifically up-regulated 17 genes, including a different itgal paralogue (encoding CD11A) to that noted as a marker for T3 (4.2-fold up-regulated; ENSSSAG00000046996), encoding a component of LFA-1 essential to the immune response of mice to Mycobacterium tuberculosis, supporting T cell- mediated activation and recognition of infected macrophages (77). My1 also specifically up-regulated olfm4 (4.1-fold upregulated; ENSSSAG00000046003), encoding a glycoprotein induced in mice macrophages by Helicobacter pylori infection, which regulates inflammatory responses (78). My1 further specifically up-regulated cats (ENSSSAG00000070942) and ctsd (ENSSSAG00000027269), encoding cathepsin S and D, proteolytic enzymes with established macrophage roles in bacterial killing and antigen processing (79). My2 specifically up-regulated 32 genes, including cath2 (6.6-fold upregulated ENSSSAG00000049319), encoding an antimicrobial peptide that increased in abundance in response to Aeromonas infection in salmon plasma (80) and abr (6.6-fold upregulated; ENSSSAG00000080204), encoding a GTPase-activating protein that down-regulates the inflammatory actions of macrophages (81). Immune cell-specific responses to Aeromonas insult To uncover the role of hepatic immune cell heterogeneity in the response to Aeromonas challenge, we performed differential expression tests comparing nuclei from control and infected fish within each immune sub-cluster (Fig. 4c, d). 819 genes showed significant up-regulation in nuclei from infected fish (criteria: P < 0.05; Log2FC > 1), among which 274 (33%) and 72 (8.8%) were up-regulated by most (≥7 of the 12 sub-clusters) or all immune cell types, respectively, and 271 (33%) showed immune sub-cluster specific up-regulation (Fig. 4c; Supplementary Table 13). B cells specifically up-regulated two genes, namely stat3 (2.4-fold up-regulated; ENSSSAG00000003657), which is essential for B cell differentiation (67) and a gene annotated zgc:174904 (2.1-fold up-regulated; ENSSSAG00000070511), encoding a 304 amino acid protein with a CD209/DC-Sign-like, C-type lectin-like domain (InterPro domain: IPR033989). C-type lectin/DC-SIGN is a broad-specificity PRR that detects bacteria by binding mannose or carbohydrate structures (68). In the candidate neutrophils, the top up-regulated gene was ladderlectin (38.9-fold up-regulated; ENSSSAG00000039613), encoding a soluble lectin that bound Aeromonas in salmonids, leading to bacterial killing actions (69). The NK-like cells up-regulated few genes specifically, one of which was pglyrp2 (2.3-fold up-regulated; ENSSSAG00000054105), a peptidoglycan recognition protein with enzymatic activity targeting and limiting the inflammatory effects of bacterial peptidoglycan (70). 13 Among the T cell populations, T3, T4, and the γδ T cells (T5), showed the strongest responses to Aeromonas infection. The top up-regulated genes in T3 included csl1 (4.3-fold up-regulated; ENSSSAG00000004327), encoding a L-rhamnose-binding lectin that binds bacteria and enhances phagocytosis in salmonids (71), and c8b (4.1-fold up-regulated; ENSSSAG00000073702), encoding a core component of the complement membrane attack complex. In T4, up-regulated genes included tln1, encoding Talin-1 (4.8-fold up-regulated; ENSSSAG00000063331), which is known to regulate the integrin LFA-1 complex (defining T3; see last section), and is required for sustained interactions between APCs and T cells, as well as T cell proliferation (72). T4 also up-regulated itgb1 (encoding CD29, also called β1-integrin) (4.8-fold up-regulated; ENSSSAG00000007621), a signature marker for cytotoxic T cells in humans (73). In T5, among the top-up-regulated genes was prex1 (4.8-fold up-regulated; ENSSSAG00000044871), a signaling molecule that promotes expression of key interleukin cytokines in activated human T cells, including IL-2 (74). T5 also up-regulated catl1, encoding cathepsin L (8.0-fold up-regulated; ENSSSAG00000077309), which regulates T cell cytotoxicity (75) and an unannotated gene encoding a protein with saposin-like domains, which is annotated as prosaposin-like in NCBI (7.4-fold up-regulated; ENSSSAG00000009411). Immune cell-specific responses to Aeromonas insult My3 specifically up-regulated 16 genes, including il12p40b2 (5.7-fold up-regulated; ENSSSAG00000069633), best known as a component of IL-12 and/or IL-23 heterodimers, but that also has defined cytokine functions as a monomer protein, including promotion of DC migration in response to bacterial infection in mammals (82, 83). Another induced gene was afp4 (3.7-fold up-regulated; ENSSSAG00000072959), encoding type IV ice-structuring protein LS-12, an apolipoprotein-like molecule that dramatically increases in abundance in salmon plasma following Aeromonas infection (84). 14 My4 specifically up-regulated 64 genes, including gas7 (8.56-fold up-regulated; ENSSSAG00000076110), which has a crucial role in phagocytosis (85), dnase1l3 (8.4-fold up-regulated; ENSSSAG00000066441), which controls inflammasome-induced cytokine secretion (86), and cd82 (6.9- fold up-regulated; ENSSSAG00000052206), previously shown to be up-regulated during DC activation, where is promotes stable interactions between DCs and T cells, and MHC-II maturation (87). My4 further upregulated irf4 (6.28-fold up-regulated; ENSSSAG00000039730), a gene essential to the ability of DCs to promote Th2 differentiation and inflammation (88). Discussion This study greatly enhances our knowledge of liver function in a salmonid fish with global commercial and scientific importance. The major advancement compared to previous work comes from our application of snRNA-Seq, which, in contrast to previously past bulk transcriptomic or proteomic studies, allowed us to identify multiple hepatic cell populations, before dissecting the role of this heterogeneity in host defense following bacterial infection. Furthermore, a plethora of novel marker genes are reported for developmentally and functionally diverse hepatic cell types, which will be useful for future studies investigating traits relevant to salmonid health and immunological status. Our results demonstrate the essential contribution of hepatocytes to the antibacterial and acute phase response in Atlantic salmon. Transcriptomic heterogeneity in the dominant hepatocyte population (i.e., H1-4; Fig. 3) was inconsistent with distinct hepatocyte populations. Instead, our data supports a single hepatocyte population that can exist in radically distinct transcriptional states dependent on infection status. At one extreme are the hepatocytes that dominate the liver of healthy fish (H1/2), which appear to be performing routine functions controlled by master hepatic transcription factors and signaling pathways. Conversely, the liver of bacterially infected Atlantic salmon was dominated by hepatocytes (H3/4) that downregulated master hepatic pathways (e.g. controlling metabolic functions), and up-regulated a suite of genes encoding APPs and innate immune molecules. This includes many APPs routinely detected in Atlantic salmon plasma by proteomics (89) implying extremely high abundance. As most plasma proteins derive from liver, this aligns with our finding that these ‘defense-specialized’ hepatocytes strongly up- regulated mTOR, its target ribosomal protein-coding genes, and genes from protein secretion pathways, presumably to boost APP translation and secretion rates during the acute phase response. This striking repurposing of hepatocyte function upon infection illustrates the vital role these cells play as a hub for cross-talk between metabolism and immunity, presumably allowing energetic resources to be allocated towards clearing a pathogen at the short-term cost of limiting investment into routine hepatic functions e.g. supporting growth (14). 15 Interestingly, a sub-population of defense-specialized hepatocytes (i.e. H5; Fig. 3), almost exclusively derived from infected fish, specifically expressed genes associated with NF-κB signaling (e.g. relb was ~10-fold more highly expressed in this population compared to the average across the eight other hepatocyte subpopulations). Discussion In addition, H5, among all hepatocyte sub-populations, showed the highest expression of stat3, which is indispensable for activation of APP and protein secretion pathway gene expression during bacterial infection in mice, acting downstream of NF-κB (37, 90). As APP and secretory protein pathway genes were strongly up-regulated in the dominant sub-populations of defense-specialized hepatocytes (H3/4), which lacked significant relb expression, H5 may represent an intermediate hepatocyte state, where the activation of the APP response and associated secretory pathway is first initiated. A previous scRNA-Seq study of zebrafish liver failed to identify hepatocytes showing any equivalent specialization towards host defense (18), while another identified a minor hepatocyte population enriched for immune functions (17), which may be analogous to H3-H5. Differences with these past zebrafish studies may reflect the fact that both studies utilized control zebrafish lacking immune stimulation. However, the fact that H3-5 comprised a significant fraction of hepatocytes in our control fish could also be explained by differences in liver function, potentially indicative of a higher baseline inflammatory state in the liver of the Atlantic salmon population we studied. While representing a small population of liver nuclei, we also offer evidence for polyploid hepatocytes in Atlantic salmon. The functional role of polyploidy in mammalian liver remains ill-defined, despite extensive study over decades (43). Work done over 40 years ago showed that the liver of several teleosts contained polyploid hepatocytes (91), so our result is perhaps not unexpected. Polyploidy increases with aging in mammals (43), which may explain why this hepatocyte population was so limited in the fish used in our study, which were juveniles. However, polyploid hepatocytes were not reported in past scRNA-Seq studies of zebrafish liver, which included an 18-month-old adult population (17), representing half the adult lifespan for this species. More work is required to understand the role of hepatocyte polyploidy in teleost health and disease. It is important to acknowledge that comparing our results with liver scRNA-Seq studies in other species has limitations due to fundamental differences in experimental execution, which, for example, is linked to striking differences in the composition of the cells captured. For example, a preprint by ref (18) surprisingly identified T cells as the dominant liver population in zebrafish, with hepatocytes reflecting a smaller proportion than expected. Discussion While underrepresentation of hepatocytes has been observed in several mammalian liver scRNA-Seq studies, a separate zebrafish scRNA-Seq study identified hepatocytes as the most abundant liver cell type (17), albeit at a smaller fraction than for our snRNA-Seq atlas in Atlantic salmon. This perhaps illustrates a recognized benefit of snRNA-Seq compared to scRNA-Seq; a more accurate representation of the true tissue cell diversity (e.g. 27-28, 92). Considering our limited 16 knowledge of cellular diversity in most fish species, including salmonids, careful comparisons of results from scRNA-Seq and snRNA-Seq will be required to establish baseline expectations for future studies. This is the first single cell transcriptomic study to characterize immune cell heterogeneity in the liver of a salmonid fish, and the first single cell study in any teleost to characterize transcriptomic responses of specific hepatic immune cell subtypes to infection. Past scRNA-Seq studies of zebrafish liver paint very distinct pictures of lymphocyte diversity, with one reporting no B cells, and little T cell heterogeneity (17). Conversely, a recent preprint reported a small B cell population, and six T cell sub-clusters capturing distinct CD8 and CD4 subsets (18). While we also identified a single B cell population and multiple T cell sub-clusters, the identity of T cells was markedly less clear in our data, due partly to a general lack of cd4 and cd8 expression, which may reflect a limitation of snRNA-Seq, or the lower sequencing depth in our study compared to (18). However, unlike these previous studies, we identified a small population of sox13+ γδ T cells, an ancient vertebrate T cell subtype with roles bridging adaptive and innate immunity. In zebrafish, γδ T cells possess phagocytic ability and act as APCs that activate CD4+ T cells, inducing B cell proliferation (93). In mammals, γδ T cells increase dramatically in liver during inflammatory conditions (94) and produce IL-17 essential for the innate response to bacterial infection (95). While γδ T cells have not been reported among the plethora of immune cells reported to date in scRNA-Seq studies spanning different teleost species (96), we find them readily identifiable in Atlantic salmon by sox13 expression in multiple tissues (not shown). Consistent with their known functions, salmon liver γδ T cells up-regulated genes with roles spanning innate and adaptive immunity during the early response to Aeromonas infection. Discussion Our study also identified evidence for myeloid heterogeneity within the Atlantic salmon liver, including two candidate DC populations that showed a strong response to bacterial infection, up-regulating genes required for interactions with T cells, phagocytosis and inflammasome activation, suggesting conserved roles between mammals and salmonids, as shown elsewhere (97). DCs were reported in a recent scRNA- Seq study of zebrafish liver (31), but not in two other liver scRNA-Seq analyses from the same species (17-18). DCs have also been reported in Atlantic cod spleen (98). We observed two distinct candidate macrophage populations, matching the level of heterogeneity recently reported in two previous zebrafish studies (17-18). A natural question relates to the relationship of teleost macrophages and mammalian Kupffer cells (KCs). Past work largely agrees that KCs are amniote-specific and hence not present in teleost liver (99), including for Atlantic salmon (100). Consistent with this notion, a recent study in zebrafish identified that resident macrophages were not located in the sinusoidal space, and lacked phagocytic ability (101). However, a recent cross-vertebrate scRNA-Seq analysis of liver cell repertoires reported putative KCs in the liver of zebrafish based on a conserved transcriptomic signature with mammals (31). This study did not report any heterogeneity within the macrophages beyond the putative KCs, notably lacking any monocyte-derived macrophages known to make up a large proportion of 17 macrophages in mammalian livers (2, 3), and which share marker genes with our Atlantic salmon macrophage clusters. We failed to identify conserved expression of the signature marker genes for KCs (in addition to transcription factors associated with KCs) defined in this past study (31) in any of our myeloid sub-clusters from Atlantic salmon (Supplementary Figure 9). While KCs were not expected among our macrophage populations, it is worth noting the expression of the predicted Atlantic salmon orthologues of mammalian Clec4f (ENSSSAG00000040735 and ENSSSAG00000076658). Clec4f encodes a C-Type Lectin (also known as Kupffer cell receptor) expressed specifically in KCs to the exclusion of non-liver macrophages (102, 103). In our Atlantic salmon dataset, one predicted orthologue of Clec4f is expressed across all four myeloid populations (ENSSSAG00000076658), showing highest expression in My2, whilst the other (ENSSSAG00000040735) was expressed across all immune cell types (Supplementary Table 11). Furthermore, both genes were among the top up-regulated genes in hepatocytes following Aeromonas infection (Supplementary Table 8). Discussion These findings highlight the potential for major differences in marker gene cell-specific expression and presumably function for homologous genes shared by mammals and teleosts. Having robust marker genes for different cell-types is clearly essential for accurate studies of cell biology in any species. However, single cell transcriptomics studies performed to date, including this study, demonstrate that marker genes can vary markedly across species, especially for immune cells (96, 104). This demands a broader uptake of single cell transcriptomics in more species to define conserved from non-conserved marker genes, and to separate true biological or evolutionary differences in cell heterogeneity and associated marker genes from differences introduced by technical reasons discussed earlier, including the use of snRNA-Seq vs. scRNA-Seq. In salmonid fishes, the presence of an ancestral whole genome duplication, and the associated retention of numerous duplicated genes that have diverged extensively in tissue expression (22, 105-107), further challenges the transfer of knowledge on marker genes from model species - more work is required in this area. In summary, our comprehensive dissection of the Atlantic salmon liver using single cell transcriptomics has generated many species-specific marker genes for a range of immune and non-immune cell types that contribute to health and immunological traits of relevance to sustainable aquaculture. Our results can further be used to extract cell-specific information from existing and future bulk gene expression studies. Author contributions - Designed study: DJM, RRD, RST; coordinated sampling and disease challenge experiment: SAM; performed fish sampling: SN, SAM, DJM; performed quantitative PCR: SN; optimized nuclear isolation: RRD; generated snRNA-Seq libraries: RD, RRD; provided infrastructure for snRNA-Seq: NCH; performed bioinformatics: RST; interpreted immunological data: TCC, PB; drafted manuscript: DJM, RST, RRD; made figures and tables: RST, DJM; contributed to data interpretation and finalization of manuscript: all authors. 18 Acknowledgements - This study was funded by grants from the Scottish Universities Life Sciences Alliance (Technology Seed Funding Call), the University of Edinburgh’s Data Driven Innovation Initiative (Scottish Funding Council Beacon ‘Building Back Better’ Call), and the Biotechnology and Biological Sciences Research Council, including the institutional strategic programme grants BBS/E/D/10002071 and BBS/E/D/20002174 and the responsive mode grant BB/W005859/1. NCH is supported by a Wellcome Trust Senior Research Fellowship in Clinical Science (ref. 219542/Z/19/Z). For the purpose of open access, the authors have applied a Creative Commons Attribution (CC BY) license to any Author Accepted Manuscript version arising from this submission. Data Availability All data generated in this study has been made available in the GEO database (https://www.ncbi.nlm.nih.gov/geo/ ) under the accession GSE207655. References 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 2. MacParland SA, Liu JC, Ma XZ, Innes BT, Bartczak AM, Gage BK, et al. Single cell RNA sequencing of human liver reveals distinct intrahepatic macrophage populations. Nat Commun. (2018) Oct 22;9(1):4383. doi: 10.1038/s41467-018-06318-7 3. Ramachandran P, Dobie R, Wilson-Kanamori JR, Dora EF, Henderson BE, Luu NT, et al. Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature. (2019) 575(7783):512-518. doi:10.1038/s41586-019-1631-3 3. Ramachandran P, Dobie R, Wilson-Kanamori JR, Dora EF, Henderson BE, Luu NT, et al. Resolving the fibrotic niche of human liver cirrhosis at single-cell level. Nature. (2019) 575(7783):512-518. doi:10.1038/s41586-019-1631-3 4. Dobie R, Wilson-Kanamori JR, Henderson BE, Smith JR, Matchett KP, Portman JR, et al. Single-Cell Transcriptomics Uncovers Zonation of Function in the Mesenchyme during Liver Fibrosis. Cell Rep. (2019) 29(7):1832-1847.e8. doi:10.1016/j.celrep.2019.10.024 5. Ben-Moshe S, Itzkovitz S. Spatial heterogeneity in the mammalian liver. Nat Rev Gastroenterol Hepatol. (2019) 16(7):395-410. doi:10.1038/s41575-019-0134-x 5. Ben-Moshe S, Itzkovitz S. Spatial heterogeneity in the mammalian liver. Nat Rev Gastroenterol Hepatol. (2019) 16(7):395-410. doi:10.1038/s41575-019-0134-x 5. Ben-Moshe S, Itzkovitz S. Spatial heterogeneity in the mammalian liver. Nat Rev Gastroenterol Hepatol. (2019) 16(7):395-410. doi:10.1038/s41575-019-0134-x 6.Ramachandran P, Matchett KP, Dobie R, Wilson-Kanamori JR, Henderson NC. Single-cell technologies in hepatology: new insights into liver biology and disease pathogenesis. Nat Rev Gastroenterol Hepatol. (2020) 17(8):457-472. doi:10.1038/s41575-020-0304-x 19 7. Stamataki Z, Swadling L. The liver as an immunological barrier redefined by single-cell analysis. Immunology. (2020) 160(2):157-170. doi:10.1111/imm.13193 8. Robinson MW, Harmon C, O'Farrelly C. Liver immunology and its role in inflammation and homeostasis. Cell Mol Immunol. (2016) 13(3):267-276. doi:10.1038/cmi.2016.3 9. Zhao J, Zhang S, Liu Y, He X, Qu M, Xu G, et al.. Single-cell RNA sequencing reveals the heterogeneity of liver-resident immune cells in human. Cell Discov. (2020) Apr 28;6:22. doi: 10.1038/s41421-020-0157-z. 10. Heymann F, Tacke F. Immunology in the liver--from homeostasis to disease. Nat Rev Gastroenterol Hepatol. (2016) 13(2):88-110. doi:10.1038/nrgastro.2015.200 11. Causey DR, Pohl MAN, Stead DA, Martin SAM, Secombes CJ, Macqueen DJ. High-throughput proteomic profiling of the fish liver following bacterial infection. BMC Genomics. (2018) 19(1):719. doi:10.1186/s12864-018-5092-0 12 Popescu DM, Botting RA, Stephenson E, Green K, Webb S, Jardine L, et al. Decoding human fetal liver haematopoiesis. Nature. (2019) Oct;574(7778):365-371. doi: 10.1038/s41586-019-1652-y. 13 Mahony CB, Bertrand JY. How HSCs Colonize and Expand in the Fetal Niche of the Vertebrate Embryo: An Evolutionary Perspective. Front Cell Dev Biol. (2019) Mar;12;7:34. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 doi: 10.3389/fcell.2019.00034 14. Rauw WM. Immune response from a resource allocation perspective. Front Genet. (2012) 3:267. doi:10.3389/fgene.2012.00267 15. Houston RD, Macqueen DJ. Atlantic salmon (Salmo salar L.) genetics in the 21st century: taking leaps forward in aquaculture and biological understanding. Anim Genet. (2019) 50(1):3-14. doi:10.1111/age.12748 16. Moraleda CP, Robledo D, Gutiérrez AP, Del-Pozo J, Yáñez JM, Houston RD. Investigating mechanisms underlying genetic resistance to Salmon Rickettsial Syndrome in Atlantic salmon using RNA sequencing. BMC Genomics. (2021) 22(1):156. doi:10.1186/s12864-021-07443-2 20 17. Morrison JK, DeRossi C, Alter IL, Nayar S, Giri M, Zhang C, et al. Single-cell transcriptomics reveals conserved cell identities and fibrogenic phenotypes in zebrafish and human liver. Hepatol Commun. (2022) Jul;6(7):1711-1724. doi: 10.1002/hep4.1930. 17. Morrison JK, DeRossi C, Alter IL, Nayar S, Giri M, Zhang C, et al. Single-cell transcriptomics reveals conserved cell identities and fibrogenic phenotypes in zebrafish and human liver. Hepatol Commun. (2022) Jul;6(7):1711-1724. doi: 10.1002/hep4.1930. 17. Morrison JK, DeRossi C, Alter IL, Nayar S, Giri M, Zhang C, et al. Single-cell transcriptomics reveals conserved cell identities and fibrogenic phenotypes in zebrafish and human liver. Hepatol Commun. (2022) Jul;6(7):1711-1724. doi: 10.1002/hep4.1930. 18. Huang Y, Liu X, Wang H-Y, Chen JY, Zhang X, Li Y, et al. Single-cell transcriptome landscape of zebrafish liver reveals hepatocytes and immune cell interactions in understanding nonalcoholic fatty liver disease. (2022) bioRxiv: https://doi.org/10.1101/2022.02.06.479276 19. Macqueen DJ, Kristjánsson BK, Johnston IA. Salmonid genomes have a remarkably expanded akirin family, coexpressed with genes from conserved pathways governing skeletal muscle growth and catabolism. Physiol Genomics. (2010) 42(1):134-148. doi:10.1152/physiolgenomics.00045.2010 19. Macqueen DJ, Kristjánsson BK, Johnston IA. Salmonid genomes have a remarkably expanded akirin family, coexpressed with genes from conserved pathways governing skeletal muscle growth and catabolism. Physiol Genomics. (2010) 42(1):134-148. doi:10.1152/physiolgenomics.00045.2010 20. Ruijter JM, Ramakers C, Hoogaars WM, Karlen Y, Bakker O, van den Hoff MJ, et al. Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data. Nucleic Acids Res. (2009) 37(6):e45. doi:10.1093/nar/gkp045 20. Ruijter JM, Ramakers C, Hoogaars WM, Karlen Y, Bakker O, van den Hoff MJ, et al. Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data. Nucleic Acids Res. (2009) 37(6):e45. doi:10.1093/nar/gkp045 21. Slyper M, Porter CB, Ashenberg O, Waldman J, Drokhlyansky E, Wakiro I, et al. A single-cell and single-nucleus RNA-Seq toolbox for fresh and frozen human tumors. Nat Med. (2020) May;26(5):792- 802. doi: 10.1038/s41591-020-0844-1. 22. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Lien S, Koop BF, Sandve SR, Miller JR, Kent MP, Nome T, et al. The Atlantic salmon genome provides insights into rediploidization. Nature. (2016) May 12;533(7602):200-5. doi: 10.1038/nature17164. 23. Kaminow B, Yunusov D, Dobin A. STARsolo: accurate, fast and versatile mapping/quantification of single-cell and single-nucleus RNA-seq data (2021) bioRxiv: https://doi.org/10.1101/2021.05.05.442755 24. Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM, et al. Comprehensive Integration of Single-Cell Data. Cell. (2019) Jun 13;177(7):1888-1902.e21. doi: 10.1016/j.cell.2019.05.031. 25. Hafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression. Genome Biol. (2019) Dec 23;20(1):296. doi: 10.1186/s13059- 019-1874-1. 21 26. Korsunsky I, Millard N, Fan J, Slowikowski K, Zhang F, Wei K, et al. Fast, sensitive and accurate integration of single-cell data with Harmony. (2019) Nat Methods. Dec;16(12):1289-1296. doi: 10.1038/s41592-019-0619-0. 26. Korsunsky I, Millard N, Fan J, Slowikowski K, Zhang F, Wei K, et al. Fast, sensitive and accurate integration of single-cell data with Harmony. (2019) Nat Methods. Dec;16(12):1289-1296. doi: 10.1038/s41592-019-0619-0. 27. Wu H, Kirita Y, Donnelly EL, Humphreys BD. Advantages of Single-Nucleus over Single-Cell RNA Sequencing of Adult Kidney: Rare Cell Types and Novel Cell States Revealed in Fibrosis. J Am Soc Nephrol. (2019) 30(1):23-32. doi:10.1681/ASN.2018090912 28. Denisenko E, Guo BB, Jones M, Hou R, de Kock L, Lassmann T, et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome Biol. (2020) Jun 2;21(1):130. doi: 10.1186/s13059-020-02048-6. 28. Denisenko E, Guo BB, Jones M, Hou R, de Kock L, Lassmann T, et al. Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows. Genome Biol. (2020) Jun 2;21(1):130. doi: 10.1186/s13059-020-02048-6. 29. Park SY, Han JE, Kwon H, Park SC, Kim JH. Recent Insights into Aeromonas salmonicida and Its Bacteriophages in Aquaculture: A Comprehensive Review. J Microbiol Biotechnol. (2020) 30(10):1443- 1457. doi:10.4014/jmb.2005.05040 29. Park SY, Han JE, Kwon H, Park SC, Kim JH. Recent Insights into Aeromonas salmonicida and Its Bacteriophages in Aquaculture: A Comprehensive Review. J Microbiol Biotechnol. (2020) 30(10):1443- 1457. doi:10.4014/jmb.2005.05040 30. Andrews TS, Atif J, Liu JC, Perciani CT, Ma XZ, Thoeni C, et al. Single-Cell, Single-Nucleus, and Spatial RNA Sequencing of the Human Liver Identifies Cholangiocyte and Mesenchymal Heterogeneity. Hepatol Commun. (2022) Apr;6(4):821-840. doi: 10.1002/hep4.1854. 31. Guilliams M, Bonnardel J, Haest B, Vanderborght B, Wagner C, Remmerie A, et al. Spatial proteogenomics reveals distinct and evolutionarily conserved hepatic macrophage niches. Cell. 34. Saxton RA, Sabatini DM. mTOR Signaling in Growth, Metabolism, and Disease. Cell. (2017) 169(2):361-371. doi:10.1016/j.cell.2017.03.035 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 (2022) Jan 20;185(2):379-396.e38. doi: 10.1016/j.cell.2021.12.018. 32. Aizarani N, Saviano A, Sagar, Mailly L, Durand S, Herman JS, et al. A human liver cell atlas reveals heterogeneity and epithelial progenitors. Nature. (2019) Aug;572(7768):199-204. doi: 10.1038/s41586- 019-1373-2. 33. Kalucka J, de Rooij LP, Goveia J, Rohlenova K, Dumas SJ, Meta E, et al. Single-Cell Transcriptome Atlas of Murine Endothelial Cells. Cell. (2020) Feb 20;180(4):764-779.e20. doi: 10.1016/j.cell.2020.01.015 34. Saxton RA, Sabatini DM. mTOR Signaling in Growth, Metabolism, and Disease. Cell. (2017) 169(2):361-371. doi:10.1016/j.cell.2017.03.035 22 35. Causey DR, Kim JH, Stead DA, Martin SA, Devlin RH, Macqueen DJ. Proteomic comparison of selective breeding and growth hormone transgenesis in fish: Unique pathways to enhanced growth. J Proteomics. (2019) 192:114-124. doi:10.1016/j.jprot.2018.08.013. 35. Causey DR, Kim JH, Stead DA, Martin SA, Devlin RH, Macqueen DJ. Proteomic comparison of selective breeding and growth hormone transgenesis in fish: Unique pathways to enhanced growth. J Proteomics. (2019) 192:114-124. doi:10.1016/j.jprot.2018.08.013. 35. Causey DR, Kim JH, Stead DA, Martin SA, Devlin RH, Macqueen DJ. Proteomic comparison of selective breeding and growth hormone transgenesis in fish: Unique pathways to enhanced growth. J Proteomics. (2019) 192:114-124. doi:10.1016/j.jprot.2018.08.013. 35. Causey DR, Kim JH, Stead DA, Martin SA, Devlin RH, Macqueen DJ. Proteomic comparison of selective breeding and growth hormone transgenesis in fish: Unique pathways to enhanced growth. J Proteomics. (2019) 192:114-124. doi:10.1016/j.jprot.2018.08.013. 36. Ahyi AN, Quinton LJ, Jones MR, Ferrari JD, Pepper-Cunningham ZA, Mella JR,et al. Roles of STAT3 in protein secretion pathways during the acute-phase response. Infect Immun. (2013) May;81(5):1644-53. doi: 10.1128/IAI.01332-12. 36. Ahyi AN, Quinton LJ, Jones MR, Ferrari JD, Pepper-Cunningham ZA, Mella JR,et al. Roles of STAT3 in protein secretion pathways during the acute-phase response. Infect Immun. (2013) May;81(5):1644-53. doi: 10.1128/IAI.01332-12. 37. He G, Karin M. NF-κB and STAT3 - key players in liver inflammation and cancer. Cell Res. (2011) 21(1):159-168. doi:10.1038/cr.2010.183 38. Herzig S, Shaw RJ. AMPK: guardian of metabolism and mitochondrial homeostasis. Nat Rev Mol Cell Biol. (2018) 19(2):121-135. doi:10.1038/nrm.2017.95 38. Herzig S, Shaw RJ. AMPK: guardian of metabolism and mitochondrial homeostasis. Nat Rev Mol Cell Biol. (2018) 19(2):121-135. doi:10.1038/nrm.2017.95 39. Lau HH, Ng NH, Loo LSW, Jasmen JB, Teo AK. The molecular functions of hepatocyte nuclear factors - In and beyond the liver. J Hepatol. (2018) 68(5):1033-1048. doi:10.1016/j.jhep.2017.11.026 39. Lau HH, Ng NH, Loo LSW, Jasmen JB, Teo AK. The molecular functions of hepatocyte nuclear factors - In and beyond the liver. J Hepatol. (2018) 68(5):1033-1048. doi:10.1016/j.jhep.2017.11.026 40. 44. Barrett MT, Pritchard D, Palanca-Wessels C, Anderson J, Reid BJ, Rabinovitch PS. Molecular phenotype of spontaneously arising 4N (G2-tetraploid) intermediates of neoplastic progression in Barrett's esophagus. Cancer Res. (2003) 63(14):4211-4217. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Kawaguchi T, Osatomi K, Yamashita H, Kabashima T, Uyeda K. Mechanism for fatty acid "sparing" effect on glucose-induced transcription: regulation of carbohydrate-responsive element-binding protein by AMP-activated protein kinase. J Biol Chem. (2002) 277(6):3829-3835. doi:10.1074/jbc.M107895200 40. Kawaguchi T, Osatomi K, Yamashita H, Kabashima T, Uyeda K. Mechanism for fatty acid "sparing" effect on glucose-induced transcription: regulation of carbohydrate-responsive element-binding protein by AMP-activated protein kinase. J Biol Chem. (2002) 277(6):3829-3835. doi:10.1074/jbc.M107895200 41. Tikhanovich I, Cox J, Weinman SA. Forkhead box class O transcription factors in liver function and disease. J Gastroenterol Hepatol. (2013) Aug;28 Suppl 1(0 1):125-31. doi: 10.1111/jgh.12021. 41. Tikhanovich I, Cox J, Weinman SA. Forkhead box class O transcription factors in liver function and disease. J Gastroenterol Hepatol. (2013) Aug;28 Suppl 1(0 1):125-31. doi: 10.1111/jgh.12021. 41. Tikhanovich I, Cox J, Weinman SA. Forkhead box class O transcription factors in liver function and disease. J Gastroenterol Hepatol. (2013) Aug;28 Suppl 1(0 1):125-31. doi: 10.1111/jgh.12021. 42. Mueller KM, Themanns M, Friedbichler K, Kornfeld JW, Esterbauer H, Tuckermann JP, et al. Hepatic growth hormone and glucocorticoid receptor signaling in body growth, steatosis and metabolic liver cancer development. Mol Cell Endocrinol. (2012) Sep 25;361(1-2):1-11. doi: 10.1016/j.mce.2012.03.026. 43. Donne R, Saroul-Aïnama M, Cordier P, Celton-Morizur S, Desdouets C. Polyploidy in liver development, homeostasis and disease. Nat Rev Gastroenterol Hepatol. (2020) 17(7):391-405. doi:10.1038/s41575-020-0284-x 23 45. D'Agostino M, Tornillo G, Caporaso MG, Barone MV, Ghigo E, Bonatti S, et al. Ligand of Numb proteins LNX1p80 and LNX2 interact with the human glycoprotein CD8α and promote its ubiquitylation and endocytosis. J Cell Sci. (2011) Nov 1;124(Pt 21):3545-56. doi: 10.1242/jcs.081224. 45. D'Agostino M, Tornillo G, Caporaso MG, Barone MV, Ghigo E, Bonatti S, et al. Ligand of Numb proteins LNX1p80 and LNX2 interact with the human glycoprotein CD8α and promote its ubiquitylation and endocytosis. J Cell Sci. (2011) Nov 1;124(Pt 21):3545-56. doi: 10.1242/jcs.081224. 45. D'Agostino M, Tornillo G, Caporaso MG, Barone MV, Ghigo E, Bonatti S, et al. Ligand of Numb proteins LNX1p80 and LNX2 interact with the human glycoprotein CD8α and promote its ubiquitylation and endocytosis. J Cell Sci. (2011) Nov 1;124(Pt 21):3545-56. doi: 10.1242/jcs.081224. 46. Parker J, Guslund NC, Jentoft S, Roth O. Characterization of Pipefish Immune Cell Populations Through Single-Cell Transcriptomics. Front Immunol. (2022) Jan 27;13:820152. doi: 10.3389/fimmu.2022.820152. 47. Bradley LM, Douglass MF, Chatterjee D, Akira S, Baaten BJ. Matrix metalloprotease 9 mediates neutrophil migration into the airways in response to influenza virus-induced toll-like receptor signaling. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 PLoS Pathog. (2012) 8(4):e1002641. doi: 10.1371/journal.ppat.1002641 48. Takizawa F, Magadan S, Parra D, Xu Z, Korytář T, Boudinot P, et al. Novel Teleost CD4-Bearing Cell Populations Provide Insights into the Evolutionary Origins and Primordial Roles of CD4+ Lymphocytes and CD4+ Macrophages. J Immunol. (2016) Jun 1;196(11):4522-35. doi: 10.4049/jimmunol.1600222. 48. Takizawa F, Magadan S, Parra D, Xu Z, Korytář T, Boudinot P, et al. Novel Teleost CD4-Bearing Cell Populations Provide Insights into the Evolutionary Origins and Primordial Roles of CD4+ Lymphocytes and CD4+ Macrophages. J Immunol. (2016) Jun 1;196(11):4522-35. doi: 10.4049/jimmunol.1600222. 49. Glinos DA, Soskic B, Williams C, Kennedy A, Jostins L, Sansom DM, et al. Genomic profiling of T- cell activation suggests increased sensitivity of memory T cells to CD28 costimulation. Genes Immun. (2020) Dec;21(6-8):390-408. doi: 10.1038/s41435-020-00118-0. 49. Glinos DA, Soskic B, Williams C, Kennedy A, Jostins L, Sansom DM, et al. Genomic profiling of T- cell activation suggests increased sensitivity of memory T cells to CD28 costimulation. Genes Immun. (2020) Dec;21(6-8):390-408. doi: 10.1038/s41435-020-00118-0. 50. Mou D, Espinosa J, Lo DJ, Kirk AD. CD28 negative T cells: is their loss our gain?. Am J Transplant. (2014) 14(11):2460-2466. doi:10.1111/ajt.12937 50. Mou D, Espinosa J, Lo DJ, Kirk AD. CD28 negative T cells: is their loss our gain?. Am J Transplant. (2014) 14(11):2460-2466. doi:10.1111/ajt.12937 51. Zhao X, Shan Q, Xue HH. TCF1 in T cell immunity: a broadened frontier. Nat Rev Immunol. (2022) 22(3):147-157. doi:10.1038/s41577-021-00563-6 52. Szabo PA, Levitin HM, Miron M, Snyder ME, Senda T, Yuan J, et al. Single-cell transcriptomics of human T cells reveals tissue and activation signatures in health and disease. Nat Commun. (2019) Oct 17;10(1):4706. doi: 10.1038/s41467-019-12464-3. 52. Szabo PA, Levitin HM, Miron M, Snyder ME, Senda T, Yuan J, et al. Single-cell transcriptomics of human T cells reveals tissue and activation signatures in health and disease. Nat Commun. (2019) Oct 17;10(1):4706. doi: 10.1038/s41467-019-12464-3. 53. Feng X, Wang H, Takata H, Day TJ, Willen J, Hu H. Transcription factor Foxp1 exerts essential cell- intrinsic regulation of the quiescence of naive T cells. Nat Immunol. (2011) 12(6):544-550. doi:10.1038/ni.2034 53. Feng X, Wang H, Takata H, Day TJ, Willen J, Hu H. Transcription factor Foxp1 exerts essential cell- intrinsic regulation of the quiescence of naive T cells. Nat Immunol. (2011) 12(6):544-550. doi:10.1038/ni.2034 54. Schwartzberg PL, Mueller KL, Qi H, Cannons JL. SLAM receptors and SAP influence lymphocyte interactions, development and function. Nat Rev Immunol. /(2009) Jan;9(1):39-46. doi: 10.1038/nri2456. 54. 73. Nicolet BP, Guislain A, van Alphen FP, Gomez-Eerland R, Schumacher TN, van den Biggelaar M, et al. CD29 identifies IFN-γ-producing human CD8+ T cells with an increased cytotoxic potential. Proc Natl Acad Sci U S A. (2020) Mar 24;117(12):6686-6696. doi: 10.1073/pnas.1913940117. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Bunin A, Sisirak V, Ghosh HS, Grajkowska LT, Hou ZE, Miron M, et al. Protein Tyrosine Phosphatase PTPRS Is an Inhibitory Receptor on Human and Murine Plasmacytoid Dendritic Cells. Immunity. (2015) Aug 18;43(2):277-88. doi: 10.1016/j.immuni.2015.07.009. 64. Bunin A, Sisirak V, Ghosh HS, Grajkowska LT, Hou ZE, Miron M, et al. Protein Tyrosine Phosphatase PTPRS Is an Inhibitory Receptor on Human and Murine Plasmacytoid Dendritic Cells. Immunity. (2015) Aug 18;43(2):277-88. doi: 10.1016/j.immuni.2015.07.009. 64. Bunin A, Sisirak V, Ghosh HS, Grajkowska LT, Hou ZE, Miron M, et al. Protein Tyrosine Phosphatase PTPRS Is an Inhibitory Receptor on Human and Murine Plasmacytoid Dendritic Cells. Immunity. (2015) Aug 18;43(2):277-88. doi: 10.1016/j.immuni.2015.07.009. 65. Worah K, Mathan TS, Vu Manh TP, Keerthikumar S, Schreibelt G, Tel J, et al. Proteomics of Human Dendritic Cell Subsets Reveals Subset-Specific Surface Markers and Differential Inflammasome Function. Cell Rep. (2016) Sep 13;16(11):2953-2966. doi: 10.1016/j.celrep.2016.08.023. 65. Worah K, Mathan TS, Vu Manh TP, Keerthikumar S, Schreibelt G, Tel J, et al. Proteomics of Human Dendritic Cell Subsets Reveals Subset-Specific Surface Markers and Differential Inflammasome Function. Cell Rep. (2016) Sep 13;16(11):2953-2966. doi: 10.1016/j.celrep.2016.08.023. 66. Grajkowska LT, Ceribelli M, Lau CM, Warren ME, Tiniakou I, Nakandakari Higa S, et al. Isoform- Specific Expression and Feedback Regulation of E Protein TCF4 Control Dendritic Cell Lineage Specification. Immunity. (2017) Jan 17;46(1):65-77. doi: 10.1016/j.immuni.2016.11.006. 66. Grajkowska LT, Ceribelli M, Lau CM, Warren ME, Tiniakou I, Nakandakari Higa S, et al. Isoform- Specific Expression and Feedback Regulation of E Protein TCF4 Control Dendritic Cell Lineage Specification. Immunity. (2017) Jan 17;46(1):65-77. doi: 10.1016/j.immuni.2016.11.006. 67. Deenick EK, Pelham SJ, Kane A, Ma CS. Signal Transducer and Activator of Transcription 3 Control of Human T and B Cell Responses. Front Immunol. (2018) 9:168. doi:10.3389/fimmu.2018.00168 67. Deenick EK, Pelham SJ, Kane A, Ma CS. Signal Transducer and Activator of Transcription 3 Control of Human T and B Cell Responses. Front Immunol. (2018) 9:168. doi:10.3389/fimmu.2018.00168 68. van Kooyk Y, Geijtenbeek TB. DC-SIGN: escape mechanism for pathogens. Nat Rev Immunol. (2003) 3(9):697-709. doi:10.1038/nri1182 68. van Kooyk Y, Geijtenbeek TB. DC-SIGN: escape mechanism for pathogens. Nat Rev Immunol. (2003) 3(9):697-709. doi:10.1038/nri1182 69. Young KM, Russell S, Smith M, Huber P, Ostland VE, Brooks AS, et al. Bacterial-binding activity and plasma concentration of ladderlectin in rainbow trout (Oncorhynchus mykiss). Fish Shellfish Immunol. (2007) Aug;23(2):305-15. doi: 10.1016/j.fsi.2006.10.014. 69. Young KM, Russell S, Smith M, Huber P, Ostland VE, Brooks AS, et al. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Schwartzberg PL, Mueller KL, Qi H, Cannons JL. SLAM receptors and SAP influence lymphocyte interactions, development and function. Nat Rev Immunol. /(2009) Jan;9(1):39-46. doi: 10.1038/nri2456. 24 55. Walling BL, Kim M. LFA-1 in T Cell Migration and Differentiation. Front Immunol. (2018) 9:952. doi:10.3389/fimmu.2018.00952 56. Bartsch LM, Damasio MPS, Subudhi S, Drescher HK. Tissue-Resident Memory T Cells in the Liver Unique Characteristics of Local Specialists. Cells. (2020) 9(11):2457. https://doi.org/10.3390/cells9112457 57. Kang SM, Tsang W, Doll S, Scherle P, Ko HS, Tran AC, et al. Induction of the POU domain transcription factor Oct-2 during T-cell activation by cognate antigen. Mol Cell Biol. (1992) 12(7):3149- 3154. doi:10.1128/mcb.12.7.3149-3154.1992 58. Burns GF, Triglia T, Werkmeister JA, Begley CG, Boyd AW. TLiSA1, a human T lineage-specific activation antigen involved in the differentiation of cytotoxic T lymphocytes and anomalous killer cells from their precursors. J Exp Med. (1985) 161(5):1063-1078. doi:10.1084/jem.161.5.1063 58. Burns GF, Triglia T, Werkmeister JA, Begley CG, Boyd AW. TLiSA1, a human T lineage-specific activation antigen involved in the differentiation of cytotoxic T lymphocytes and anomalous killer cells from their precursors. J Exp Med. (1985) 161(5):1063-1078. doi:10.1084/jem.161.5.1063 59. Takeuchi A, Badr Mel S, Miyauchi K, Ishihara C, Onishi R, Guo Z, et al. CRTAM determines the CD4+ cytotoxic T lymphocyte lineage. J Exp Med. (2016) Jan 11;213(1):123-38. doi: 10.1084/jem.20150519. 60. Melichar HJ, Narayan K, Der SD, Hiraoka Y, Gardiol N, Jeannet G, et al. Regulation of gammadelta versus alphabeta T lymphocyte differentiation by the transcription factor SOX13. Science. (2007) 315(5809):230-233. doi:10.1126/science.1135344 61. Karsunky H, Merad M, Cozzio A, Weissman IL, Manz MG. Flt3 ligand regulates dendritic cell development from Flt3+ lymphoid and myeloid-committed progenitors to Flt3+ dendritic cells in vivo. J Exp Med. (2003) 198(2):305-313. doi:10.1084/jem.20030323 62. Leone DA, Peschel A, Brown M, Schachner H, Ball MJ, Gyuraszova M, et al. Surface LAMP-2 Is an Endocytic Receptor That Diverts Antigen Internalized by Human Dendritic Cells into Highly Immunogenic Exosomes. J Immunol. (2017) 199(2):531-546. doi:10.4049/jimmunol.1601263 63. Rocha-Perugini V, Martínez Del Hoyo G, González-Granado JM, Ramírez-Huesca M, Zorita V, Rubinstein E, et al. CD9 Regulates Major Histocompatibility Complex Class II Trafficking in Monocyte- Derived Dendritic Cells. Mol Cell Biol. (2017) Jul 14;37(15):e00202-17. doi: 10.1128/MCB.00202-17. 25 64. Bunin A, Sisirak V, Ghosh HS, Grajkowska LT, Hou ZE, Miron M, et al. Protein Tyrosine Phosphatase PTPRS Is an Inhibitory Receptor on Human and Murine Plasmacytoid Dendritic Cells. Immunity. (2015) Aug 18;43(2):277-88. doi: 10.1016/j.immuni.2015.07.009. 64. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Bacterial-binding activity and plasma concentration of ladderlectin in rainbow trout (Oncorhynchus mykiss). Fish Shellfish Immunol. (2007) Aug;23(2):305-15. doi: 10.1016/j.fsi.2006.10.014. 70. Royet J, Dziarski R. Peptidoglycan recognition proteins: pleiotropic sensors and effectors of antimicrobial defences. Nat Rev Microbiol. (2007) 5(4):264-277. doi:10.1038/nrmicro1620 70. Royet J, Dziarski R. Peptidoglycan recognition proteins: pleiotropic sensors and effectors of antimicrobial defences. Nat Rev Microbiol. (2007) 5(4):264-277. doi:10.1038/nrmicro1620 71. Watanabe Y, Tateno H, Nakamura-Tsuruta S, Kominami J, Hirabayashi J, Nakamura O, et al. The function of rhamnose-binding lectin in innate immunity by restricted binding to Gb3. Dev Comp Immunol. (2009) Feb;33(2):187-97. doi: 10.1016/j.dci.2008.08.008. 72. Wernimont SA, Wiemer AJ, Bennin DA, Monkley SJ, Ludwig T, Critchley DR, et al. Contact- dependent T cell activation and T cell stopping require talin1. J Immunol. (2011) Dec 15;187(12):6256- 67. doi: 10.4049/jimmunol.1102028. 73. Nicolet BP, Guislain A, van Alphen FP, Gomez-Eerland R, Schumacher TN, van den Biggelaar M, et al. CD29 identifies IFN-γ-producing human CD8+ T cells with an increased cytotoxic potential. Proc Natl Acad Sci U S A. (2020) Mar 24;117(12):6686-6696. doi: 10.1073/pnas.1913940117. 26 74. Kremer KN, Dinkel BA, Sterner RM, Osborne DG, Jevremovic D, Hedin KE. TCR-CXCR4 signaling stabilizes cytokine mRNA transcripts via a PREX1-Rac1 pathway: implications for CTCL. Blood. (2017) 130(8):982-994. doi:10.1182/blood-2017-03-770982 74. Kremer KN, Dinkel BA, Sterner RM, Osborne DG, Jevremovic D, Hedin KE. TCR-CXCR4 signaling stabilizes cytokine mRNA transcripts via a PREX1-Rac1 pathway: implications for CTCL. Blood. (2017) 130(8):982-994. doi:10.1182/blood-2017-03-770982 75. Perišić Nanut M, Sabotič J, Jewett A, Kos J. Cysteine cathepsins as regulators of the cytotoxicity of NK and T cells. Front Immunol. (2014) 5:616. doi:10.3389/fimmu.2014.00616 75. Perišić Nanut M, Sabotič J, Jewett A, Kos J. Cysteine cathepsins as regulators of the cytotoxicity of NK and T cells. Front Immunol. (2014) 5:616. doi:10.3389/fimmu.2014.00616 76. Darmoise A, Maschmeyer P, Winau F. The immunological functions of saposins. Adv Immunol. (2010) 105:25-62. doi:10.1016/S0065-2776(10)05002-9 76. Darmoise A, Maschmeyer P, Winau F. The immunological functions of saposins. Adv Immunol. (2010) 105:25-62. doi:10.1016/S0065-2776(10)05002-9 77. Ghosh S, Chackerian AA, Parker CM, Ballantyne CM, Behar SM. The LFA-1 adhesion molecule is required for protective immunity during pulmonary Mycobacterium tuberculosis infection. J Immunol. (2006) 176(8):4914-4922. doi:10.4049/jimmunol.176.8.4914 77. Ghosh S, Chackerian AA, Parker CM, Ballantyne CM, Behar SM. The LFA-1 adhesion molecule is required for protective immunity during pulmonary Mycobacterium tuberculosis infection. J Immunol. (2006) 176(8):4914-4922. doi:10.4049/jimmunol.176.8.4914 78. Liu W, Yan M, Liu Y, Wang R, Li C, Deng C, et al. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Olfactomedin 4 down-regulates innate immunity against Helicobacter pylori infection. Proc Natl Acad Sci U S A. (2010) Jun 15;107(24):11056-61. doi: 10.1073/pnas.1001269107. 78. Liu W, Yan M, Liu Y, Wang R, Li C, Deng C, et al. Olfactomedin 4 down-regulates innate immunity against Helicobacter pylori infection. Proc Natl Acad Sci U S A. (2010) Jun 15;107(24):11056-61. doi: 10.1073/pnas.1001269107. 78. Liu W, Yan M, Liu Y, Wang R, Li C, Deng C, et al. Olfactomedin 4 down-regulates innate immunity against Helicobacter pylori infection. Proc Natl Acad Sci U S A. (2010) Jun 15;107(24):11056-61. doi: 10.1073/pnas.1001269107. 79. Szulc-Dąbrowska L, Bossowska-Nowicka M, Struzik J, Toka FN. Cathepsins in Bacteria-Macrophage Interaction: Defenders or Victims of Circumstance?. Front Cell Infect Microbiol. (2020) 10:601072. doi:10.3389/fcimb.2020.601072 80. Sun B, van Dissel D, Mo I, Boysen P, Haslene-Hox H, Lund H. Identification of novel biomarkers of inflammation in Atlantic salmon (Salmo salar L.) by a plasma proteomic approach. Dev Comp Immunol. (2022) 127:104268. doi:10.1016/j.dci.2021.104268 81. Cunnick JM, Schmidhuber S, Chen G, Yu M, Yi SJ, Cho YJ, et al. Bcr and Abr cooperate in negatively regulating acute inflammatory responses. Mol Cell Biol. (2009) Nov;29(21):5742-50. doi: 10.1128/MCB.00357-09. 82. Cooper AM, Khader SA. IL-12p40: an inherently agonistic cytokine. Trends Immunol. (2007) 28(1):33-38. doi:10.1016/j.it.2006.11.002 82. Cooper AM, Khader SA. IL-12p40: an inherently agonistic cytokine. Trends Immunol. (2007) 28(1):33-38. doi:10.1016/j.it.2006.11.002 82. Cooper AM, Khader SA. IL-12p40: an inherently agonistic cytokine. Trends Immunol. (2007) 28(1):33-38. doi:10.1016/j.it.2006.11.002 (2007) 83. Khader SA, Partida-Sanchez S, Bell G, Jelley-Gibbs DM, Swain S, Pearl JE, et al. Interleukin 12p40 is required for dendritic cell migration and T cell priming after Mycobacterium tuberculosis infection. J Exp Med. (2006) Jul 10;203(7):1805-15. doi: 10.1084/jem.20052545. 83. Khader SA, Partida-Sanchez S, Bell G, Jelley-Gibbs DM, Swain S, Pearl JE, et al. Interleukin 12p40 is required for dendritic cell migration and T cell priming after Mycobacterium tuberculosis infection. J Exp Med. (2006) Jul 10;203(7):1805-15. doi: 10.1084/jem.20052545. 27 84. Russell S, Hayes MA, Simko E, Lumsden JS. Plasma proteomic analysis of the acute phase response of rainbow trout (Oncorhynchus mykiss) to intraperitoneal inflammation and LPS injection. Dev Comp Immunol. (2006) 30(4):393-406. doi:10.1016/j.dci.2005.06.002 84. Russell S, Hayes MA, Simko E, Lumsden JS. Plasma proteomic analysis of the acute phase response of rainbow trout (Oncorhynchus mykiss) to intraperitoneal inflammation and LPS injection. Dev Comp Immunol. (2006) 30(4):393-406. doi:10.1016/j.dci.2005.06.002 85. Hanawa-Suetsugu K, Itoh Y, Ab Fatah M, Nishimura T, Takemura K, Takeshita K, et al. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Phagocytosis is mediated by two-dimensional assemblies of the F-BAR protein GAS7. Nat Commun. (2019) Oct 18;10(1):4763. doi: 10.1038/s41467-019-12738-w. 86. Shi G, Abbott KN, Wu W, Salter RD, Keyel PA. Dnase1L3 Regulates Inflammasome-Dependent Cytokine Secretion. Front Immunol. (2017) May 8;8:522. doi: 10.3389/fimmu.2017.00522. 86. Shi G, Abbott KN, Wu W, Salter RD, Keyel PA. Dnase1L3 Regulates Inflammasome-Dependent Cytokine Secretion. Front Immunol. (2017) May 8;8:522. doi: 10.3389/fimmu.2017.00522. 87. Jones EL, Wee JL, Demaria MC, Blakeley J, Ho PK, Vega-Ramos J, et al. Dendritic Cell Migration and Antigen Presentation Are Coordinated by the Opposing Functions of the Tetraspanins CD82 and CD37. J Immunol. (2016) 1;196(3):978-87. doi: 10.4049/jimmunol.1500357. 88. Williams JW, Tjota MY, Clay BS, Vander Lugt B, Bandukwala HS, Hrusch CL, et al. Transcription factor IRF4 drives dendritic cells to promote Th2 differentiation. Nat Commun. (2013) 4:2990. doi: 10.1038/ncomms3990. 89. Bakke FK, Monte MM, Stead DA, Causey DR, Douglas A, Macqueen DJ, et al. Plasma Proteome Responses in Salmonid Fish Following Immunization. Front Immunol. (2020) 8;11:581070. doi: 10.3389/fimmu.2020.581070. 90. Quinton LJ, Blahna MT, Jones MR, Allen E, Ferrari JD, Hilliard KL, et al. Hepatocyte-specific mutation of both NF-κB RelA and STAT3 abrogates the acute phase response in mice. J Clin Invest. (2012) May;122(5):1758-63. doi: 10.1172/JCI59408. 91. Brasch K. Endopolyploidy in vertebrate liver: an evolutionary perspective. Cell Biol Int Rep. (1980) 4(2):217-226. doi:10.1016/0309-1651(80)90077-6 92. Bakken TE, Hodge RD, Miller JA, Yao Z, Nguyen TN, Aevermann B, et al. Single-nucleus and single- cell transcriptomes compared in matched cortical cell types. PLoS One. (2018) Dec 26;13(12):e0209648. doi: 10.1371/journal.pone.0209648. 92. Bakken TE, Hodge RD, Miller JA, Yao Z, Nguyen TN, Aevermann B, et al. Single-nucleus and single- cell transcriptomes compared in matched cortical cell types. PLoS One. (2018) Dec 26;13(12):e0209648. doi: 10.1371/journal.pone.0209648. 93. Wan F, Hu CB, Ma JX, Gao K, Xiang LX, Shao JZ. Characterization of γδ T Cells from Zebrafish Provides Insights into Their Important Role in Adaptive Humoral Immunity. Front Immunol. (2017) 7:675. doi:10.3389/fimmu.2016.00675 93. Wan F, Hu CB, Ma JX, Gao K, Xiang LX, Shao JZ. Characterization of γδ T Cells from Zebrafish Provides Insights into Their Important Role in Adaptive Humoral Immunity. Front Immunol. (2017) 7:675. doi:10.3389/fimmu.2016.00675 28 94. Hammerich L, Tacke F. Role of gamma-delta T cells in liver inflammation and fibrosis. World J Gastrointest Pathophysiol. (2014) 5(2):107-113. doi:10.4291/wjgp.v5.i2.107 94. Hammerich L, Tacke F. Role of gamma-delta T cells in liver inflammation and fibrosis. World J Gastrointest Pathophysiol. (2014) 5(2):107-113. doi:10.4291/wjgp.v5.i2.107 95. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Diversity of Rainbow Trout Blood B Cells Revealed by Single Cell RNA Sequencing. Biology (Basel). (2021) 10(6):511. doi:10.3390/biology10060511 105. Macqueen DJ, Johnston IA. A well-constrained estimate for the timing of the salmonid whole genome duplication reveals major decoupling from species diversification. Proc Biol Sci. (2014) 281(1778):20132881. doi:10.1098/rspb.2013.2881 106. Gundappa MK, To TH, Grønvold L, Martin SA, Lien S, Geist J, et al. Genome-Wide Reconstruction of Rediploidization Following Autopolyploidization across One Hundred Million Years of Salmonid Evolution. Mol Biol Evol. (2022) 7;39(1):msab310. doi: 10.1093/molbev/msab310. 107. Gillard GB, Grønvold L, Røsæg LL, Holen MM, Monsen Ø, Koop BF, et al. Comparative regulomics supports pervasive selection on gene dosage following whole genome duplication. Genome Biol. (2021) Apr 13;22(1):103. doi: 10.1186/s13059-021-02323-0. 1. Trefts E, Gannon M, Wasserman DH. The liver. Curr Biol. (2017) 27(21):R1147-R1151. doi:10.1016/j.cub.2017.09.019 Hamada S, Umemura M, Shiono T, Tanaka K, Yahagi A, Begum MD, et al. IL-17A produced by gammadelta T cells plays a critical role in innate immunity against listeria monocytogenes infection in the liver. J Immunol. (2008) Sep 1;181(5):3456-63. doi: 10.4049/jimmunol.181.5.3456. 95. Hamada S, Umemura M, Shiono T, Tanaka K, Yahagi A, Begum MD, et al. IL-17A produced by gammadelta T cells plays a critical role in innate immunity against listeria monocytogenes infection in the liver. J Immunol. (2008) Sep 1;181(5):3456-63. doi: 10.4049/jimmunol.181.5.3456. 96. Chan JT, Kadri S, Köllner B, Rebl A, Korytář T. RNA-Seq of Single Fish Cells - Seeking Out the Leukocytes Mediating Immunity in Teleost Fishes. Front Immunol. (2022) 13:798712. doi:10.3389/fimmu.2022.798712 97. Bassity E, Clark TG. Functional identification of dendritic cells in the teleost model, rainbow trout (Oncorhynchus mykiss). PLoS One. (2012) 7(3):e33196. doi:10.1371/journal.pone.0033196 98. Guslund NC, Solbakken MH, Brieuc MS, Jentoft S, Jakobsen KS, Qiao SW. Single-Cell Transcriptome Profiling of Immune Cell Repertoire of the Atlantic Cod Which Naturally Lacks the Major Histocompatibility Class II System. Front Immunol. (2020) 11:559555. doi:10.3389/fimmu.2020.559555 99. Seternes T, Bøgwald J, Dalmo RA. Scavenger endothelial cells of fish, a review. J Fish Dis. (2021) Sep;44(9):1385-1397. doi: 10.1111/jfd.13396. 100. Speilberg L, Evensen O, Nafstad P. Liver of juvenile Atlantic salmon, Salmo salar L.: a light, transmission, and scanning electron microscopic study, with special reference to the sinusoid. Anat Rec. (1994) Nov;240(3):291-307. doi: 10.1002/ar.1092400302. 101. Cheng D, Morsch M, Shami GJ, Chung RS, Braet F. Observation and characterisation of macrophages in zebrafish liver. Micron. (2020) May;132:102851. doi: 10.1016/j.micron.2020.102851 102. Haltiwanger RS, Lehrman MA, Eckhardt AE, Hill RL. The distribution and localization of the fucose-binding lectin in rat tissues and the identification of a high affinity form of the mannose/N- acetylglucosamine-binding lectin in rat liver. J Biol Chem. (1986) Jun 5;261(16):7433-9. PMID: 3711095. 102. Haltiwanger RS, Lehrman MA, Eckhardt AE, Hill RL. The distribution and localization of the fucose-binding lectin in rat tissues and the identification of a high affinity form of the mannose/N- acetylglucosamine-binding lectin in rat liver. J Biol Chem. (1986) Jun 5;261(16):7433-9. PMID: 3711095. 103. Yang CY, Chen JB, Tsai TF, Tsai YC, Tsai CY, Liang PH, et al. CLEC4F is an inducible C-type lectin in F4/80-positive cells and is involved in alpha-galactosylceramide presentation in liver. PLoS One. (2013) Jun 6;8(6):e65070. doi: 10.1371/journal.pone.0065070. 29 104. Perdiguero P, Morel E, Tafalla C. Figure Legends: Figure 1. Major cell types in the Atlantic salmon liver defined by 47,982 nuclear transcriptomes. (a) UMAP highlighting five main liver cell-type clusters according to a priori defined marker genes (see Supplementary Table 2). (b) Bubble plots showing the expression of a priori marker genes for the five main liver cell types, including mean expression (bubble intensity) and percentage of nuclei expressing gene (bubble size). (c) Heatmap of the top 20 differentially expressed genes per each liver cell type defined against the background of all other cell types. (d) Violin plots demonstrating the expression of the most specific marker genes per each of the main liver cell types (colours are matched to the colours of the 5 cell lineages defined in part a). 30 Figure 2. Higher resolution atlas of Atlantic salmon liver cells defined by snRNA-Seq. (a) Unbiased graph-based clustering reveals varying heterogeneity levels across the major liver cell types, presented on a UMAP. Each cell population retains the gene signature of the parent lineage (Supplementary Fig. 2), while also displaying its own distinct transcriptomic profile, presented here as a heatmap (b), inclusive of the top 10 marker genes based on differential gene expression against all other cell clusters. The colour 30 bars above columns on the heatmap illustrate the cell types to which the genes shown were identified as markers (matched to part a). Figure 3. Striking remodeling of the hepatocyte transcriptome in response to bacterial infection. (a) UMAP visualisation of 41,792 hepatocyte nuclei, with sub-clustering performed using the most variable genes restricted to this cell lineage. (b) Shows the same UMAP with nuclei coloured by infection state (top) and the proportion of nuclei originating from each sample after normalising for different nuclei numbers across samples (bottom). (c) Pearson correlation of the expression values for the top 2,000 most variable genes across the nine hepatocytes populations H1-H9. (d) Violin plots of mean expression for the 20 most down-regulated (left) and up-regulated (right) genes based on log-fold change in H4 vs. H1. (e) Heatmap of the top 20 most upregulated and top 20 most downregulated genes in H4 relative to H1, illustrating a gradient of expression from H1 → H2 → H3 → H4 (f) Example enriched GO terms in H4 based on all up-regulated and all down-regulated genes in H4 vs. H1 (full data provided in Supplementary Tables 10 and 11) Figure 4. Figure Legends: Heterogeneity in Atlantic salmon immune cells. (a) UMAP visualization of 1,620 immune nuclei. (b) Heatmap of top 5 markers genes for each immune sub-cluster, sorted by log2 fold change. (c) Number of cell-specific genes up-regulated by infection in immune sub-clusters (d) Examples of genes showing cell-specific up-regulation in response to Aeromonas infection across the breath of immune cell heterogeneity identified. 31
https://openalex.org/W2164187131	https://seer.ufrgs.br/anos90/article/download/51319/34951	Portuguese	null	Sobrevoando histórias: sobre índios e historiadores no Brasil e nos Estados Unidos	Anos 90	2,015	cc-by	14,871	* Doutoranda em História Social (UNICAMP). Bolsista CNPq. Contato: soraiasdornelles@gmail.com. ** Doutoranda em História Social (UNICAMP). Bolsista CNPq. Contato: karinaemelo@hotmail.com. Introdução: sobrevoando os EUA O surgimento do campo de estudos que aborda aspectos históricos sobre os grupos nativos nos Estados Unidos pode ser identificado dentro do contexto da expansão para o Oeste, no século XIX, pautada pela ideia de um Destino Manifesto e pela concomi- tante profissionalização de disciplinas como História, Etnologia e Antropologia1. 174 Em maio de 1830, durante a presidência de Andrew Jackson, o Congresso americano aprovara o Indian Removal Act (Lei de Remoção dos Índios), que determinava que as terras indígenas poderiam ser compradas somente pelo Estado e permitia ao poder executivo nego- ciar com os grupos do Sul, chamados de Cinco Tribos Civilizadas g p , – Chickasaw, Choctaw, Muscogee-Creek, Seminole e Cherokee –, sua remoção para um território federal a oeste do rio Mississipi em troca de suas terras originais. Conforme a política anterior, de Thomas Jefferson, os nativos que se adaptassem ao modo de vida civilizado, isto é, voltado ao trabalho agrícola, poderiam viver ao leste do Mississipi e manter a autonomia de suas nações; caso não aceitassem a assimilação, poderiam ser punidos. Essas terras eram amplamente cobiçadas pelos colonizadores sulistas, que já as dispu- tavam com os índios, principalmente na Geórgia, e, assim, apoiaram as mudanças propostas por Andrew Jackson. Algumas remoções ocorreram por meio de tratados que implicaram a venda ou troca de territórios. Outros grupos nunca firmaram acordos e foram perseguidos em longas guerras (McDONNELL, 1991, p. 06-19)2. g p – Chickasaw, Choctaw, Muscogee-Creek, Seminole e Cherokee –, sua remoção para um território federal a oeste do rio Mississipi em troca de suas terras originais. Conforme a política anterior, de Thomas Jefferson, os nativos que se adaptassem ao modo de vida civilizado, isto é, voltado ao trabalho agrícola, poderiam viver ao leste do Mississipi e manter a autonomia de suas nações; caso não aceitassem a assimilação, poderiam ser punidos. Essas terras eram amplamente cobiçadas pelos colonizadores sulistas, que já as dispu- tavam com os índios, principalmente na Geórgia, e, assim, apoiaram as mudanças propostas por Andrew Jackson. Algumas remoções ocorreram por meio de tratados que implicaram a venda ou troca de territórios. Outros grupos nunca firmaram acordos e foram perseguidos em longas guerras (McDONNELL, 1991, p. 06-19)2. Sobrevoando histórias: sobre índios e historiadores no Brasil... Sobrevoando histórias: sobre índios e historiadores no Brasil... Sobrevoando histórias: sobre índios e historiadores no Brasil e nos Estados Unidos Soraia Sales Dornelles* Karina Moreira Ribeiro da Silva e Melo Resumo: As historiografias brasileira e norte-americana compartilham diversos aspectos quando o assunto é a questão indígena. Em ambas, as histórias dos grupos nativos modificaram as formas de produzir conhecimento sobre eles, criando e transformando políticas públicas. Jogos de influências complexos pautam as formas de lidar com os saberes sobre as relações interétnicas. Em muitos casos, tais saberes serviram como sustentáculo para a sobrevivência ou para o padecimento daqueles grupos. As trajetórias historiográficas, lá e cá, são repletas de convergências, divergências, dinamismo e complexidade política. Isto posto, o objetivo deste artigo é apresentar uma visão paralela dos dois processos de construção de discursos históricos sobre os índios e traçar, a partir daí, as agendas e possibilidades de contribuições mútuas. Palavras-chave: Índios. História Indígena Brasileira. História Indígena Norte- Americana. Historiografia. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... (1851), no qual os índios das planícies e do norte das Montanhas Rochosas permitiam a passagem de imigrantes, a construção de estradas e acampamentos militares (Idem, p. 291). Um primeiro grande boom de imigração para as regiões centrais esteve relacio- nado com “as corridas pelo ouro”, iniciadas por volta de 1859. As possibilidades de enriquecimento rápido e fácil atraíram milhares de colonos e de empresários das ferrovias, que rapidamente iniciaram a construção de transcontinentais. Mas, foi somente no decorrer e após a Guerra Civil (1861-1865) que podemos identificar o período dos maiores e mais sangrentos conflitos envolvendo índios, colonos e o exército norte-americano, conhecido oficialmente como Guerras Indígenas. Apesar de todos os grupos, em um momento ou outro, terem estado em conflito direto com os recém-chegados, Apaches e Sioux foram certamente os mais “celebrados” em sua resistência, fazendo com que alguns de seus líderes, como Red ClouD e Crazy Horse ficassem conhecidos e tivessem suas imagens coladas aos mais variados ícones culturais norte-americanos. Com o fim da Guerra Civil, o Estado iniciou uma política de confinamento dos grupos em reservas controladas principalmente pelo exército. Nelas, a sua segurança estaria garantida; o que, de fato, não ocorreu. Em todos os estados, houve batalhas militares longas para “acomodar” os grupos resistentes e submete-los às novas regras. Dois marcos desse momento ficaram conhecidos como símbolos da violência da conquista e, ao mesmo tempo, do fechamento do processo3. Um é o episódio conhecido como “o massacre de Sand Creek” (1864), no estado do Colorado, onde a milícia local, mesmo com ordens contrárias, assassinou brutalmente um grupo Cheyenne e Arapaho. O outro é “o massacre de Wounded Knee” (1890), em Dakota do Sul, onde, também de forma bárbara, padeceu um grupo Lakota. l d d d (1851), no qual os índios das planícies e do norte das Montanhas Rochosas permitiam a passagem de imigrantes, a construção de estradas e acampamentos militares (Idem, p. 291). Um primeiro grande boom de imigração para as regiões centrais esteve relacio- nado com “as corridas pelo ouro”, iniciadas por volta de 1859. As possibilidades de enriquecimento rápido e fácil atraíram milhares de colonos e de empresários das ferrovias, que rapidamente iniciaram a construção de transcontinentais. Mas, foi somente no decorrer e após a Guerra Civil (1861-1865) que podemos identificar o período dos maiores e mais sangrentos conflitos envolvendo índios, colonos e o exército norte-americano, conhecido oficialmente como Guerras Indígenas. Introdução: sobrevoando os EUA Enquanto isso, no Oeste, em estados mais antigos como o Texas, Novo México, Utah, Oregon, Califórnia e Washington, diversos grupos indígenas e colonos, de origem hispânica majoritariamente, mantinham relações de certa forma amistosas de comércio desde os tempos coloniais. A maior parte dos conflitos podia ser identificada nas regiões fronteiriças, bastante móveis no período, conforme verificamos no mapa a seguir (BARNES, 2010, p. 75). Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Distribuição das Nações Indígenas no começo do século XVI* Distribuição das Nações Indígenas no começo do século XVI* no começo do século XVI Fonte: BARNES, 2010, p. 75 175 Fonte: BARNES, 2010, p. 75 Para as regiões ainda não exploradas por colonizadores, alguns tratados mantinham relações amistosas, como o Treaty of Fort Laramie Para as regiões ainda não exploradas por colonizadores, alguns tratados mantinham relações amistosas, como o Treaty of Fort Laramie * As autoras deste artigo fazem uso de um mapa sobre o século XVI para inteirar, grosso modo, a disposição espacial da presença de diferentes grupos étnicos no t l t itó i d E t d U id C dá t d Mé i Para as regiões ainda não exploradas por colonizadores, alguns tratados mantinham relações amistosas, como o Treaty of Fort Laramie *** As autoras deste artigo fazem uso de um mapa sobre o século XVI para inteirar, grosso modo, a disposição espacial da presença de diferentes grupos étnicos no atual território dos Estados Unidos, Canadá e parte do México. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo confederação de cinco nações, em verdade, provinha do fato de viver próximo a uma reserva em Rochester, estado de New York. Com o auxílio do filho de um chefe Seneca, seu tradutor, escreveu seu pri- meiro livro em 1851. Nesse texto, trazia suas considerações sobre o “destino manifesto dos índios” dentro da chave da impossibilidade de sua sobrevivência fora dos parâmetros civilizadores, motivo pelo qual sua rudeza deveria ser arrancada de forma branda, através da edu cação e da religião. Entre seus questionamentos e suas preocupações, apontava a dúvida sobre o compromisso do governo com a tutela das ‘raças’ indígenas para além das questões concernentes àquele tempo, mas com a “elevação final aos direitos e privilégios de cidadãos norte- -americanos” (MORGAN, 1851). Publicado em 1871, período ápice da violência no Oeste, Systems of consanguinity and affinity of the human Family, compilava os estudos comparativos de parentesco que ele havia empreendido em aproximadamente duas décadas envolvendo grupos dos EUA, Índia, Oceania, África e Oriente Médio. Os dados coletados nos EUA provinham através de questionários enviados para missões e agentes oficiais do interior. Posteriormente, produziu uma análise sobre a evolução da sociedade em que a dividia em estágios: selvageria, civilização e barbárie. Essa visão também estava refletida no conceito de fronteira de Frederick Jackson Turner, expoente simbólico e teórico da conquista do Oeste. Coincidentemente, ou não, Turner apresentou seu trabalho de doutoramento em História em 1890, ano do “fechamento do Oeste”, com um estudo sobre o comércio de peles em Wisconsin: The Character and Influence of the Indian Trade in Wisconsin, anterior aos artigos em que as teses sobre fronteira apareceriam. Também pode ser considerado bastante expressivo da ligação entre o papel dos indígenas e a História do Oeste, a apresentação do artigo The significance of the frontier in American History, durante a Feira Mundial de Chicago, mesmo palco onde Buffalo Bill encenava seu espetáculo sobre a conquista do Oeste e a derrota dos índios.4 i confederação de cinco nações, em verdade, provinha do fato de viver próximo a uma reserva em Rochester, estado de New York. Com o auxílio do filho de um chefe Seneca, seu tradutor, escreveu seu pri- meiro livro em 1851. Sobrevoando histórias: sobre índios e historiadores no Brasil... Apesar de todos os grupos, em um momento ou outro, terem estado em conflito direto com os recém-chegados, Apaches e Sioux foram certamente os mais “celebrados” em sua resistência, fazendo com que alguns de seus líderes, como Red ClouD e Crazy Horse ficassem conhecidos e tivessem suas imagens coladas aos mais variados ícones culturais norte-americanos. Com o fim da Guerra Civil, o Estado iniciou uma política de confinamento dos grupos em reservas controladas principalmente pelo exército. Nelas, a sua segurança estaria garantida; o que, de fato, não ocorreu. Em todos os estados, houve batalhas militares longas para “acomodar” os grupos resistentes e submete-los às novas regras. Dois marcos desse momento ficaram conhecidos como símbolos da violência da conquista e, ao mesmo tempo, do fechamento do processo3. Um é o episódio conhecido como “o massacre de Sand Creek” (1864), no estado do Colorado, onde a milícia local, mesmo com ordens contrárias, assassinou brutalmente um grupo Cheyenne e Arapaho. O outro é “o massacre de Wounded Knee” (1890), em Dakota do Sul, onde, também de forma bárbara, padeceu um grupo Lakota. Naquele momento, os primeiros discursos produzidos tentavam j ifi b i dí é d 176 76 76 Naquele momento, os primeiros discursos produzidos tentavam justificar o avanço sobre as terras indígenas através de sua represen- tação como selvagens, entraves à civilização, receptores passivos da religião, seres condenados ao passado e à extinção (SHEPHERD, s/d, p. 02). Foi nesse contexto que Henry Louis Morgan, pai fundador da antropologia americana, a partir de seus estudos sobre os Iroquois, estruturou os parâmetros de análise de parentesco através da criação do método comparativo. Seu interesse por esse grupo específico, uma Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Sobrevoando histórias: sobre índios e historiadores no Brasil.. originando um gênero que ficou conhecido como antropologia de resgate (SHEPHERD, s/d, p. 2), condizente com a fase assimilatória vivida na realidade. Somente com a publicação de And Still the Waters Run (1940), da historiadora Angie Debo, inaugurava-se um novo momento, em que as denúncias revelavam a pilhagem das terras indígenas por diversos setores, com ampla conivência do Estado, e a violência desses processos. Seu texto, audaciosamente, explicava a expansão para o Oeste calcada na exploração dos índios e de suas terras, e não baseada no destino manifesto. Entretanto, o estilo denúncia colocava os índios em um papel de vítimas passivas e ela mesma, nas margens acadêmicas, visto que nunca pôde firmar seu lugar em instituições de ensino superior. Na mesma linha de Angie Debo, os trabalhos de Alvin Josephy’s (1965) e de Donald Berthrong’s (1976)5 contribuíram para mostrar como a ação do Estado e da Igreja, em tentativas agressivas de incorporação, prejudicaram a autonomia das comunidades. Porém, não incluíram as vozes indígenas e suas interpretações não foram capazes de incorporar criticamente como os índios se adaptaram e incorporaram as mudanças que ocorriam ao seu redor, dando uma visão única de seus passados (SHEPHERD, s/d, p. 04). 78 Muito do que foi escrito posteriormente, enquadrava-se dentro dessa escola do “índio vitimizado”, mas as perspectivas ampliavam-se através do intenso diálogo com os Movimentos dos Direitos Civis e de contestação à guerra do Vietnã. Como atesta Denise Bates (2012), é impressionante perceber como o papel dos ativistas indígenas nas discussões raciais e de discriminação, lutas por educação e melhorias materiais para grupos em situação crítica de pobreza dos anos 1950 e 1960 foi minimizada ora perante os movimentos raciais negros, ora pela invisibilidade dessas pessoas para a sociedade americana6. Também é preciso dizer que, embora vitimizados na historiografia, durante os anos 1960 e 1970, elementos culturais indígenas protago- nizavam formas de protestar da contracultura: vestuário, penteados e costumes da cultura hippie eram inspirados na cultura indígena7. pp p g Com os trabalhos de Vine Deloria, Custer Died for Your Sins: An Indian Manifesto (1969) e de Dee Brown, Bury My Heart at Wounded Knee (1971), a ideia de resistência somava-se à vitimização. No primeiro, observamos o resultado prático da crescente organização do movi- mento indígena e de seus intelectuais orgânicos no final dos anos Anos 90, Porto Alegre, v. 22, n. 41, p. Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Nesse texto, trazia suas considerações sobre o “destino manifesto dos índios” dentro da chave da impossibilidade de sua sobrevivência fora dos parâmetros civilizadores, motivo pelo qual sua rudeza deveria ser arrancada de forma branda, através da edu cação e da religião. Entre seus questionamentos e suas preocupações, apontava a dúvida sobre o compromisso do governo com a tutela das ‘raças’ indígenas para além das questões concernentes àquele tempo, mas com a “elevação final aos direitos e privilégios de cidadãos norte- -americanos” (MORGAN, 1851). Publicado em 1871, período ápice da violência no Oeste, Systems of consanguinity and affinity of the human Family, compilava os estudos comparativos de parentesco que ele havia empreendido em aproximadamente duas décadas envolvendo grupos dos EUA, Índia, Oceania, África e Oriente Médio. Os dados coletados nos EUA provinham através de questionários enviados para missões e agentes oficiais do interior. Posteriormente, produziu uma análise sobre a evolução da sociedade em que a dividia em estágios: selvageria, civilização e barbárie. Essa visão também estava refletida no conceito de fronteira de Frederick Jackson Turner, expoente simbólico e teórico da conquista do Oeste. Coincidentemente, ou não, Turner apresentou seu trabalho de doutoramento em História em 1890, ano do “fechamento do Oeste”, com um estudo sobre o comércio de peles em Wisconsin: The Character and Influence of the Indian Trade in Wisconsin, anterior aos artigos em que as teses sobre fronteira apareceriam. Também pode ser considerado bastante expressivo da ligação entre o papel dos indígenas e a História do Oeste, a apresentação do artigo The significance of the frontier in American History, durante a Feira Mundial de Chicago, mesmo palco onde Buffalo Bill encenava seu espetáculo sobre a conquista do Oeste e a derrota dos índios.4 As produções que se seguiram fizeram crer que a história dos As produções que se seguiram fizeram crer que a história dos índios terminava em 1890, com as últimas batalhas militares contra os Apache, e a derrota simbólica dos Sioux no Massacre de Wounded Knee. Como resultado, antropólogos como Franz Boaz, James Mooney, and Alfred Kroeber assumiram o papel de catalogação dos grupos fadados ao desaparecimento para museus e universidades do país, Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Sobrevoando histórias: sobre índios e historiadores no Brasil... Sobrevoando histórias: sobre índios e historiadores no Brasil... então se presumia. James Axtell (1985), por exemplo, enfatiza como as políticas indigenistas coloniais foram delineadas pelas ações indígenas, que orquestraram regras comerciais e de ocupação territorial sob as quais os não índios tiveram que se submeter para garantir o domínio das áreas e, muitas vezes, os indígenas puderam impedir, mesmo que temporariamente, os projetos expansionistas coloniais. Axtel demonstra, ainda, como os índios foram capazes de incluir elementos culturais europeus e manter, ao mesmo tempo, suas referências prévias. Essa ideia também é trabalhada por James Merrill em The Indians New World: The Catawbas and Their Neighbors from European Contact to Removal (1989), em que apresenta o conceito de “novo mundo” para europeus e índios. Em seguida, em New Worlds for All: Indians, Europeans, and the Remaking of Early America (1997), Collin Calloway também se debruça sobre essa problemática. Esses autores mostram como as doenças, o comércio, os colonos, os assentamentos, as coroas, o mercado, os recursos e a religião interagiram para criar um “novo mundo”, que não era nem a Europa nem a América pré-contato. i 180 Até aí, podemos verificar que a “Nova História Indígena” teria uma tendência a enfocar aspectos positivos do passado indígena, mesmo que suas análises não neguem os resultados práticos das conjunturas em questão, em que os grupos perderam sua autonomia e indepen- dência, ou foram aniquilados. Assim, ao abordar esses pontos, Richard White, em The Roots of Dependency: Subsistence, Environment, and Social Change Among the Choctaws, Pawnees, and Navajos (1983), apresenta uma nova perspectiva historiográfica aos estudos indígenas. Esse impac- tante trabalho abordou temas como a possibilidade de uma visão utilitarista do meio ambiente por parte dos índios em contraposição ao mito do “índio-ecologista” e também se debruçou sobre como esses grupos participaram do comércio colonial, tendo em vista os limites do controle que eles tinham sobre sua própria inserção. Já em seu trabalho mais conhecido, The Middle Ground: Indians, Empires, and Republics in the Great Lakes Region, 1650-1815 (1991), Richard White explora como europeus e índios forjaram um universo no qual as forças eram equiparadas, onde cada um podia ver ao “outro” como estrangeiro e criar, sobre esses parâmetros, um universo compreensível que se sustentou por duzentos anos, e como esse mesmo mundo se desfez após a era das Revoluções. O Middle Ground seria um lugar (as Anos 90, Porto Alegre, v. 22, n. Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo 1960 (Red Power). Escrito pelo ativista Sioux, além de tratar dos assuntos concernentes à história do grupo ao qual pertence, o livro abordava temas como discriminação racial e consequências da desapropriação em um tom de acusação contundente à cultura americana, aos antro- pólogos e ao cristianismo (SHEPHERD, s/d, p. 4-5). Já o best seller de Dee Brown, além de inserir-se na chave vitimização-resistência, acabou contribuindo para o reforço do mito do nobre selvagem. 179 Após a segunda guerra mundial, antropólogos também estavam engajados na produção de conhecimento sobre o passado indígena: ao combinarem os métodos de pesquisa antropológica e histórica, majori- tariamente, a ida aos arquivos no caso da última, terminaram por criar uma metodologia híbrida conhecida como etno-história. Inaugurava-se um período de grande interesse pela perspectiva dos grupos nativos sobre os eventos (agency), sobre como as mudanças e as permanências processaram-se ao longo do tempo (AXTELL, 1979, p. 03). Dos usos desta renovação metodológica, na década de 1980, uma nova geração de estudiosos sobre a temática indígena emergiu. Influenciados pelos trabalhos da década de 1960, combinaram o método etno-histórico, advindos da teoria da história social, com estudos e teorias do simbo lismo cultural de Clifford Geertz e Marshall Sahlins, no intuito de colocar os povos indígenas no centro da história como agentes ativos das mudanças. Surgia uma “Nova História Indígena”, denominação que congrega um corpo muito diverso de temas e objetivos. Ao colocar a perspectiva dos próprios indígenas em foco, em análises cuja chave era a agência, e não um entendimento de suas ações como reações, esses estudos abriram espaço para questionar as narrativas tradicionais da história americana. Nas palavras de Shepherd, os “[...] índios tiveram um impacto sobre o curso do desenvolvimento americano, não como um obstáculo, mas como um formador ativo da história” (s./d., p. 7). São expoentes da “Nova História Indígena” James Merrill, James Axtell, Gregory E. Dowd, RD Edmunds, Richard White, Colin Calloway, Jean M. O’Brien, Daniel Richter e Theda Perdue8. A maior parte desses estudos trata do período colonial e do século XIX, contendo discussões sobre gênero, trabalho, meio ambiente e processos de urbanização. Numa espécie de reinterpretação das histórias coloniais mainstrem, mostram que, ao incluir as populações nativas, pode-se ter entendimentos completamente distintos do que até Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo “bordas do império”, nesse caso a região dos Grandes Lagos) onde a frouxidão das diretrizes coloca-se ora pela distância, ora pela falta de vontade dos personagens de colocar em prática as regras; e ao mesmo tempo um estado de espírito, no qual ajustes e incompreensões produziram um modelo novo para resolver situações complicadas, não havendo, segundo White, distinção entre história indígena e branca, nem prevalência demarcada de forças de uns sobre os outros. “bordas do império”, nesse caso a região dos Grandes Lagos) onde a frouxidão das diretrizes coloca-se ora pela distância, ora pela falta de vontade dos personagens de colocar em prática as regras; e ao mesmo tempo um estado de espírito, no qual ajustes e incompreensões produziram um modelo novo para resolver situações complicadas, não havendo, segundo White, distinção entre história indígena e branca, nem prevalência demarcada de forças de uns sobre os outros. 181 Com a colaboração de outros autores, a “Nova História Indígena” deu conta de diversas regiões e períodos, como os primórdios da fase Nacional e o que ficou conhecido com “Indian Removal”. Temas antigos foram sendo revisitados, como as políticas estatais e as suas implicações jurídicas, as intenções assimilacionistas e o respaldo científico que as embasavam. Questões como qual a responsabilidade do Estado pela perda da terra pelos indígenas, e quais efeitos as realocações trouxeram para as comunidades estiveram na pauta nos anos de 1980 e 1990. Foi também neste período que surgiram muitos estudos voltados para questões de gênero, educação e fronteira, além do formidável avanço no conhecimento da história do século XX mais recente, que solapa, de uma vez por todas, a ideia de desaparecimento ou assimilação total dos índios. Em vez disso, mostram o crescimento das comunidades e a continuidade de sua busca por direitos e reconhecimento. Também como parte desse cenário, temos a considerável produção histórica elaborada por estudiosos nativos9. Para Ned Blackhawk (2005, p. 15), durante os anos 1990, no contexto de comemorações dos quinhentos anos da Conquista e as revisitações do passado, os estudos indígenas tiveram grande destaque, e afirma que “[...] se foram os dias em que os historiadores podiam casualmente descartar ou ignorar a poderosa e diversa constelação de populações nativas do continente”. Sobrevoando histórias: sobre índios e historiadores no Brasil... 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Segundo Blackhawk, os estudos mostraram como houve, em finais do século XIX e início do XX, um crescimento da capacidade do governo americano de aprovar, policiar e colocar em prática as políticas indigenistas. Outros historiadores aprofundaram o entendimento sobre as políticas assimilacionistas focadas em crianças, confinadas em internatos religiosos (2005, p. 15). Segundo Pekka Hämäläinen (2012), os historiadores da, hoje velha, Nova História Indígena, buscavam colocar os indígenas no centro das coisas, na tentativa de reformular as narrativas oficiais, que Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Sobrevoando histórias: sobre índios e historiadores no Brasil... haviam “trancado as populações nativas no lado errado da moderni- dade”, tendo sido bastante bem sucedidos nessa empreitada. Estudos revolucionaram a história dos EUA durante o último quarto do século XX, sendo o período colonial aquele em que o impacto desses estudos foi maior a ponto de os intelectuais utilizarem o termo “indigenous turn” (virada indígena). Entretanto, o autor coloca em consideração que, para alguns estudiosos, esse sucesso seria, justamente, o primeiro sinal do esgotamento “do momento criativo” do campo. A história indígena tornou-se mainstream, juntamente com estudos sobre fron- teiras, história ambiental e do imperialismo; conta com um já sólido corpo de estudiosos e há perspectivas de estudos que aprofundem os conhecimentos sobre temas de grande interesse como gênero, escravidão, sexualidade e casamentos mistos, para dar alguns exemplos. 182 esc av dão, se ua dade e casa e tos stos, pa a da a gu s e e p os. Durante as últimas décadas, as comunidades indígenas ame- ricanas foram capazes de reestruturar sua organização e alcançar avanços demográficos, econômicos e sociais. Esse fortalecimento, acompanhado de um aumento substancial na formação de intelectuais nativos, fez com que suas vozes pudessem alçar voos mais amplos, reivindicando espaço na construção de narrativas históricas e mesmo contestando a legitimidade de outras. O campo tornou-se um terreno de contestações, em que a produção do conhecimento e sua recepção tornaram-se tão importantes quanto o impacto que esses estudos produzem na vida prática das comunidades. Para Hämäläinen, esse diálogo já está gerando resultados através de estudos, produzidos por indígenas e não indígenas, que superam as lacunas entre as expectativas das comunidades indígenas e das práticas acadêmicas estabelecidas. Em suas palavras, “[...] os estudiosos agora escrevem diretamente sobre a dor e o trauma psicológico dos povos nativos, nos sensibili- zando para as dimensões humanas do racismo e da expropriação e nos obrigando a chegar a um acordo sobre a violência generalizada da história americana”. De fato, temas tabu têm sido postos em pauta, como a relação entre escravidão africana e indígena, limpeza étnica e genocídio cultural. Estas e outras linhas de trabalho têm realinhando a história indígena americana e podem modificar a forma como entendemos o multiculturalismo, a democracia, a soberania, bem como o papel do Estado-nação na sociedade americana moderna. Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo algum tempo, principalmente após os quinhentos anos da chegada de Colombo, os temas indígenas têm recebido atenção de intelectuais, ativistas e governos no mundo todo, alimentando a expectativa de que os diálogos sobre o tema serão cada vez mais comparativos e transcontinentais. A criação da Native American and Indigenous Studies Association, em 2008, é um grande símbolo disso. 183 g Mesmo com visões bastante otimistas, ainda existe espaço para insatisfação dos estudiosos da temática, principalmente com a falta de reconhecimento e inclusão dos resultados das pesquisas nos programas das Universidades Americanas, sem falar nos currículos escolares. Em maio de 2013, a Newberry Library – um dos maiores centros de documentação sobre historia dos índios no mundo – promoveu o evento Why You Can’t Teach U.S. History without American Indians (Por que você não pode ensinar a história dos Estados Unidos sem os índios). O encontro reuniu pesquisadores nativos e não índios de todas as regiões do país em diversos períodos de estudo. Sua queixa comum: a falta de interesse de outros pesquisadores em reconhecer a importância da inclusão da História dos Índios nas narrativas sobre a História Norte-americana10. O cartaz do evento traz uma conhecida pintura de Benjamin West. Ela retrata uma cena de interação entre índios e europeus em 1683, quando William Penn negociou um acordo de paz com o chefe do grupo Lenape (Delaware) perto de uma aldeia na Pennsylvania (BARNES, 2010, p. 145). Concluída quase um século depois (1771- 1772), a pintura foi inspirada em um retrato feito de memória por uma testemunha ocular do encontro. Num jogo de imagens feito pelos idealizadores do evento, os índios foram ‘deletados’ da obra, fazendo com que ela ficasse incompleta de tal modo que a história retratada perde o sentido. A constatação de que sem os índios não é possível explicar e entender o cenário é imediata. Por outro lado, sua deleção deixa marcas tão profundas na obra de arte original que, apenas com as suas sombras, é possível dar conta não só do significado da ausên- cia dos índios na história, como também da presença de indícios que indicam que algo indispensável à sua compreensão foi dado como inexistente. Sobrevoando histórias: sobre índios e historiadores no Brasil... Assim, a história indígena tornou-se um tema internacional e, há p g p Durante as últimas décadas, as comunidades indígenas ame- ricanas foram capazes de reestruturar sua organização e alcançar avanços demográficos, econômicos e sociais. Esse fortalecimento, acompanhado de um aumento substancial na formação de intelectuais nativos, fez com que suas vozes pudessem alçar voos mais amplos, reivindicando espaço na construção de narrativas históricas e mesmo contestando a legitimidade de outras. O campo tornou-se um terreno de contestações, em que a produção do conhecimento e sua recepção tornaram-se tão importantes quanto o impacto que esses estudos produzem na vida prática das comunidades. Para Hämäläinen, esse diálogo já está gerando resultados através de estudos, produzidos por indígenas e não indígenas, que superam as lacunas entre as expectativas das comunidades indígenas e das práticas acadêmicas estabelecidas. g p Em suas palavras, “[...] os estudiosos agora escrevem diretamente sobre a dor e o trauma psicológico dos povos nativos, nos sensibili- zando para as dimensões humanas do racismo e da expropriação e nos obrigando a chegar a um acordo sobre a violência generalizada da história americana”. De fato, temas tabu têm sido postos em pauta, como a relação entre escravidão africana e indígena, limpeza étnica e genocídio cultural. Estas e outras linhas de trabalho têm realinhando a história indígena americana e podem modificar a forma como entendemos o multiculturalismo, a democracia, a soberania, bem como o papel do Estado-nação na sociedade americana moderna. Assim, a história indígena tornou-se um tema internacional e, há Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Isso faz com que estudiosos pensem em estratégias sobre como lidar com situações em que, não apenas hipoteticamente, mas, de fato, os índios são tidos como desimportantes no processo histórico. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Figura 1: Cartaz de divulgação do evento comemorativo dos quarenta anos do D’Arcy McNickle Center, realizado nos dias três e quatro de maio de 2013, na Newberry Library, em Chicago Fonte: Disponível em: http://www.newberry.org/search/site/why%20you%20 can%26%23039%3Bt%20teach. Acesso em: 14 set. 2014. 8 Figura 1: Cartaz de divulgação do evento comemorativo dos quarenta anos do D’Arcy McNickle Center, realizado nos dias três e quatro de maio de 2013, na Newberry Library, em Chicago na Newberry Library, em Chicago ponível em: http://www.newberry.org/search/site/why%20you%20 23039%3Bt%20teach. Acesso em: 14 set. 2014. na Newberry Library, em Chicago /search/site/why%20you%20 et. 2014. 184 Fonte: Disponível em: http://www.newberry.org/search/site/why%20you%20 can%26%23039%3Bt%20teach. Acesso em: 14 set. 2014. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Figura 2: O Tratado de Penn com os índios, óleo sobre tela. (1771-1772). Fonte: Disponível em: http://withart.visitphilly.com/artworks/penns-treaty-with- -the-indians/. Acesso em: 21 jul. 2015. Figura 2: O Tratado de Penn com os índios, óleo sobre tela. (1771-1772). Figura 2: O Tratado de Penn com os índios, óleo sobre tela. (1771-1772). 185 Fonte: Disponível em: http://withart.visitphilly.com/artworks/penns-treaty-with- -the-indians/. Acesso em: 21 jul. 2015. Sobrevoando as Terras Baixas Foi no século XIX que floresceu a historiografia brasileira. A questão indígena confundiu-se com as práticas e a produção dos discursos. Em 1839, ao inaugurar o Instituto Histórico e Geográfico do Brasil (IHGB)11, o Estado assumia o papel de “criador” de uma história nacional e oficial. Seu primeiro grande e simbólico ato foi o lançamento do Concurso Como Escrever a História do Brasil. O vencedor foi o botânico e etnógrafo Carl Friedrich Philippe von Martius, que havia feito pesquisas sobre língua e cultura indígenas em parte do Brasil e afirmava que os índios estavam fadados ao desaparecimento. Na história do Brasil de von Martius, os índios foram apresentados de forma romântica, quase confundidos à natureza, felizmente aco- lhidos pelo colonizador português em sua inteligência civilizadora. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo sobre o que fazer com os índios do presente: se formariam o corpo de trabalhadores tão caros à nação; se seriam entregues às missões religiosas e devidamente trazidos ao seio da civilização (aqui também por meio do trabalho); ou ainda, se serviriam ao papel de guardas das fronteiras, conforme projeto do Gen. Couto de Magalhães, através das colônias militares; e, claro, as propostas de extermínio. Em São Paulo, por exemplo, ainda imperava a perseguição de índios e sua escravização, de acordo com a Carta régia de 1808, e mesmo com a lei de proibição, em 1831, à revelia do Direito, muitos índios encontravam-se nessa condição. Em outras situações, como no Rio de Janeiro e em Pernambuco, eram grandes as disputas entre ex-colonos e índios sobre as terras de aldeamentos formados em tempos coloniais. 187 p Em 24 de julho de 1845, foi aprovado o Decreto 426, conhecido como “Regulamento acerca das Missões de Catequese e Civilização dos Índios” ou, simplesmente, Regulamento das Missões, tornando-se a Lei Indígena do Império. Atribuía aos Diretores Gerais a função de administrar antigos aldeamentos e criar novos para atrair os indígenas hostis dos sertões. Aos padres, capuchinhos em sua maioria, cabia a missão civilizadora perante o indígena, em que ele seria instruído e assimilado à nação. Juridicamente incapaz, a cidadania era negada como lhe fora também a participação na História pelos pais fundadores. Com a Lei de Terras de 1850, os territórios indígenas passavam a ser de propriedade do Estado, bem como a definição de terras ocupadas ou devolutas, e seus direitos atrelavam-se à condição de “assentados”. Estava posto o cenário histórico das maiores expropriações territoriais e confrontos entre a nação, os índios e imigrantes estrangeiros12. É este o caso da Companhia de Comércio e Navegação do Vale do Mucuri, fundada no interior de Minas Gerais entre 1850 e 1860, por Teófilo Ottoni. Deputado por aquela província, Ottoni incentivou o desenvolvimento e a colonização na região onde, segundo ele, “construirei minha Philadelphia”. Ele escreveu pelo menos cinco livros e dois relatórios sobre a Companhia, a importância de seus privilégios, o escopo de seu projeto e a relação com os ‘selvagens’13. Depois da Lei de Terras de 1850, o governo passou a exercer mais controle sobre as companhias de colonização. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Coube a ele mostrar como a influência moral do branco sobre os índios engrandeceu a nação, e em menos de cinquenta páginas ele expos ideias como a de que “[...] o sangue português, em um pode- roso rio, deverá absorver os pequenos confluentes das raças índia e etiópica” (MARTIUS, [1854] 1982, p. 88). O projeto premiado de escrita da história de von Martius, fundado na presença das três raças na constituição do Brasil não vingou. Entre 1854 e 1857, Francisco Adolpho de Varnhagen publi- cou, em cinco volumes, a obra História geral do Brasil, sob o patro- cínio imperial. No primeiro volume, de 1854, ele assentiu: para os índios, “[...] povos na infância, não há história: há só etnografia” (VARNHAGEN, 1962 [1854], v. 1, p. 42). Tratava-se de uma noção que excluía os índios da historiografia, os detratava quando os mencionava e indicava sua escravidão e seu extermínio físico. O projeto com relação à história dos índios foi fruto de uma cultura de época e serviu a interesses muito distintos. Além disso, não foi uma unanimidade e entrou em confronto com visões mais amenas. Obras como as de João Francisco Lisboa, Capistrano de Abreu, Couto de Magalhães, Gonçalves de Magalhães e Gonçalves Dias produziram trabalhos etnográficos e de defesa do “salvamento” daquelas populações ou resultaram em comemoradas produções literárias sobre os índios. Gonçalves de Magalhães, por exemplo, é autor de A Confederação dos Tamoyos (1857), uma epopeia que toma o sacrifício e o heroísmo indígena como marco fundador da nação brasileira através da morte de Aimberê, o último sobrevivente da nação tamoia, massacrada pelo colonizador português. Inspirada na obra literária, a pintura O último tamoio, de Rodolfo Amoêdo (1857-1941), foi exibida na Exposição Geral de Belas Artes no Rio de Janeiro, em 1884 (SCRICH, 2009). 186 Embora visões e projetos sobre os índios fossem diferentes no Oitocentos, partilhavam os mesmos pressupostos de excluí- -los da historia. Desde o final do período colonial, a monarquia portuguesa precisou realizar uma série de ajustes políticos no que dizia respeito aos indígenas, visto que havia não apenas diferentes projetos teóricos, mas distintas práticas espalhadas pelas províncias, segundo as histórias das relações de ocupação territorial, mestiçagem e utilização da mão de obra. Foram amplas as frentes de discussões Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Mesmo assim, os planos de Teófilo Ottoni foram executados sem barreiras opressivas. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Essa constância quanto ao “problema indígena”, parecia con- ciliar violência e filantropia, ambos pertinentes ao projeto de civi- lização, encontrando representantes nos sertões e nos gabinetes (MONTEIRO, 2001, p. 147-157). Todos esses dilemas encontram-se nos artigos da Revista Trimestral do IHGB, nas pesquisas sistemáticas sobre as raças indígenas realizadas pelo Museu Nacional a partir dos anos 1870 e, claro, na realização, em 1882, da Exposição Antropológica, no Rio de Janeiro. Ali, as novas e internacionais discussões sobre raças e evolução encontraram eco profundo em um já bastante complexo debate sobre o significado dos índios do passado e do presente: como se acomodaria essa classe de homens de valores tão indefinidos na construção da nação? Através de estudos linguísticos, craniométricos, etnográficos ia-se cumprindo o desinteresse pela participação histórica nos mais diversos processos. 188 p p ç p Ao contrário do que muitos pensam, foi durante os séculos XIX e início do XX, e não no tempo colonial, que se deu a expansão e ocupação da maior parte do Brasil interior. Durante a década de 1840, iniciou-se um trabalho de demarcação das fronteiras através do uso de documentação espalhada pelas províncias com a participação dos membros do IHGB, fronteiras essas, que tiveram sua fragilidade evidenciada durante a Guerra do Paraguai. Mas, em 1891, um dos estados mais ricos da nação, São Paulo, ainda apresentava nos mapas, sob um terço do seu território, a expressão Terrenos desconhecidos, para desgosto das autoridades (CAVENAGHI, 2004, p. 193). O mesmo tipo de decepção pode ser percebida quando ocorreu a revolta dos indígenas do aldeamento do Itambacuri (MG), em 1893, “[...] colocando em questão a eficácia da política de aldeamentos, estabelecida com tanta expectativa na legislação de 1845 e combatida com tanto furor ao longo do Segundo Reinado pelos descrentes que achavam que a civilização dos índios se resolvia com bala e pólvora” (MONTEIRO, 2001, p. 165). Nos sertões desconhecidos de todo o país, ocorriam enfrentamentos entre as populações das frentes de expansão e os grupos indígenas. Nacionais e imigrantes, pequenos e grandes pro- prietários – o caso das lavouras de café em São Paulo e no Paraná, e da extração da borracha no Norte são emblemáticos –, e o advento de empreendimentos públicos e privados, como a construção de estradas de ferro e a instalação de linhas telegráficas, motivaram chacinas e perseguições de fins verdadeiramente trágicos para os índios. Anos 90, Porto Alegre, v. 22, n. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Dentro desse contexto de instabilidade total, as alternativas para o “problema indígena” novamente se colocavam. Uma delas seria a busca de um tratamento mais humanitário desses grupos, surgida no seio do exército – principalmente daqueles que participavam da exploração dos sertões através da instalação de telégrafos. Essa linha, também podia ser observada nas propostas do Apostolado Positivista, que via no indígena, um ser na fase inferior do progresso humano, mas que poderia alcançar a civilização com o auxílio dos que já a viviam. O expoente intelectual e sintetizador das novas ideias era Cândido Mariano Rondon, que se tornou figura cativa da mídia na virada do século, apresentando resultados práticos da viabilidade de tratar o índio sem violência. A outra possibilidade ainda era a do extermínio: em 1908, a publicação do artigo de Herman Von Hering na Revista do Museu Paulista evidenciava mais uma vez que a vida e o futuro dos índios eram incertos. O certo é que, naquele momento, houve uma crescente pressão para a criação de um órgão de pro- teção e civilização dos índios. Alguns aspectos gerais constituíam a demanda em que claramente prevaleceu a influência positivista: a demarcação das terras indígenas; a proteção e a garantia de suas vidas; a punição dos crimes contra eles praticados; a laicização da civilização. O Estado republicano finalmente assumiu a responsa- bilidade sobre o destino dos índios ao criar, em 1910, o Serviço de Proteção aos Índios e Localização de Trabalhadores Nacionais (SPILTN). Conforme Gagliardi, praticamente sem recursos e vivendo sob constantes ameaças de fechamento, logrou não colocar em prática a maioria dos seus intuitos. O órgão contribuiu para o confinamento dos grupos em territórios pequenos e impróprios, que continuaram sendo disputados com outros segmentos. Foram vítimas de traba- lhos forçados, doenças, prostituição e estigmas sociais como a sua incapacidade para o trabalho. 189 Nos meios intelectuais, por um lado, permanecia uma vontade geral de entender e firmar o que seria a identidade nacional; por outro, um esforço em conhecer melhor sociedades indígenas tidas como em vias de extinção. Isso pode ser exemplificado através da Semana da Arte Moderna de 1922 e através da trajetória profissional de Curt Unckel Nimuendajú e sua rede de relações. Na Semana da Arte Moderna, que aconteceu em São Paulo, a grande questão foi Tupi or not Tupi? Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo A população indígena, em contato com os brancos, vai sendo progressivamente eliminada e repetindo mais uma vez um fato que sempre ocorreu em todos os lugares e em todos os tempos em que se verificou a presença, uma ao lado da outra, de raças de níveis culturais muito apartadas: a inferior e dominada desaparece. E não fosse o cruzamento, praticado em larga escala entre nós e que permitiu a perpetuação do sangue indígena, este estaria fortemente condenado à extinção total (Apud Freitas). 191 Em 1933, também era publicado Casa Grande & Senzala, de Gilberto Freyre, que novamente analisava a formação do Brasil a partir da fusão das três raças, positivando esse aspecto. Um dado interessante para este artigo, diz respeito à grande influência de Franz Boas nos trabalhos de Freyre, visto que foi nos EUA que obteve sua forma- ção superior. Em 1933, também era publicado Casa Grande & Senzala, de Gilberto Freyre, que novamente analisava a formação do Brasil a partir da fusão das três raças, positivando esse aspecto. Um dado interessante para este artigo, diz respeito à grande influência de Franz Boas nos trabalhos de Freyre, visto que foi nos EUA que obteve sua forma- ção superior. p O segundo caso é representado pelo historiador Sérgio Buarque de Holanda. Em seus estudos sobre expansão e ocupação territo- riais, ele adotou uma abordagem que incluiu os índios na história e mostrou ser central o seu papel na formação e concretização do projeto colonial luso. Em Raízes do Brasil (1936), Monções (1945), Caminhos e Fronteiras (1957), Visão do Paraíso (1959) e no inacabado O Extremo Oeste (1986), o autor parece ter incluído em um crescente os grupos indígenas nas análises: trata de redes de comércio, da mestiçagem, da utilização de caminhos e modos indígenas na relação com o meio, de sua participação na exploração dos sertões e em bandeiras. O autor também participou da elaboração da História Geral da Civilização Brasileira durante os anos 1960 e 1970, e com a colaboração de outros professores da Universidade de São Paulo (USP) – antropólogos e sociólogos como Egon Schaden e Florestan Fernandes – que apresentavam os índios, mas de forma pouco relevante, não correspondendo aos resultados da obra do próprio Sérgio Buarque. Sobrevoando histórias: sobre índios e historiadores no Brasil... Conectado a ela, esteve O Manifesto Antropofágico, apresentado por Oswald de Andrade em 1928. Nele, o escritor exalta a identidade e a criatividade brasileiras através das manifestações culturais. O docu- mento é considerado um marco do Modernismo no Brasil e “Tupi, or not tupi that is the question” é uma das metáforas do manifesto, segundo o qual “[...] sem nós a Europa não teria sequer a sua pobre declaração dos direitos do homem”14. Também em 1928, Tarsila do Amaral retratou a brasilidade em Abopuru, a pintura símbolo do Modernismo nas terras baixas. A figura com pés gigantes fincados na terra e a cabeça minúscula que alcança a altura do sol parecem valorizar o trabalho braçal e sugere a ideia de homem selvagem, nativo, plantado em chão brasileiro, antropófago, que se alimenta ritualmente da carne do outro para adquirir sua força, como indica o próprio nome Abopuru, em tupi-guarani, o homem que come carne humana. Etnólogo alemão, Curt Unckel Nimuendajú dedicou- -se à militância indigenista e à pesquisa etnológica ao longo das quatro primeiras décadas do século XX. Sua obra produziu valiosos registros etnográficos tanto sobre os grupos tupi-guarani, um dos quais o batizou com o nome indígena15, quanto sobre os grupos Jê. Nimuendajú manteve diálogo com o americanismo que nascia nos museus de etnologia na Europa e nos EUA, ao mesmo tempo em que fazia um árduo trabalho de campo coletando e catalogando objetos e artefatos da cultura material dos índios, dados sobre sua língua e relações de parentesco. Assumiu em seus escritos um tom trágico em relação ao futuro das populações indígenas. Mesmo assim, “[...] seu empenho em etnografar a complexa organização social dos grupos Jê contribuiu, por sua vez, para que a moderna reflexão etnológica de David Maybury-Lewis e Claude Lévi-Strauss pudessem avançar sobre bases seguras nas décadas de 1950 e 1960”16. i g Já nas décadas de 1930 e 1940, verificamos continuidades e certa renovação na forma de tratar a participação dos índios na História. No primeiro caso, foi exemplar a postura de Caio Prado Jr., que, em Evolução Política do Brasil (1933), e em Formação do Brasil Contemporâneo (1942), relegava aos índios uma visão de vítimas dos processos de disputas territoriais e de mão de obra e, por vezes, apresentava preconceitos explícitos sobre eles: Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Ainda em aberto está o questionamento sobre por que, mesmo os índios ocupando um papel central nos estudos de um dos historiadores mais importantes do Brasil, esse fato não foi capaz de colocar os indígenas em evidência. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... O trabalho mais significativo sobre os índios no Brasil nos anos de 1950, foi escrito e publicado nos EUA, The Origin and Establishment of Brazil’s Indian Service, 1889 – 1910, de David Hall Stauffer. Com valioso trabalho documental e narrativo, analisa o processo de expan- são para interior no século XIX, além dos intensos debates sobre extermínio dos índios e a criação do SPI, valorizando amplamente a atuação de Rondon. No Brasil, sua publicação se deu de forma incompleta, tendo alguns capítulos traduzidos pela Revista de História da USP nos anos de 1960. Durante a década de 1970 e boa parte dos anos 1980, os estudos históricos sobre os índios equivaliam a “crônicas de sua extinção”, fruto de uma real situação de diminuição das populações. Exemplo deste tipo de abordagem é o importantíssimo estudo de Silvio Coelho dos Santos, Índios e brancos no sul do Brasil: a dramática experiência dos Xokleng (1973), no qual o autor, em um tom de denún- cia, relata a história de confrontos entre colonizadores europeus, predominantemente de origem germânica, e os índios da etnia Xokleng, no estado de Santa Catarina. Àquela altura, o Brasil dos estados do sul e sudeste já vivia a negação da existência dos índios, vinculando-os com a floresta Amazônica ou ao recém-inaugurado Parque Nacional do Xingu17. Nos mesmos moldes, entretanto com uma análise menos refinada, por apresentar valiosíssimas fontes histó- ricas, porém sem discuti-las, Ítala Irene Basile Becker apresenta uma história do grupo Kaingang no Rio Grande do Sul (1976)18. Outras produções nesse modelo são Red Gold (1978), de John Hemming, que fez um estudo sobre a América Portuguesa, e Carlos Moreira Neto, em Índios da Amazônia: de maioria a minoria (1988). Ambos tratam da exploração e destruição dos índios pelos processos de dominação colonial, resultando em abordagens cristalizadoras, que negativam as mudanças vividas – como integração ou assimilação – em termos de aculturação, diminuindo o “grau de indianidade”. Mas foi durante a década de 1990 que o Brasil também viu nascer a sua “Nova História Indígena”. Fruto de um esforço coletivo para colocar os índios no centro dos processos estudados, procurou enten- der a sua “agência” como sujeitos, não como vítimas passivas e enges- sadas. O momento de renascimento dos movimentos sociais perante a ditadura militar – que havia sido denunciada internacionalmente Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo sobre os atos cometidos contra populações indígenas – resultou na desmontagem do SPI e na criação da Fundação Nacional de Proteção ao Índio (FUNAI), em 1967. Essa virada acompanhou a emergên- cia do movimento indígena em todas as partes do país, envolvendo militantes “orgânicos” e intelectuais, principalmente, nas áreas de antropologia e história. Com ampla participação na Assembleia Constituinte e aprovação do compromisso com as demarcações de terras assumidas pela Constituição Cidadã, tomou fôlego a produ- ção de laudos e dossiês baseados em documentação – sobretudo do período colonial –, dada a reconfiguração dos direitos indígenas como históricos (MONTEIRO, 2001, p. 05). Nas universidades, novas orientações teóricas ampliaram as possibilidades de estudos sobre os índios dentro da antropologia, como visões mais dinâmicas de cultura, o uso de fontes orais, as possibilidades analíticas da etno-história. 193 A publicação de As Muralhas dos Sertões: os povos indígenas no rio Branco e a colonização (1991), de Nádia Farage, simboliza a renovação no fazer histórico: sua análise documental dotada de sensibilidade etnográfica apresenta como os índios da região de Rio Branco, dispu tada por portugueses e holandeses, atuaram segundo intenções próprias na construção do universo colonial. Outros estudos publi- cados na década de 1990 corroboraram com a colocação dos índios no cenário historiográfico e são exemplos importantes: A heresia dos Índios (1995), de Ronaldo Vaifas, tratou de um processo da Inquisição sobre um movimento profético sincrético (cristão e tupi) do interior baiano baseado na perspectiva da história cultural; Ronald Raminelli, em Imagens da Colonização. Representação do Índio de Caminha a Vieira (1996), debruçou-se sobre a temática da representação indígena com uso de fontes iconográficas e apontou um “descompasso” entre imagens e os relatos coloniais. A publicação de História dos Índios no Brasil (1992), organizado pela antropóloga Manuela Carneiro da Cunha foi, sem dúvida, um marco importante. O trabalho reuniu resultados de pesquisa em história, arqueologia e linguística vinculados ao Núcleo de História Indígena e do Indigenismo da USP. Embora o livro seja uma contribuição importante para preencher vazios sobre a historia dos índios no Brasil, a própria organizadora reconhece que se trata de um conjunto fragmentado de estudos de casos, “[...] que permitem imaginar, mas Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... não preencher as lacunas de um quadro [...]” que poderia ser mais completo. Em que se pese o pouco conhecimento sobre o assunto, o fato é que “[...] hoje está mais claro, pelo menos, a extensão do que não se sabe” (CUNHA, 2006, p. 11). q ( p ) Em 1994, outro grande esforço para o estímulo de trabalhos em história indígena foi publicado: o Guia de Fontes para a História Indígena e do Indigenismo em arquivos brasileiros. Acervos das Capitais. Organizado por John Manuel Monteiro19, o instrumento de pes- quisa agrega informações documentais e propõe um intercâmbio de informações sobre os índios. No projeto, uniram-se pesquisadores ligados a universidades em cada estado (apenas Tocantins, Amapá e Roraima não puderam ser contemplados). Assim como História dos Índios no Brasil, o Guia era uma tentativa de elaborar uma agenda engajada com o passado e com o futuro dos povos indígenas, recha- çando a tese do desaparecimento e unindo esforços em conjunto com as demandas dos próprios índios. Nas palavras do organizador, tratava-se de “um passo inicial – tímido, no caso de alguns estados, forte e decidido, em outros” (1994, p. 07), mas que impossibilitaria, daquele momento em diante, a explicação da ausência de história por carência de documentação. Éi 194 É bastante significativa a publicação, no mesmo ano do Guia, do principal livro sobre história indígena no Brasil: Negros da Terra: índios e bandeirantes nas origens de São Paulo, de John Manuel Monteiro. O trabalho assumia, na prática, os compromissos e desafios da agenda anunciada anteriormente. Originado da tese defendida pelo autor em 1985, na Universidade de Chicago, o livro foi enriquecido por seus estudos nos anos seguintes. Negros da Terra consiste em uma interpretação renovada e crítica da história social do período colonial em São Paulo. O estudo mostra como praticamente a totalidade de aspectos da “[...] formação da sociedade e economia paulista durante seus primeiros dois séculos confundem-se de modo essencial com os processos de integração, exploração e destruição de populações indígenas trazidas de outras regiões” (MONTEIRO, 1994, p. 09). O livro aborda aspectos caros à historiografia brasileira, como a relação do trabalho dos índios com a economia colonial. Ainda maculada por uma espécie de mito, um equívoco largamente reproduzido em livros escolares, de que o índio não era apto para o trabalho, a Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo historiografia carecia de um estudo exaustivo, com uso de fontes diversificadas, sobre escravidão indígena. Estampada no título do livro, está também a problemática sobre a relação entre a história dos índios e a história dos negros no Brasil. Negros da terra era uma designação usada para os índios, remetendo à equivalência e ao mesmo tempo a uma diferenciação entre eles e os negros da Guiné. Esses termos reportavam-se simultaneamente a uma equivalência de status servil e a uma diferenciação de origem (LANGFUR, 2005). 195 ç g ( , ) Outro estudo importante é a tese de livre docência do mesmo autor, do ano de 2001, intitulada Tupis, Tapuias e Historiadores: estudos de história indígena e do indigenismo. Nela, John Monteiro apresenta um conjunto elaborado de artigos e ensaios que tratam de temas e problemas seminais da historiografia indígena brasileira, muitos deles colocados nas discussões apresentadas anteriormente. Entre os temas centrais do conjunto estão a “defasagem e os deslizamen- tos temporais” quanto ao uso das fontes históricas; as refratadas construções etnográficas dos passados indígenas feitas em tempos imperiais; o estabelecimento de um “padrão bipolar” de análise do passado indígena, o padrão Tupi/Tapuia, muitas vezes travestido em outros nomes e colocado em distintos tempos históricos e historiográficos, ora analiticamente, ora como política para e pelos índios (2001, p. 08). Em análises astutas, o historiador tece ligações e desconstrói mitos sobre os períodos colonial e imperial brasileiro. O tema do extermínio historiográfico e real dos índios também perpassa todo o trabalho, mas ganha destaque no capítulo Entre o etnocídio e a etnogênese, em que o autor chama a atenção dos historiado- res para a complexidade de formação de novas identidades surgidas no período colonial. Tais transformações podem ser vistas como estratégias políticas dos grupos ou indivíduos – o autor destaca a atuação das lideranças indígenas na chave da agência – “perante a catástrofe demográfica, as estratégias militares, evangelizadoras e econômicas dos europeus” (2001, p. 75). Assim, ao localizar o pro- blema do vazio historiográfico sobre os indígenas no século XIX como fruto da relação entre suas próprias histórias e o surgimento da História como disciplina, John Monteiro esclarece problemáticas e contextos, bem como oferece pistas para o futuro da historia dos índios no Brasil. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... A produção de John Monteiro influenciou as gerações seguin- tes de historiadores dos índios. Sob sua orientação direta20 e em outros centros de pesquisa, diversos estudos avançaram em apontar a participação indígena na história e cobriram distintas temáticas e períodos.21 Essa primeira leva de pesquisadores, através de extensas pesquisas documentais, debruçaram-se especialmente sobre o período colonial. Temas como os conflitos gerados pela interiorização da conquista, a dependência dos colonizadores perante os índios e a consequente gama de alianças geradas e as disputas pela força de trabalho foram os principais pontos debatidos por Pedro Puntoni (2002) e Glóri Kok (2004). Outro problema trazido à tona pelos estudiosos dos índios nesse momento foi a relação entre a religio- sidade e a vida em missões, evidenciando a complexidade do tema da tradução cultural, autores como Cristina Pompa (2003) e Almir Diniz de Carvalho Jr. (2005) são exemplos disso. A utilização de documentação produzida pelos índios foi explorada por Eduardo dos Santos Neumann (2005), demonstrando como os Guaranis se apropriaram com propriedade das instituições coloniais. Já a tese de Mariana de Campos Françozo (2009), apresenta como foram construídos, concomitantemente, os saberes coloniais por meio de análise da circulação de pessoas, objetos e discursos entre o Brasil e a Holanda. A Maria Regina Celestino de Almeida cabe menção especial, visto que seu trabalho teve um importante alcance e colaborou para a divulgação da temática. Além disso, ela é responsável por toda uma outra leva de orientações na Universidade Federal Fluminense, um dos principais centros de produção do conhecimento histórico do país. Em Metamorfoses Indígenas: identidade e cultura nas aldeias coloniais do Rio de Janeiro (2003), a autora trata do lugar dos índios nas aldeias cariocas e as disputas pelo uso de seu trabalho e terras, bem como sua autoria quanto às decisões sobre esses processos. Esse traba- lho modifica a visão estática e passiva sobre os índios ao oferecer uma nova leitura dos seus papéis de liderança, mediadores, aliados e por vezes inimigos dos portugueses na construção do universo colonial. Ao fazer uso de fontes variadas e provenientes de diversos lugares, algumas das quais bastante conhecidas, a autora mostra como interessantes processos de reconstrução das identidades e Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo cultura indigenas foram forjados em espaços, onde, à primeira vista, não existiriam brechas para romper com a rígida hierarquia social dos tempos coloniais. Para o período imperial e republicano, os trabalhos tornam- -se um tanto mais escassos se comparados ao colonial. Mas alguns mostram a riqueza da gama de possibilidades, seja sob o aspecto das fontes, metodologia ou recortes: a discussão sobre o tema dos distintos regimes de historicidade proposto por Izabel Missagia de Mattos (2004) no contexto dos aldeamos capuchinhos; os problemas da imigração do século XIX e as possibilidades que levam em conta trajetórias de indivíduos nos contextos de contato através da adoção ou do rapto de indígenas são apresentados por Luisa Wittmann (2007). Lucybeth Camargo de Arruda (2012) mostra como o uso de fotografias pode ser uma interessante fonte para compreender a atuação indígena nos Postos do SPI, para além de uma visão que os colocava como receptáculos de políticas governamentais durante a década de 1940. Outros estudos têm se dedicado à compreensão do papel das lideranças indígenas nas lutas pelo direito territorial, como a tese de Edson Hely Silva (2008), que demonstra a força do uso de fontes orais para construção de novas narrativas históricas. O uso e a discussão sobre a memória indígena foi tratada por Giovani José da Silva (2009), na fronteira Brasil-Bolívia, sob uma atenta ótica transnacional dos processos vividos entre as décadas de 1930 e 1980. Também tem crescido o número de trabalhos sobre o Movimento indígena brasileiro, que em breve deverá contar com a produção de intelectuais “orgânicos”. i 197 Em toda essa reconfiguração ocorrida nas últimas décadas no campo da historiografia brasileira sobre os índios, os trabalhos e a atuação profissional de John Monteiro tiveram papel decisivo. Como vimos anteriormente, o pesquisador teve sua formação vinculada às transformações que o campo sofria nos EUA, e incorporou as inovações teóricas e metodológicas, principalmente oriundas da etno-história. Seu trabalho é fruto de um profundo conhecimento e diálogo entre a historiografia norte-americana e brasileira sobre os índios em ambos os países, tema central deste artigo. Atento às imagens dos índios na historia e na historiografia, o pesquisador fazia uso constante da analise delas nas aulas, cursos, palestras e Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... conferências que ministrava. Em mais de uma ocasião, usou a imagem intitulada “Dança Tapuia”, de Albert Eckhout. O pintor holandês acompanhou a corte de Mauricio de Nassal quando houve a ocupação holandesa de parte do atual nordeste brasileiro. Como funcionário da Companhia das Indias Ocidentais, desembarcou no Recife, em 1637 (BRIENEN, 2006). As pinturas de Eckout trazem consigo um conjunto importante de informações sobre o contexto da cultura visual da época, sobre os imaginários classificatórios dos holande- ses e as maneiras escolhidas para retratar os índios, sobretudo os índios Tapuia, relativamente marginais do conjunto de representa- ções coloniais voltadas para os grupos Tupi. A análise das imagens possibilita um melhor entendimento sobre tipos de as experiencias, materiais, ideias e práticas utilizadas pelos europeus para retratar o novo mundo. Além disso, oferece caminhos para repensar as dinâ- micas coloniais à medida que recompõe relevantes relações entre a interferência editorial e a utilização política dessas imagens. Porém, para alcançar todo um conjunto de informações e questões que uma fonte histórica pode trazer, é preciso saber ler além das palavras, advertiu o historiador durante o I Fórum Internacional da Temática indígena, realizado em junho de 2010, em Porto Alegre.22 Atento às temporalidades, ao momento da produção da obra, ao momento de sua publicação e aos usos feitos da imagem, John Monteiro chamava a atenção para o fato de que a Dança Tapuia aparecia em sites e em capa de livro de maneira incompleta, com parte dos personagens ausentes da obra. Tal operação de apagamento era feita através de um recorte, que por sua vez descontextualizava e tornava impossível perceber a totalidade da história retratada. Assim, parte da com- preensão da fonte feita literalmente pelas metades, John Monteiro enfatizava que, com os recortes, o historiador perde informações importantes sobre zoologia, botânica, técnicas de arte e contextos socioculturais, além da “devolução do olhar” de uma das personagens. Em outras palavras, perde a interação entre a fonte e o espectador proposta pelo autor da obra; interação essa tão fundamental para captar elementos da cultura de contato quanto aspectos sobre as ações dos índios 198 198 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Figura 3: Dança Tapuia, óleo sobre tela, ca. 1643. Albert Eckout (ca. 1607- ca. Sobrevoando histórias: sobre índios e historiadores no Brasil... Para as autoras, foi impactante perceber que, cerca de três anos mais tarde, numa ponta oposta das Américas, historiadores da nova historia indígena usavam o mesmo argumento para demonstrar o que às vezes parece tão difícil de ser entendido, que não se pode ensinar a história das Américas sem a história dos indios. 200 Sobrevoando histórias: sobre índios e historiadores no Brasil... 1666) Figura 3: Dança Tapuia, óleo sobre tela, ca. 1643. Albert Eckout (ca. 1607- ca. 1666) Fonte: Disponível em: https://upload.wikimedia.org/wikipedia/commons/5/54/ Dan%C3%A7a_dos_Tapuias.jpg. Acesso em: 21 jul. 2015. ( ) Fonte: Disponível em: https://upload.wikimedia.org/wikipedia/commons/5/54/ Dan%C3%A7a_dos_Tapuias.jpg. Acesso em: 21 jul. 2015. 199 Fonte: Disponível em: https://upload.wikimedia.org/wikipedia/commons/5/54/ Dan%C3%A7a_dos_Tapuias.jpg. Acesso em: 21 jul. 2015. Figura 4: Dança Tapuia, dividida aleatoriamente em duas partes, para ilustrar como a obra costuma aparecer em websites e livros didáticos. Fonte: Disponível em: https://upload.wikimedia.org/wikipedia/commons/5/54/ Dan%C3%A7a_dos_Tapuias.jpg. Acesso em: 21 jul. 2015. Fonte: Disponível em: https://upload.wikimedia.org/wikipedia/commons/5/54/ Dan%C3%A7a_dos_Tapuias.jpg. Acesso em: 21 jul. 2015. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo ONE FLIGHT OVER HISTORIES: ABOUT INDIANS AND HISTO- RIANS IN BRAZIL AND AMERICA Abstract: Brazilian and North American historiography share many aspects when it comes to indigenous issues. In both cases, the histories of native groups changed the ways of producing knowledge about them, creating and transforming public policy. Games of complex influences guided the ways of dealing with the knowledge about inter-ethnic relations. In many cases, such knowledge served as a fulcrum for the survival of the implicated groups. Historiographical trajecto- ries, here and there, are full of convergence, divergence, dynamism and political complexity. That said, the purpose of this article is to present a vision of the two parallel processes of construction of historical discourses about Indians and trace from there, agendas and possibilities of mutual contributions. 201 Keywords: Indians. Brazilian Indigenous History. North American Indian History. Historiography. À guisa de conclusão Escrever e ensinar sobre a história dos índios nas Américas não é uma tarefa fácil. Em 1987, Michael Dorris, pesquisador indígena da etnia Modoc, e então diretor do Núcleo de Estudos Indígenas em Dartmouth College, em New Hampshire, apontou caminhos estratégicos para que os pesquisadores pudessem encarar, talvez com maior êxito, esse desafio. Escreveu que uma séria consideração sobre história indígena deve começar com uma “inicial, abrupta e dolorosa ‛desmitologização’”. De acordo com a antropóloga Jennifer Brown e a historiadora Elizabeth Vibert, o que Dorris chamou de “desmi- tologização”, alguns estudiosos têm chamado de “descolonização do conhecimento” (BROWN; VIBERT, 2003, p. 12). Descolonizar o conhecimento significa entender melhor o projeto de colonização, ou seja, buscar compreender da maneira mais profunda possível, os contextos históricos específicos em que diversos tipos de narrativas sobre os europeus e o outro foram produzidas. É preciso confrontar a lógica de ideias arraigadas sobre cultura e diferença cultural para romper com o que Daniel Usner considerou um dos legados mais duradouros e enigmáticos do colonialismo, o discurso infeliz que prende os índios entre dois polos opostos, o da autencidade e o da aniquilação e oculta a maneira pela qual sociedades dinamicas e duraveis buscam conciliar tradição e inovação em sua acomodação e resistência diante de forças poderosas (USNER, 2009). Romper com esse discurso na prática do fazer históriografico abre novos caminhos para tornar a presença dos índios menos discreta – que, neste caso, é o mesmo que dizer mais real – no âmbito dos estudos sobre historia social de distintas partes das Américas. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... Pós-Graduação em História – Departamento de História, UFRGS, Porto Alegre, 2006. Para informações sobre a institucionalização da Antropologia e Etnologia, ver: ERIKSEN, Thomas Hylland; NIELSEN, Finn Sivert. História da Antropologia. Petrópolis: Vozes, 2007. Pós-Graduação em História – Departamento de História, UFRGS, Porto Alegr 2 Para mais informações sobre o Indian Removal Act, consultar: SPENCE, Mark David. Dispossessing the Wilderness: Indian Removal and the Making of the National Parks. Oxford University Press, 2000. Esse livro também traz uma interessante análise sobre a formação dos parques nacionais não tratados neste artigo.i 202 3 Conforme afirmamos anteriormente, outros grandes massacres podem ser identi- ficados na historiografia Americana: Sacramento river, Moutain Meadows, Marias River, Camp Grant, entre outros. Para mais informações sobre esses eventos, consultar McMurtry, Larry. On what a slaughter. Massacres in the American West: 1846-1890. New York: Simon & Schuster, 2005. 4 De maio a outubro de 1893, a cidade de Chicago sediou A Feira Mundial que comemorava os 400 anos da chegada de Colombo a América e colocava os EUA simbolicamente dentro do grupo de países que podiam celebrar e apontar o futuro, apresentando novidades tecnológicas, científicas, gastronômicas e socioculturais. Entre as atrações, o já famoso espetáculo do empresário e artista William Frederick Cody, o Buffalo Bill, O oeste selvagem, destacava-se. Contava com a participação de indígenas de diversos grupos, inclusive a liderança Touro Sentado, bem como atiradores, militares, turcos, mongóis, cossacos, árabes. Em encenações que contavam uma história do Oeste, o espetáculo levou aos EUA e a Europa, aquilo que os intelectuais tentavam resolver: do que era feita a América. p q q q Paralelamente à Feira, ocorria um congresso de historiadores no qual Frederick Turner apresentava suas teses sobre a fronteira. Uma excelente análise sobre esse “encontro” é feita por Richard White no ensaio Frederick Jackson Turner and Buffalo Bill. In. The Frontier in American Culture. ed. James R. Grossman. Berkeley: University of California Press, 1994. 5 JOSEPHY, Alvin M. The Nez Perce and the Opening of the West. Boston: Mariner Books, 1997. BERTHRONG, Donald J. The Cheyenne and Arapaho Ordeal: Reservation and Agency Life in the Indian Territory, 1875-1907. London: University of Oklahoma Press, 1992. 6 A autora concentra seu trabalho no sul dos Estados Unidos (Lousiana e Alabama), onde os embates raciais eram claramente polarizados entre brancos e negros. Notas 1 Segundo Artur Lima de Avila, o processo de instituicionalização da História como disciplina nos EUA está imbricada com a atuação direta de Frederick Jackson Turner e da Western History como motivo temático central, “Em suas mãos, a história do Oeste, ou melhor, da fronteira, era a história dos Estados Unidos” (2006, p. 36). O autor argumenta que Turner fora responsável pela for- mação de grande parte dos historiadores graduados nos EUA no final do século XIX e início do XX, atuando na Universidade Estadual de Wisconsin até 1910 e em Harvard, posteriormente. A História Americana era fruto de um processo contínuo de ocupação do Atlântico ao Pacífico, “[…] A tese de Turner deu movimento ao espaço nacional, conferindo-lhe uma identidade através deste processo. Os personagens de sua história eram homens e mulheres anônimos, que moviam-se sobre o continente americano, impelidos por um destino que lhes era anterior [o chamado Destino Manifesto]. Do choque dialético entre a selvageria, representada pela natureza indomada, e a civilização, simbolizada pelos colonos, nasceu a excepcional democracia norte-americana. A condição material para o surgimento deste regime político sem igual eram as chamadas “terras livres” do Oeste, territórios apenas esperando para serem ocupados por seres humanos estóicos e dedicados à árdua labuta na terra. A expansão, contudo, havia chegado ao fim em 1890, quando o Censo Nacional declarou estarem ocupadas todas estas terras livres. O país estava fechado em si mesmo” (2006, p. 37). Para mais informações sobre Turner, fronteira Americana e Destino Manifesto, ver: AVILA, Arthur Lima de. E da Fronteira veio um pioneiro... a frontier thesis de Frederick Jackson Turner (1861-1932). Dissertação (Mestrado). Programa de Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo 8 Segue uma lista com algumas obras dos autores mencionados que não puderam ser tratadas no corpo do artigo: AXTELL, James. The invasion within: the contest of cultures in Colonial North America. New York: Oxford University Press, 1985.; _________. Natives and newcomers: the cultural origins of North America. 8 Segue uma lista com algumas obras dos autores mencionados que não puderam ser tratadas no corpo do artigo: AXTELL, James. The invasion within: the contest of cultures in Colonial North America. New York: Oxford University Press, 1985.; _________. Natives and newcomers: the cultural origins of North America. New York: Oxford University Press, 2001; DOWD, Gregory E. A spirited resistance the North American Indian struggle for unity, 1745-1815. Baltimore: Johns Hopkins University Press, 1992; EDMUNDS, R. David. Tecumseh and the Quest for Indian Leadership. Boston: Little, Brown, 1984; O’BRIEN, Jean M. After King Philip’s War: presence and persistence in Indian New England. Hanover University Press of New England, 1997; RICHTER, Daniel. The Ordeal of the Longhouse: The Peoples of the Iroquois League in the Era of European Colonization. Chapel Hill: University of North Carolina Press, 1992; PERDUE, Theda. Cherokee women: gender and culture change, 1700-1835. Lincoln: University of Nebraska Press, 1998. 203 9 A publicação, em 2011, de Native historians write back: decolonizing American Indian history, reunindo ensaios e artigos de autores nativos de distintos grupos sintetiza como as perspectivas nativas podem apontar para caminhos renova- dores da história americana. MILLER, Susan A; RIDING, James (Org.). Native historians write back: decolonizing American Indian history. Lubbock: Texas Tech University Press, 2011. 10 Todas as impressões sobre este evento são fruto dos entendimentos das autoras que tiveram a oportunidade de participar do mesmo como pesquisadoras associadas à Newberry Library naquele momento. A imagem do cartaz do evento encontra-se 10 Todas as impressões sobre este evento são fruto dos entendimentos das autor que tiveram a oportunidade de participar do mesmo como pesquisadoras associad 10 Todas as impressões sobre este evento são fruto dos entendimentos das autoras que tiveram a oportunidade de participar do mesmo como pesquisadoras associadas à Newberry Library naquele momento. A imagem do cartaz do evento encontra-se disponível em: http://www.newberry.org/past-american-indian-studies-seminars. Acesso em: 15 set. 2014. 11 Para mais informações sobre o IHGB, consultar SCHWARCZ, Lilia Katri Moritz. Os Institutos Históricos e Geográficos. Os guardiões de nossa história oficial. São Paulo: Editora Vértice/Idesp, 1989. Sobrevoando histórias: sobre índios e historiadores no Brasil... Uma das principais críticas coladas na época e de consequências permanentes, é que as questões raciais nos EUA tenderam a respostas negras, para problemas e demandas que eram muito mais diversas. 7 Para mais informações sobre as conecções entre os movimentos de contracultura e as questões indígenas, consultar: SMITH, Sherry L. Hippies, Indians & the figth for Red Power. New York: Oxford University Press, 2012. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo 12 Para mais informações sobre o tema da Lei de Terras, ver SILVA, Lígia Osório. Terras Devolutas e Latifúndio. Campinas: Editora da Unicamp, 2008. Embora a autora concentre-se no processo paulista, pode-se ter uma visão geral do processo em termos judiciais e de direitos sobre terras dos índios.i 13 OTTONI, Teófilo Benedito; OTTONI, Honório Benedito. Condições para Incorporação de uma Companhia de Comércio e Navegação do Rio Mucuri, precedidas de uma exposição das vantagens da empresa. Rio de Janeiro: Tipografia Imperial e Constitucional de J. Velleneuve e Companhia, 1847; _____. Notícia Sobre os Selvagens do Mucuri. DUARTE, Regina Horta (Org.). Belo Horizonte: Editora UFMG, 2002; _______. Circular dedicada aos Srs. Eleitores de senadores pela província de Minas Gerais no quadriênio atual e especialmente dirigida aos Srs. eleitores de deputados pelo 2º distrito eleitoral da mesma Província para a próxima legislatura. Revista do Instituto Histórico e Geográfico Brasileiro. Tomo LXXVIII, Parte 2. Rio de Janeiro, 1916; ________. A Colonização Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... do Mucuri. Rio de Janeiro: Tipografia Brasiliense de Maximiano Gomes Ribeiro, 1859. In: ARAUJO, Valdei Lopes de (Org.). Teófilo Benedito Ottoni e a Companhia do Mucuri: a modernidade possível. Belo Horizonte: Secretaria de Estado da Cultura; Arquivo Público Mineiro, 2007; _______. Companhia do Mucuri. História da empresa, importância de seus privilégios, alcance dos seus projetos. Rio de Janeiro: Tipografia Imperial e Constitucional de J. Villeneuve e Companhia, 1856. In.: ARAUJO, Valdei Lopes de (Org.). Teófilo Benedito Ottoni e a Companhia do Mucuri: a modernidade possível. Belo Horizonte: Secretaria de Estado da Cultura; Arquivo Público Mineiro, 2007; ______. Breve Resposta ao Relatório de Liquidação da Cia. do Mucuri, por parte do Governo. Rio de Janeiro: Tip. de M. Barreto, Mendes Campos e Comp., 1862; _______. Discursos Parlamentares. Seleção e introdução de Paulo Pinheiro Chagas. Brasília: Câmara dos Deputados, 1979. _______. Relatório apresentado aos acionistas da Companhia do Mucuri por Teófilo Benedito Ottoni em 15 de outubro de 1857. Rio de Janeiro: Tipografia Imperial e Constitucional de J. Villeneuve e Companhia, 1857. (Coleção Assuntos Mineiros). In. ARAUJO, Valdei Lopes de (Org.). Teófilo Benedito Ottoni e a Companhia do Mucuri: a modernidade possível. Belo Horizonte: Secretaria de Estado da Cultura; Arquivo Público Mineiro, 2007; _______. Relatório apresentado aos acionistas da Companhia do Mucuri no dia 15 de maio de 1860. Rio de Janeiro: Tipografia do Correio Mercantil, 1860. (Coleção Assuntos Mineiros) In. ARAUJO, Valdei Lopes de (Org.). Teófilo Benedito Ottoni e a Companhia do Mucuri: a modernidade possível. Belo Horizonte: Secretaria de Estado da Cultura; Arquivo Público Mineiro, 2007. 204 14 ANDRADE, Osvald de. O Manifesto Antropofágico. Disponível em http:// serantropofagia.wordpress.com/about/. Acesso em: 15 set. 2014. 15 Curt Unckel foi naturalizado brasileiro e batizado pelos Guarani-Ñandeva com o nome Nimuendajú, aquele que faz a sua morada, em tupi-guarani. 16 AMOROSO, Marta Rosa. Nimuendajú às voltas com a história. Revista de Antropologia, São Paulo: USP, 2001, v. 44, n. 2. Disponível em http://www.ifch. unicamp.br/ihb/HZ868-06/MartaNimuendajuRA.pdf. Acesso em: 22 out. 2014. 17 Para mais informações, FREIRE, Carlos Augusto da Rocha (Org.). Memória do SPI. Textos, imagens e documentos sobre o serviço de proteção aos índios (1919-1967). Rio de Janeiro M se do Índio FUNAI 2011 16 AMOROSO, Marta Rosa. Nimuendajú às voltas com a história. Revista de Antropologia, São Paulo: USP, 2001, v. 44, n. 2. Disponível em http://www.ifch. unicamp.br/ihb/HZ868-06/MartaNimuendajuRA.pdf. Acesso em: 22 out. 2014. Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo 20 John Manuel Monteiro lecionou na UNESP (Araraquara, Assis e Franca) entre 1986 e 1991, e na Universidade Estadual de Campinas (UNICAMP) desde 1994, vinculado tanto ao Departamento de Antropologia quanto ao de História. Foi professor visitante na Harvard University (2003-2004), University of Michigan (1997) e University of North Carolina-Chapel Hill (1985-1986). Além de sua experiência como docente, foi “Directeur d’Études Invité”, na EHESS, em Paris (1999) e pesquisador do CEBRAP de 1991 a 1998. Sua atuação junto à Associação Brasileira de Antropologia (ABA), a Associação Nacional de História (ANPUH) e a Associação Nacional de Pós-Graduação e Pesquisa em Ciências Sociais (ANPOCS) são demonstrativos de sua circulação e abertura para diálogo e articulação. Era também coordenador do Grupo de Trabalho sobre temática indígena na ANPUH que reunia pesquisadores de todo o país. John Monteiro faleceu em março de 2013, aos 56 anos, em um acidente de trânsito. 205 21 Felizmente, por os índios terem se tornado tema de estudo de um grande e eficiente grupo de historiadores e antropólogos, este artigo não poderá contem- plar a totalidade de trabalhos antigos e novos produzidos após a década de 1990. Nosso critério foi priorizar os estudos publicados ou aqueles que se destacaram em temáticas específicas. Uma relação praticamente completa pode ser encon- trada no site criado por John Monteiro, Os Índios na História do Brasil. Informações, estudos, imagens. Ali, além das indicações, são apresentadas resenhas dos trabalhos. Disponível em http://www.ifch.unicamp.br/ihb/. Acesso em: 15 set. 2014. 21 Felizmente, por os índios terem se tornado tema de estudo de um grande e eficiente grupo de historiadores e antropólogos, este artigo não poderá contem- plar a totalidade de trabalhos antigos e novos produzidos após a década de 1990. Nosso critério foi priorizar os estudos publicados ou aqueles que se destacaram em temáticas específicas. Uma relação praticamente completa pode ser encon- trada no site criado por John Monteiro, Os Índios na História do Brasil. Informações, estudos, imagens. Ali, além das indicações, são apresentadas resenhas dos trabalhos. Disponível em http://www.ifch.unicamp.br/ihb/. Acesso em: 15 set. 2014. 22 John Monteiro realizou a conferência de encerramento do I Fórum Internacional da Temática Indígena, a convite das autoras deste artigo, também idealizadoras e organizadoras do referido evento. Sobrevoando histórias: sobre índios e historiadores no Brasil... 17 i i f C l A d h ( ) M d 17 Para mais informações, FREIRE, Carlos Augusto da Rocha (Org.). Memória do SPI. Textos, imagens e documentos sobre o serviço de proteção aos índios (1919-1967). Rio de Janeiro: Museu do Índio-FUNAI, 2011. 18 O Índios Kaingang no Rio Grande do Sul. Pesquisas, Antropologia, n. 29, 1976. 19 A história de criação do Guia de Fontes mescla-se com a do Núcleo de Pesquisas da História dos Índios e do Indigenismo da USP. Manuela Carneiro da Cunha coordenava um projeto interinstitucional e interdisciplinar na tentativa de elaborar um Guia para o campo em crescimento aos moldes do Guia de fontes para a história da África, da escravidão negra e do negro na sociedade atual, publicado pelo Arquivo Nacional em 1988, centenário da abolição. Por razões diversas, o projeto não se concretizou por essas vias, mas deu origem ao Núcleo (MONTEIRO, 1994, p. 07-08). Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Referências ALMEIDA, Maria Regina Celestino de. Metamorfoses indígenas: identidade e cultura nas aldeias coloniais do Rio de Janeiro. Rio de Janeiro: Arquivo Nacional, 2003. ALMEIDA, Maria Regina Celestino de. Metamorfoses indígenas: identidade e cultura nas aldeias coloniais do Rio de Janeiro. Rio de Janeiro: Arquivo Nacional, 2003. AVILA, Arthur Lima de. Território Contestado: a reescrita da história do Oeste norte- -americano (c. 1985-c. 1995). Tese (Doutorado). Programa de Pós-Graduação em História – UFRGS, Porto Alegre, 2010. AVILA, Arthur Lima de. Território Contestado: a reescrita da história do Oeste norte- -americano (c. 1985-c. 1995). Tese (Doutorado). Programa de Pós-Graduação em História – UFRGS, Porto Alegre, 2010. ______. E da Fronteira veio um pioneiro... a frontier thesis de Frederick Jackson Turner (1861-1932). Dissertação (Mestrado). Programa de Pós-Graduação em História – UFRGS, Porto Alegre, 2006. AXTELL, James. Ethnohistory: An Historian’s Viewpoint. Ethnohistory, v. 26, n. 1, 1979, p.1-13. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... AXTELL, James. The invasion within: the contest of cultures in Colonial North America. New York: Oxford University Press, 1985. ______. Natives and newcomers: the cultural origins of North America. New York: Oxford University Press, 2001 BARNES, Ian. The Historical Atlas of Native Americans. New York: Chartwell Books, 2010. 206 BATES, Denise E. The Other Movement: Indian Rights and Civil Rights in the Deep South. Alabama: The University of Alabama Press, 2012. BLACKHAWK, Ned. Look how far we’ve come: how American Indian History changed the study of American History in the 1990s. OAH Magazine of History. v. 19, n. 6, American West, Nov., 2005, p. 13-17. BRIENEN, Rebecca Parker. Visions of Savage Paradise: Albert Eckhout, Court Painter in Colonial Dutch Brazil. Amsterdam: Amsterdam Press, 2006. BROWN, Dee. Bury My Heart at Wounded Knee; an Indian history of the American West. New York: Holt, Rinehart & Winston, 1971. BROWN, Jennifer S. H.; VIBERT, Elizabeth. Reading beyond words: contexts for native history. Canada: Broadview Press Ltd, 2003. CALLOWAY, Colin G. New worlds for all: Indians, Europeans, and the remaking of early America. Baltimore: Johns Hopkins University Press, 1997. CARVALHO Jr., Almir Diniz de. Índios cristãos: a conversão dos gentios na Ama- zônia portuguesa (1653-1769). Doutorado (Tese) Campinas, 2005. CAVENAGHI, A. J. Uma guerra, dois mapas e duas fotografias. O Sertão do Noroeste Paulista e a aventura do registro iconográfico ao final do século XIX. Revista Projeto História. São Paulo, n. 32, p. 191-219, jun. 2006. CUNHA, Manuela Carneiro da. Política Indigenista no século XIX. In. História dos índios no Brasil. São Paulo: Companhia das Letras, 1992. DEBO, Angie. And still the Waters Run. The betrayal of the five civilized tribes. New Jersey: Princeton University Press, [1940] 1968. DELORIA Jr., Vine. Custer Died for Your Sins: An Indian Manifesto. New York: Macmillan, [1969] 1988. ERIKSEN, Thomas Hylland; NIELSEN, Finn Sivert. História da Antropologia. Petrópolis: Vozes, 2007. FREITAS, Edinaldo Bezerra de. A construção do imaginário nacional: entre repre- sentações e ocultamentos. As populações indígenas e a historiografia. Revista Labi- rinto. 2000. Disponível em: http://www.cei.unir.br/artigo103.html. Acesso em: 02 out. 2014. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Soraia Sales Dornelles, Karina Moreira Ribeiro da Silva e Melo HÄMÄLÄINEN, Pekka. The Futures of Native American History in the United States. In: Perspectives on History. The Newmagazine of American Historical Association. Dezembro de 2012. Disponível em: http://www.historians.org/publications-and- directories/perspectives-on-history/december-2012/the-future-of-the-discipline/ the-futures-of-native-american-history-in-the-united-states. Acesso em: 03 set. 2014. HÄMÄLÄINEN, Pekka. The Futures of Native American History in the Unit States. In: Perspectives on History. The Newmagazine of American Historical Associati KOK, Glória. O Sertão Itinerante: expedições da Capitania de São Paulo no Século XVIII. São Paulo: Hucitec, 2004. 207 MARTIUS, Carl Friedrich Phillip von. O Estado de Direito entre os autóctones do Brasil. Belo Horizonte: Itatiaia, 1982. McDONNELL, Janet A. The Dispossession of the American Indian, 1887-1934. Bloo- mington: Indiana University Press, 1991. MERRILL, James. The Indians’ New World: Catawbas and Their Neighbors from European Contact through the Era of Removal. Williamsburg: University of North Carolina Press, 1989 MONTEIRO, John Manuel. Negros da terra: índios e bandeirantes nas origens de São Paulo. São Paulo: Companhia das Letras, 1994. ______. Tupis, tapuias e historiadores. Estudos de história indígena e do indigenismo. Tese apresentada para o concurso de Livre Docência em Antropologia na Universi- dade Estadual de Campinas. Campinas, 2001. MORGAN, Lewis Henry. League of the Hode ‘nosaunee, or Iroquois. vol. 2. Rochester/ New York, 1851; rept. New York: Dodd, Mead, 1901, p. 10813. Disponível em: http://www.fordham.edu/halsall/mod/1851morgan.asp. Acesso em 03 set. 2014. O’BRIEN, Jean M. After King Philip’s War: presence and persistence in Indian New England. Hanover University Press of New England, 1997. POMPA, Cristina. Religião Como Tradução: missionários, Tupi e Tapuia no Brasil colonial. Bauru: EDUSC, 2003. PUNTONI, Pedro. A Guerra dos barbarous: povos indígenas e a colonização do sertão nordeste do Brasil, 1650-1720. São Paulo: Hucitec/Edusp, 2002. SHEPHERD, Jeffrey P. From Savages to Sovereigns: A General Historiography of American Indian History. University of Texas at El Paso. s/d. Disponível em: https://www.google.com.br/urlsa=t&rct=j&q=&esrc=s&source=web&cd= 1&cad=rja&uact=8&ved=0CB0QFjAA&url=https%3A%2F%2Ffaculty.utep. edu%%2FI. Acesso em: 15 set. 2014. SMITH, Sherry L. Hippies, Indians & the figth for Red Power. New York: Oxford University Press, 2012. Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015 Sobrevoando histórias: sobre índios e historiadores no Brasil... SPENCE, Mark David. Dispossessing the Wilderness: Indian Removal and the Making of the National Parks. Oxford University Press, 2000. USNER, Daniel H. Indian Work: Language and Livelihood in Native American History.Cambridge: Harvard University Press, 2009. VAINFAS, Ronaldo. A Heresia dos Índios. Catolicismo e Rebeldia no Brasil Colonial. São Paulo: Companhia das Letras, 1995. WHITE, Richard. The roots of dependency: subsistence, environment, and social change among the Choctaws, Pawnees, and Navajos. Lincoln: University of Nebraska Press, 1983. ______. The Middle Ground: Indians, Empires, and Republics in the Great Lakes Region, 1650-1815. New York: Cambridge University Press, 1991. ______. Frederick Jackson Turner and Buffalo Bill. In. WHITE, R., LIMERICK, P. The Frontier in American Culture. An exhibition at the Newberry Library, August 26, 1994-January 7, 1995. Barkeley, Los Angeles, London: University of California Press, 1994. p. 07-66. WITTMANN, Luisa Tombini. O Vapor e o Botoque: imigrantes alemães e índios Xokleng no Vale do Itajaí/SC (1850-1926). Florianópolis: Letras Contemporâ- neas, 2007. Recebido em: 31/10/2014 Aprovado em: 18/05/2015 Recebido em: 31/10/2014 Aprovado em: 18/05/2015 Anos 90, Porto Alegre, v. 22, n. 41, p. 173-208, jul. 2015
https://openalex.org/W1903443724	https://ru.historicas.unam.mx/bitstream/20.500.12525/1562/1/3089-Texto%20del%20trabajo-3040-1-10-20091023.pdf	es		Luis Fernando Granados, Sueñan las piedras. Alzamiento ocurrido en la Ciudad de México, 14, 15 y 16 de septiembre de 1847	Estudios de historia moderna y contemporánea de México	2,003	cc-by	1,086	RESEÑAS BIBLIOGRÁFICAS 161 Luis Fernando Granados, Sueñan las piedras. Alzamiento ocurrido en la ciudad de México, 14, 15 y 16 de septiembre de 1847, México, Era/Consejo Nacional para la Cultura y las Artes, Instituto Nacional de Antropología e Historia, 2003 , 173 p., ils., planos (Colección Problemas de México). Sueñan las piedras es el pormenorizado relato que hace Luis Fernando Granados de la sublevación popular contra el ejército de los Estados Unidos en los primeros días de la ocupación militar de la ciudad de México. En definitiva, no es una obra acerca de la guerra de 18461848; es, más bien, un estudio de historia urbana y social. También es un ensayo de microhistoria y un extraño ejercicio narrativo. Extraño, porque nunca había visto una narración hecha a base de preguntas. El autor se propuso hacer una reconstrucción minuciosa de acontecimientos difíciles de precisar, a partir de testimonios muchas veces encontrados. De ahí que buena parte del relato se halle entre signos de interrogación que permiten ir avanzando en el desarrollo de los sucesos. De ahí también la pertinencia de la primera cuestión: “¿Existen los hechos?” No estoy muy seguro de la respuesta de Luis Fernando Granados, porque después de leer el libro, tampoco sé qué responder. Para empezar, el autor hace un recuento breve de las pocas miradas historiográficas sobre la rebelión popular de la ciudad de México contra el ejército estadounidense. Resalta que pocos autores se hayan ocupado del asunto, sobre todo si se considera que el episodio bien pudo haber sido interpretado como hizo Ernesto Lemoine en su tesis profesional, esto es, como una muestra espontánea del patriotismo de las clases populares de la ciudad de México, ofendidas al ver el pabellón de las barras y las estrellas ondeando en el Palacio Nacional. Esta falta de atención puede deberse, asegura el autor, al temor de las elites políticas del país al populacho y, también, al desdén de los historiadores por este tipo de fenómenos. Son, por cierto, estas razones las que proporcionan 162 ESTUDIOS DE HISTORIA MODERNA Y CONTEMPORÁNEA DE MÉXICO al autor las principales hipótesis y modo de abordar el tema. Luis Fernando Granados decidió tratar la sublevación plebeya de 14, 15 y 16 de septiembre de 1847 desde abajo y desde la ciudad, no como un episodio más de la guerra con los Estados Unidos sino como uno de la larga historia de amotinamientos y sublevaciones de los léperos capitalinos. Al mismo tiempo, procuró no encasillar a los pobres de la ciudad de México en los rígidos estancos de las clases sociales, lo cual le hubiera permitido explicar, de un modo más o menos fácil, los hechos como lucha de clases. De seguro, habría algo de verdad en una interpretación así, aunque presentaría algunos problemas, tanto teóricos como empíricos. Sin embargo, el principal inconveniente de una explicación basada en el comportamiento de las clases sociales de mediados del siglo XIX hubiera sido la deshumanización de esta historia, y Luis Fernando Granados está muy interesado en poner nombre y apellido a esos pobres individuos que arengaron a las piedras para que salieran de su sueño. Por esto, ha procurado hacer una reconstrucción tan detallada; por eso mismo, las dudas acerca de la existencia de los hechos. De tal detalle es el relato de Sueñan las piedras que sería necio intentar un resumen aquí. Cada uno de los capítulos (salvo el último) se refiere a cada día y noche de la ciudad de México a partir de la ocupación del ejército estadounidense. Se presentan las versiones de los testigos y aquellas que el autor considera más verosímiles. Es un verdadero juego de detectives: ¿quién lanzó la primera piedra?, ¿dónde se hallaba tal regimiento?, ¿qué barrios participaron en los amotinamientos? La lectura se vuelve más interesante a cada momento y la trama se va complicando hasta que se hacen necesarios un alto y una explicación. De esto trata el último apartado, titulado “Las piedras, los pobres”. Aquí, el autor procura hacer un poco de sociología de la capital de la república mexicana, no tanta como para borrar los rostros que, con tanto cuidado (casi forense), ha dibujado en los capítulos anteriores. Con el apoyo de autores como Richard Warren, Frederick Shaw y Torcuato S. di Tella, Luis Fernando Granados muestra las delgadas líneas que separaban los estratos sociales y propone una geografía de la pobreza que bien explica la de los enfrentamientos de mediados de septiembre. Si el libro tuviera notas a pie de página (pero las tiene al final) diría —tomando una frase de Álvaro Matute —que se lee casi por entero en el piso de abajo. En efecto, el lector poco interesado en dar RESEÑAS BIBLIOGRÁFICAS 163 seguimiento a las notas tendrá una lectura menos interesante y una explicación menos satisfactoria de la historia que nos cuenta el autor. La dificultad de cambiar de páginas a cada instante (las notas son muchas, como no podía ser de otra manera en un ejercicio como el que propone Granados) hubiera hecho muy recomendable que los comentarios del autor se hallaran en el mismo texto y dejara para las notas sólo el aparato erudito. Salvo por esto y por algunas ilustraciones a las que se remite en el texto pero que no existen, el libro puede considerarse una obra que no por breve es fundamental para comprender los motivos de los léperos capitalinos y también los de una elite que prefirió colaborar con el ejército de los Estados Unidos en el apaciguamiento de la ciudad a arriesgar su posición privilegiada. Después de todo, al autor no se le escapa que el amotinamiento popular se presentó en el momento en que las autoridades mexicanas iban desapareciendo y todavía no se hacían tan presentes las estadounidenses. Para los criollos de la ciudad de México (y del país entero, si recordamos las rebeliones en el Sur, la Sierra Gorda y Yucatán), estaba abierta la posibilidad de la pérdida de su posición hegemónica frente a los amplios grupos sociales hasta entonces oprimidos. Hay que decir que este temor, manifiesto, por ejemplo, en los escritos de José María Luis Mora, estaba fundado; la guerra sólo debilitó más a un Estado que se hallaba sobre una olla de presión. Episodios como los ocurridos en la capital en septiembre de 1847 mostraron a la elite la fragilidad de su dominio y la capacidad de los líderes populares y de los léperos para actuar en condiciones más o menos propicias para mostrar su enojo y su fuerza. Alfredo ÁVILA Instituto de Investigaciones Históricas, UNAM
https://openalex.org/W4296012619	https://dash.harvard.edu/bitstream/1/2770519/2/Rolian_WhyToesTiny.pdf	English	null	Abstracts of AAPA poster and podium presentations	American journal of physical anthropology	2,007	cc-by	573	Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA Permanent link http://nrs.harvard.edu/urn-3:HUL.InstRepos:2770519 Published Version http://dx.doi.org/10.1002/ajpa.20577 Published Version http://dx.doi.org/10.1002/ajpa.20577 Citation Rolian, Campbell, Daniel E. Lieberman, John W. Scott. 2007. Why are our toes so tiny? Walking, running and the evolution of a short forefoot in the genus Homo. Abstracts of AAPA poster and podium presentations. American Journal of Physical Anthropology 132(S44): 202. Rolian, Campbell, Daniel E. Lieberman, John W. Scott. 2007. Why are our toes so tiny? Walking, running and the evolution of a short forefoot in the genus Homo. Abstracts of AAPA poster and podium presentations. American Journal of Physical Anthropology 132(S44): 202. Why are our toes so tiny? Walking, running and the evolution of a short forefoot in the genus Homo Campbell Rolian, Daniel E. Lieberman, John W. Scott Campbell Rolian, Daniel E. Lieberman, John W. Scott Campbell Rolian, Daniel E. Lieberman, John W. Scott Share Your Story The Harvard community has made this article openly available. Please share how this access benefits you. Submit a story . Accessibility Why are our toes so tiny? Walking, running and the evolution of a short forefoot in the genus Homo Campbell Rolian, Daniel E. Lieberman, John W. Scott Why are our toes so tiny? Walking, running and the evolution of a short forefoot in the genus Homo Campbell Rolian Daniel E Lieberman John W Scott Abstract: Humans have an extremely short forefoot relative to total foot length. The derived pedal proportions of humans are thought to have evolved in the context of committed bipedalism, but the benefits of shorter toes for walking and/or running have not previously been tested. Short toes are typically associated with cursorial digitigrade mammals, where they improve the ability of the digital flexor apparatus – the muscles, tendons and ligaments that collectively flex and resist extension of the Humans have an extremely short forefoot relative to total foot length. The derived pedal proportions of humans are thought to have evolved in the context of committed bipedalism, but the benefits of shorter toes for walking and/or running have not previously been tested. Short toes are typically associated with cursorial digitigrade mammals, where they improve the ability of the digital flexor apparatus – the muscles, tendons and ligaments that collectively flex and resist extension of the metatarsophalangeal (MTP) joints – to support the body and generate propulsion at the end of stance. We tested the hypothesis that in humans a shorter forefoot similarly improves locomotor performance by decreasing the force, power and work outputs of the digital flexor apparatus (DFA) during late stance, especially in running, when only one foot provides support and propulsion against high ground reaction forces. Kinematic, force and plantar pressure data were collected from a sample representing normal variation in toe length (n=12). Hindlimb kinematics, DFA force, power and work outputs were compared during barefoot walking and running in subjects with short, average and long forefeet in relation to body mass. Results suggest that individuals with relatively longer forefeet experience higher MTP joint moments, and their DFA generates more force, power and work than subjects with shorter forefeet, at both walking and running speeds. Contrary to our prediction, however, the difference between groups in DFA performance is not greater at running speeds. Implications for the evolution of endurance running in the genus Homo are discussed.
https://openalex.org/W3042710344	https://tc.copernicus.org/articles/12/189/2018/tc-12-189-2018.pdf	English	null	Modelling debris transport within glaciers by advection in a full-Stokes ice flow model	null	2,017	cc-by	14,310	Received: 18 May 2017 – Discussion started: 1 June 2017 Revised: 10 November 2017 – Accepted: 20 November 2017 – Published: 19 January 2018 Received: 18 May 2017 – Discussion started: 1 June 2017 Revised: 10 November 2017 – Accepted: 20 November 2017 – Published: 19 January 2018 Received: 18 May 2017 – Discussion started: 1 June 2017 Revised: 10 November 2017 – Accepted: 20 November 2017 – Published: 19 January 2018 Abstract. Glaciers with extensive surface debris cover respond differently to climate forcing than those with- out supraglacial debris. In order to include debris-covered glaciers in projections of glaciogenic runoff and sea level rise and to understand the paleoclimate proxy recorded by such glaciers, it is necessary to understand the manner and timescales over which a supraglacial debris cover devel- ops. Because debris is delivered to the glacier by processes that are heterogeneous in space and time, and these de- bris inclusions are altered during englacial transport through the glacier system, correctly determining where, when and how much debris is delivered to the glacier surface requires knowledge of englacial transport pathways and deforma- tion. To achieve this, we present a model of englacial de- bris transport in which we couple an advection scheme to a full-Stokes ice ﬂow model. The model performs well in numerical benchmark tests, and we present both 2-D and 3- D glacier test cases that, for a set of prescribed debris in- puts, reproduce the englacial features, deformation thereof and patterns of surface emergence predicted by theory and observations of structural glaciology. In a future step, cou- pling this model to (i) a debris-aware surface mass balance scheme and (ii) a supraglacial debris transport scheme will enable the co-evolution of debris cover and glacier geometry to be modelled. or basal erosion (Benn and Evans, 2010). Rock and dust de- bris deposited onto the surface of a glacier in the accumula- tion zone is buried by subsequent snowfall and transported englacially with the glacier ice as it ﬂows downslope. In the ablation zone of a glacier, ice ﬂow transports debris towards the glacier surface and surface ice ablation leaves behind a residue of rock material (Fig. 1a). If debris supply and abla- tion is sufﬁciently high, and transport of rock material out of the glacier system is inefﬁcient, a debris-covered glacier can develop, where a large portion of the ablation zone is covered with a continuous layer of rock material (Kirkbride, 2011). Received: 18 May 2017 – Discussion started: 1 June 2017 Revised: 10 November 2017 – Accepted: 20 November 2017 – Published: 19 January 2018 A surface debris cover more than a few centimetres thick inhibits surface ablation of ice and thus alters glacier runoff, local water resources and contribution to sea level change. It also affects glacier dynamics and geometry such that stag- nating, low-angled debris-covered ice can survive for longer at lower altitudes than neighbouring clean-ice glaciers (Benn et al., 2012; Anderson and Anderson, 2016). Thus the pa- leoclimatic signal represented by sediment deposits from a debris-covered glacier is not the same as one from a clean- ice glacier. Fluxes of ice and debris change over time in response to climatic variations and in space due to differences in local site characteristics. Kirkbride (1989) proposes that variations in ice mass inﬂux serve to unify a process continuum of deformational geomorphological features of mixed ice and debris composition. The implication of this process contin- uum is that glaciers can transition between rock glaciers, debris-covered glaciers and clean-ice glaciers through space or time as a result of the varying ice inﬂux (Kirkbride, 1989; Ackert Jr., 1998; Clark et al., 1998). Accordingly, the extent of a debris cover varies inversely with the glacier 1 Introduction All mountain glaciers carry rock and dust material within the ice. This can originate from gravitational mass move- ments from the surrounding valley walls, aeolian deposition The Cryosphere, 12, 189–204, 2018 https://doi.org/10.5194/tc-12-189-2018 © Author(s) 2018. This work is distributed under the Creative Commons Attribution 3.0 License. The Cryosphere, 12, 189–204, 2018 https://doi.org/10.5194/tc-12-189-2018 © Author(s) 2018. This work is distributed under the Creative Commons Attribution 3.0 License. Modelling debris transport within glaciers by advection in a full-Stokes ice ﬂow model Anna Wirbel1, Alexander H. Jarosch2, and Lindsey Nicholson1 1Institute of Atmospheric and Cryospheric Sciences, University of Innsbruck, Innsbruck, Austria 2Institute of Earth Sciences, University of Iceland, Reykjavík, Iceland Correspondence: Anna Wirbel (anna.wirbel@uibk.ac.at) Correspondence: Anna Wirbel (anna.wirbel@uibk.ac.at) Correspondence: Anna Wirbel (anna.wirbel@uibk.ac.at) Received: 18 May 2017 – Discussion started: 1 June 2017 Revised: 10 November 2017 – Accepted: 20 November 2017 – Published: 19 January 2018 A. Wirbel et al.: Modelling debris advection in glaciers This requires tackling many component parts to model the full debris- covered glacier system. To date, existing numerical mod- els of debris-covered glaciers either restrict debris inputs to the ablation zone (Konrad and Humphrey, 2000; Menounos et al., 2013; Vacco et al., 2010), prescribe an englacial debris concentration (Bozhinskiy et al., 1986) or use empirical rela- tionships to describe accumulation of debris on the glacier surface (Jouvet et al., 2011). Recent studies apply simpli- ﬁed treatment of englacial transport (Rowan et al., 2015; Anderson and Anderson, 2016), but as yet no model ex- plicitly resolves fully 3-D (three-dimensional) time-evolving transport of debris within the ice ﬂow ﬁeld of the glacier body. This is a signiﬁcant omission because, as surface de- bris mainly originates from localized debris inputs (rockfall or mixed avalanche events) in the accumulation zone, mod- elling englacial transport is crucial to predict the location and timing of surface emergence of debris, as well as its con- centration and its spatial extent, all of which are required to constrain the nature of the developing debris cover and its resultant impact on glacier behaviour. mass balance, whereby the debris cover extent is governed by transport-dominant conditions (higher ice velocities and lower surface ablation) during periods of positive mass bal- ance and ablation-dominant conditions (lower ice velocities and higher ablation) during periods of negative mass balance (Kirkbride, 2000). In addition, debris covers can be formed instantaneously by isolated events such as ash fall or a large rockfall onto the glacier ablation zone (e.g. Nield et al., 2013; Hewitt, 2009; Shugar et al., 2012; Reznichenko et al., 2011). The complex interplay between debris supply and ice supply and their variation in space and time mean that the thickness and character of the debris cover, and its resultant impact on the glacier behaviour is also strongly space and time depen- dent. In order to resolve this, it is necessary to understand how the debris co-evolves with the glacier. This requires tackling many component parts to model the full debris- covered glacier system. To date, existing numerical mod- els of debris-covered glaciers either restrict debris inputs to the ablation zone (Konrad and Humphrey, 2000; Menounos et al., 2013; Vacco et al., 2010), prescribe an englacial debris concentration (Bozhinskiy et al., 1986) or use empirical rela- tionships to describe accumulation of debris on the glacier surface (Jouvet et al., 2011). A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers Figure 1. (a) Debris-covered Kennicott Glacier, Wrangell Mountains, Alaska, USA, and (b) deformed englacial bands emerging at the surface of Skeiðarárjökull, an outlet glacier of Vatnajökull ice cap, Iceland. Source: ESRI Basemap DigitalGlobe imagery. Figure 1. (a) Debris-covered Kennicott Glacier, Wrangell Mountains, Alaska, USA, and (b) deformed englacial bands emerging at the surface of Skeiðarárjökull, an outlet glacier of Vatnajökull ice cap, Iceland. Source: ESRI Basemap DigitalGlobe imagery. Individual clasts are considered to be predominantly pas- sively transported by glacier ice, unless within the basal trac- tion zone, and so their shape remains fundamentally unal- tered by transport. For a static velocity ﬁeld, the pathway of such a feature is identical to a streamline within the glacier, but with the evolving glacier geometry these will change. However, debris inputs from rock, mixed snow–ice avalanches and other gravitational mass movements tend to be deposited as bodies of polymictic ice–sediment mixtures, which become severely deformed in the course of trans- port through the glacier, as revealed by studies of structural glaciology (e.g. Fig. 1b; Jennings et al., 2014; Mackay et al., 2014). Hence, the initial shape of the deposit will be changed signiﬁcantly, and this englacial deformation will affect the pattern of debris emergence at the glacier surface (e.g. Good- sell et al., 2005). In order to numerically model transport and deformation of sediment inclusions, a full representation of 3-D velocity ﬁelds resolving all spatial gradients is essential, which calls for a full-Stokes ice ﬂow modelling approach. mass balance, whereby the debris cover extent is governed by transport-dominant conditions (higher ice velocities and lower surface ablation) during periods of positive mass bal- ance and ablation-dominant conditions (lower ice velocities and higher ablation) during periods of negative mass balance (Kirkbride, 2000). In addition, debris covers can be formed instantaneously by isolated events such as ash fall or a large rockfall onto the glacier ablation zone (e.g. Nield et al., 2013; Hewitt, 2009; Shugar et al., 2012; Reznichenko et al., 2011). The complex interplay between debris supply and ice supply and their variation in space and time mean that the thickness and character of the debris cover, and its resultant impact on the glacier behaviour is also strongly space and time depen- dent. In order to resolve this, it is necessary to understand how the debris co-evolves with the glacier. Published by Copernicus Publications on behalf of the European Geosciences Union. Published by Copernicus Publications on behalf of the European Geosciences Union. 190 2.2 Advection of material within a glacier To describe transport and associated deformation of advected material within a glacier, we employ the linear transient advection–diffusion equation: This paper is structured as follows: Sect. 2 provides details about the equations governing ice ﬂow and how the transport problem is addressed from a mathematical perspective. Sec- tion 3 describes the numerical schemes employed. Section 4 details the test simulations performed, the results of which are presented in Sect. 5 and discussed in terms of model per- formance, limitations and applicability in Sect. 6. Conclu- sions and outlook are presented in Sect. 7. ∂c ∂t = ∇· (D∇c) −∇· (uc) + r in (5a) c = 0 on ∂0, (5b) (5a) (5a) (5b) (5b) where c is the concentration of the material, D ≥0 is the diffusion coefﬁcient, u is the divergence-free velocity ﬁeld and r represents any internal sources or sinks. In the case of transport of debris through a glacier, it is reasonable to assume r = 0 and that material is predominantly transported by advection; therefore we currently neglect diffusion by set- ting D sufﬁciently small. As we focus on englacial transport in this study, we set c = 0 at the domain boundaries (∂0) except the parts of the boundary where an input location is assigned. Based on these assumptions and deﬁning a con- stant diffusion coefﬁcient D, Eq. (5) becomes A. Wirbel et al.: Modelling debris advection in glaciers The computa- tional domain is conﬁned by a free surface boundary at the ice–air interface (∂top) satisfying where ε = 1/2(∇u+(∇u)T) is the strain rate tensor and ˙ϵ = p0.5εijεji is the effective strain rate. A represents the Glen rate factor and n the Glen ﬂow law exponent. The computa- tional domain is conﬁned by a free surface boundary at the ice–air interface (∂top) satisfying 2ηε · n −pn = 0 on ∂top, (3) (3) where n is the outward pointing surface normal. At the ice– bedrock interface (∂bed) either a no-slip (Dirichlet) bound- ary condition (Eq. 4a) or, for glaciers where basal sliding contributes to total movement, an alternative ice–bedrock (Neumann) boundary condition can be applied (Eq. 4b) in conjunction with an appropriate sliding law conditioning the interface parallel ice velocity components at the glacier base. The model presented here forms part of an envisaged fully integrated model framework that, by incorporating a (1) free surface evolution scheme including debris-aware mass bal- ance subroutines and (2) transport model for debris at the glacier surface interacting with the mass balance subrou- tines, will be capable of simulating the transient response of debris-covered glaciers, with predetermined debris inputs, to a changing climate. (4a) (4b) u = 0 on ∂bed (4a) u · n = 0 on ∂bed (4b) (4b) A. Wirbel et al.: Modelling debris advection in glaciers Recent studies apply simpli- ﬁed treatment of englacial transport (Rowan et al., 2015; Anderson and Anderson, 2016), but as yet no model ex- plicitly resolves fully 3-D (three-dimensional) time-evolving transport of debris within the ice ﬂow ﬁeld of the glacier body. This is a signiﬁcant omission because, as surface de- bris mainly originates from localized debris inputs (rockfall or mixed avalanche events) in the accumulation zone, mod- elling englacial transport is crucial to predict the location and timing of surface emergence of debris, as well as its con- centration and its spatial extent, all of which are required to constrain the nature of the developing debris cover and its resultant impact on glacier behaviour. Here we present a new model that simulates transport, and resultant deformation, of material within a glacier coupled to 3-D resolved ice ﬂow, and we demonstrate the capabilities and performance of the model through a series of evalua- tion simulations. Although dynamics of debris–ice mixtures can differ from clean-ice dynamics depending on several pa- rameters such as concentration of debris, particle size and temperature (Moore, 2014), in this work we assume that sed- iment inclusions within the glacier do not affect ice rheology due to the small total amount of transported material in com- parison to overall ice volume. The model is coded in python and relies on the FEniCS framework, an open-source soft- ware for automated solution of partial differential equations (PDEs) (Alnæs et al., 2015; Logg et al., 2012a). The model employs an existing benchmarked full-Stokes ice ﬂow model (icetools; Jarosch, 2008, now implemented in FEniCS) to compute 3-D velocity ﬁelds that govern an advection algo- rithm used to describe debris transport. Prevailing stress conditions, and the resulting strain and velocity ﬁelds, control sediment transport within an ice body. www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers 191 Assuming that ice is an incompressible ﬂuid, and conse- quently that the ice ﬂow ﬁelds must be divergence-free, any deformational patterns inducing horizontal elongation must, at the same time, cause vertical compression. In the context of englacial debris transport, this implies that the initial de- bris concentration is constant for an initial control volume of ice being tracked (i.e. seen from Lagrangian perspective). The incompressibility assumption demands also the abso- lute values of concentration to remain constant during trans- port when following the initial control volume of ice as it is becoming deformed during transport. To solve the trans- port problem mathematically, we take an Eulerian approach. The accuracy of the results is directly related to mesh size. If the mesh size chosen was inﬁnitesimally small, the con- centration features would recover over the entire transport path, their initial values and sharp layers at their boundaries (if initial debris inputs are delineated by sharp boundaries). As a consequence of a ﬁxed mesh, the bigger the mesh size, the greater the amount of numerical smearing in the simula- tions which results in a decrease of maximum and marginal concentration values and a smearing towards the edges of the concentration features. This problem is inherent to the method chosen but can be dealt with by applying an appro- priate mesh size, such that numerical smearing is minimized according to the application of interest. −∇· [η(∇u + (∇u)T)] + ∇p = ρiceg in (1a) ∇· u = 0 in . (1b) (1a) (1b) (1b) Here u is the 3-D velocity ﬁeld, η is the non-linear viscosity, p is the pressure, ρice is the density of ice and g is the ac- celeration due to gravity. The density of ice is assumed to be constant in time and space. By including the standard rheol- ogy of ice (Glen, 1955; Nye, 1957), its non-linear viscosity can be described by η = 1 2A −1 n ˙ϵ (1−n) n , (2) η = 1 2A −1 n ˙ϵ (1−n) n , (2) where ε = 1/2(∇u+(∇u)T) is the strain rate tensor and ˙ϵ = p0.5εijεji is the effective strain rate. A represents the Glen rate factor and n the Glen ﬂow law exponent. 3 Numerical schemes and model software The model consists of three main components, (1) ice de- formation (icetools), (2) adaptive mesh reﬁnement accord- ing to concentration patterns (reﬁne_gl) and (3) debris trans- port (advect_gl). All model components are individual open- source modules coded in python and utilizing the FEniCS framework (Alnæs et al., 2015; Logg et al., 2012a). Com- putations are performed on unstructured meshes (triangles in 2-D (two dimensions) and tetrahedrons in 3-D), which al- low for variable mesh size according to local requirements in spatial resolution and geometry complexity. Computational meshes are generated with gmsh (Geuzaine and Remacle, 2009), an open-source ﬁnite element mesh generator. For 2-D simulations, mesh reﬁnement is implemented en- tirely in python using the FEniCS software framework (Al- næs et al., 2015; Logg et al., 2012a). A function representing the coordinates (Fig. 2b) where the concentration exceeds a threshold (Fig. 2a) and Rcells is deﬁned on the domain-wide coarse mesh (Fig. 2c). Using this function, the cells to be re- ﬁned can be marked at any stage of mesh reﬁnement. The marked cells are reﬁned uniformly until all affected cells have an area smaller than a threshold cvol (Fig. 2d). In this study, cvol is set to 0.075 m2 (equivalent to an equilateral triangle of edge length 0.416 m), which, according to the ﬁndings in Sect. 5.1, is suitable to successfully represent englacial debris features originating from surface layer de- posits of several metres in thickness. For 3-D simulations, gmsh (Geuzaine and Remacle, 2009) is used to create a new reﬁned mesh at every reﬁnement time step. Therefore, a domain-wide coarse mesh is updated with information on the coordinates where concentration exceeds a threshold. To create the new mesh, the cell size within the radius Rcells of these coordinates is set to Lcsize. This parameter is rep- resentative of the average cell size within this area. In order to further reduce the number of required cells, the mesh is primarily reﬁned in streamline direction. Therefore, the co- ordinate points are ﬁrst shifted using the present velocity and a time step of 0.75dtADV. As the reﬁnement is based on the same domain-wide coarse mesh for each reﬁnement time step, both approaches (2-D and 3-D mesh reﬁnement) prevent over-reﬁnement and, at the same time, the need for mesh coarsening. 3.2 Ice deformation Three-dimensional ice velocities are computed using ice- tools, a parallelized, open-source full-Stokes model for ice ﬂow (Jarosch, 2008) that solves Eqs. (1)–(4). A mixed func- tion space of continuous piecewise quadratics and linears is used to compute ice velocity and pressure. The capability of the model to simulate 3-D velocity ﬁelds for complex ice bodies has been demonstrated in previous studies (Jarosch, 2008; Jarosch and Gudmundsson, 2012). Initial versions of icetools accounted for stress-dependent ice viscosity using a Picard iteration scheme, but here we employ an updated version, where the non-linear problem of including stress- dependent viscosity is solved by the Newton method. Adaptive mesh reﬁnement strategies often employ a pos- teriori error estimation (e.g. John, 2000). The PDE is solved and the assigned error estimators and indicators are used to mark the cells for reﬁnement and potentially coarsen- ing. Subsequently, the marked cells become modiﬁed and 3.1 FEniCS software FEniCS is an open-source project designed for automated solution of PDEs by ﬁnite element methods (FEM) (https: //fenicsproject.org, Alnæs et al., 2015; Logg et al., 2012a). It includes several components such as DOLFIN (Logg and Wells, 2010; Logg et al., 2012c), FFC (Kirby and Logg, 2006; Logg et al., 2012b; Ølgaard and Wells, 2010) and FIAT (Kirby, 2004, 2012), which enable automatic solution of lin- ear and non-linear problems once the variational forms of the PDEs are expressed in the Uniﬁed Form Language (UFL; Al- næs et al., 2014; Alnæs, 2012). A. Wirbel et al.: Modelling debris advection in glaciers At starting time t0, a known initial concentration is given for all grid points on locations (x) by umented since 1900. For the documented events, they found mean deposit lengths (n = 55) and ﬁnal thickness (n = 20) of 6.3 (max. 18.2/min. 1.4) km and 3.5 (max. 22.0/min. 1.0) m respectively. These values represent the upper limits on the likely dimensions of individual rockfall events onto glaciers, although megaslides could have larger dimensions. In order to resolve debris or ash deposits covering the range of these observations, computational meshes are required to have a spatial resolution in the submeter scale. In the case of simu- lating glaciers several kilometres long, this would lead to im- mense computational costs. Therefore, we take the approach of reﬁning the mesh locally, i.e. only those areas where con- centration is present. In order to avoid mesh reﬁnement at every computation time step, we increase the area of reﬁne- ment by a spatial radius Rcells = umaxdtADVcref surrounding the concentration features, where umax is the maximum ve- locity in the reﬁned area, dtADV is the reﬁnement time step (see Sect. 3.5) and cref is a positive deﬁned constant. In this manner, the mesh is reﬁned in an area that covers the ac- tual concentration feature and the distance it can possibly be transported within the reﬁnement time step. (7) c(x,t = t0) = c0(x) in . Apart from an initial concentration, material can enter the domain at the boundaries as a single input at time tinput or by a deﬁned rate as a function of time and location. 2.1 Full-Stokes formulation for ice ﬂow Ice is treated as an incompressible, non-linear viscous ﬂuid, whose velocity and pressure distribution can be described by the incompressible stationary Stokes equations on a spatial domain ϵR3, representing the ice body: ∂c ∂t = D∇2c −u · ∇c in (6a) c = 0 on ∂0. (6b) (6a) (6b) The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 192 192 A. Wirbel et al.: Modelling debris advection in glaciers 193 Figure 2. Illustration of mesh reﬁnement. (a) Initial concentration ﬁeld where bright colours indicate high concentration values. (b) Coor- dinate points of grid locations where concentration exceeds a threshold of 0.01. (c) Function that shows high values, indicated by bright colours, at all cells of the coarse mesh that lie within the radius Rcells of the respective coordinate points in (b). (d) Representation of the ﬁnal reﬁned mesh. Figure 2. Illustration of mesh reﬁnement. (a) Initial concentration ﬁeld where bright colours indicate high concentration values. (b) Coor- dinate points of grid locations where concentration exceeds a threshold of 0.01. (c) Function that shows high values, indicated by bright colours, at all cells of the coarse mesh that lie within the radius Rcells of the respective coordinate points in (b). (d) Representation of the ﬁnal reﬁned mesh. the PDE is solved on the newly reﬁned mesh. This process is repeated until the error estimators and indicators fall be- low a user-deﬁned tolerance within every cell. This proce- dure can require many iterations, resulting in high computa- tional costs. In the case of debris inclusions within glaciers, we deal with rather smooth concentration ﬁelds except for sparse areas of high concentration that often, initially or over time exhibit a band-like shape. Therefore, rather than focus- ing only on sharp layers that are the main contributors to high errors on too-coarse meshes, we perform the reﬁnement on the entire area of high concentration. Instead of using error estimators and indicators to locate the cells for mesh reﬁne- ment, our methods use a cell-based, concentration-threshold reﬁnement indicator. This indicator is computed just once per reﬁnement time step and all affected cells and those within a velocity-based distance (Rcells) are (a) reﬁned until a problem-speciﬁc mesh size tolerance (cvol) is achieved (2-D) or (b) created with a cells size prescribed by Lcsize (3-D). The results of the benchmark test in Sect. 5.1 derived with our ap- proach compare well with that derived using adaptive mesh reﬁnement based on a posteriori error estimation (de Frutos et al., 2014), which demonstrates that our approach is an ac- ceptable balance between accuracy and savings in compu- tational costs. The 3-D benchmark test in Sect. 5.1 demon- strates the suitability of the 3-D mesh reﬁnement approach. Brooks, 1982). 3.4 Material transport The transient advection–diffusion equation is discretized in time by an implicit Euler scheme and a standard continu- ous Galerkin FEM is used for the space discretization, sep- arating the temporal and spatial discretizations. The con- centration is expressed as a scalar function in a continuous piecewise linear function space. In the case of advection- dominated transport, solving Eq. (6) via standard continuous Galerkin FEM leads to non-physical spurious oscillations (e.g. Bochev et al., 2004). In order to inhibit these spurious oscillations and ensure stability, we employ the streamline- upwind Petrov–Galerkin (SUPG) approach (Hughes and A. Wirbel et al.: Modelling debris advection in glaciers In this method, a residual-based stabilization term is added to the variational form and in this way intro- duces artiﬁcial diffusion to the system in streamline direc- tion. The stabilization term is based on the residual of Eq. (6) including the time derivative and a mesh-size dependent sta- bilization parameter τ. Following John and Novo (2011) and Bochev et al. (2004), in the advection-dominated case (i.e. Peclet numbers greater than 3) we use a stabilization term of O(hK) and deﬁne it as τ = hK 2\|\|u\|\|, where hK is a measure of the local cell size and u is the divergence-free velocity ﬁeld. 3.5 Time stepping The reﬁnement time step prescribes how often the reﬁnement has to be performed and, in conjunction with the velocity ﬁeld, deﬁnes the total number of cells in the mesh. For ex- ample, for a given englacial debris concentration, the total number of cells in the reﬁned mesh increases with increasing reﬁnement time step as the distance the debris inclusions are transported within this time step also increases. In order to minimize computational effort, the reﬁnement time step has to be chosen according to the characteristics of the computer system that is used to run the computations. The computation time step for the advection module is de- rived using the Courant–Friedrich–Lewy condition, applied on the smallest cell size and the maximum velocity within the reﬁned region. In this study, we apply Courant numbers ranging from 0.5 to 1.5. The work of Bochev et al. (2004) combined with the tests in the Supplement, show that the SUPG stabilization scheme coupled to a Crank–Nicholson or an implicit Euler scheme for time-dependent advection- dominated advection–diffusion problems is stable for this choice of Courant numbers. 4 Model simulations Direct evaluation of our advection model against real-world glacier cases is not possible at present because (i) compre- 3.3 Mesh reﬁnement Deline et al. (2015) collated statistics of the dimensions of deposits from massive rock slope failures onto glaciers doc- The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers 194 hensive ﬁeld measurements of englacial debris transport are not available and (ii) simulating the full glacier system would require further model development as outlined in Sect. 1. in Example 4 in de Frutos et al. (2014) for an advection- dominated case (D = 10−6 m2 s−1). However, we employ the mesh reﬁnement and time stepping described in Sect. 3 and redeﬁne the model domain as = (0,100)×(0,100) m. By doing this enlargement, the size of concentration features becomes comparable to the size of debris inputs in the glacier simulations. Initializing the mesh reﬁnement with the same cell area threshold as is used in the glacier simulations allows us to estimate the level of accuracy that we can achieve in the glacier cases. Here, we present results of computations using two different reﬁnement time steps: (a) a small reﬁnement time step of 0.01π s and (b) a larger reﬁnement time step of 0.1π s. This results in (a) 200 and (b) 20 reﬁnement time steps for a full rotation of 2π s (ttotal). The computation time step is derived using a Courant number of 0.5. In order to evaluate the chosen cell area threshold, we perform conver- gence tests where (a) ∥ch −ce∥L2 = qPn k=1(chk −cek)2 the L2 norm of the error between the computed ﬁnite element so- lution ch and the exact solution ce, where n is the number of computation locations, and (b) the root mean square (RMS) error between the computed ﬁnite element solution and the exact solution for different cell size thresholds is computed. Therefore, we ﬁrst compute the exact solution on the same mesh that is used in the ﬁnite element solution. To subject the model to an even more severe test, a second set of simu- lations is performed where the velocity ﬁeld is prescribed as a swirling ﬂow (LeVeque, 1996), but all other settings remain identical. Due to the swirling ﬂow, the shapes of the three bodies become deformed, but at total time (ttotal, at t = 1.5 s) the three bodies recover their initial shape. An animation of the swirling ﬂow is included in the Supplement. q p Nevertheless, an evaluation of how well the model (a) per- forms and (b) reproduces structures observed in glaciers is important. 4.1 Benchmark tests The numerical Examples 1 and 2 in Bochev et al. (2004) are chosen to demonstrate the effect of the SUPG stabilization approach in terms of reducing non-physical spurious oscil- lations that are a known problem for standard continuous Galerkin FEM schemes in the case of advection-dominated problems and the stability of this stabilization scheme for a wide range of Courant numbers that control the time step- ping. Details of this set of numerical tests are presented in the Supplement and here we present only the results of the most demanding numerical test to which we subjected the model, which is Example 4 in de Frutos et al. (2014). This test is known as the “rotating three body problem” (LeVeque, 1996; John and Novo, 2011; de Frutos et al., 2014) and is a standard test for computing advection of a scalar quantity in an incom- pressible ﬂow ﬁeld using the transient advection–diffusion equation in the advection-dominated case. Furthermore, in the study of de Frutos et al. (2014), the capabilities of a pos- teriori error-based adaptive mesh reﬁnement are evaluated. By comparing the published results to those reproduced with our implementation of adaptive mesh reﬁnement, we evalu- ate our method. In the “rotating three body problem”, a slot- ted cylinder, a hump and a conical body undergo clockwise rotation in a divergence-free velocity ﬁeld. A visualization of the velocity ﬁeld is given in Fig. S4a in the Supplement. In order to reproduce the results with our model, we set up the velocity ﬁeld, initial and boundary conditions as described To test the model capabilities in 3-D, we reproduce the nu- merical test described in Christensen (1993), where a sphere of high concentration undergoes rotation. In this test, the ve- locity ﬁeld is constructed in such a manner that the shape of the sphere is deformed throughout the rotation, but after a full rotation of 2π s (ttotal) the sphere recovers its initial shape. A visualization of the velocity ﬁeld is given in Fig. S4b. The model domain is deﬁned as = (0,32)×(0,32)×(0,40) m and the mesh reﬁnement is initialized with Lcsize = 0.15 m. The reﬁnement time step is set to 0.04π s and the Courant number to 0.5. By comparing the results to the analytical so- lution of the problem presented in Christensen (1993), model performance and chosen reﬁnement settings can be evalu- ated. A. Wirbel et al.: Modelling debris advection in glaciers To this end, we present results from speciﬁc nu- merical tests that benchmark the advection module, followed by glacier simulations for a 2-D proﬁle of an alpine val- ley glacier and an idealized 3-D glacier geometry. In these glacier simulations the ﬂow ﬁelds are computed by solving Eqs. (1)–(4) for given geometries and are kept ﬁxed as no mass balance routine is coupled to the model yet. The bench- mark tests were performed to quantitatively ascertain that the numerics of our model adequately meet the requirements of the task in terms of mass conservation, numerical stability and prevention of non-physical spurious oscillations and nu- merical smearing. By comparing the results to those of the published tests, the suitability, stability and general perfor- mance of the advection module are evaluated. The glacier simulations are used to qualitatively evaluate how well the coupled ice ﬂow–advection model reproduces glacial struc- tures related to idealized debris inputs of various dimen- sions illustratively representing rockfall deposits, extensive ash layer or debris avalanche deposits and crevasse ﬁll in comparison to the structures predicted by theory or observed in the ﬁeld. www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 The Cryosphere, 12, 189–204, 2018 4.2 Glacier tests The purpose of these tests is to demonstrate the character- istics of debris transport within mountain glaciers, not to re- produce a particular event on a speciﬁc glacier. Hence, all ve- locity computations are initialized with a no-slip condition at the glacier–bedrock boundary, the ﬂow law exponent n is set to 3 and the Glen rate factor A is set to 2.4×10−24 s−1 Pa−3, The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers 195 Figure 3. Velocity components computed with icetools for the 2-D long proﬁle of Haute Glacier d’Arolla. (a) Horizontal velocity (m year−1) and (b) vertical velocity (m year−1) and contour line of zero vertical velocity in grey. 195 Figure 3. Velocity components computed with icetools for the 2-D long proﬁle of Haute Glacier d’Arolla. (a) Horizontal velocity (m year−1) and (b) vertical velocity (m year−1) and contour line of zero vertical velocity in grey. Figure 3. Velocity components computed with icetools for the 2-D long proﬁle of Haute Glacier d’Arolla. (a) Horizontal velocity (m year−1) and (b) vertical velocity (m year−1) and contour line of zero vertical velocity in grey. Figure 4. Debris concentration at time step t = 0 years, where C0 indicates the circular debris inclusion, D1–D3 the surface debris layer deposits and CRV the crevasse ﬁlls. The horizontal lines indicate the location and extent of the additional debris layers deposited at the surface at the stated times. L indicates the horizontal distance of the glacier surface where debris is deposited. Figure 4. Debris concentration at time step t = 0 years, where C0 indicates the circular debris inclusion, D1–D3 the surface debris layer deposits and CRV the crevasse ﬁlls. The horizontal lines indicate the location and extent of the additional debris layers deposited at the surface at the stated times. L indicates the horizontal distance of the glacier surface where debris is deposited. a standard value for temperate ice (Cuffey and Paterson, 2010). The density of ice ρice is set to 917 kg m−3. The transport simulations are initialized with a debris concen- tration ﬁeld c0. 4.2 Glacier tests X, Y and Z coordinates are used to iden- tify a debris-deposition zone characterizing (a) a part of the glacier–atmosphere, glacier–bedrock or glacier–sidewall in- terface that receives instant, continuous or variable debris in- put or (b) a localized debris inclusion as e.g. a remnant of a rockfall event or a crevasse ﬁll. This is implemented by as- signing desired values of initial concentration at the respec- tive locations to the function c0, which is set to 0 everywhere else on the entire domain. In the presented glacier simula- tions, all debris inclusions have been deposited in a single event, and hence they are all initialized as inclusions within the glacier; i.e. the entire glacier–atmosphere boundary be- longs to ∂0. In the 3-D cases, the concentration is initialized with a smoothed function at the boundaries of the feature. This is done to represent it most efﬁciently in a continuous function space. Debris concentration that is transported be- yond the boundaries of the glacier domain is removed from the system. The concentration itself is a scalar function able to have arbitrary numbers. It can be converted into actual de- bris mass as a function of the actual debris density and con- centration of the initial debris deposit, i.e. the percentage of debris vs. ice or snow in the initial volume of the deposit. In this study, we present model simulations for initial debris concentrations of the value 100 that can be scaled according to the case-relevant initial proportions of debris and ice. For example, in the case of an ash layer deposit, the initial ash concentration will likely make up almost 100 %, compared to a mixed avalanche deposit that is likely to have much lower initial concentrations of debris vs. snow or ice. 4.2.1 Two-dimensional glacier test For the 2-D glacier test simulation, a 100 m spatial reso- lution longitudinal proﬁle of bedrock and glacier surface for Haute Glacier d’Arolla was downloaded from the Ice Sheet Model Intercomparison Project (ISMIP) website (http: //homepages.ulb.ac.be/~fpattyn/ismip/). These data represent the glacier in 1930, based on digitization of data from Blatter et al. (1998) and further described in Pattyn (2002). The lon- gitudinal proﬁle of 1930 is 5 km long. Haute Glacier d’Arolla is an alpine valley glacier with supraglacial debris cover- ing approximately 10 % of the glacier surface in 2012 (Reid et al., 2012). In this test, we apply idealized debris inputs in the glacier accumulation area and track the evolution of the internal de- bris concentration while being transported in a ﬁxed veloc- ity ﬁeld. For the 2-D glacier proﬁle, horizontal and verti- cal velocity components are shown in Fig. 3. In the Supple- ment, Fig. S6 shows the surface velocity computed with a www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers 196 196 A. Wirbel et al.: Modelling debris advection in glaciers 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] m a-1 (a) Down-valley velocity component 2.50 1.25 0.00 -1.25 -2.50 3.12 0.36 -2.41 -5.18 -7.95 320 235 150 65 -20 m a.s.l. (e) Glacier bedrock Elevation [m a.s.l.] 0 200 400 600 100 300 Distance along ﬂow line [m] (c) Vertical velocity component (d) Proﬁle along central ﬂow line m a-1 m a-1 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] 0.00 4.40 8.80 13.20 17.60 (b) Cross-valley velocity component Figure 5. Idealized geometry and computed velocity components (m year−1) for the 3-D glacier case. (a) Down-valley direction (along x axis), (b) cross-valley direction (along y axis) and (c) vertical direction (along z axis), (d) 2-D proﬁle along a central ﬂow line and (e) glacier bedrock elevation and ﬂow line from (d) in blue. 96 A. Wirbel et al.: Modelling debris advection in glaciers 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] m a-1 (a) Down-valley velocity component 2.50 1.25 0.00 -1.25 -2.50 3.12 0.36 -2.41 -5.18 -7.95 (c) Vertical velocity component m a-1 m a-1 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] 0.00 4.40 8.80 13.20 17.60 (b) Cross-valley velocity component 300 0 200 400 600 100 -200 200 0 Distance cross-valley [m] Distance down-valley [m] Elevation [m a.s.l.] Dista 320 235 150 65 -20 m a.s.l. (e) Glacier bedrock Elevation [m a.s.l.] 0 200 400 600 100 300 Distance along ﬂow line [m] (d) Proﬁle along central ﬂow line Figure 5. Idealized geometry and computed velocity components (m year−1) for the 3-D glacier case. www.the-cryosphere.net/12/189/2018/ (a) Down-valley direction (along x axis), (b) cross-valley direction (along y axis) and (c) vertical direction (along z axis), (d) 2-D proﬁle along a central ﬂow line and (e) glacier bedrock elevation and ﬂow line from (d) in blue. Glen rate factor A = 10−16 years−1 Pa−3 as used in ISMIP, demonstrating that our model reproduces the ISMIP results in Pattyn et al. (2008). Five debris features of different size and input location are prescribed at varying time intervals (Fig. 4). These debris deposits of varying size, shape and lo- cation of deposition were chosen to facilitate analysis of the interplay between debris input location, deformation during transport and the zone of emergence. At t = 0, the prescribed initial debris concentration ﬁeld consists of a circular debris inclusion centred at x = 500 m and z = 3052 m with a radius of 25 m (C0), a group of three crevasse ﬁlls of 2–5 m width and 50–75 m length between x = 1800 m and x = 1900 m (CRV), and a ca. 4 m thick debris layer covering the glacier surface between x = 296 m and x = 854 m of its length (D1). Subsequently, further debris layers are prescribed as follows: after 25 years, a ca. 9.5 m thick layer is deposited on the glacier surface between x = 1000 m and x = 1600 m (D2), and after 50 years another layer of ca. 5 m thickness is de- posited between x = 300 m and x = 1950 m (D3). The circu- lar inclusion and the debris layer deposits were prescribed to provide tight constraints on the shapes, whereas the vertical inclusions were initialized having irregularly shaped bound- aries considered more representative for actual crevasse ﬁlls of variably sized debris material. The entire simulation, dis- played in the video of the Supplement, covers 90 years. The mesh reﬁnement is initialized with cvol = 0.075 m2, the re- ﬁnement time step is set to 0.2 years and the Courant number to deﬁne the computation time step is set to 0.5. www.the-cryosphere.net/12/189/2018/ 4.2.2 Idealized 3-D glacier test In the 3-D glacier test, we perform simulations for an ide- alized glacier geometry. The geometry represents a valley glacier, including topographically induced complexities such as a wide accumulation basin leading to a narrow valley, a bump in the bedrock geometry and a turn of the valley it- self. In this manner, topographic features that control ice ﬂow in an alpine setting are represented, though idealized to re- duce computational effort. The idealized ice geometry and the computed 3-D velocity ﬁeld is shown in Fig. 5. In this test, the mesh reﬁnement is initialized with Lcsize = 0.4 m, the reﬁnement time step is set to 0.4 years and the Courant number to 1.5. The simulation is initialized with a spherical debris in- clusion centred at x = 0 m, y = 0 m and z = 270 m in the accumulation area with a radius of 9.5 m. This initial con- centration is chosen to aid visualization of the transport www.the-cryosphere.net/12/189/2018/ www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers 197 Figure 6. Results of 2-D three body rotation test for reﬁnement time step 0.1π s. (a) Initial condition, (b) solution after one full rotation and (c) solution after one full rotation on the underlying mesh. (d) Overshoots (values greater than maximum initial value, as deviation from max(cinitial) = 1.0) and (e) undershoots (values smaller than minimum initial value, as deviation from min(cinitial) = 0.0) at ttotal. (f) Results of the convergence test as a function of mesh reﬁnement parameter cvol. Colour scales show concentration values. Figure 6. Results of 2-D three body rotation test for reﬁnement time step 0.1π s. (a) Initial condition, (b) solution after one full rotation and (c) solution after one full rotation on the underlying mesh. (d) Overshoots (values greater than maximum initial value, as deviation from max(cinitial) = 1.0) and (e) undershoots (values smaller than minimum initial value, as deviation from min(cinitial) = 0.0) at ttotal. (f) Results of the convergence test as a function of mesh reﬁnement parameter cvol. Colour scales show concentration values. Figure 7. Results of 2-D swirling ﬂow three body test for reﬁnement time step 0.01π s. (a) Initial condition, (b) solution at ttotal/2 and (c) solution at ttotal. 4.2.2 Idealized 3-D glacier test (d) Overshoots (values greater than maximum initial value, as deviation from max(cinitial) = 1.0), (e) undershoots (values smaller than minimum initial value, as deviation from min(cinitial) = 0.0) at ttotal and (f) solution at ttotal/2 on the underlying mesh, shown in plan view. Colour scales show concentration values. Figure 7. Results of 2-D swirling ﬂow three body test for reﬁnement time step 0.01π s. (a) Initial condition, (b) solution at ttotal/2 and (c) solution at ttotal. (d) Overshoots (values greater than maximum initial value, as deviation from max(cinitial) = 1.0), (e) undershoots (values smaller than minimum initial value, as deviation from min(cinitial) = 0.0) at ttotal and (f) solution at ttotal/2 on the underlying mesh, shown in plan view. Colour scales show concentration values. and deformation rather than to best represent likely en- and supraglacial debris deposits. spurious oscillations in the solutions and allow us to choose suitable Courant numbers to ensure numerical stability. The results of the “rotating three body problem” (de Fru- tos et al., 2014) are shown in Fig. 6, for the reﬁnement time step 0.1π s. The results as well as animations for all sets of tests can be found in the Supplement. In Fig. 6a–c, the initial condition, the solution after a full rotation of 2π s and the so- lution on the underlying mesh are shown. The shapes of the concentration features are well recovered in the case of both reﬁnement time steps (see Fig. 6b for reﬁnement time step 0.1π s and Fig. S5b for reﬁnement time step 0.01π s). Pos- itive and negative oscillations in the solution are shown in Fig. 6d–e. The highest oscillations occur where the gradients 5.1 Benchmark tests Our results of reproducing Examples 1 and 2 in the numer- ical results in Bochev et al. (2004) are shown in Figs. S2 and S3. These simulations demonstrate the efﬁciency of our SUPG algorithm implementation for reducing non-physical 5.2 Two-dimensional glacier test in concentration are strongest, i.e. at the walls of the slotted cylinder. To measure the magnitude of remaining spurious oscillations in the solution, the difference of the maximum and minimum value of the solution is given in de Frutos et al. (2014). In our results, for case (a) max(c)−min(c) = 1.2526 with 87 773 cells in the ﬁnal mesh, whereas for case (b) max(c) −min(c) = 1.2524 with 142 792 cells in the ﬁnal mesh. These oscillations are slightly higher, but comparable to the values of 1.1010–1.1301 reported in de Frutos et al. (2014). The total number of cells in the ﬁnal meshes is larger in our computations, as the mesh reﬁnement is performed in an interval of (a) 0.01π s or (b) 0.1π s and not individu- ally for every computation time step. Mass loss is < 0.009 % for both reﬁnement time steps. Results of the convergence test for decreasing cell area thresholds, which are required to drive the mesh reﬁnement, are shown in Fig. 6f. The cho- sen cell area threshold of 0.075 m2 yields acceptable results. A further decrease leads to a drastic increase in computa- tional costs, with only a small increase in model accuracy. Also, when the initial concentration pattern is subjected to a more complex, swirling ﬂow (LeVeque, 1996), the results of these more challenging test simulations again show satisfac- tory model performance, as can be seen in Figs. 7 and S6. The upper boundary of the circular inclusion is initially lo- cated approximately 30 m beneath the glacier surface. Dur- ing transport, it becomes severely elongated, as the upper part of the feature is transported faster with the ice ﬂow than its lower part (Fig. 9a). After some travel time and sustained elongation, the vertical distance between the initially circular inclusion and the debris layer deposit D1 (Fig. 4) gradually decreases due to the vertical gradients in velocity (Fig. 9b). The initial surface emergence of the circular inclusion oc- curs later and further down glacier than any of the other im- posed concentration features. It travels the longest distance and reaches the greatest depths within the glacier ﬂow ﬁeld. The crevasse ﬁlls are initially quasi-perpendicular to the glacier surface. As they are transported through the glacier, the vertical inclusions become deformed and exhibit a pro- gressively more arcuate shape (Fig. 9d). A. Wirbel et al.: Modelling debris advection in glaciers 198 g g Concentration Analytical - y Analytical - z Analytical - x Figure 8. Results of 3-D rotation test. Isosurfaces of concentration 10, 30, 50, 70 and 90 for the FEM solution (in colour) and the analytical solution (in solid grey) at (a) ttotal/2 and at (b) ttotal. Note that for the FEM solution the isosurface 90 is missing as the maximum values of concentration decreased to 87.5 (at ttotal/2) and 83 (at ttotal). (c) Cross proﬁles parallel to the x, y and z axis for the analytical and the FEM solution at ttotal. Figure 8. Results of 3-D rotation test. Isosurfaces of concentration 10, 30, 50, 70 and 90 for the FEM solution (in colour) and the analytical solution (in solid grey) at (a) ttotal/2 and at (b) ttotal. Note that for the FEM solution the isosurface 90 is missing as the maximum values of concentration decreased to 87.5 (at ttotal/2) and 83 (at ttotal). (c) Cross proﬁles parallel to the x, y and z axis for the analytical and the FEM solution at ttotal. The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers 199 A. Wirbel et al.: Modelling debris advection in glaciers 199 → D2 D1 C0 → → → → → → → D3 0.2 years 4 years 8 years 20 years 26 years C0 CRV D1 → → → 24 years 83 years 1750 1950 1950 2150 2100 2750 2950 2750 2950 2950 2900 2700 2750 2700 2900 2700 2750 3000 3200 2300 1000 2000 2700 0 Distance down-valley [m] Elevation [m a.s.l.] 3000 (a) 3000 4000 5000 50 25 100 0 75 Concentration 62 years 1000 2000 2700 0 Distance down-valley [m] Elevation [m a.s.l.] 3000 (b) 3000 4000 5000 1000 2000 2700 0 Distance down-valley [m] Elevation [m a.s.l.] 3000 (c) 3000 4000 5000 50 25 100 0 75 Concentration 50 25 100 0 75 Concentration 50 25 100 0 75 Concentration Distance down-valley [m] Elevation [m a.s.l.] D3 D2 D1 C0 (d) Figure 9. Results of the debris transport simulations for the 2-D long proﬁle of Haute Glacier d’Arolla, where C0 indicates the circular debris inclusion, D1–D3 the surface debris layer deposits and CRV the crevasse ﬁlls. Debris concentration (a) at 24 years, (b) at 62 years and (c) at 83 years after start of the simulations. Concentrations are displayed in the range of 0 to 100 and numerical oscillations as excursions beyond the initial values of 0 or 100 are of magnitude less than ±17 and are truncated to the data limits. (d) Zoom of the crevasse ﬁlls at 0.2, 4, 8, 20 and 26 years after the start of the simulations. years 8 years 20 years 26 years 2150 2100 2950 2900 2700 2750 2700 2900 2700 2750 3000 3200 2300 Distance down valley [m] 50 25 100 0 75 Concentration Distance down-valley [m] 0.2 years 1750 1950 2750 2950 Elevation [m a.s.l.] (d) 4 years 1950 2150 2750 2950 Figure 9. Results of the debris transport simulations for the 2-D long proﬁle of Haute Glacier d’Arolla, where C0 indicates the circular debris inclusion, D1–D3 the surface debris layer deposits and CRV the crevasse ﬁlls. Debris concentration (a) at 24 years, (b) at 62 years and (c) at 83 years after start of the simulations. 6 Discussion served in the modelling results, is a reduction in the distance between the debris bands further down glacier, coinciding with decreasing ice velocities in this part of the glacier. The dip angle at the point of emergence to the surface differs be- tween the three debris layer deposits and also changes as each layer feature is advected further down glacier. 6.1 Model capabilities and applicability The debris transport and deformation modelled here repro- duces structures analogous to those observed in structural glaciology, where elongated and sometimes cross-cutting de- bris layers outcrop with a range of dip angles at the glacier surface (Jennings et al., 2014; Goodsell et al., 2005). Not only can these structures be reproduced, but the 2-D glacier simulations indicate that these elongated, band-shaped de- bris layers can form from initially fundamentally different debris deposits. In these simulations, ash fall or avalanche events that uniformly cover wider portions of the accumu- lation area are included as layer-shaped debris deposits at the glacier surface. Rockfall events that result in a locally thick debris deposit are represented by a circular inclusion, as an end-member case of possible remnants thereof. Both distinctly different debris inputs become severely elongated and band-like shaped during transport. The degree of elon- gation depends on the input location and, hence, the trajec- tory through the glacier. In addition to horizontal stretching, A. Wirbel et al.: Modelling debris advection in glaciers Concentrations are displayed in the range of 0 to 100 and numerical oscillations as excursions beyond the initial values of 0 or 100 are of magnitude less than ±17 and are truncated to the data limits. (d) Zoom of the crevasse ﬁlls at 0.2, 4, 8, 20 and 26 years after the start of the simulations. 5.3 Idealized 3-D glacier test The deformation of englacial features shown in 2-D is also represented in the 3-D cases. The initially spherical inclusion becomes severely elongated in down-glacier direction, form- ing a “comet-like” tail as it is transported through the glacier. In addition to the down-glacier elongation, where the glacier becomes narrower, the orographically left side of the debris inclusion is tilted upwards, and the centre of concentration is displaced laterally due to unequal lateral compression in the glacier ﬂow ﬁeld as it rounds the bend in the idealized valley (Fig. 10). www.the-cryosphere.net/12/189/2018/ 5.2 Two-dimensional glacier test These features reach the glacier surface and emerge ﬁrst at x = 2715 m. They are removed from the glacier domain over the course of 31 years and over a distance of 580 m of the glacier surface. As they progressively emerge to the surface, the angle of outcrop ro- tates from vertical to up-glacier dipping bands (Fig. 9d). The results of the 3-D test following Christensen (1993) are illustrated in Fig. 8. An animation of the full rotation is given in the Supplement. During the full rotation mass loss/gain is less than 0.01 %. The shape of the concentration feature is in good agreement with the analytical solution as indicated by the isosurfaces of concentration (Fig. 8), just the highest concentrations (isosurface 90) are not captured well due to numerical smearing. This also causes the maximum value of concentration in the ﬁnal solution to reduce to 83 (initially 100) and the spreading of very low concentrations (< 5) over a larger volume compared to the analytical solu- tion. The layer-shaped debris inputs (D1–3 in Fig. 4) not only have different characteristics such as length and thickness but are also deposited at different locations on the glacier surface. D1 ﬁrst reaches the glacier surface at x = 3450 m and over a 90-year simulation period emerges over 109 m of the glacier surface. D2 ﬁrst reaches the glacier surface at x = 3190 m and over a 90-year simulation period emerges over 281 m of the glacier surface. D3 ﬁrst reaches the glacier surface at x = 2645 m and over a 90-year simulation period emerges over 645 m of the glacier surface. In comparison to D1 and D3, the upper limit of D2 is located ca. 700 m further down glacier. The zone of emergence is signiﬁcantly shorter, it becomes less elongated and is exhumed in a shorter period of time, compared to D1 and D3. Another characteristic, ob- The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ A. Wirbel et al.: Modelling debris advection in glaciers www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers A. Wirbel et al.: Modelling debris advection in glaciers 200 200 igure 10. Results of the debris transport simulations for the 3-D idealized glacier. Debris concentration isosurfaces of concentration 1, 0–100 in steps of 10 are displayed for a cut through (a) the x–z plane and (b) the y–z plane, as labelled in the ﬁgure. The isosurfaces shown efer to time 0–50 years after the start of the simulation in an interval of 10 years. Concentrations are displayed in the range of 0 to 100, umerical oscillations as excursions beyond the initial values of 0 or 100 are of magnitude less than ±4 and are truncated to the data limits. Figure 10. Results of the debris transport simulations for the 3-D idealized glacier. Debris concentration isosurfaces of concentration 1, 10–100 in steps of 10 are displayed for a cut through (a) the x–z plane and (b) the y–z plane, as labelled in the ﬁgure. The isosurfaces shown refer to time 0–50 years after the start of the simulation in an interval of 10 years. Concentrations are displayed in the range of 0 to 100, numerical oscillations as excursions beyond the initial values of 0 or 100 are of magnitude less than ±4 and are truncated to the data limits. the effect of lateral compressional ﬂow is shown in the 3-D glacier simulations. detail. By explicitly modelling changes of debris concentra- tion distribution as a result of transport within the glacier, local concentration changes (Eulerian perspective) caused by the deformation of debris deposit shape (Lagrangian perspec- tive; see Fig. 10) during englacial transport can be captured. Hence, for a given debris deposition event and glacier ge- ometry, we can quantify the exact amount of debris concen- tration at any point in space and time. As a result, we can track the englacial debris transport and quantify the timing, location and debris concentration of a debris band emerging in the ablation zone, as well as quantifying how the location of maximum debris emergence from a debris band, and its dip angle, will change over time. This is all critical informa- tion for determining how the spatial pattern of surface debris thickness will develop and evolve in time. www.the-cryosphere.net/12/189/2018/ www.the-cryosphere.net/12/189/2018/ is due to the impact of debris cover on ice melt (Östrem, 1959; Mattson et al., 1993), which mainly depends on its thickness (Nicholson and Benn, 2006; Reid and Brock, 2010). In the case of negative mass balance conditions, the emergence of thin debris cover at the upper end of the abla- tion zone can lead to locally enhanced melting, lower the sur- face slope and alter the dynamic regime of the glacier (Benn et al., 2012). When and where those transitions occur is also related to the location and rate of debris emergence. vary depending on the dimensions of the debris inputs. Nev- ertheless, we have demonstrated that the model performs sat- isfactorily in comparison to benchmark standards in the liter- ature and that this performance quality applies to the given model set-up and thresholds used to simulate the glacier cases presented. cases p ese ted. The ratio of debris input size vs. the total size of the glacier requires very ﬁne mesh sizes to actually resolve the debris inputs, their transport and associated deformation. For exam- ple, for the presented model set-up in the 3-D glacier exam- ple, the total number of cells is in the order of 108. This leads to high computational costs and the available computing re- sources impose constraints on the size of debris inputs the model is able to adequately simulate. In the case of local- ized debris inputs, our mesh reﬁnement approach has the po- tential to reduce the total number of cells substantially com- pared to a mesh that is globally reﬁned. In the simulations presented here, the parameters in the reﬁnement module are chosen to produce a mesh that is as coarse as possible while guaranteeing mass conservation (> 99 %), numerical stabil- ity and limiting numerical oscillations and numerical smear- ing to the levels presented in the Results (Sect. 5). Although numerical instabilities such as non-physical spurious oscilla- tions and numerical diffusion are reduced efﬁciently by the approaches described in Sect. 3, numerical smearing cannot be eliminated completely. Its magnitude is controlled by the mesh resolution (see Sect. 1). Decreasing cell size reduces numerical diffusion, but limitations of computing power will in practice impose a lower bound on cell size. Therefore, the magnitude of numerical diffusion expected for a given model set-up should be taken into account when interpreting model results. www.the-cryosphere.net/12/189/2018/ The exact location and orientation of the proﬁles is indicated in Fig. S8. A. Wirbel et al.: Modelling debris advection in glaciers 201 g g 0 20 40 60 80 0 40 80 Along cross-proﬁle distance [m] Concentration D1 t = 0.2 yrs D1 t = 83 yrs D2 t = 25.2 yrs D2 t = 83 yrs C0 t = 0.2 yrs C0 t = 83 yrs → → → → → → D1 t = 83 yrs → Figure 11. Vertical cross sections at different model times for debris features D1, D2 and CO from the 2-D glacier test (see Fig. 9). The verti- cal cross sections are taken at the initial feature geometries in blue and evolved feature geometries at the simulation end in red. Corresponding cross sections are centred at the same along cross-proﬁle distance for ease of comparison. The vertical cross proﬁles are taken at a horizontal distance of 601.5 m (D1, t = 0.2 years) and 3148.5 m (D1, t = 83 years), 1100 m (D2, t = 25.2 years) and 3270 m (D2, t = 83 years), 502 m (C0, t = 0.2 years) and 2062 m (C0, t = 83 years). The exact location and orientation of the proﬁles is indicated in Fig. S8. 0 20 40 60 80 0 40 80 Along cross-proﬁle distance [m] Concentration D1 t = 0.2 yrs D1 t = 83 yrs D2 t = 25.2 yrs D2 t = 83 yrs C0 t = 0.2 yrs C0 t = 83 yrs → → → → → → D1 t = 83 yrs → Figure 11. Vertical cross sections at different model times for debris features D1, D2 and CO from the 2-D glacier test (see Fig. 9). The verti- cal cross sections are taken at the initial feature geometries in blue and evolved feature geometries at the simulation end in red. Corresponding cross sections are centred at the same along cross-proﬁle distance for ease of comparison. The vertical cross proﬁles are taken at a horizontal distance of 601.5 m (D1, t = 0.2 years) and 3148.5 m (D1, t = 83 years), 1100 m (D2, t = 25.2 years) and 3270 m (D2, t = 83 years), 502 m (C0, t = 0.2 years) and 2062 m (C0, t = 83 years). The exact location and orientation of the proﬁles is indicated in Fig. S8. www.the-cryosphere.net/12/189/2018/ The work of Kirkbride and Deline (2013) illustrates how both the thickness and angle of emergence of a debris band play a role in determining the initial thickness of an emer- gent debris deposit. This, in combination with the location of emergence is an essential prerequisite to predict the de- velopment and further evolution of debris cover. The sim- ulations highlight that for spatially restricted debris deposi- tion events, distinct debris bands form within the glacier that will lead to initially delimited areas of debris cover on the surface. Hence, an assumption of a uniform englacial de- bris distribution of constant englacial debris concentration (Naito et al., 2000) that would result in a continuously debris- covered glacier surface where surface ablation is occurring might not reﬂect reality adequately in order to capture the geometrical response of the glacier to the developing debris cover. The model presented here allows us to simulate the advection of debris concentration through a glacier in great These results are also important in the context of the re- sponse of debris-covered glaciers to changes in climatic forc- ing or debris supply. Debris-covered glaciers are known to show distinctly different behaviour to clean-ice glaciers. This www.the-cryosphere.net/12/189/2018/ www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers 201 0 20 40 60 80 0 40 80 Along cross-proﬁle distance [m] Concentration D1 t = 0.2 yrs D1 t = 83 yrs D2 t = 25.2 yrs D2 t = 83 yrs C0 t = 0.2 yrs C0 t = 83 yrs → → → → → → D1 t = 83 yrs → Figure 11. Vertical cross sections at different model times for debris features D1, D2 and CO from the 2-D glacier test (see Fig. 9). The verti- cal cross sections are taken at the initial feature geometries in blue and evolved feature geometries at the simulation end in red. Corresponding cross sections are centred at the same along cross-proﬁle distance for ease of comparison. The vertical cross proﬁles are taken at a horizontal distance of 601.5 m (D1, t = 0.2 years) and 3148.5 m (D1, t = 83 years), 1100 m (D2, t = 25.2 years) and 3270 m (D2, t = 83 years), 502 m (C0, t = 0.2 years) and 2062 m (C0, t = 83 years). www.the-cryosphere.net/12/189/2018/ The 2-D benchmark tests show that the chosen cell area threshold of 0.075 m2 (for the 2-D simulations) effec- tively limits non-physical spurious oscillations and numeri- cal smearing below the levels presented in Sect. 5.1. g The model presented here, which resolves the governing physical processes without parameterization and is based on a comprehensive numerical framework, offers a powerful tool with which to examine the validity of assumptions made in simpler models. For example, this model can be used to explore how the manner of prescribing debris (localized or distributed, spatially variable or constant, frequent or rare) affects the manner in, and timescales over, which a surface debris cover develops. This is valuable in the study of poorly understood earth systems like debris-covered glaciers, which evolve over timescales too long to allow real-world observa- tions to answer these questions. The model presented here can be used to track the passage of any material through the glacier – under the assumption that the transported mate- rial itself is not signiﬁcantly altering the glacier ﬂow ﬁeld. It therefore has potential applications not only for under- standing the development of supraglacial debris layers but also for interpreting observed structures in glaciers related to speciﬁc tephra deposits or rockfall events, for example. This model also offers the possibility to test the ﬁndings of studies that use patterns of englacial debris distribution on Antarc- tic debris-covered glaciers to infer climate information at or- bitally paced timescales (Mackay and Marchant, 2017). 6.2 Model performance and limitations To demonstrate the concentration changes during trans- port within debris features for the 2-D glacier test and its The numerical accuracy of the presented model is dictated by the reﬁnement and stability thresholds selected and will also The Cryosphere, 12, 189–204, 2018 7 Conclusions and outlook relation to numerical diffusion controlled by the cell area threshold (i.e. mesh reﬁnement parameter), we plot concen- tration cross sections for different debris features and model times in Fig. 11. The preservation of a central concentration peak over time is a good measure for identifying minimal nu- merical diffusion cases, as the redistribution of concentration by numerical diffusion would instantly lower that peak. In contrast, redistribution of debris concentration by transport preserves that peak due to the incompressibility of ice (see Sect. 1). Debris feature D1 experiences a strong amount of numerical diffusion over its 82.8 years of transport through the glacier and its central concentration decreases by ∼62 %. This performance deﬁcit is due to the choice of a cell area threshold being too large to preserve this initially thin debris feature. In contrast to D1, the central concentrations of de- bris features D2 and C0 are clearly well preserved (maximum concentration values above 100) and their change in concen- tration distribution (thinning of the cross-sectional length) is mostly due to transport. As a measure for debris band thick- ness over time we take full width at half maximum (FWHM) of the plotted debris concentration cross sections in Fig. 11. D2 decreases from an initial FWHM of 9.3 m to a FWHM of 4.4 m over 57.8 years of transport. C0 decreases from an ini- tial FWHM of 48.8 m to a FWHM of 2.2 m over 82.8 years of transport. Both cases clearly demonstrate the debris concen- tration changes caused by the transport through the glacier with minimal numerical diffusion. This highlights that a suit- able cell area threshold choice is paramount to correctly sim- ulate debris transport, as demonstrated with features D2 and C0. To improve the simulation results for feature D1, the only requirement is to lower the cell area threshold to a suitable value. We developed a model to simulate debris transport within glaciers based on an advection algorithm that is coupled to a full-Stokes ice ﬂow model. To facilitate computations and provide the spatial resolution required to accurately represent observed debris inputs, a localized mesh reﬁnement strategy is employed. In this manner, the deformation of debris inputs, arising from gradients in the glacier’s velocity ﬁeld, can be modelled explicitly. 7 Conclusions and outlook This is crucial, as the location of emer- gence as well as amount and rate of debris emergence on the glacier surface depend on the deposition location of debris inputs and are subsequently controlled by englacial transport and deformation. This is the ﬁrst model capable of resolving transport and deformation of debris inputs in this detail. The advection algorithm combined with the full-Stokes approach offers the potential to model englacial transport of various debris inputs and for complex glacier geometries. In a fu- ture step, coupling the englacial transport model presented here to a (i) debris-aware surface mass balance scheme and (ii) supraglacial debris transport scheme will enable us to fully model the co-evolution of debris cover and glacier ge- ometry and the behaviour of debris-covered glacier systems in general. Additionally, this will offer a powerful means by which to evaluate simpler representations of debris cover de- velopment within glacier systems. Code and data availability. The corresponding version of the model used to produce the tests in this study can be found at https://doi.org/10.5281/zenodo.1139804 as well as under https:// github.com/awirbel/debadvect/tree/v1.0.0. In 3-D, the constraints on cell size are even more restrictive in terms of numerical stability and numerical diffusion. How- ever, increase in computational costs in 3-D is non-linear. The results we show here in the case of an idealized 3-D glacier geometry will be subject to some numerical smear- ing, but the 3-D benchmark test shows that numerical smear- ing can be minimized by choosing a suitable mesh reﬁnement parameter. Higher accuracy for representing sharp concentra- tion variations can easily be achieved by changing mesh re- ﬁnement cell size variables at increased computational cost. A. Wirbel et al.: Modelling debris advection in glaciers relation to numerical diffusion controlled by the cell area threshold (i.e. mesh reﬁnement parameter), we plot concen- tration cross sections for different debris features and model times in Fig. 11. The preservation of a central concentration peak over time is a good measure for identifying minimal nu- merical diffusion cases, as the redistribution of concentration by numerical diffusion would instantly lower that peak. In contrast, redistribution of debris concentration by transport preserves that peak due to the incompressibility of ice (see Sect. 1). Debris feature D1 experiences a strong amount of numerical diffusion over its 82.8 years of transport through the glacier and its central concentration decreases by ∼62 %. This performance deﬁcit is due to the choice of a cell area threshold being too large to preserve this initially thin debris feature. In contrast to D1, the central concentrations of de- bris features D2 and C0 are clearly well preserved (maximum concentration values above 100) and their change in concen- tration distribution (thinning of the cross-sectional length) is mostly due to transport. As a measure for debris band thick- ness over time we take full width at half maximum (FWHM) of the plotted debris concentration cross sections in Fig. 11. D2 decreases from an initial FWHM of 9.3 m to a FWHM of 4.4 m over 57.8 years of transport. C0 decreases from an ini- tial FWHM of 48.8 m to a FWHM of 2.2 m over 82.8 years of transport. Both cases clearly demonstrate the debris concen- tration changes caused by the transport through the glacier with minimal numerical diffusion. This highlights that a suit- able cell area threshold choice is paramount to correctly sim- ulate debris transport, as demonstrated with features D2 and C0. To improve the simulation results for feature D1, the only requirement is to lower the cell area threshold to a suitable value. The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 202 References Goodsell, B., Hambrey, M. J., and Glasser, N. F.: Debris trans- port in a temperate valley glacier: Haut Glacier d’Arolla, Valais, Switzerland, J. Glaciol., 51, 139–146, 2005. Ackert Jr., R. P.: A rock glacier/debris-covered glacier system at Galena Creek, Absaroka Mountains, Wyoming, Geogr. Ann. A, 80, 267–276, https://doi.org/10.1111/j.0435-3676.1998.00042.x, 1998. Hewitt, K.: Rock avalanches that travel onto glaciers and related de- velopments, Karakoram Himalaya, Inner Asia, Geomorphology, 103, 66–79, https://doi.org/10.1016/j.geomorph.2007.10.017, 2009. Alnæs, M. S.: UFL: a ﬁnite element form language, in: Auto- mated Solution of Differential Equations by the Finite Element Method, vol. 84, Lecture Notes in Computational Science and Engineering, edited by: Logg, A., Mardal, K.-A., and Wells, G. N., chap. 17, Springer, 2012. Hughes, T. J. R. and Brooks, A.: A theoretical framework for Petrov-Galerkin methods with discontinuous weighting func- tions: application to the streamline-upwind procedure, in: Finite Element in Fluids, edited by: Gallagher, R. H., vol. 4, Wiley & Sons, 1982. Alnæs, M. S., Logg, A., Ølgaard, K. B., Rognes, M. E., and Wells, G. N.: Uniﬁed Form Language: A Domain-Speciﬁc Language for Weak Formulations of Partial Differential Equations, ACM T. Math. Software, 40, 9, https://doi.org/10.1145/2566630, 2014. Jarosch, A. H.: Icetools: A full Stokes ﬁnite element model for glaciers, Comput. Geosci., 34, 1005–1014, https://doi.org/10.1016/j.cageo.2007.06.012, 2008. Alnæs, M. S., Blechta, J., Hake, J., Johansson, A., Kehlet, B., Logg, A., Richardson, C., Ring, J., Rognes, M. E., and Wells, G. N.: The FEniCS Project Version 1.5, Arch. Numerical Softw., 3, 9– 23, https://doi.org/10.11588/ans.2015.100.20553, 2015. Jarosch, A. H. and Gudmundsson, M. T.: A numerical model for meltwater channel evolution in glaciers, The Cryosphere, 6, 493– 503, https://doi.org/10.5194/tc-6-493-2012, 2012. Jennings, S. J., Hambrey, M. J., and Glasser, N. F.: Ice ﬂow-unit inﬂuence on glacier structure, debris entrainment and transport, Earth Surf. Proc. Land., 39, 1279–1292, https://doi.org/10.1002/esp.3521, 2014. Anderson, L. S. and Anderson, R. S.: Modeling debris-covered glaciers: response to steady debris deposition, The Cryosphere, 10, 1105–1124, https://doi.org/10.5194/tc-10-1105-2016, 2016. Benn, D., Bolch, T., Hands, K., Gulley, J., Luckman, A., Nicholson, L., Quincey, D., Thompson, S., Toumi, R., and Wiseman, S.: Response of debris-covered glaciers in the Mount Everest region to recent warming, and implications for outburst ﬂood hazards, Earth-Sci. Rev., 114, 156–174, https://doi.org/10.1016/j.earscirev.2012.03.008, 2012. John, V.: A numerical study of a posteriori error estimators for convection–diffusion equations, Comput. Method. Appl. M., 190, 757–781, https://doi.org/10.1016/S0045-7825(99)00440-5, 2000. John, V. References and Novo, J.: Error analysis of the SUPG ﬁnite element discretization of evolutionary convection-diffusion- reaction equations, SIAM J. Numer. A., 49, 1149–1176, https://doi.org/10.1137/100789002, 2011. Benn, D. I. and Evans, D. J. A.: Glaciers and Glaciation, Hodder Education, 2010. Blatter, H., Clarke, G. K. C., and Colinge, J.: Stress and velocity ﬁelds in glaciers: Part II, Sliding and basal stress distribution, J. Glaciol., 44, 457–466, 1998. Jouvet, G., Huss, M., Funk, M., and Blatter, H.: Modelling the re- treat of Grosser Aletschgletscher, Switzerland, in a changing cli- mate, J. Glaciol., 57, 1033–1045, 2011. Bochev, P. B., Gunzburger, M. D., and Shadid, J. N.: Stability of the SUPG ﬁnite element method for transient advection– diffusion problems, Comput. Method. Appl. M., 193, 2301– 2323, https://doi.org/10.1016/j.cma.2004.01.026, 2004. Kirby, R. C.: Algorithm 839: FIAT, A New Paradigm for Comput- ing Finite Element Basis Functions, ACM T. Math. Software, 30, 502–516, 2004. Kirby, R. C.: FIAT: numerical construction of ﬁnite element basis functions, in: Automated Solution of Differential Equations by the Finite Element Method, vol. 84, Lecture Notes in Compu- tational Science and Engineering, edited by: Logg, A., Mardal, K.-A., and Wells, G. N., chap. 13, Springer, 2012. Bozhinskiy, A., Krass, M., and Popovnin, V.: Role of debris cover in the thermal physics of glaciers, J. Glaciol., 32, 255–266, 1986. Christensen, J.: Testing Advection Schemes in a Three-Dimensional Air Pollution Model, Math. Comput. Model., 18, 75–88, 1993. Clark, D. H., Steig, E. J., Potter Jr., N., and Gillespie, A. R.: Ge- netic variability of rock glaciers, Geogr. Ann. A, 80, 175–182, https://doi.org/10.1111/j.0435-3676.1998.00035.x, 1998. Kirby, R. C. and Logg, A.: A Compiler for Varia- tional Forms, ACM T. Math. Software, 32, 417–444, https://doi.org/10.1145/1163641.1163644, 2006. Cuffey, K. M. and Paterson, W. S. B.: The Physics of Glaciers, Aca- demic Press, Burlington, MA, 4th edn., 2010. Kirkbride, M.: About the concepts of continuum and age, Boreas, 18, 87–88, https://doi.org/10.1111/j.1502-3885.1989.tb00376.x, 1989. de Frutos, J., García-Archilla, B., John, V., and Novo, J.: An adaptive SUPG method for evolutionary convection– diffusion equations, Comput. Method. Appl. M., 273, 219–237, https://doi.org/10.1016/j.cma.2014.01.022, 2014. Kirkbride, M.: Ice-marginal geomorphology and Holocene expan- sion of debris-covered Tasman Glacier, New Zealand, Proceed- ings of an International Workshop, University of Washington in Seattle, Washington, USA, 13–15 September 2000, IAHS Publi- cation, 264, 211–217, 2000. Deline, P., Hewitt, K., Reznichenko, N., and Shugar, D.: Rock Avalanches onto Glaciers, in: Landslide Hazards, Risks and Disasters, chap. 9, Elsevier, https://doi.org/10.1016/B978-0-12- 396452-6.00009-4, 2015. A. Wirbel et al.: Modelling debris advection in glaciers 203 The Supplement related to this article is available online at https://doi.org/10.5194/tc-12-189-2018-supplement. Competing interests. The authors declare that they have no conﬂict of interest. In practice, application details dictate the constraints on model accuracy required to be able to adequately resolve the problem at hand. In this respect, by performing multiple sim- ulations, the model can also be used to quantify the smear- ing of concentration features that arises from the choice of a coarser, but computationally feasible, cell size. Thus the introduced error by coarse mesh size choices can be easily quantiﬁed. Acknowledgements. This work was funded by the Austrian Science Fund (FWF), projects P28521 and V309. Thanks to Christoph Mayer for comments on an early draft of this paper. Edited by: Andreas Vieli Reviewed by: Garry K. C. Clarke and one anonymous referee Edited by: Andreas Vieli Reviewed by: Garry K. C. Clarke and one anonymous referee The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/ A. Wirbel et al.: Modelling debris advection in glaciers 204 debris bands, Earth Surf. Proc. Land., 38, 1779–1792, https://doi.org/10.1002/esp.3416, 2013. Nicholson, L. and Benn, D. I.: Calculating ice melt beneath a debris layer using meteorological data, J. Glaciol., 52, 463–470, 2006. Nicholson, L. and Benn, D. I.: Calculating ice melt beneath a debris layer using meteorological data, J. Glaciol., 52, 463–470, 2006. Nield, J. M., Chiverrell, R. C., Darby, S. E., Leyland, J., Vircavs, L. H., and Jacobs, B.: Complex spatial feedbacks of tephra re- distribution, ice melt and surface roughness modulate ablation on tephra covered glaciers, Earth Surf. Proc. Land., 38, 95–102, https://doi.org/10.1002/esp.3352, 2013. Konrad, S. K. and Humphrey, N. F.: Steady-state ﬂow model of debris-covered glaciers (rock glaciers), Proceedings of an Inter- national Workshop Held at the University of Washington in Seat- tle, Washington, USA, 13–15 September 2000, IAHS Publica- tion, 264, 255–266, 2000. LeVeque, R.: High-resolution conservative algorithms for advec- tion in incompressible ﬂow, SIAM J. Numer. A., 33, 627–665, https://doi.org/10.1137/0733033, 1996. Nye, J. F.: The distribution of stress and velocity in glaciers and ice-sheets, P. Roy. Soc. Lond. A Mat., 239, 113–133, https://doi.org/10.1098/rspa.1957.0026, 1957. Ølgaard, K. B. and Wells, G. N.: Optimisations for Quadra- ture Representations of Finite Element Tensors through Au- tomated Code Generation, ACM T. Math. Software, 37, 8, https://doi.org/10.1145/1644001.1644009, 2010. Logg, A. and Wells, G. N.: DOLFIN: Automated Finite El- ement Computing, ACM T. Math. Software, 37, 1–28, https://doi.org/10.1145/1731022.1731030, 2010. Logg, A., Mardal, K.-A., and Wells, G. (Eds.): Automated Solu- tion of Differential Equations by the Finite Element Method, vol. 84 of Lecture Notes in Computational Science and Engi- neering, Springer, 731 pp., 2012a. Östrem, G.: Ice melting under a thin layer of moraine, and the ex- istence of ice cores in the moraine ridges, Geogr. Ann., 41, 228– 230, 1959. Logg, A., Ølgaard, K. B., Rognes, M. E., and Wells, G. N.: FFC: the FEniCS form compiler, in: Automated Solution of Differential Equations by the Finite Element Method, vol. 84, Lecture Notes in Computational Science and Engineering, edited by: Logg, A., Mardal, K.-A., and Wells, G. N., chap. 11, Springer, 2012b. Pattyn, F.: Transient glacier response with a higher-order numerical ice-ﬂow model, J. Glaciol., 48, 467–477, 2002. Pattyn, F., Perichon, L., Aschwanden, A., Breuer, B., de Smedt, B., Gagliardini, O., Gudmundsson, G. H., Hindmarsh, R. C. A., Hubbard, A., Johnson, J. V., Kleiner, T., Konovalov, Y., Martin, C., Payne, A. A. Wirbel et al.: Modelling debris advection in glaciers J., Pollard, D., Price, S., Rückamp, M., Saito, F., Soucek, O., Sugiyama, S., and Zwinger, T.: Benchmark experi- ments for higher-order and full-Stokes ice sheet models (ISMIP- HOM), The Cryosphere, 2, 95–108, https://doi.org/10.5194/tc-2- 95-2008, 2008. Logg, A., Wells, G. N., and Hake, J.: DOLFIN: a C++/Python ﬁnite element library, in: Automated Solution of Differential Equations by the Finite Element Method, vol. 84, Lecture Notes in Compu- tational Science and Engineering, edited by: Logg, A., Mardal, K.-A., and Wells, G. N., chap. 10, Springer, 2012c. Mackay, S. L. and Marchant, D. R.: Obliquity-paced climate change recorded in Antarctic debris-covered glaciers, Nat. Comm., 8, 14194, https://doi.org/10.1038/ncomms14194, 2017. Reid, T. D. and Brock, B. W.: An energy-balance model for debris- covered glaciers including heat conduction through the debris layer, J. Glaciol., 56, 903–916, 2010. y Reid, T. D., Carenzo, M., Pellicciotti, F., and Brock, B. W.: Including debris cover effects in a distributed model of glacier ablation, J. Geophys. Res.-Atmos., 117, D18105, https://doi.org/10.1029/2012JD017795, 2012. Mackay, S. L., Marchant, D. R., Lamp, J. L., and Head, J. W.: Cold- based debris-covered glaciers: Evaluating their potential as cli- mate archives through studies of ground-penetrating radar and surface morphology, J. Geophys. Res.-Earth, 119, 2505–2540, https://doi.org/10.1002/2014JF003178, 2014. Reznichenko, N. V., Davies, T. R., and Alexander, D. J.: Effects of rock avalanches on glacier behaviour and moraine formation, Geomorphology, 132, 327–338, https://doi.org/10.1016/j.geomorph.2011.05.019, 2011. Mattson, L. E., Gardner, G. S., and Young, G. J.: Ablation on Debris Covered Glaciers: an Example from the Rakhiot Glacier, Pun- jab, Himalaya, Proceedings of a Symposium Held at Kathmandu, Nepal 1992 – Snow and Glacier Hydrology, IAHS publication, 218, 289–296, 1993. Rowan, A. V., Egholm, D. L., Quincey, D. J., and Glasser, N. F.: Modelling the feedbacks between mass balance, ice ﬂow and de- bris transport to predict the response to climate change of debris- covered glaciers in the Himalaya, Earth Planet. Sc. Lett., 430, 427–438, https://doi.org/10.1016/j.epsl.2015.09.004, 2015. Menounos, B., Clague, J. J., Clarke, G. K., Marcott, S. A., Osborn, G., Clark, P. U., Tennant, C., and Novak, A. M.: Did rock avalanche deposits modulate the late Holocene ad- vance of Tiedemann Glacier, southern Coast Mountains, British Columbia, Canada?, Earth Planet. Sc. Lett., 384, 154–164, https://doi.org/10.1016/j.epsl.2013.10.008, 2013. Shugar, D. H., Rabus, B. T., Clague, J. J., and Capps, D. M.: The response of Black Rapids Glacier, Alaska, to the Denali earth- quake rock avalanches, J. Geophys. Res.-Earth, 117, F01006, https://doi.org/10.1029/2011JF002011, 2012. Moore, P. References Kirkbride, M. P.: Debris-Covered Glaciers, in: Encyclopedia of Snow, Ice and Glaciers, edited by: Singh, V. P., Singh, P., and Haritashya, U. K., Springer Netherlands, Dordrecht, 180–182, 2011. Geuzaine, C. and Remacle, J.-F.: Gmsh: a three-dimensional ﬁnite element mesh generator with built-in pre-and post-processing fa- cilities, Int. J. Numer. Meth. Eng., 79, 1309–1331, 2009. Kirkbride, M. P. and Deline, P.: The formation of supraglacial debris covers by primary dispersal from transverse englacial Glen, J. W.: The creep of polycrystalline ice, P. Roy. Soc. Lond. A Mat., 228, 519–538, https://doi.org/10.1098/rspa.1955.0066, 1955. www.the-cryosphere.net/12/189/2018/ The Cryosphere, 12, 189–204, 2018 A. Wirbel et al.: Modelling debris advection in glaciers L.: Deformation of debris-ice mixtures, Rev. Geophys., 52, 435–467, https://doi.org/10.1002/2014RG000453, 2014. Vacco, D. A., Alley, R. B., and Pollard, D.: Glacial advance and stagnation caused by rock avalanches, Earth Planet. Sc. Lett., 294, 123–130, https://doi.org/10.1016/j.epsl.2010.03.019, 2010. Naito, N., Nakawo, M., Kadota, T., and Raymond, C. F.: Numer- ical simulation of recent shrinkage of Khumbu glacier, Nepal Himalayas, Proceedings of an International Workshop, Univer- sity of Washington, Seattle, Washington, USA, 13–15 September 2000, IAHS Publication, 264, 245–254, 2000. The Cryosphere, 12, 189–204, 2018 www.the-cryosphere.net/12/189/2018/
https://openalex.org/W1982014742	https://bmcbioinformatics.biomedcentral.com/counter/pdf/10.1186/1471-2105-7-S2-S16	English	null	Analysis of gene expression changes in relation to toxicity and tumorigenesis in the livers of Big Blue transgenic rats fed comfrey (Symphytum officinale)	BMC bioinformatics	2,006	cc-by	11,100	BioMed Central BioMed Central © 2006 Mei et al; licensee BioMed Central Ltd. © 2006 Mei et al; licensee BioMed Central Ltd. ; This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BMC Bioinformatics 2006, 7(Suppl 2):S16 doi:10.1186/1471-2105-7-S2-S16 BMC Bioinformatics 2006, 7(Suppl 2):S16 doi:10.1186/1471-2105-7-S2-S16 Open Acc Proceedings Analysis of gene expression changes in relation to toxicity and tumorigenesis in the livers of Big Blue transgenic rats fed comfrey (Symphytum officinale) Nan Mei1, Lei Guo2, Lu Zhang3,4, Leming Shi2, Yongming Andrew Sun3, Chris Fung3, Carrie L Moland2, Stacey L Dial2, James C Fuscoe2 and Tao Chen1 Open Access Address: 1Division of Genetic and Reproductive Toxicology, National Center for Toxicological Research, FDA, Jefferson, AR 72079, USA, 2Division of Systems Toxicology, National Center for Toxicological Research, FDA, Jefferson, AR 72079, USA, 3Molecular Biology-SDS/Arrays, Applied Biosystems, Foster City, CA 94404, USA and 4Solexa, Inc., 25861 Industrial Boulevard, Hayward, CA 94545, USA mail: Nan Mei - nan.mei@fda.hhs.gov; Lei Guo - lei.guo@fda.hhs.gov; Lu Zhang - lzhang@solexa.com; i Shi l i i hi fd hh Y i A d S li dbi Ch i F Email: Nan Mei* - nan.mei@fda.hhs.gov; Lei Guo - lei.guo@fda.hhs.gov; Lu Zhang - lzhang@solexa.com; Leming Shi - leiming.shi@fda.hhs.gov; Yongming Andrew Sun - sunya@appliedbiosystems.com; Chris Fung - Chris.Fung@ucsf.edu; Carrie L Moland - carrie.moland@fda.hhs.gov; Stacey L Dial - Stacey.dial@fda.hhs.gov; James C Fuscoe - james.fuscoe@fda.hhs.gov; Tao Chen - tao.chen@fda.hhs.gov * Corresponding author from The Third Annual Conference of the MidSouth Computational Biology and Bioinformatics Society Baton Rouge, Louisiana. 2–4 March, 2006 Published: 26 September 2006 Mutant frequency (MF) in the liver cII gene of comfrey-fed rats The results of cII MF analyses in the comfrey-fed and con- trol rats are shown in Figure 2. DNA from each liver was packaged 2–4 times either to confirm the MF or to obtain a minimum of 2 × 105 plaque-forming units for mutant detection. The MF for rats fed with 8% comfrey was 139 ± 35 (SD) × 10-6, which was similar to the MF previously detected in 2% comfrey-fed rats [10] and significantly increased over the control group (30 ± 16 × 10-6, P < 0.001). In addition to essential nutrients, comfrey also contains pyrrolizidine alkaloids (PAs). PAs are constituents of over 6000 plants, and many of them are hepatotoxic and carci- nogenic in humans and animals [3]. In the liver, PAs are transformed to pyrroles by the mixed-function oxidases. Pyrroles exert their toxic effect by reacting with cellular macromolecules, including proteins and DNA [4]. There- fore, comfrey's therapeutic use might increase the risk of liver toxicity. Many countries including Canada, Ger- many, and the UK, have restricted its availability. In 2001, the US Food and Drug Administration requested volun- tary compliance for the removal of products containing comfrey [5]. Results g Comfrey belongs to the family Boraginaceae. Three plant species in the genus Symphytum contribute to the crop known as comfrey, Symphytum officinale L. (wild or com- mon comfrey; the major comfrey species), S. asperum Lep- echin (prickly or rough comfrey), and S. x uplandicum Nyman (quaker, Russian, or blue comfrey; a natural hybrid of S. officinale L. and S. asperum Lepechin). Symphy- tum officinale L. is a tall perennial with large hairy leaves and small purple flowers [1]. Comfrey has been used as an herbal medicine for more than two thousand years for the treatment of broken bones, tendon damage, ulcerations in the gastrointestinal tract, and lung congestion, as well as for wound healing and/or reducing joint inflammation when it is applied externally [2]. Growth curve of rats fed with 8% comfrey root Growth curve of rats fed with 8% comfrey root Growth curve of rats fed with 8% comfrey root Male Big Blue transgenic rats were fed with 8% comfrey root for 12 weeks. The mean body weight of the comfrey- fed rats was less than that of the vehicle controls through- out the study (Figure 1). Rats fed with comfrey weighed 5%, 23%, and 35% less than the control rats after 1, 6, and 12 weeks of the study, respectively, and displayed little weight gain after 6 weeks of feeding with comfrey. Abstract Background: Comfrey is consumed by humans as a vegetable and a tea, and has been used as an herbal medicine for more than 2000 years. Comfrey, however, is hepatotoxic in livestock and humans and carcinogenic in experimental animals. Our previous study suggested that comfrey induces liver tumors by a genotoxic mechanism and that the pyrrolizidine alkaloids in the plant are responsible for mutation induction and tumor initiation in rat liver. Results: In this study, we identified comfrey-induced gene expression profile in the livers of rats. Groups of 6 male transgenic Big Blue rats were fed a basal diet and a diet containing 8% comfrey roots, a dose that resulted in liver tumors in a previous carcinogenicity bioassay. The animals were treated for 12 weeks and sacrificed one day after the final treatment. We used a rat microarray containing 26,857 genes to perform genome-wide gene expression studies. Dietary comfrey resulted in marked changes in liver gene expression, as well as in significant decreases in the body weight and increases in liver mutant frequency. When a two-fold cutoff value and a P-value less than 0.01 were selected, 2,726 genes were identified as differentially expressed in comfrey-fed rats compared to control animals. Among these genes, there were 1,617 genes associated by Ingenuity Pathway Analysis with particular functions, and the differentially expressed genes in comfrey-fed rat livers were involved in metabolism, injury of endothelial cells, and liver injury and abnormalities, including liver fibrosis and cancer development. Conclusion: The gene expression profile provides us a better understanding of underlying mechanisms for comfrey-induced hepatic toxicity. Integration of gene expression changes with known pathological changes can be used to formulate a mechanistic scheme for comfrey-induced liver toxicity and tumorigenesis. Page 1 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Mutation spectrum in the liver cII gene from comfrey-fed rats Comfrey-induced mutations in the liver cII gene were evaluated by DNA sequence analysis of 106 mutants iso- lated from 6 rats fed a diet containing 8% comfrey root. Since mutations that were found more than once among the mutants isolated from a single animal were assumed to be siblings and to represent only one independent mutation, a total of 99 independent mutations were iden- tified. Table 1 summarizes the types of cII mutations observed in the livers of rats fed 8% comfrey compared The major hepatotoxic manifestation in humans ingesting comfrey is the hepatic veno-occlusive lesion (VOD) [6,7], also called sinusoidal obstruction syndrome (SOS) [8]. Comfrey is also carcinogenic in rats, which suggests the potential tumorigenic effects of the plant [9]. Hepatocel- lular adenomas were induced in rats receiving a diet con- taining comfrey. Feeding rats comfrey leaves produced a dose-dependent reduction in survival and an increase in liver tumor incidence [9]. Comfrey roots are much more toxic than the leaves. Page 2 of 15 (page number not for citation purposes) Mean body weight of Big Blue rats fed with 8% comfrey root for 12 weeks Figure 1 Mean body weight of Big Blue rats fed with 8% comfrey root for 12 weeks. The data represent the means ± SD of 6 rats from the control (■) and the 8% comfrey root (●) groups. 100 150 200 250 300 350 400 0 1 2 3 4 5 6 7 8 9 10 11 12 Weeks on study Mean body weight (g) 100 150 200 250 300 350 400 0 1 2 3 4 5 6 7 8 9 10 11 12 Weeks on study Mean body weight (g) In a previous study, we developed evidence indicating that the liver tumors induced by feeding rats 2% comfrey root were generated by a genotoxic mechanism and that the PAs in the plant were responsible for mutation induction and tumor initiation in rat liver [10]. Considering the fact that rats tolerate diets containing up to 33% comfrey leaves and 8% comfrey roots for relatively long periods of time (at least 6 months) [9], in the present study, we eval- uated the mutagenicity of 8% comfrey root. Using a toxi- cogenomic approach, we analyzed the changes in global gene expressions in the liver of rats following comfrey- treatment. Mutation spectrum in the liver cII gene from comfrey-fed rats Weeks on study Mean body weight of Big Blue rats fed with 8% comfrey ro for 12 weeks Figure 1 Mean body weight of Big Blue rats fed with 8% comfrey root for 12 weeks. The data represent the means ± SD of 6 rats from the control (■) and the 8% comfrey root (●) groups. BMC Bioinformatics 2006, 7(Suppl 2):S16 Liver cII mutant frequencies in control and 8% comfrey-fed rats Figure 2 Liver cII mutant frequencies in control and 8% comfrey-fed rats. The data represent the means ± SD from groups of 6 rats. Significantly differed from control group (P < 0.001). Gene expression regulated by comfrey-treatment Gene expression regulated by comfrey-treatment In order to determine gene expression changes associated with comfrey exposure, we employed the Genome Survey Microarray, which contains 26,857 verified rat genes. We isolated total RNA from the livers of 6 control and 6 comfrey-fed rats. For the purpose of visualizing the data, the intensities of the whole rat gene data were analyzed by Principal Components Analysis within ArrayTrack (Figure 3). It demonstrates that one array of sample 6 from the control group appears to be quite different from the rest of the arrays, which was further identified as an outlier array by the Pearson's correlation coefficient of pair-wise log2 intensity correlation (data not shown). This outlier was excluded from further data analysis. A separation between control and comfrey-treated groups was clearly observed, suggesting that there was a clear comfrey-treat- ment effect on liver gene expression (Figure 3). To select significant genes, minimum requirements were estab- lished for both a two-fold change in the gene expression compared to the controls and a P-value less than 0.01 for the difference. A total of 2,726 genes satisfied the require- ments; 1,235 genes were up-regulated and 1,491 genes were down-regulated in response to comfrey treatment (Figure 4). Among the regulated genes, 1,671 were identi- fied by Ingenuity Pathway Analysis. In this study, we focused on genes involved in metabolism, injury of endothelial cells, and liver injury and abnormalities. Liver cII m rats Figure 2 Liver cII mutant frequencies in control and 8% comfrey fed rats Figure 2 Liver cII mutant frequencies in control and 8% comfrey-fed rats. The data represent the means ± SD from groups of 6 rats. Significantly differed from control group (P < 0.001). Mutation spectrum in the liver cII gene from comfrey-fed rats with mutation spectra of control and 2% comfrey-fed rats that we reported previously [10]. The overall pattern of mutations in 8% comfrey-fed rats differed significantly from controls (P < 0.0001), but did not differ from 2% comfrey-fed rats. Among the independent mutations, about 85% from both the comfrey-treated and control rats were base pair substitutions. G:C → T:A transversion (41%) was the major type of mutation in the 8% comfrey- fed rats, whereas G:C → A:T transition was the predomi- nant mutation in the controls. In addition, a 13% fre- quency of tandem base substitutions was observed among the mutations from the 8% comfrey-fed rats. The 8% comfrey-fed rats had more large deletions than control rats, although the percentages of frameshift mutations were the same. Genes associated with drug metabolizing genes Since the PAs in comfrey require metabolic activation to exert their biological effects, we investigated the gene expression changes of drug metabolizing genes. We observed the up- or down-regulation of many cytochrome P450 genes (e.g., Cyp2c12, Cyp3a18, Cyp4a12, Cyp26), glutathione S-transferases (Gsta3, Gstm3, and Gstp1), ATP-binding cassette transporters (e.g., Abcb9 and Abcc3), and other metabolism-associated genes, includ- Table 1: Summary of independent mutations in the liver cII gene from comfrey-fed and control Big Blue rats Type of mutation Control * 8% comfrey • 2% comfrey •, * Number % Number % Number % G:C → C:G 5 11 6 6 11 6 G:C → A:T 20 43 11 11 24 12 G:C → T:A 9 20 40 41 83 42 A:T → T:A 1 2 4 4 5 2 A:T → G:C 1 2 6 6 9 4 A:T → C:G 3 7 4 4 7 3 Frameshift 7 15 14 14 26 13 Complex 0 0 1 1 2 1 Tandem-base substitution 0 0 13 13 33 17 Total mutants screened 46 100 99 100 200 100 •Spectra for 2% and 8% comfrey-fed rats were significantly different from the controls (P < 0.001). There was no significant difference between the spectra for 2% and 8% comfrey. Data are from literature [10]. dent mutations in the liver cII gene from comfrey-fed and control Big Blue rats e 1: Summary of independent mutations in the liver cII gene from comfrey-fed and control Big Blue rats •Spectra for 2% and 8% comfrey-fed rats were significantly different from the controls (P < 0.001). There was no significant difference between the spectra for 2% and 8% comfrey. Data are from literature [10]. BMC Bioinformatics 2006, 7(Suppl 2):S16 Principal component analysis for gene expression profiles from livers of control and 8% comfrey-fed rats Figure 3 Principal component analysis for gene expression profiles from livers of control and 8% comfrey-fed rats. No specific cut off was applied and the intensity of whole rat genome data was used. The red and blue dots indicate control and comfrey-treated samples, respectively. Figure 3. Second principal component First principal component Second principal component First principal component First principal component First principal component First principal component First principal component Principal component analysis for gene expression profiles from livers of control and 8% comfrey-fed rats Figure 3 Principal component analysis for gene expression profiles from livers of control and 8% comfrey-fed rats. Genes associated with drug metabolizing genes No specific cut off was applied and the intensity of whole rat genome data was used. The red and blue dots indicate control and comfrey-treated samples, respectively. ing NAD(P)H oxidoreductase (Nqo1) and aldehyde dehy- drogenase (Aldh1a1). Table 2 shows the significant changes in gene expression related to phase I, II, and III drug metabolizing genes. Genes involved in injury of endothelial cells Genes involved in injury of endothelial cells Considering the involvement of sinusoidal endothelial cells and sometimes hepatic venular endothelial cells in VOD/SOS [8], we focused on genes expressed in endothe- Page 4 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Volcano plots (log2 fold change vs. -log2 P-value) Figure 4 Volcano plots (log2 fold change vs. -log2 P-value). A gene was identified as significantly changed if the fold change was greater than 2 (up or down) and the P-value was less than 0.01 in comparison to the control group. Each group consisted of 5– 6 replicates. o ca o p ots ( og o c a ge vs. og va ue) gu e Volcano plots (log2 fold change vs. -log2 P-value). A gene was identified as significantly changed if the fold change was greater than 2 (up or down) and the P-value was less than 0.01 in comparison to the control group. Each group consisted of 5– 6 replicates. lial cells. Table 3 summarizes the alteration of genes related to cell death, apoptosis, and cell growth of endothelial cells. For example, we observed the up-regula- tion of previously reported tumor necrosis factors (Tnfsf10, Tnfrsf6, and Tnfrsf12a), transforming growth factor (Tgfb1 and Tgfb2), and other genes (such as, Mmp2, PAI-1 or Serpine1, Plaur, Col4a2, and Edn1) [8,11,12]. Several novel genes were responsive to comfrey treatment, including activating transcription factor 3 (Atf3), B-cell cLL/lymphoma 2 (Bcl2), caspase 1 (Casp1), mitogen-activated protein kinase 9 (Mapk9), and secreted phosphoprotein 1 (Spp1). Genes involved in liver injury and abnormalities Genes involved in liver injury and abnormalities appeared to have a role in liver fibrosis (Table 4). Most of them encode cytokines (Hgf, Igf1, Lif, Tgfb1, and Tnfrsf6), angiogenic molecules (Agt, and Agtr1a), chemokines and their receptors (Ccr2, Cxcl4, and Cxcl12), coagulation cas- cade genes (PAI-1/Serpine1 and F11), and cytoskeletal molecules (Krt2-8). Moreover, eight genes related to the development of liver cancer were among the most dis- criminating genes involved in cancer development, such as growth arrest and DNA damage inducible 45 alpha (Gadd45a), B-cell cLL/lymphoma 2 (Bcl2), N-myc down- stream regulated 1 (Ndr1), and xeroderma pigmentosum complementation group C (Xpc). Endothelial damage can lead to hepatocellular injury, and fibrotic reaction in the sinusoids is characterized as the later stages of VOD/SOS. Chronic comfrey treatment also resulted in the induction of a number of genes involved in liver injury and abnormalities. Significantly changed genes were divided into subsets based on functionality, and categories included cell death, growth, fibrosis, mor- phology, and liver cancer development (Table 4). Sixteen genes were functionally categorized in liver cell death. Besides the tumor necrosis factors and transforming growth factor, comfrey treatment resulted in strong up- regulation of cyclin-dependent kinase inhibitor 1A (Cdkn1a) and heme oxyenase 1 (Hmox1), as well as down-regulation of epidermal growth factor (Egf) and CCAAT/enhancer binding protein beta (Cebpb). Of note, genes involved in cell growth were up-regulated along with the genes involved in liver cell death, such as cyclin E1 (Ccne1), hepatocyte growth factor (Hgf), and tissue inhibitor of metalloproteinase 1 (Timp1). Genes involved in injury of endothelial cells Page 5 of 15 (page number not for citation purposes) Page 5 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Table 2: Genes involved in drug metabolism altered by comfrey treatment in liver Gene symbol Gene description Locus link ID Fold change P-value Phase I metabolism Cyp2c cytochrome P450, family 2, subfamily c 29277 0.03 0.00000 Cyp2c12 cytochrome P450, family 2, subfamily c 25011 6.52 0.00000 Cyp2c39 cytochrome P450, family 2, subfamily c 29298 0.35 0.00001 Cyp2d1 cytochrome P450, family 2, subfamily d 266684 0.38 0.00000 Cyp2d2 cytochrome P450, family 2, subfamily d 25053 0.43 0.00002 Cyp2d3 cytochrome P450, family 2, subfamily d 24303 0.39 0.00000 Cyp2d5 cytochrome P450, family 2, subfamily d 286963 0.44 0.00000 Cyp3a18 cytochrome P450, family 3, subfamily a 252931 0.38 0.00060 Cyp4a3 cytochrome P450, family 4, subfamily a 298423 0.46 0.00010 Cyp4a12 cytochrome P450, family 4, subfamily a 266674 0.36 0.00015 Cyp4b1 cytochrome P450, family 4, subfamily b 24307 0.32 0.00008 Cyp4f5 cytochrome P450, family 4, subfamily f 286905 4.71 0.00000 Cyp4f14 cytochrome P450, family 4, subfamily f 56266 0.48 0.00005 Cyp4f18 cytochrome P450, family 4, subfamily f 72054 0.47 0.00014 Cyp7a1 cytochrome P450, family 7, subfamily a 25428 6.16 0.00067 Cyp7b1 cytochrome P450, family 7, subfamily b 13123 2.15 0.00169 Cyp11b3 cytochrome P450, family 11, subfamily b 353498 0.43 0.00032 Cyp26 cytochrome P450, family 26 154985 18.54 0.00005 Cyp39a1 cytochrome P450, family 39, subfamily a 56050 0.29 0.00315 Phase II metabolism Gsta3 glutathione S-transferase, alpha 3 14859 14.47 0.00001 Gstm3 glutathione S-transferase, mu 3 81869 0.39 0.00173 Gstp1 glutathione S-transferase, pi 1 14869 3.16 0.00001 Phase III metabolism Abcb9 ATP-binging cassette, subfamily b (MDR/TAP) 63886 0.23 0.00292 Abcc1 ATP-binging cassette, subfamily c (CFTR/MRP) 24565 2.65 0.00132 Abcc3 ATP-binging cassette, subfamily c (CFTR/MRP) 140668 21.20 0.00006 Abcc6 ATP-binging cassette, subfamily c (CFTR/MRP) 81642 0.42 0.00000 Abcc8 ATP-binging cassette, subfamily c (CFTR/MRP) 25559 0.10 0.00000 Table 2: Genes involved in drug metabolism altered by comfrey treatment in liver Locus link ID Fold change P-value Locus link ID Fold change appeared to have a role in liver fibrosis (Table 4). Most of them encode cytokines (Hgf, Igf1, Lif, Tgfb1, and Tnfrsf6), angiogenic molecules (Agt, and Agtr1a), chemokines and their receptors (Ccr2, Cxcl4, and Cxcl12), coagulation cas- cade genes (PAI-1/Serpine1 and F11), and cytoskeletal molecules (Krt2-8). Genes involved in injury of endothelial cells Moreover, eight genes related to the development of liver cancer were among the most dis- criminating genes involved in cancer development, such as growth arrest and DNA damage inducible 45 alpha (Gadd45a), B-cell cLL/lymphoma 2 (Bcl2), N-myc down- stream regulated 1 (Ndr1), and xeroderma pigmentosum complementation group C (Xpc). Page 6 of 15 (page number not for citation purposes) Discussion Comfrey was one of the most popular herbal teas in the world, including the United States. Although its popular- ity has declined due to the understanding of its dangers, it is still available commercially in several forms. The regular use of comfrey is a potential health risk owing to the pres- ence of PAs. Comfrey contains as many as nine PAs, including acetyl intermedine, acetyl lycopsamine, echimi- dine, intermedine, lasiocarpine, lycopsamine, symland- ine, symphytine, and symviridine [13-15]. The PA content of comfrey is less than 1% and depends on the plant part Liver fibrosis is the hallmark of all chronic liver diseases, irrespective of their origin. Discussion Ingenuity Pathway Analysis showed that sixteen significantly regulated genes Page 6 of 15 (page number not for citation purposes) Page 6 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Table 3: Genes involved in endothelial cells altered by comfrey treatment in liver Gene symbol Gene description Locus link ID Fold change P-value Cell death and apoptosis , • Angpt2 angiopoietin 2 11601 3.45 0.00121 App amyloid beta (A4) precursor protein 54226 4.69 0.00000 Atf3 activating transcription factor 3 25389 6.22 0.00010 Bcl2 B-cell cLL/lymphoma 2 24224 2.41 0.00500 Bcl2a1 Bcl2-related protein A1 170929 3.43 0.00022 Bcl2l Bcl2-like 12048 2.53 0.00000 Casp1 caspase 1, apoptosis-related cysteine protease (interleukin 1, beta, convertase) 25166 2.26 0.00551 Cnp 2',3'-cyclic nucleotide 3' phosphodiesterase 1267 2.37 0.00010 • Col4a2 collagen, type IV, alpha 2 1284 2.81 0.00000 • Cxcr4 chemokine receptor (LCR1) 60628 3.22 0.00186 Dusp6 dual specificity phosphatase 6 116663 2.31 0.00036 Edn1 endothelin 1 24323 4.87 0.00033 Fgf1 fibroblast growth factor 1 25317 2.05 0.00057 Gpr9 G protein-coupled receptor 9 (Cxcr3) 84475 2.02 0.00944 • Hgf hepatocyte growth factor 24446 3.18 0.00065 Hmox1 heme oxygenase 1 24451 3.23 0.00064 Il2 interleukin 2 116562 0.29 0.00010 Ins1 insulin 1 24505 0.22 0.00000 Lrp5 low density lipoprotein receptor-related protein 5 16973 0.35 0.00010 Mapk9 mitogen-activated protein kinase 9 50658 0.49 0.00009 Pparg peroxisome proliferator activated receptor, gamma 25664 0.32 0.00033 Serpinf1 serine (or cysteine) proteinase inhibitor, clade F), member 1 287526 0.49 0.00009 Spp1 secreted phosphoprotein 1 25353 11.63 0.00003 , • Tgfb1 transforming growth factor, beta 1 59086 2.11 0.00002 Tnfrsf6 tumor necrosis factor receptor superfamily, member 6 246097 3.40 0.00007 Tnfsf10 tumor necrosis factor (ligand) superfamily, member 10 246775 2.19 0.00366 Adhesion Anxa1 annexin 1 25380 2.50 0.00009 Anxa5 annexin 5 25673 4.29 0.00000 Cd44 CD44 antigen 25406 2.65 0.00004 Cx3cl1 chemokine (C-X3-C motif) ligand 1 89808 2.55 0.00333 • Cxcl4 chemokine (C-X-C motif) ligand 4 56744 3.12 0.00008 Cyr61 cysteine rich protein 61 83476 4.54 0.00000 •Itga4 integrin alpha 4 16401 5.28 0.00002 •Itgav integrin alpha V 16410 2.00 0.00113 Selplg selectin P ligand 6404 2.26 0.00088 Cell movement Agt angiotensinogen 24179 0.36 0.00000 Ccl2 chemokine (C-C motif) ligand 2 24770 4.27 0.00026 Cd151 CD151 antigen 64315 2.13 0.00000 Cd9 CD9 antigen 12527 2.10 0.00127 Ceacam1 carcinoembryonic antigen-related cell adhesion molecule 1 81613 0.49 0.00138 Cxcl12 chemokine (C-X-C motif) ligand 12 24772 0.19 0.00000 Edil3 EGF-like repeats and discoidin I-like domains 3 10085 3.57 0.00015 Efnb2 ephrin B2 13642 0.49 0.00006 Fgfr1 Fibroblast growth factor receptor 1 79114 2.25 0.00002 Id1 Inhibitor of DNA binding 1, helix-loop-helix protein (splice variation) 25261 2.89 0.00068 Lgals3 lectin, galactose binding, soluble 3 83781 2.31 0.00003 Mmp2 matrix metalloproteinase 2 17390 4.31 0.00000 Plaur urokinase plasminogen activator receptor 50692 2.93 0.00164 Ptk2b protein tyrosine kinase 2 beta 50646 3.85 0.00000 S100a4 S100 calcium-binding protein A4 24615 3.61 0.00001 Serpine1 serine (or cysteine) proteinase inhibitor, clade E, member 1 (plasminogen activator inhibitor type 1) 24617 30.94 0.00003 Stc1 stanniocalcin 1 81801 3.50 0.00306 TEK TEK tyrosine kinase, endothelial (venous malformations, multiple cutaneous and mucosal) 7010 0.34 0.00012 Tgfb2 transforming growth factor, beta 2 81809 2.31 0.00458 Tnfrsf12a tumor necrosis factor receptor superfamily, member 12a 302965 2.91 0.00090 Genes also involved in adhesion* and cell movement• of endothelial cells. Discussion Table 3: Genes involved in endothelial cells altered by comfrey treatment in liver Locus link ID Fold change P-value Cell death and apoptosis , • Angpt2 angiopoietin 2 App amyloid beta (A4) precursor protein Atf3 activating transcription factor 3 Bcl2 B-cell cLL/lymphoma 2 Bcl2a1 Bcl2-related protein A1 Bcl2l Bcl2-like Casp1 caspase 1, apoptosis-related cysteine protease (interleukin Cnp 2',3'-cyclic nucleotide 3' phosphodiesterase • Col4a2 collagen, type IV, alpha 2 • Cxcr4 chemokine receptor (LCR1) Dusp6 dual specificity phosphatase 6 Edn1 endothelin 1 Fgf1 fibroblast growth factor 1 Gpr9 G protein-coupled receptor 9 (Cxcr3) • Hgf hepatocyte growth factor Hmox1 heme oxygenase 1 Il2 interleukin 2 Ins1 insulin 1 Lrp5 low density lipoprotein receptor-related protein 5 Mapk9 mitogen-activated protein kinase 9 Pparg peroxisome proliferator activated receptor, gamma Serpinf1 serine (or cysteine) proteinase inhibitor, clade F), membe Spp1 secreted phosphoprotein 1 , • Tgfb1 transforming growth factor, beta 1 Tnfrsf6 tumor necrosis factor receptor superfamily, member 6 Tnfsf10 tumor necrosis factor (ligand) superfamily, member 10 Adhesion Anxa1 annexin 1 Anxa5 annexin 5 Cd44 CD44 antigen Cx3cl1 chemokine (C-X3-C motif) ligand 1 • Cxcl4 chemokine (C-X-C motif) ligand 4 Cyr61 cysteine rich protein 61 •Itga4 integrin alpha 4 •Itgav integrin alpha V Selplg selectin P ligand Cell movement Agt angiotensinogen Ccl2 chemokine (C-C motif) ligand 2 Cd151 CD151 antigen Cd9 CD9 antigen Ceacam1 carcinoembryonic antigen-related cell adhesion molecule Cxcl12 chemokine (C-X-C motif) ligand 12 Edil3 EGF-like repeats and discoidin I-like domains 3 Efnb2 ephrin B2 Fgfr1 Fibroblast growth factor receptor 1 Id1 Inhibitor of DNA binding 1, helix-loop-helix protein (splic Lgals3 lectin, galactose binding, soluble 3 Mmp2 matrix metalloproteinase 2 Plaur urokinase plasminogen activator receptor Ptk2b protein tyrosine kinase 2 beta S100a4 S100 calcium-binding protein A4 Serpine1 serine (or cysteine) proteinase inhibitor, clade E, member inhibitor type 1) Stc1 stanniocalcin 1 TEK TEK tyrosine kinase, endothelial (venous malformations, m mucosal) Tgfb2 transforming growth factor, beta 2 Tnfrsf12a tumor necrosis factor receptor superfamily, member 12a Genes also involved in adhesion* and cell movement• of endothelial cells. Discussion BMC Bioinformatics 2006, 7(Suppl 2):S16 Table 4: Genes involved in liver injury and abnormalities altered by comfrey-treatment Locus link ID Fold change P-value y p g Cell death of liver cells Bcl2 B-cell cLL/lymphoma 2 24224 2.41 0.00500 Bcl2l Bcl2-like 12048 2.53 0.00000 Ccr2 chemokine receptor CCR2 gene 60463 2.47 0.00591 Cdkn1a cyclin-dependent kinase inhibitor 1A (p21) 114851 4.83 0.00014 Cebpb CCAAT/enhancer binding protein (C/EBP), beta 24253 0.46 0.00000 Egf epidermal growth factor 25313 0.36 0.00000 Hmox1 heme oxygenase 1 24451 3.23 0.00064 Ier3 immediate early response 3 294235 2.24 0.00002 Itgav integrin alpha V 16410 2.00 0.00113 Krt2-8 keratin complex 2, basic, gene 8 25626 2.82 0.00000 Nos2 nitric oxide synthase 2, inducible 24599 3.38 0.00491 , • Tgfb1 transforming growth factor, beta 1 59086 2.11 0.00002 Tnfrsf5 tumor necrosis factor receptor superfamily, member 5 21939 3.58 0.00009 , • Tnfrsf6 Tumor necrosis factor receptor superfamily, member 6 246097 3.40 0.00007 Tnfsf10 tumor necrosis factor (ligand) superfamily, member 10 246775 2.19 0.00366 Ugcg UDP-glucose ceramide glycosyltransferase 83626 2.08 0.00000 Growth of liver cells Ccne1 cyclin E1 898 7.93 0.00000 Edn1 endothelin 1 24323 4.87 0.00033 , • Hgf hepatocyte growth factor 24446 3.18 0.00065 Inhba inhibin beta-A (activin A) 29200 0.08 0.00000 Tgfb1 transforming growth factor, beta 1 59086 2.11 0.00002 Timp1 tissue inhibitor of metalloproteinase 1 116510 4.25 0.00000 Liver fibrosis Adrb2 adrenergic receptor, beta 2 24176 3.58 0.00000 Agt angiotensinogen 24179 0.36 0.00000 Agtr1a angiotensin II receptor, type 1 (AT1A) 24180 0.38 0.00000 Atf3 activating transcription factor 3 25389 6.22 0.00010 Cav caveolin 25404 3.15 0.00006 Cxcl4 chemokine (C-X-C motif) ligand 4 56744 3.12 0.00008 Cxcl12 chemokine (C-X-C motif) ligand 12 24772 0.19 0.00000 F11 coagulation factor XI 109821 0.43 0.00000 Igf1 insulin-like growth factor 1 24482 0.48 0.00000 Lif leukemia inhibitory factor 60584 0.41 0.00215 Serpine1 serine (or cysteine) proteinase inhibitor, clade E member 1 (plasminogen activator inhibitor type 1) 24617 30.94 0.00003 Liver morphology Ccnd1 cyclin D1 58919 5.65 0.00002 Ins1 insulin 1 24505 0.22 0.00000 Junb Jun-B oncogene 24517 3.77 0.00001 Liver cancer development Bcl2 B-cell cLL/lymphoma 2 24224 2.41 0.00500 Ccne1 cyclin E1 898 7.93 0.00000 Gadd45a growth arrest and DNA-damage-inducible 45 alpha 25112 2.15 0.00326 Gjb1 gap junction membrane channel protein beta 1 29584 0.48 0.00003 Hgf hepatocyte growth factor 24446 3.18 0.00065 Ndr1 N-myc downstream regulated 1 (Ndrg1) 10397 2.26 0.00020 Timp1 tissue inhibitor of metalloproteinase 1 116510 4.25 0.00000 Xpc xeroderma pigmentosum, complementation group C 7508 0.42 0.00001 Five genes also involved in liver fibrosis and • three genes also involved in liver morphology. Discussion be derived from comfrey roots [16,17]. PAs are the leading plant toxins associated with disease in humans and ani- mals. [15]. Higher PA concentrations occur in the roots of comfrey than in the leaves, and commercially available comfrey tablets containing high levels of PAs are likely to [15]. Higher PA concentrations occur in the roots of comfrey than in the leaves, and commercially available comfrey tablets containing high levels of PAs are likely to Page 8 of 15 (page number not for citation purposes) Page 8 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Several cases of VOD/SOS associated with comfrey inges- tion have been reported in humans [6,7,18-20], as well as in experimental animals [21]. Deleve and her colleagues [22] have developed a reproducible animal model of hepatic VOD, in which the rats are gavaged with a single dose of PA, monocrotaline. The model exhibits the char- acteristic clinical and histological features of hepatic VOD, including the earliest manifestations (progressive injury to the sinusoidal endothelial cells and central vein endothelium), early VOD (centrilobular coagulative necrosis and severe sinusoidal injury, hemorrhage, and central vein endothelial damage), and late VOD (fibrotic occulusion of the central veins) [22]. Comfrey and the specific alkaloids in comfrey (e.g., symphytine and lasio- carpine) induce hepatoadenoma and hemangioendothe- lial sarcomas in rats [9,23,24]. The mechanisms by which toxicity and carcinogenicity are produced are still not fully understood. Gene expression profiling offers a powerful approach for identifying differentially expressed genes and identifying mechanisms. ers. It is known that many herbal/dietary constituents form reactive intermediates capable of irreversibly inhib- iting various Cyps (reviewed in [25]). The resultant metabolites lead to Cyp inactivation by chemical modifi- cation of the heme, the apoprotein, or both, as a result of covalent binding of modified heme to the apoprotein. Phase II consists of conjugating enzymes, such as glutath- ione S-transferases (GSTs), UDP-glucuronosyltransferases (UGTs), and sulfatases. GSTs also participate in oxidative stress release pathways. The altered expression of phase I and phase II enzymes along with altered drug transport proteins (phase III) could contribute to the increased sus- ceptibility of rats to carcinogenic chemicals, such as comfrey. Sinusoidal endothelial cells are more susceptible than hepatocytes to PAs that cause VOD/SOS [22]. Discussion Functional annotation extracted from Ingenuity Pathway Analysis revealed that many of the transcriptional responses were associated with the apoptosis, cell death, adhesion, and cell movement of endothelial cells (Table 3). The genes in these pathways were highly expressed in comfrey-treated livers; these genes included the endothelial cell markers plasminogen activator inhibitor type 1 (PAI-1, also known as Serpine1 and Serpine2) and tissue plasminogen activator (Plat), and the cytokine tumor necrosis factors (Tnfrsf6, Tnfrsf12a, and Tnfsf10), and transforming growth factors (Tgfb1 and Tgfb2). Induction of a number of genes involved in the injury of endothelial cells was also detected, including endothelin 1 (Edn1), urokinase plasminogen activator receptor (Plaur), collagen type IV alpha 2 (Col4a2), matrix metalloproteinase 2 (Mmp2), mitogen-activated protein kinase 9 (Mapk9), and secreted phosphoprotein 1 (Spp1). It has been reported that Edn1 is a mediator of hepatic sinusoidal constriction, and increased activity of matrix metalloproteinases is respon- sible for changes of sinusoidal endothelial cells [11]. Ele- vated plasma PAI-1 levels are useful in distinguishing VOD/SOS [26]. Endothelial injury is the initiating event in the cascade of events leading to the hepatic changes and clinical manifestation of VOD/SOS [27]. Our results offer a more comprehensive overview of the molecular responses to comfrey exposure by expression of multiple genes in liver endothelial cells. Gene expression was markedly affected (Figure 3) in the livers of rats exposed to 8% comfrey root, a dose that resulted in significant decreases in body weight (Figure 1) and increases in liver MF (Figure 2). Out of 26,857 genes evaluated, the expression of 4,132 (15%) and 9,937 (37%) genes were altered more than 2-fold and 1.5-fold, respectively. At P-values of 0.01 and 0.05, 7,518 (28% of expressed genes) and 1,0341 (39%) genes, respectively, displayed a significant effect after comfrey-treatment com- pared to control group. In this study, differential gene expression was considered significant for genes showing at least a 2-fold up- or down-change, and a P < 0.01. In total, 2,726 genes (10%) satisfied the requirements and about half of them were down-regulated and half up-reg- ulated in response to comfrey exposure (Figure 4). Such a large number of significantly altered genes may partly reflect the therapeutic effects of comfrey exerted through plant components other than PAs [15]. In the present study, we concentrated on the analysis of genes involved in metabolism, injury of endothelial cells, and liver injury and abnormalities. Discussion Gap junction membrane channel protein beta 1 (Gjb1), also called connexin 32 (Cx32), is the main gap junction protein in hepatocytes and plays an important role in the regulation of signal transfer and growth control in the liver. It has been reported that Cx32 expression decreases gradually as liver disease progresses to cirrhosis and HCC [43], and a low expression of Cx32 mRNAs in HCC tissues is also predictive of the postoperative recur- rence of HCCs [44]. Bcl-2 is characterized as an antiapop- totic/oncogenic protein and also functions as an antioxidant. Increased Bcl-2 expression in cirrhotic patients correlates with the development of HCC [45]. Bcl-2 is also expressed in HCC tissues and the increasing Bcl-2 expression associated with HCC progression sug- gests that the Bcl-2 protein takes part in the formation of HCC [46]. Hgf, identified originally as the most potent mitogen for hepatocytes, is now known to be a cytokine with numerous functions in a wide variety of cells [47]. It is up-regulated in inflammatory liver diseases and stimu- lates DNA synthesis preferentially in initiated hepato- cytes, presumably resulting in tumour promotion [48]. N- myc downstream-regulated gene 1 (Ndr1, or Ndrg1) plays a role in growth arrest and cell differentiation, is induced PA-induced DNA damage in the liver (endothelial cells and hepatocytes), if not repaired prior to DNA synthesis, might produce replication errors and mutations, which eventually could result in the development of neoplasmas in the treated animals. We determined MFs in the liver cII gene of Big Blue transgenic rats. After feeding with 8% comfrey root for 12 weeks, we observed a 4-fold higher MF in the liver cII gene compared to the controls (Figure 2). The induction of mutation was similar to that reported previously for rats fed with 2% comfrey root [10]. These observations suggest that the rats could not tolerate the feeding of roots in concentrations over 2%, in terms of mutation induction. Furthermore, the overall pattern of mutations induced by 8% comfrey in liver was similar to that in the livers of rats fed 2% comfrey root (Table 1), whereas both the 2% and 8% comfrey-induced mutation spectra were significantly different from liver controls. Discussion known that Hmox1 induction is a protective mechanism against the oxidative stress associated with liver injury [28], and that elevated Hgf could protect hepatocytes from injury or promote hepatocellular regeneration [29]. Since Inhba, so-called activin A, is a negative regulator of hepatocyte cell growth [30], the decreased expression of Inhba observed after comfrey treatment suggests the induction of hepatic growth. known that Hmox1 induction is a protective mechanism against the oxidative stress associated with liver injury [28], and that elevated Hgf could protect hepatocytes from injury or promote hepatocellular regeneration [29]. Since Inhba, so-called activin A, is a negative regulator of hepatocyte cell growth [30], the decreased expression of Inhba observed after comfrey treatment suggests the induction of hepatic growth. Necrosis of hepatocytes and mesenchymal cells follows comfrey-induced liver cell injury, and functional cells are replaced by fibrotic tissues [27,31,32]. In the present study, we observed 16 genes involved in the function of liver fibrosis by Ingenuity Pathway Analysis (Table 4). Liver fibrosis is characterized by cell proliferation and the accumulation of extracellular matrix components and is mediated by cytokines and growth factors, of which TGF- β1 appears to be a key mediator [32]. The up-regulation of cytokines Hgf and Tnfrsf6 and down-regulation of Igf1 and Lif play an important role in the pathogenesis of liver injury and fibrosis. Decreased serum Igf-1 levels provide a useful index of hepatocellular dysfunction and impaired nutritional status, and increased Hgf appears to limit liver fibrosis [33]. Agt and Agtr1a, cytokines with vasoactive properties, also regulate liver fibrogenesis. Chemokines have a much wider biological role including angiogenesis, carcinogenesis, and cell cycle control [34]. Chemokines in the liver (Cxcl4 and Cxcl12) may modulate the progres- sion of liver fibrosis through their actions on hepatic stel- late cells. Ingenuity Pathway Analysis found 8 genes (2 down- and 6 up -regulated) involved in liver cancer development that were altered due to comfrey treatment (Table 5). In mam- mals, the nucleotide excision repair process is the most important repair pathway for elimination of DNA damage caused by exogenous agents, including UV light, DNA- reactive carcinogens, and some endogenously generated oxidative lesions [41]. Xeroderma pigmentosum group C (Xpc) is implicated in the early steps of this repair path- way. A significantly higher incidence of chemically induced liver and lung tumors is observed in Xpc null mice [42]. Discussion Liver is the major organ for biotransformation of xenobi- otics and drugs. PAs are metabolically activated to toxic, alkylating pyrroles by mixed-function oxidases. The cyto- chrome P450 (Cyp) superfamily contains 57 genes and plays a critical role in the phase I metabolism of a variety of xenobiotics including drugs, carcinogens, steroids and eicosanoids [25]. In the present study, comfrey exposure resulted in changes in the expression of 19 Cyp genes (Table 2). Among phase I, II, and III drug metabolizing genes, Cyp2c12, Cyp7a1, Cyp26, Gsta3, and Abcc3 were increased 6-21-fold. In contrast, Cyp2c, Cyp39a1, Gstm3, Abcc8, and others were reduced in the comfrey-treated liv- PAs in comfrey can reach the hepatocytes via the sinusoi- dal blood, and their toxic metabolites lead to immediate damage to the hepatocytes [4]. Genes involved in liver cell death and growth were also induced or repressed in response to comfrey treatment (Table 4), including tumor necrosis factors, transforming growth factor β, chemokine receptor CCR2 gene (Ccr2), heme oxygenase 1 (Hmox1), immediate early response 3 (Ier3), cyclin-dependent kinase inhibitor (Cdkn1a), inhibin beta A (Inhba), and tissue inhibitor of metalloproteinase 1 (Timp1). It is well Page 9 of 15 (page number not for citation purposes) Page 9 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 tion, which has been suggested as a mutational signature for the genetic damage of PAs [36]. G:C → T:A transver- sion may cause the initiation of tumors in the liver of rats fed with comfrey, because it has been reported that more than half of riddelliine-induced liver hemangiosarcomas have a G → T mutation at K-ras codon 12 [37]. p53 muta- tion also has been detected at an early stage of riddelliine exposure [38]. Mutations are thought to be involved in carcinogenesis because the transition from a normal somatic cell to a cancer cell is due to mutations in pro- tooncogenes, tumor suppressor genes and/or genes that function in the maintenance of genomic stability [39,40]. Comfrey-induced neoplasms in the rat were mostly found in the liver. Hepatocellular adenomas were induced in all experimental groups that received diets containing 1–8% comfrey root or 8–33% comfrey leaves. In addition, a few rats bearing hepatocellular adenomas simultaneously had hemangioendothelial sarcoma of the liver [9]. The first hepatocellular carcinoma (HCC) appearance was 7 months after initiating the 8% comfrey diet. Page 10 of 15 (page number not for citation purposes) Conclusion h The integration of gene expression changes with a mecha- nistic pathway analysis suggests a scheme for comfrey treatment leading to tumorigenesis via mutation induc- tion (Figure 5). The available evidence suggests that active metabolites of PAs in comfrey interact with endothelial and hepatocyte DNA, causing damage to hepatic endothe- lial cells and hepatocytes. This may result in liver fibrosis and increases in mutation induction, which may be asso- ciated with the development of HCC and hemagiosar- coma. We have identified 2,726 genes in the livers of comfrey-fed rats that were differentially expressed. Some of the gene changes are associated with the metabolism, injury of endothelial cells, and liver injury and abnormal- ities that are postulated to occur as a result of comfrey exposure. This approach provides further insight into the mechanisms involved in the development of VOD/SOS and tumorigenesis after exposure to comfrey. Sequence analysis of the cII mutants h d d The mutants were sequenced according to the method of Mei et al. [36]. The cII mutant plaques were selected at random from different animals and replated at low den- sity to verify the mutant phenotype. Single, well-isolated plaques were selected from these plates and transferred to a microcentrifuge tube containing 100 μl of sterile dis- tilled water. The tube was heated at 100°C for 5 min and centrifuged at 12,000 g for 3 min. The cII target DNA for sequencing was amplified by PCR using primers 5'- AAAAAGGGCATCAAATTAACC-3' (upstream) and 5'- CCGAAGTTGAGTATTTTTGCTG-3' (downstream). For Materials and methods Plant material and animals Comfrey roots (Symphytum officinale) were purchased from Camas Prairie Products (Trout Lake, WA). Male Big Blue Fisher 344 transgenic rats were obtained from Taconic Laboratories (Germantown, NY) through pur- chase from Stratagene (La Jolla, CA). All animal proce- dures followed the recommendations of the NCTR Institutional Animal Care and Use Committee for the handling, maintenance, treatment, and sacrifice of the rats. cII mutation assay i h l l y High-molecular-weight genomic DNA was extracted from rat livers using the RecoverEase DNA Isolation Kit (Strata- gene) and stored at 4°C until DNA packaging was per- formed. The packaging of the phage, plating the packaged DNA samples, and determination of MF were carried out following the manufacturer's instructions for the λ Select- cII Mutation Detection System for Big Blue Rodents (Strat- agene). The shuttle vector containing the cII target gene was rescued from total genomic DNA with phage packag- ing extract (Transpack, Stratagene). The plating was per- formed with the Escherichia coli host strain G1250. To determine the total titer of packaged phages, G1250 bac- teria were mixed with 1:3000 dilutions of phage, plated on TB1 plates, and incubated overnight at 37°C (nonse- lective conditions). For mutant selection, the packaged phages were mixed with G1250, plated on TB1 plates, and incubated at 24°C for about 42 h (conditions for cII-selec- tion). Under these conditions, phages with wild-type cII genes undergo lysogenization and become part of the developing bacterial lawn, whereas phages with mutated cII genes undergo lytic growth and give rise to plaques. When incubated at 37°C, phages with wild-type cII genes also undergo a lytic cycle, resulting in plaque formation. Assays were repeated until a minimum of 2 × 105 plaque- forming units from each sample were examined for muta- tion. The cII MF is defined as the total number of mutant plaques (determined at 24°C) divided by the total number of plaques screened (determined at 37°C). Discussion Tissue inhibitor of metalloprotei- nase 1 (Timp1) is a contributory factor to fibrosis of a vari- ety of organs including the liver. Timp1 and other extracellular matrix remodeling genes are implicated in the transition from mild to moderate fibrosis in patients with chronic hepatitis C [55]. temperature until use. NIH-31 pellets (Purina Mills Inter- national, Brentwood, MO) were autoclaved and ground into meal form. The base diet was blended with comfrey root powder in a Hobart Mixer to make 8% comfrey root dosed diet, which was provided in rat feeders. The treat- ment schedule was based on the protocol used in a car- cinogenesis assay [9]. Male, 6-week-old Big Blue rats were fed without (vehicle control group) or with 8% comfrey roots for 12 weeks. Six rats from each treatment group were sacrificed at the end of the treatment. The livers were isolated, frozen quickly in liquid nitrogen, and stored at - 80°C. Discussion In contrast to the G:C → A:T transition that was the predom- inant mutation in the controls, the major type of muta- tion in the 8% comfrey-fed rats was G:C → T:A transversion (41%), a mutation that was also induced by riddelliine, a representative genotoxic PA that is tumori- genic for rat liver [35]. In addition, 13% of mutations from the 8% comfrey-fed rats were tandem base substitu- Page 10 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 by several stress conditions, and is overexpressed in many cancers [49]. Expression of Ndrg1 was significantly up- regulated in HCC tissues compared to that of noncancer- ous and normal liver tissues [50]. Cyclin E (Ccne1), a reg- ulatory subunit of cyclin-dependent kinase 2, is an important regulator for entry into the S phase of the mam- malian cell cycle. Overexpression of Ccne1 has been observed in many tumors including primary HCCs [51]; overexpression results in chromosome instability and thus may contribute to tumorigenesis [52]. Growth arrest and DNA damage 45 alpha (Gadd45a) is a nuclear pro- tein involved in the maintenance of genomic stability, DNA repair, and the suppression of cell growth [53]. Gadd45a protein levels are higher in liver cirrhotic and neoplastic tissues [54]. Tissue inhibitor of metalloprotei- nase 1 (Timp1) is a contributory factor to fibrosis of a vari- ety of organs including the liver. Timp1 and other extracellular matrix remodeling genes are implicated in the transition from mild to moderate fibrosis in patients with chronic hepatitis C [55]. by several stress conditions, and is overexpressed in many cancers [49]. Expression of Ndrg1 was significantly up- regulated in HCC tissues compared to that of noncancer- ous and normal liver tissues [50]. Cyclin E (Ccne1), a reg- ulatory subunit of cyclin-dependent kinase 2, is an important regulator for entry into the S phase of the mam- malian cell cycle. Overexpression of Ccne1 has been observed in many tumors including primary HCCs [51]; overexpression results in chromosome instability and thus may contribute to tumorigenesis [52]. Growth arrest and DNA damage 45 alpha (Gadd45a) is a nuclear pro- tein involved in the maintenance of genomic stability, DNA repair, and the suppression of cell growth [53]. Gadd45a protein levels are higher in liver cirrhotic and neoplastic tissues [54]. Page 11 of 15 (page number not for citation purposes) Comfrey diet and treatments The comfrey roots were ground into powder using a Wiley Mill, and the comfrey root powder was stored at room Page 11 of 15 (page number not for citation purposes) Page 11 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 Integration of gene expression with the proposed mechanistic pathway initiated by comfrey treatment leading to tumorigenesi Figure 5 Integration of gene expression with the proposed mechanistic pathway initiated by comfrey treatment leading to tumorigene- sis. Responses in boxes represent gene expression observations and arrows in the boxes indicate direction of indicated response. Mutation induction in liver e.g., ras, p53 Development of liver tumors Hepatocellular carcinoma Hemangiosarcoma Reactive metabolites, e.g., pyrroles Comfrey including PAs Hepatocyte necrosis !" !# !$ % & ' ( ( # () * * ( + + (, * + - Endothelial injury +(( , & ( ./ 0$$, 0( 0(( ! !$ (- !&1 2 ," ! ( 3 4( Sinosoidal fibrosis + + , 1 56 0( ! !# + 7 DNA damage Liver injury Comfrey including PAs Integration of gene expression with the proposed mechanistic pathway initiated by comfrey treatment leading to tumorigenesis Figure 5 Integration of gene expression with the proposed mechanistic pathway initiated by comfrey treatment leading to tumorigene- sis. Responses in boxes represent gene expression observations and arrows in the boxes indicate direction of indicated response. Page 12 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 PCR amplification, 10 μl of the supernatant were added to 10 μl of a PCR Master Mix (Promega, Madison, WI) and the primers. The final concentrations of the reagents were: 1× Taq polymerase reaction buffer, 0.2 μM of each primer, 200 μM of each dNTP, 1.5 mM MgCl2, and 0.25 U of Taq DNA polymerase. The PCR reactionwas performed using a PCR System 9700 (Applied Biosystems, Foster City, CA), with the following cycling parameters: a 3 min denatura- tion at 95°C, followed by 35 cycles of 30 s at 95°C, 1 min at 60°C, and 1 min at 72°C, with a final extension of 10 min at 72°C. The PCR products were isolated using a PCR purification kit (Qiagen, Chatsworth, CA). The cII mutant DNA was sequenced with a CEQ Dye Terminator Cycle Sequencing Kit and a CEQ 8000 Genetic Analysis System (Beckman Coulter, Fullerton, CA). The primer for cII mutation sequencing was the upstream primer used for the PCR. scription was performed by incubation of the cDNA product with T7 RNA polymerase, 0.75 mM Digoxigenin- 11-UTP (Roche Applied Science, Indianapolis, IN) and all other NTPs for 9 h. Labeled cRNA was purified according to the RT-IVT kit protocol and analyzed for quality and quantity using standard UV spectrometry and the Bioana- lyzer. Statistical analyses for mutagenicity data Analyses were performed using the SigmaStat 2.03 pro- gram (SPSS, Chicago, IL). All of the MF data were expressed as the mean ± standard deviation (SD) from 6 rats per group. Statistical significance was determined by one-way analysis of variance (ANOVA) followed by the Holm-Sidak test. Mutational spectra were compared using the computer program written by Cariello and colleagues [56] for the Monte Carlo analysis developed by Adams and Skopek [57]. Microarray data analysis f Gene expression data from the Applied Biosystems' Rat Genome Survey Microarray were input to ArrayTrack, a software system developed by the FDA's National Center for Toxicological Research for the management, analysis, visualization and interpretation of microarray data [58]. Raw microarray intensity data were normalized per chip to the same median intensity value of 500. Chemilumi- nescent signals from 1529 control probes that track sys- tem performance through each experiment were not used in normalization. The identification of differentially expressed genes based on t-tests and fold-change cutoffs, and Principal Component Analysis were conducted within ArrayTrack. Ingenutity Pathway Analysis (Moun- tain View, CA) was used for pathway and function analy- sis. Preparation of digoxigenin labeled in vitro transcribed cRNA All RNA targets were labeled using the Applied Biosystems RT-IVT Labeling Kit Version 2.0. Briefly, 1.5 μg of total RNA was reverse transcribed via 2 h incubation at 42°C with ArrayScript RT enzyme (Ambion, Austin, TX) and oligo dT-T7 primer. Double stranded cDNA was produced following 2 h incubation with E. coli DNA polymerase and RNase H at 16°C. Double-stranded cDNA was puri- fied according to the RT-IVT kit protocol. In vitro tran- Hybridization of labeled cRNA to microarrays and microarray imaging y g g Digoxigenin labeled cRNA targets were hybridized to Applied Biosystems Rat Whole Genome Survey Microar- rays using the Applied Biosystems Chemiluminescent Detection Kit. Briefly, 15 μg of labeled cRNA targets were fragmented via incubation with fragmentation buffer pro- vided in the kit for 30 min at 60°C. Fragmented targets were hybridized to microarrays during a 16 h incubation at 55°C with buffers and reagents from the Chemilumi- nescent Detection Kit. Post-hybridization washes and anti-Digoxigenin-Alkaline Phosphatase binding were per- formed according to the protocol of the kit. Chemilumi- nescence detection, image acquisition and analysis were performed using Applied Biosystems Chemiluminescence Detection Kit and Applied Biosystems 1700 Chemilumi- nescent Microarray Analyzer following the manufacturer's protocols. Images were auto-gridded and the chemilumi- nescent signals were quantified, corrected for background, and finally, spot- and spatially-normalized using the Applied Biosystems 1700 Chemiluminescent Microarray Analyzer software version 1.1. RNA isolation and quality control Total RNA was isolated from liver tissues of 6 control and 6 comfrey-fed rats using an RNeasy system (Qiagen). The yield of the extracted RNA was determined spectrophoto- metrically by measuring the optical density at 260 nm. The purity and quality of extracted RNA were evaluated using the RNA 6000 LabChip and Agilent 2100 Bioana- lyzer (Agilent Technologies, Palo Alto, CA). RNA samples with RNA integrity numbers (RINs) greater than 7.5 were used for microarray experiments performed using Applied Biosystems' Rat Genome Survey Microarray platform, which is a one channel microarray with chemilumines- cence detection, and contains 26,857 probes (60-mer) for the interrogation of 27,088 genes and 1592 controls that track system performance through each experiment. Page 13 of 15 (page number not for citation purposes) References Mei N, Guo L, Fu PP, Heflich RH, Chen T: Mutagenicity of comfrey (Symphytum Officinale) in rat liver. Br J Cancer 2005, 92(5):873-875. 33. Sanz S, Pucilowska JB, Liu S, Rodriguez-Ortigosa CM, Lund PK, Bren- ner DA, Fuller CR, Simmons JG, Pardo A, Martinez-Chantar ML, et al.: Expression of insulin-like growth factor I by activated hepatic stellate cells reduces fibrogenesis and enhances regenera- tion after liver injury. Gut 2005, 54(1):134-141. ( ) 11. DeLeve LD, Wang X, Kanel GC, Ito Y, Bethea NW, McCuskey MK, Tokes ZA, Tsai J, McCuskey RS: Decreased hepatic nitric oxide production contributes to the development of rat sinusoidal obstruction syndrome. Hepatology 2003, 38(4):900-908. 34. Marra F: Chemokines in liver inflammation and fibrosis. Front Biosci 2002, 7:d1899-1914. 12. Paradis V, Bieche I, Dargere D, Cazals-Hatem D, Laurendeau I, Saada V, Belghiti J, Bezeaud A, Vidaud M, Bedossa P, et al.: Quantitative gene expression in Budd-Chiari syndrome: a molecular approach to the pathogenesis of the disease. Gut 2005, 54(12):1776-1781. 35. Mei N, Chou MW, Fu PP, Heflich RH, Chen T: Differential muta- genicity of riddelliine in liver endothelial and parenchymal cells of transgenic big blue rats. Cancer Lett 2004, 215(2):151-158. ( ) 36. Mei N, Heflich RH, Chou MW, Chen T: Mutations induced by the carcinogenic pyrrolizidine alkaloid riddelliine in the liver cII gene of transgenic big blue rats. Chem Res Toxicol 2004, 17(6):814-818. ( ) 13. Kim NC, Oberlies NH, Brine DR, Handy RW, Wani MC, Wall ME: Isolation of symlandine from the roots of common comfrey (Symphytum officinale) using countercurrent chromatogra- phy. J Nat Prod 2001, 64(2):251-253. 37. Hong HL, Ton TV, Devereux TR, Moomaw C, Clayton N, Chan P, Dunnick JK, Sills RC: Chemical-specific alterations in ras, p53, and beta-catenin genes in hemangiosarcomas from B6C3F1 mice exposed to o-nitrotoluene or riddelliine for 2 years. Toxicol Appl Pharmacol 2003, 191(3):227-234. p y J ( ) 14. Schaneberg BT, Molyneux RJ, Khan IA: Evaporative light scatter- ing detection of pyrrolizidine alkaloids. Phytochem Anal 2004, 15(1):36-39. ( ) 15. Stickel F, Seitz HK: The efficacy and safety of comfrey. Public Health Nutr 2000, 3(4A):501-508. pp ( ) 38. Nyska A, Moomaw CR, Foley JF, Maronpot RR, Malarkey DE, Cum- mings CA, Peddada S, Moyer CF, Allen DG, Travlos G, et al.: The hepatic endothelial carcinogen riddelliine induces endothe- lial apoptosis, mitosis, S phase, and p53 and hepatocytic vas- cular endothelial growth factor expression after short-term exposure. References 1. Winship KA: Toxicity of comfrey. Adverse Drug React Toxicol Rev 1991, 10(1):47-59. 2. Rode D: Comfrey toxicity revisited. Trends Pharmacol Sci 2002, 23(11):497-499. ( ) 27. Kumar S, DeLeve LD, Kamath PS, Tefferi A: Hepatic veno-occlu- sive disease (sinusoidal obstruction syndrome) after hemat- opoietic stem cell transplantation. Mayo Clin Proc 2003, 78(5):589-598. ( ) 3. Fu PP, Xia Q, Lin G, Chou MW: Pyrrolizidinealkaloids – genoto- xicity, metabolism enzymes, metabolic activation, and mechanisms. Drug Metab Rev 2004, 36(1):1-55. g ( ) 4. Prakash AS, Pereira TN, Reilly PE, Seawright AA: Pyrrolizidine alkaloids in human diet. Mutat Res 1999, 443(1–2):53-67. ( ) 28. Bauer M, Bauer I: Heme oxygenase-1: redox regulation and role in the hepatic response to oxidative stress. Antioxid Redox Signal 2002, 4(5):749-758. ( ) 5. USFDA, Center for Food Safety and Applied Nutrition: FDA advises dietary supplement manufacturers to remove comfrey products from the market. FDA 2001 [http://vm.cfsan.fda.gov/ ~dms/dspltr06.html]. g ( ) 29. Oe H, Kaido T, Mori A, Onodera H, Imamura M: Hepatocyte growth factor as well as vascular endothelial growth factor gene induction effectively promotes liver regeneration after hepatectomy in Solt-Farber rats. Hepatogastroenterology 2005, 52(65):1393-1397. p ] 6. Weston CF, Cooper BT, Davies JD, Levine DF: Veno-occlusive dis- ease of the liver secondary to ingestion of comfrey. Br Med J (Clin Res Ed) 1987, 295(6591):183. ( ) 30. Gold EJ, Francis RJ, Zimmermann A, Mellor SL, Cranfield M, Ris- bridger GP, Groome NP, Wheatley AM, Fleming JS: Changes in activin and activin receptor subunit expression in rat liver during the development of CCl4-induced cirrhosis. Mol Cell Endocrinol 2003, 201(1–2):143-153. ( ) ( ) 7. Ridker PM, Ohkuma S, McDermott WV, Trey C, Huxtable RJ: Hepatic venocclusive disease associated with the consump- tion of pyrrolizidine-containing dietary supplements. Gastro- enterology 1985, 88(4):1050-1054. gy ( ) 8. DeLeve LD, Shulman HM, McDonald GB: Toxic injury to hepatic sinusoids: sinusoidal obstruction syndrome (veno-occlusive disease). Semin Liver Dis 2002, 22(1):27-42. 31. Yeong ML, Wakefield SJ, Ford HC: Hepatocyte membrane injury and bleb formation following low dose comfrey toxicity in rats. Int J Exp Pathol 1993, 74(2):211-217. ) ( ) 9. Hirono I, Mori H, Haga M: Carcinogenic activity of Symphytum officinale. J Natl Cancer Inst 1978, 61(3):865-869. J p ( ) 32. Bataller R, Brenner DA: Liver fibrosis. J Clin Invest 2005, 115(2):209-218. J ( ) 10. Authors' contributions NM performed the animal treatment and mutagenicity analysis, was involved in the analysis of microarray data, and wrote the manuscript. LG, LS, and JCF helped con- ceive the experiments, analysis of the data, and writing of the manuscript. LZ, YAS, and CF conducted the micro- array experiment and generated the raw data. CLM and Page 13 of 15 (page number not for citation purposes) Page 13 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 21. Yeong ML, Clark SP, Waring JM, Wilson RD, Wakefield SJ: The effects of comfrey derived pyrrolizidine alkaloids on rat liver. Pathology 1991, 23(1):35-38. SLD performed technical support. TC had the original idea for this study, and was involved in designing the experiment and writing the manuscript. All authors approved the final version of manuscript. gy ( ) 22. DeLeve LD, McCuskey RS, Wang X, Hu L, McCuskey MK, Epstein RB, Kanel GC: Characterization of a reproducible rat model of hepatic veno-occlusive disease. Hepatology 1999, 29(6):1779-1791. Acknowledgements ( ) 23. Svoboda DJ, Reddy JK: Malignant tumors in rats given lasio- carpine. Cancer Res 1972, 32(5):908-913. We thank Drs. Robert H. Heflich and Tao Han from NCTR for their helpful discussions, comments, and criticisms. 24. Hirono I, Haga M, Fujii M, Matsuura S, Matsubara N, Nakayama M, Furuya T, Hikichi M, Takanashi H, Uchida E, et al.: Induction of hepatic tumors in rats by senkirkine and symphytine. J Natl Cancer Inst 1979, 63(2):469-472. The views presented in this article do not necessarily reflect those of the Food and Drug Administration. ( ) 25. Zhou S, Koh HL, Gao Y, Gong ZY, Lee EJ: Herbal bioactivation: the good, the bad and the ugly. Life Sci 2004, 74(8):935-968. Food and Drug Administration. g g y ( ) 26. Salat C, Holler E, Kolb HJ, Reinhardt B, Pihusch R, Wilmanns W, Hiller E: Plasminogen activator inhibitor-1 confirms the diag- nosis of hepatic veno-occlusive disease in patients with hyperbilirubinemia after bone marrow transplantation. Blood 1997, 89(6):2184-2188. BMC Bioinformatics 2006, 7(Suppl 2):S16 42. Cheo DL, Burns DK, Meira LB, Houle JF, Friedberg EC: Mutational inactivation of the xeroderma pigmentosum group C gene confers predisposition to 2-acetylaminofluorene-induced liver and lung cancer and to spontaneous testicular cancer in Trp53-/-mice. Cancer Res 1999, 59(4):771-775. p ( ) 43. Nakashima Y, Ono T, Yamanoi A, El-Assal ON, Kohno H, Nagasue N: Expression of gap junction protein connexin32 in chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma. J Gas- troenterol 2004, 39(8):763-768. ( ) 44. Sheen IS, Jeng KS, Wang PC, Shih SC, Chang WH, Wang HY, Chen CC, Shyung LR: Are gap junction gene connexins 26, 32 and 43 of prognostic values in hepatocellular carcinoma? A prospec- tive study. World J Gastroenterol 2004, 10(19):2785-2790. 45. Frommel TO, Yong S, Zarling EJ: Immunohistochemical evalua- tion of Bcl-2 gene family expression in liver of hepatitis C and cirrhotic patients: a novel mechanism to explain the high incidence of hepatocarcinoma in cirrhotics. Am J Gastroenterol 1999, 94(1):178-182. ( ) 46. Duan XX, Ou JS, Li Y, Su JJ, Ou C, Yang C, Yue HF, Ban KC: Dynamic expression of apoptosis-related genes during devel- opment of laboratory hepatocellular carcinoma and its rela- tion to apoptosis. World J Gastroenterol 2005, 11(30):4740-4744. 47. Kaido T, Imamura M: Hepatocyte growth factor: clinical impli- cations in hepatobiliary pancreatic surgery. J Hepatobiliary Pan- creat Surg 2001, 8(1):65-75. 48. Drucker C, Parzefall W, Teufelhofer O, Grusch M, Ellinger A, Schulte- Hermann R, Grasl-Kraupp B: Non-parenchymal liver cells sup- port the growth advantage in the first stages of hepatocar- cinogenesis. Carcinogenesis 2006, 27(1):152-161. 49. Cangul H: Hypoxia upregulates the expression of the NDRG1 gene leading to its overexpression in various human cancers. BMC Genet 2004, 5:27. 50. He J, Zhou R: [Correlation of NDRG1 gene with liver tissue differentiation and hepatocarcinogenesis]. Beijing Da Xue Xue Bao 2003, 35(5):471-475. 51. Yasui K, Arii S, Zhao C, Imoto I, Ueda M, Nagai H, Emi M, Inazawa J: TFDP1, CUL4A, and CDC16 identified as targets for ampli- fication at 13q34 in hepatocellular carcinomas. Hepatology 2002, 35(6):1476-1484. ( ) 52. Spruck CH, Won KA, Reed SI: Deregulated cyclin E induces chromosome instability. Nature 1999, 401(6750):297-300. 53. Hollander MC, Sheikh MS, Bulavin DV, Lundgren K, Augeri-Henmuel- ler L, Shehee R, Molinaro TA, Kim KE, Tolosa E, Ashwell JD, et al.: Genomic instability in Gadd45a-deficient mice. Nat Genet 1999, 23(2):176-184. ( ) 54. References Toxicol Appl Pharmacol 2002, 184(3):153-164. 16. Couet CE, Crews C, Hanley AB: Analysis, separation, and bio- assay of pyrrolizidine alkaloids from comfrey (Symphytum officinale). Nat Toxins 1996, 4(4):163-167. ) ( ) 17. Huxtable RJ, Luthy J, Zweifel U: Toxicity of comfrey-pepsin prep- arations. N Engl J Med 1986, 315(17):1095. g J ( ) 18. Bach N, Thung SN, Schaffner F: Comfrey herb tea-induced hepatic veno-occlusive disease. Am J Med 1989, 87(1):97-99. 39. Hennings H, Glick AB, Greenhalgh DA, Morgan DL, Strickland JE, Tennenbaum T, Yuspa SH: Critical aspects of initiation, promo- tion, and progression in multistage epidermal carcinogene- sis. Proc Soc Exp Biol Med 1993, 202(1):1-8. p J ( ) 19. Ridker PM, McDermott WV: Comfrey herb tea and hepatic veno-occlusive disease. Lancet 1989, 1(8639):657-658. ( ) 20. Yeong ML, Swinburn B, Kennedy M, Nicholson G: Hepatic veno- occlusive disease associated with comfrey ingestion. J Gastro- enterol Hepatol 1990, 5(2):211-214. p ( ) 40. Loeb LA: A mutator phenotype in cancer. Cancer Res 2001, 61(8):3230-3239. 40. Loeb LA: A mutator phenotype in cancer. Cancer Res 2001, 61(8):3230-3239. ( ) 41. Friedberg EC: How nucleotide excision repair protects against cancer. Nat Rev Cancer 2001, 1(1):22-33. 41. Friedberg EC: How nucleotide excision repair protects against cancer. Nat Rev Cancer 2001, 1(1):22-33. Page 14 of 15 (page number not for citation purposes) BMC Bioinformatics 2006, 7(Suppl 2):S16 BMC Bioinformatics 2006, 7(Suppl 2):S16 Gramantieri L, Chieco P, Giovannini C, Lacchini M, Trere D, Grazi GL, Venturi A, Bolondi L: GADD45-alpha expression in cirrhosis and hepatocellular carcinoma: relationship with DNA repair and proliferation. Hum Pathol 2005, 36(11):1154-1162. 55. Asselah T, Bieche I, Laurendeau I, Paradis V, Vidaud D, Degott C, Mar- tinot M, Bedossa P, Valla D, Vidaud M, et al.: Liver gene expression signature of mild fibrosis in patients with chronic hepatitis C. Gastroenterology 2005, 129(6):2064-2075. gy ( ) 56. Cariello NF: Software for the analysis of mutations at the human hprt gene. Mutat Res 1994, 312(2):173-185. p g ( ) 57. Adams WT, Skopek TR: Statistical test for the comparison of samples from mutational spectra. J Mol Biol 1987, 194(3):391-396. ( ) 58. Fang H, Tong W, Perkins R, Shi L, Hong H, Cao X, Xie Q, Yim SH, Ward JM, Pitot HC, et al.: Bioinformatics approaches for cross- species liver cancer analysis based on microarray gene expression profiling. BMC Bioinformatics 2005, 6(Suppl 2S6 [http:/ /]. BMC Bioinformatics 2006, 7(Suppl 2):S16 Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Page 15 of 15 (page number not for citation purposes) Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge Page 15 of 15 (page number not for citation purposes)
https://openalex.org/W2007670091	https://revistas.ucr.ac.cr/index.php/rbt/article/download/5257/5057	Spanish; Castilian	null	Arquitectura y anatomía foliar del complejo Polypodium plesiosorum sensu Moran (Polypodiaceae)	Revista de Biología Tropical	2,009	cc-by	13,482	Arquitectura y anatomía foliar del complejo Polypodium plesiosorum sensu Moran (Polypodiaceae) J. Daniel Tejero-Díez1, Silvia Aguilar-Rodríguez1, Teresa Terrazas2 & Leticia Pacheco3 1. Facultad de Estudios Superiores Iztacala, Universidad Nacional Autónoma de México, Unidad de Morfología y Función, Av. de los Barrios 1, Los Reyes Iztacala, Tlalnepantla 54090, Edo. México, México, Apdo. Postal 314; tejero@unam.mx; siagro@campus.iztacala.unam.mx 1. Facultad de Estudios Superiores Iztacala, Universidad Nacional Autónoma de México, Unidad de Morfología y Función, Av. de los Barrios 1, Los Reyes Iztacala, Tlalnepantla 54090, Edo. México, México, Apdo. Postal 314; tejero@unam.mx; siagro@campus.iztacala.unam.mx j @ g @ p 2. Instituto de Biología, Universidad Nacional Autónoma de México, Av. Universidad, Coyoacán 04510, México D.F., Apdo. Postal 70-233; tterrazas@ibiologia.unam.mx j @ g @ p 2. Instituto de Biología, Universidad Nacional Autónoma de México, Av. Universidad, Coyoacán 04510, México D.F., Apdo. Postal 70-233; tterrazas@ibiologia.unam.mx 3. Depto. Biología, Universidad Autónoma Metropolitana-Iztapalapa, Botánica Estructural y Sistemática Vegetal, Av. San Rafael Atlixco 186, Vicentina, Iztapalapa 09340, México D. F., Apdo. Postal 55-535; pacheco@xanum.uam.mx 3. Depto. Biología, Universidad Autónoma Metropolitana-Iztapalapa, Botánica Estructural y Sistemática Vegetal, Av. San Rafael Atlixco 186, Vicentina, Iztapalapa 09340, México D. F., Apdo. Postal 55-535; pacheco@xanum.uam.mx 3. Depto. Biología, Universidad Autónoma Metropolitana-Iztapalapa, Botánica Estructural y Sistemática Vegetal, Av. San Rafael Atlixco 186, Vicentina, Iztapalapa 09340, México D. F., Apdo. Postal 55-535; pacheco@xanum.uam.mx Recibido 23-viii-2009. Corregido 07-i-2010. Aceptado 08-ii-2010. Abstract: Architecture and leaf anatomy of the Polypodium plesiosorum sensu Moran complex (Polypodiaceae). The Polypodium plesisorum complex is a heterogeneous morphology group. We studied leaf architecture and anatomy of 12 species of this group, in addition to other 23 species of Polypodium (P. dulce, P. polypodioides and P. vulgare complex sensu Moran) and related genera (Goniophlebium, Phlebodium, Pleopeltis, Serpocaulon, Synammia), for comparative purposes. Two homogeneous groups of species were established in P. plesiosorum complex. One of them is recognized based on the type of anastomosis of the veins (Type III): Polypodium conterminans (originally considered in the group of P. dulce with free venation by Moran), P. hispidulum, P. plesiosorum and P. rhodopleuron. The second group with Polypodium arcanum, P. castaneum, P. colpodes, P. eatonii, and P. flagellare were recognized based on the type of anastomosis of the veins (Type IV) and how vascular strands fuse along the petiole-leaf rachis. Characters studied helped to group the remaining species with other groups outside the complex under study. Rev. Biol. Trop. 58 (3): 955-976. Epub 2010 September 01. Key words: polypodiaceae, Polypodium plesiosorum complex, leaf architecture, leaf anatomy, Mexico. Key words: polypodiaceae, Polypodium plesiosorum complex, leaf architecture, leaf anatomy, Mexico castaneum Maxon ex Tejero, ya que el primero se encuentra fuera del área de Mesoamérica y el segundo es de reciente nombramiento. En el trabajo original (Tejero-Díez 2005), por definición, se incluyó a Synammia feuillei (Bertero) Copel. (=Polypo- dium feuillei Bertero). En este estudio se analiza la arquitectura y anatomía foliar de las especies del complejo P. plesiosorum (sensu Moran 1995) y se compara con especies de otros complejos y géneros cer- canos que crecen en México y Centroamérica, con el propósito de identificar caracteres que permitan apoyar el reconocimiento de un grupo P. plesiosorum homogéneo y reubicar taxonómi- camente al resto de las especies. Todas las especies revisadas habitan en México, excepto Polypodium kunzeanum y Synammia feuillei que presentan una distribu- ción sudamericana. Polypodium (sensu stricto) se distingue por tener tallo rizomatoso alargado y ramificado, escamas del rizoma con celdillas opacas, hojas con pecíolo surcado en la superficie adaxial y lámina pinnatisecta a 1-pinnada con los segmen- tos adnatos al raquis, esporas monoletes con la superficie típicamente verrugosa generalmente sin perisporio y número cromosómico conocido n=37. El taxón así definido considera la reciente segregación de Serpocaulon (Smith et al. 2006) e incluye 70 a 75 especies, la mayoría con dis- tribución en el continente americano y algunas en el Viejo Mundo. Key words: polypodiaceae, Polypodium plesiosorum complex, leaf architecture, leaf anatomy, Mexico 2000, Ranker et al. 2004, Schneider et al. 2004, Tejero-Díez 2005, Schneider et al. 2004, 2006, Otto 2007, Schuettpelz & Pryer 2007) y están propiciando una clasificación más sólida (Smith et al. 2006). Presl., Goniophlebium C. Presl, Marginariopsis C. Chr., Microsorium Link, Pecluma M. G. Price, Phlebodium (R. Br.) J. Sm., Pleopeltis Humb. & Bonpl. ex Willd., Serpocaulon A. R. Sm. y Synammia C. Presl (Evans 1969, Gómez 1975, Wagner & Wagner 1975, Anthony & Schlepe 1985, Mickel & Beitel 1987, Bosman 1991, Widham 1993, Smith et al. 2006, Otto 2007, Salino et al. 2008, Tejero-Díez et al. 2009). Además, con base en caracteres morfo- lógicos, se han definido varios complejos en Polypodium (Maxon 1903, 1916 a-c, Weather- by 1922, 1939, 1947, Lloyd & Lang 1964, de la Sota 1965, 1966, Tryon & Tryon 1982, Price 1983, Hennipman et al. 1990, Hensen 1990, Windham 1993, Moran 1995, Neuroth 1996). En este siglo, caracteres moleculares o morfo- anatómicos analizados por el método cladista, están esclareciendo algunas de las relaciones de parentesco entre las diferentes jerarquías taxonómicas de Polypodiaceae (Smith et al. 2000, Ranker et al. 2004, Schneider et al. 2004, Tejero-Díez 2005, Schneider et al. 2004, 2006, Otto 2007, Schuettpelz & Pryer 2007) y están propiciando una clasificación más sólida (Smith et al. 2006). especies. Moran (1995) consideró como “grupo P. plesiosorum” a: P. colpodes Kunze (=P. fuscopetiolatum A. R. Sm.), P. eatonii Baker, P. flagellare H. Christ, P. fraternum Schltdl. & Cham., P. hispidulum Bartlett, P. kunzeanum C. Chr., P. plesiosorum Kunze, P. pleurosorum Kunze ex Mett., P. rhachipterygium Liebm. y P. rhodopleuron Kunze. Además, Tejero-Díez & Pacheco (2004) incluyeron por definición a P. arcanum Maxon y P. castaneum Maxon ex Tejero, ya que el primero se encuentra fuera del área de Mesoamérica y el segundo es de reciente nombramiento. En el trabajo original (Tejero-Díez 2005), por definición, se incluyó a Synammia feuillei (Bertero) Copel. (=Polypo- dium feuillei Bertero). especies. Moran (1995) consideró como “grupo P. plesiosorum” a: P. colpodes Kunze (=P. fuscopetiolatum A. R. Sm.), P. eatonii Baker, P. flagellare H. Christ, P. fraternum Schltdl. & Cham., P. hispidulum Bartlett, P. kunzeanum C. Chr., P. plesiosorum Kunze, P. pleurosorum Kunze ex Mett., P. rhachipterygium Liebm. y P. rhodopleuron Kunze. Además, Tejero-Díez & Pacheco (2004) incluyeron por definición a P. arcanum Maxon y P. Key words: polypodiaceae, Polypodium plesiosorum complex, leaf architecture, leaf anatomy, Mexico La arquitectura y la morfo-anatomía foliar de los helechos han sido analizadas con dife- rentes fines (Wardlaw 1944, Iwatsuki 1962, de la Sota 1963, Tryon 1964, 1970, Lucansky & White 1974, Kaur 1978, Wagner 1979, Beck et al. 1982, Salgado 1982, White 1984, Saiki et al. 1988, Steeves 1988, Stewart & Rothwell 1993). En la familia Polypodiaceae, la disposición de las venas ha apoyado el establecimiento de grupos y géneros en los sistemas de clasifica- ción (Presl 1836, Maxon 1916 b, Ching 1940, Copeland 1947); sin embargo, sólo en algunos estudios se discute la relación entre los patro- nes de venación y los aspectos taxonómicos o evolutivos (Christensen 1928, Tindale 1960, de la Sota 1963, 1966, Hetterscheid & Hennipman 1984, Mitsuta 1984, Hensen 1990, Rödl-Linder 1990). Sobre la anatomía del pecíolo-raquis, de la Sota (1966) y Rödl-Linder (1990) describen la forma en que se organizan los cordones vas- culares en los taxa que estudiaron y sólo Ching (1933) evidencia la utilidad de estos cordones como elementos de valor taxonómico en Polypo- diaceae de China para discernir grupos. Por otra parte, Sen & Hennipman (1981) abordaron los tipos de aparatos estomáticos en 32 especies de Polypodium (s.l.). Sin embargo, son escasos los trabajos referentes a la anatomía de la lámina. La ubicación taxonómica de Polypodium L. es problemática (Hennipman 1986, Wagner 1986, Hennipman et al. 1990), a partir de él se han segregado otros como Campyloneurum C. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 955 Presl., Goniophlebium C. Presl, Marginariopsis C. Chr., Microsorium Link, Pecluma M. G. Price, Phlebodium (R. Br.) J. Sm., Pleopeltis Humb. & Bonpl. ex Willd., Serpocaulon A. R. Sm. y Synammia C. Presl (Evans 1969, Gómez 1975, Wagner & Wagner 1975, Anthony & Schlepe 1985, Mickel & Beitel 1987, Bosman 1991, Widham 1993, Smith et al. 2006, Otto 2007, Salino et al. 2008, Tejero-Díez et al. 2009). Además, con base en caracteres morfo- lógicos, se han definido varios complejos en Polypodium (Maxon 1903, 1916 a-c, Weather- by 1922, 1939, 1947, Lloyd & Lang 1964, de la Sota 1965, 1966, Tryon & Tryon 1982, Price 1983, Hennipman et al. 1990, Hensen 1990, Windham 1993, Moran 1995, Neuroth 1996). En este siglo, caracteres moleculares o morfo- anatómicos analizados por el método cladista, están esclareciendo algunas de las relaciones de parentesco entre las diferentes jerarquías taxonómicas de Polypodiaceae (Smith et al. MATERIALES Y MÉTODOS Se estudiaron las 12 especies reconocidas para el complejo de Polypodium plesiosorum (sensu Moran 1995) y otras 23 de Polypodium y especies representantes de géneros segregados con fines comparativos (Apéndice I). Arquitectura foliar: Se consultaron ejem- plares depositados en 27 herbarios (Tejero-Díez 2005). Las hojas se estudiaron directamente de material herborizado o fijado en FAA (formol, ácido acético glacial y alcohol etílico al 70%, 5:5:90). Los patrones de venación se observa- ron directamente al microscopio estereoscópico por transiluminación. Sólo en aquellos casos, en que el grosor del mesofilo de la lámina o el indumento de la misma no permitieron ver el El conjunto de especies de Polypodium carentes de escamas laminares, incluidas por Tryon & Tryon (1982) en el grupo de Polypo- dium plesiosorum Kunze y por Moran (1995) en los grupos P. plesiosorum y P. dulce Poir., han recibido hasta el momento poca atención taxo- nómica. Smith (1981), Stolze (1981) y Mickel & Beitel (1988) indican que subsisten pro- blemas de circunscripción de subgrupos y de Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 956 patrón de disposición de las venas, se practicó la técnica de diafanizado, la cual consistió en sumergir al segmento o parte de la lámina en una solución de hidróxido de potasio al 2% con hipoclorito de sodio diluido al 5% durante dos a cuatro días. Los términos y conceptos sobre los caracteres de las hojas y las venas se tomaron de Wagner (1979) y Rödl-Linder (1990). muestras se incluyeron en un bloque de para- fina. De los bloques, se obtuvieron secciones transversales del pecíolo, raquis y segmento laminar de 20-25µm de grosor; se tiñeron con safranina-verde rápido (Johansen 1940) y se montaron en portaobjetos con resina sintética. Las laminillas se depositaron como referencia en las colecciones de los herbarios de la Facul- tad de Estudios Superiores Iztacala (IZTA) y Nacional de México (MEXU), ambos en la Uni- versidad Nacional Autónoma de México. Anatomía: Los caracteres anatómicos se estudiaron en uno a cinco individuos recolecta- dos en el campo, con ejemplares de referencia para herbario (Apéndice I). El material biológi- co recolectado se fijó en solución FAPA (formol 40%, ácido acético glacial, ácido propiónico y alcohol al 50%, 5:2.5:2.5:90) durante un mes; posteriormente se conservó en solución GAA (glicerina-alcohol-agua, 1:2:3) hasta su proce- samiento en el laboratorio. MATERIALES Y MÉTODOS Para algunas especies el material se obtuvo a partir de ejemplares de herbario y, los fragmentos de pecíolos, raquis y/o pinnas removidos se hirvieron en una solu- ción de glicerina-agua 1:1 para rehidratarlos y posteriormente se fijaron en solución FAPA. Cordones vasculares. El número, tipo y forma en que se fusionan o dividen los cordones vasculares a lo largo del eje foliar, fueron obser- vados al efectuar cortes transversales a mano alzada en la parte basal (tanto en el filopodio como en la parte inmediata posterior a la articu- lación), media y distal del pecíolo; así como en la base (entre el segundo y tercer par de segmentos), la parte media y la distal del raquis. Las seccio- nes se montaron en gelatina-safranina (Aguilar- Rodríguez 1998). Siempre que fue posible se analizaron varias muestras de una especie dada, o incluso pecíolos con diferentes grosores del mismo ejemplar. La manera en que se disponen los cordones vasculares en el eje de la hoja sigue en parte el concepto de Rödl-Linder (1990); se analizaron dos caracteres básicos: el tamaño y la forma de fusión o división de los cordones a lo largo del eje. Para ilustrar las descripciones se tomaron fotografías y se realizaron dibujos con cámara clara a partir de un microscopio óptico Nikon Labophoto-2. Aparato estomático. Se obtuvieron réplicas negativas de la superficie epidérmica abaxial, mediante la aplicación de un barniz compuesto de celulosa acética y resina (Payne 1968), las cuales se montaron en gelatina glicerinada para su observación en el microscopio óptico. Los tipos de aparatos estomáticos se describieron de acuerdo a Sen & Hennipman (1981), excepto el tipo diacítico de Fryns-Claessens & van Cotthem (1973). Con la finalidad de evaluar el predominio de ciertos tipos de aparatos en las especies, se observaron dos o tres muestras, se hicieron 50 conteos por muestra y se calculó el porcentaje. Arquitectura foliar Preparaciones histológicas. Para lograr la descripción histológica de los ejes y láminas, las muestras fijadas en GAA, se deshidrataron por medio de una serie de cambios de con- centración creciente de etanol y ter-butanol (Johansen 1940). A partir del último cambio, se agregó en el lapso de dos o tres día pequeñas cantidades de parafina hasta la proporción 1:1 parafina/ter-butanol; posteriormente se dejó evaporar el ter-butanol en una estufa y las Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 Forma y división de la lámina En el grupo de Polypodium plesiosorum se reconocieron cuatro tipos de láminas: pinnati- secta, pinnatipartida, semipinnada y 1-pinnada; la primera se presenta en siete especies, siendo basipinnada en dos de ellas (Cuadro 1). En las especies que no pertenecen al complejo se encontraron los mismos tipos, además de lámi- nas simples y pinnatífidas. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 957 Cuadro 1 Tipos de división, segmento apical y margen de la lámina en especies del grupo Polypodium plesiosorum (sensu Moran 1995). subcom=subcomforme. * Especies incluidas por definición ya que se distribuyen fuera del área de estudio de la Flora Mesoamericana o han sido descritas recientemente Cuadro 1 Cuadro 1 Tipos de división, segmento apical y margen de la lámina en especies del grupo Polypodium plesiosorum (sensu Moran 1995). subcom=subcomforme. * Especies incluidas por definición ya que se distribuyen fuera del área de estudio de la Flora Mesoamericana o han sido descritas recientemente Table 1 Blade characteristics in species of Polypodium plesiosorum group (sensu Moran 1995). Subcom=subconform. * Species included by definition since they are distributed out of the area of study of the Mesoamerican flora or have been described recently Polypodium /Lámina División Segmento apical Margen P. arcanum Pinnatisecta graduado/subcom. crenado P. castaneum Pinnatisecta/basipinnada graduado/subcom. crenado P. colpodes Pinnatisecta/basipinnada graduado/subcom. crenado P. eatonii pinnatisecta y/o pinnatipartida graduado/subcom. entero P. flagellare Semipinnada graduado/subcom. crenado P. fraternum Pinnada conforme crenado P. hispidulum Pinnatisecta graduado/subcom. crenado P. kunzeanum Pinnada (pinnas articuladas al raquis) conforme crenado P. plesiosorum Pinnatisecta graduado/subcom. crenado P. pleurosorum Pinnada conforme crenado P. rhachipterygium Pinnatipartida graduado/subcom. entero P. rhodopleuron Pinnatisecta graduado/subcom. crenado Entre las especies del grupo estudiado los segmentos basales son tan largos o más que los segmentos medios, pero presentan una unión al eje muy disímil: Polypodium rhachiptery- gium es largamente decurrente, P. eatonii y P. kunzeanum, pueden o no tener lóbulos, en P. fraternum son estrechamente sésiles y en el resto de las especies varían entre escavados a anchamente sésiles y levemente decurrentes en su base. La mayor parte de las especies presen- tan los segmentos basales transversales al eje, sin embargo, en P. hispidulum, P. plesiosorum y P. rhodopleuron tienden a estar reflejos. proporciones, donde aparentemente cada espe- cie tiene su propia particularidad. El margen de la lámina es generalmente crenado en el grupo estudiado; con excepción de P. eatonii y P. Forma y división de la lámina rhachipterygium en donde es entero (Cuadro 1). Disposición de las venas en la lámina El sistema de venas secundarias (vs) de cada segmento es catádromo. En las venas terciarias, la basal siempre es fértil (vf) y en ocasiones también las terciarias subsecuentes. En las especies estudiadas los patrones de fusión, división o reducción de las venas terciarias o de orden superior fértiles o estériles (ves) se caracte- rizan por presentar una variación gradual que va desde un sistema de venas abiertas o libres (Figs. 1A, B) hasta anastomosadas o cerradas (Figs. 1C-H). Las venas anastomosadas se caracterizan por formar aréolas que presentan varios niveles de complejidad al conformar distintos tipos de polígonos angulares (también llamados gonio- fleboides), así como por la presencia o no de una De acuerdo a la disposición de los seg- mentos de la lámina en su porción distal, se distinguen tres tipos de ápices en todas las especies estudiadas: graduado, conforme y subconforme. Nueve especies del grupo bajo estudio presentan ápice graduado o subconfor- me, pero las especies con lámina pinnada sólo tienen ápice conforme (Cuadro 1). Por lo anterior, el contorno de la lámina en las especie del grupo estudiado es en gene- ral deltado a ovado-lanceolado en diferentes Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 958 venación y distribución en los segmentos de la lámina foliar (Cuadro 2). C F vex ACF AFAp A B C F E D G H vex vex ves soro vf vs vp 2 1 3 4 ACF AFAp ACE AEF 10 mm ADAI C vex ACF A B ves soro vf vs vp 2 1 3 4 C B Tipo I. Venas terciarias libres (Fig. 1A). El tipo I es característico del grupo de P. dulce. En el grupo estudiado, el 1% de los ejemplares revi- sados de P. plesiosorum presentaron este tipo. En P. fraternum el 60% o más de los ejemplares presentan venas libres y el resto son del Tipo III por lo que el carácter en esta especie es dimorfo. P. rhachipterygium presenta algunas poblaciones con este tipo de venas. En todas las especies del complejo en estudio, la porción distal de cada segmento muestra venación libre. Esta propor- ción es mayor (hasta ¾ partes del segmento) en Polypodium hispidulum, P. plesiosorum y P. rhodopleuron. La proporción de venas libres en el segmento es menor (hasta ½ del segmen- to) en Polypodium arcanum, P. castaneum, P. colpodes, P. eatonii y P. Disposición de las venas en la lámina flagellare y es prácti- camente inexistente (menos de la ¼ parte del segmento) en P. pleurosorum y P. kunzeanum. D G vex E H ADAI F AFAp D E F AFAp G H AEF 10 mm AEF 10 mm ACE 10 mm Fig. 1. Patrones de venación. Los tipos indican una variación gradual desde un sistema de venas abiertas hasta las anastomosadas que forman aréolas: A-B. Venas abiertas. A. Tipo I. B. Tipo II. C-H. Venas anastomosadas. C. Tipo III. D. Tipo IV. E. Tipo V. F. Tipo VI. G. Tipo VII. H. Tipo VIII. ACE=aréola costal estéril; ACF=aréola costal fértil; ADAl=fila de aréolas discoides alternas; AFAp=aréolas fértiles apiladas; AEF=aréola elíptica fértil; ves=vena estéril; vex=vena excurrente; vf=vena fértil; vp=vena primaria; vs=vena secundaria. Fig. 1. Patrones de venación. Los tipos indican una variación gradual desde un sistema de venas abiertas hasta las anastomosadas que forman aréolas: A-B. Venas abiertas. A. Tipo I. B. Tipo II. C-H. Venas anastomosadas. C. Tipo III. D. Tipo IV. E. Tipo V. F. Tipo VI. G. Tipo VII. H. Tipo VIII. Fig. 1. Patrones de venación. Los tipos indican una variación gradual desde un sistema de venas abiertas hasta las anastomosadas que forman aréolas: A-B. Venas abiertas. A. Tipo I. B. Tipo II. C-H. Venas anastomosadas. C. Tipo III. D. Tipo IV. E. Tipo V. F. Tipo VI. G. Tipo VII. H. Tipo VIII. Tipo II. Semejante al tipo I, pero con reducción de las venas terciarias en número o longitud (Fig. 1B). Esta disposición no se obser- vó en las especies del grupo de P. plesiosorum, pero lo presentaron Pecluma dispersa (A. M. Evans) M. G. Price, P. martensii Mett. y P. fur- furaceum Schltdl. et Cham. ACE=aréola costal estéril; ACF=aréola costal fértil; ADAl=fila de aréolas discoides alternas; AFAp=aréolas fértiles apiladas; AEF=aréola elíptica fértil; ves=vena estéril; vex=vena excurrente; vf=vena fértil; vp=vena primaria; vs=vena secundaria. Fig. 1. Venation patterns. The types show a gradual variation from a system of free veins to anastomosing veins which form areoles: A-B. Open veins. A. Type I. B. Type II. C-H. Anastomosed veins. C. Type III. D. Type IV. E. Type V. F. Type VI. G. Type VII. H. Type VIII. ACE=costal sterile areole; ACF=costal fertile areole; ADAl=disc-shaped alternates areoles row; AFAp=fertile piled up areoles; AEF=elliptical fertile areole; ves=sterile vein; vex=excurrent vein; vf=fertile vein; vp=primary vein; vs=secondary vein. Tipo III. Disposición de las venas en la lámina Venación terciaria anastomosada que forma una hilera a cada lado de la vena primaria (también denominada costa en este trabajo) de aréolas pentagonales que encierran a la vena fértil (ACF) y presentan una vena excurrente externa (vex) (Fig. 1C). Se observó en P. plesiosorum, P. rhodopleuron, P. hispidu- lum y P. rhachipterygium aunque en esta última especie no en todos lo ejemplares examinados. Tipo III. Venación terciaria anastomosada que forma una hilera a cada lado de la vena primaria (también denominada costa en este trabajo) de aréolas pentagonales que encierran a la vena fértil (ACF) y presentan una vena excurrente externa (vex) (Fig. 1C). Se observó en P. plesiosorum, P. rhodopleuron, P. hispidu- lum y P. rhachipterygium aunque en esta última especie no en todos lo ejemplares examinados. En P. conterminans Liebm., del grupo de P. dulce, aproximadamente el 80% de los ejem- plares presentan también este tipo. En estas especies, la aréola es alargada, de la misma forma que aquellas que se forman accidental- mente en especies de los grupos de P. dulce y P. vulgare L. La aréola pentagonal también es vena fértil en su interior. A su vez, la vena fértil es simple en los primeros niveles de complejidad, pero llega a tener varios patrones de ramificación en los niveles más altos. El análisis comparativo de las especies del grupo P. plesiosorum y de complejos o géneros relacionados estudiados, permitió establecer ocho tipos de patrones de En P. conterminans Liebm., del grupo de P. dulce, aproximadamente el 80% de los ejem- plares presentan también este tipo. En estas especies, la aréola es alargada, de la misma forma que aquellas que se forman accidental- mente en especies de los grupos de P. dulce y P. vulgare L. La aréola pentagonal también es Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. Disposición de las venas en la lámina 58 (3): 955-976, September 2010 959 Cuadro 2 Table 2 Distribution and type of veins in segments of the blade in the Polypodium groups (Moran 1995) Tipos Venas terciarias o superiores Aréola costal (fértil) Vena fértil Alimentación al soro Aréola fértil secundaria Aréolas discoidales marginales I libres desarrolladas ausente simple simple ausente ausente II libres reducidas ausente simple simple ausente ausente III anastomosadas pentagonal alargada o isodiamétrica, una hilera simple simple ausente ausente IV anastomosadas pentagonales/ hexagonales, una hilera simple simple ausente ausente o irregulares V anastomosadas poligonales (siete a doce ángulos), una hilera simple simple ausente con una o más hileras VI anastomosadas pentagonales isodiamétricas varias hileras apiladas simple simple ausente ausente o irregulares VII anastomosadas poligonales (siete a doce ángulos), una hilera dividida en ángulo subrecto dos ramas presente con una o más hileras VIII anastomosadas poligonal con más de doce ángulos, una hilera dividida primariamente en ángulo agudo por la unión de varias venas (compital) presente con una hilera Table 2 Table 2 Distribution and type of veins in segments of the blade in the Polypodium groups (Moran 1995) generalmente son estériles por carecer de veni- lla interna (Fig. 1E). Se observó en P. eatonii, P. kunzeanum y P. pleurosorum del grupo de interés, así como en Phlebodium pseudoaureum (Cav.) Lellinger, P. xhemipinnatum Tejero, Mickel & A.R. Sm. y en varias especies del grupo de Polypodium polypodioides (L.) Watt. característica de Serpocaulon A. R. Sm. aunque el pentágono formado es isodiamétrico. Tipo IV. Venación terciaria y cuaternaria anastomosada que forman una hilera a cada lado de la vena primaria de aréolas pentagonales y/o hexagonales que encierran a la vena fértil y presentan dos venas excurrentes externas (vex) (Fig. 1D). Se observó en P. arcanum, P. casta- neum, P. colpodes, P. eatonii y P. flagellare. Tipo VI. Aréolas fértiles secundarias api- ladas (AFAp), que se producen por anastomosis sucesivas de las venas terciarias, donde la vena excurrente a su vez queda encerrada y produ- ce un soro en el ápice. (Fig. 1F). Son típicas de algunas especies de Serpocaulon como S. triseriale (Sw.) A. R. Sm. Excepcionalmente se observaron aréolas apiladas infértiles en Polypodium plesiosorum y P. rhodopleuron del grupo bajo estudio. Tipo V. Venación cuaternaria o de orden superior anastomosadas que forman una hile- ra a cada lado de la vena primaria de aréolas poliangulares (de ocho o más ángulos). A su vez, las venas marginales se vuelven a anasto- mosar formando una o más hileras de aréolas discoides alternas a las costales (ADAl), que Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 960 Tipo VII. La vena interna fértil de las aréolas costales se dividen produciendo venas de cuarto o quinto orden que forman una aréola secundaria costal estéril (ACE) y una alimentación dos ramas del soro (Fig. 1G). Se observó en Phlebodium pseudoaureum y en Ph. xhemipinnatum es variable; en Polypodium pleurosorum del grupo de interés es casual en las láminas de algunos ejemplares. confluencia de varias venillas (soro compital). Las venas estériles se organizan en las distintas especies observadas del grupo de Polypodium polypodioides de acuerdo con los tipos I, III, IV y V. Este tipo no se observó en las especies del complejo de P. plesiosorum. Pecíolo-raquis Tipo VIII. La vena interna fértil de las aréolas costales se divide y fusiona nuevamen- te formando una aréola elíptica (AEF) (Fig. 1H). En Pleopeltis angusta o P. mexicana a su vez, pueden dividirse y anastomosarse nue- vamente para formar un patrón complejo de areolillas internas. El soro se encuentra en la En las especies del grupo de Polypodium plesiosorum estudiadas, la base del pecíolo esta articulado por medio de una banda de abscisión (Figs. 2A, B) que permite a la hoja desprenderse al iniciar la temporada de sequía, mientras que un remanente peciolar, el filopodio, permanece unido al rizoma. En sección transversal (Fig. Fig. 2. Filopodio-pecíolo y raquis. A, B. morfología y anatomía del filopodio, vistas superficial y longitudinal respectivamente. C. Corte transversal en la base del pecíolo en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (130 X y recuadro a 400 X). D. Corte transversal en la base del eje raquis en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (200 X). CVG=cordones vasculares grandes; CVP=cordones vasculares pequeños; e=endodermis; hd=hipodermis; mx=metaxilema; p=periciclo; px=protoxilema; tf=trazas foliares; vc=vaina circumendodermal. Flechas=zona de abscisión. Fig. 2. Phyllopodium-petiole and rachis. A, B. Morphology and anatomy of phyllopodium. C. Transverse section of petiole base in Polypodium colpodes (D. Tejero Díez 4270 IZTA) (130 X and square 400 X). D. Transverse section in the base of the rachis axis in Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (200 X). CVG=large vascular strands; CVP=small vascular strands; hd=hypodermis; vc=circumendodermal sheath; e=endodermis; p=pericycle; mx=metaxylem; px=protoxylem; tf=leaf trace. Arrows=abscission zone. A B D C CVP CVG mx mx px px tf p p e hd hd vc vc 500 µm 500 µm 500 µm 1 cm A 1 cm A C CVP CVG mx px p e hd vc vc 500 µm C D D mx px tf p hd 500 µm 500 µm 500 µm Fig. 2. Filopodio-pecíolo y raquis. A, B. morfología y anatomía del filopodio, vistas superficial y longitudinal respectivamente. C. Corte transversal en la base del pecíolo en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (130 X y recuadro a 400 X). D. Corte transversal en la base del eje raquis en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (200 X). CVG=cordones vasculares grandes; CVP=cordones vasculares pequeños; e=endodermis; hd=hipodermis; mx=metaxilema; p=periciclo; px=protoxilema; tf=trazas foliares; vc=vaina circumendodermal. Flechas=zona de abscisión. Fig. 2. Filopodio-pecíolo y raquis. A, B. morfología y anatomía del filopodio, vistas superficial y longitudinal respectivamente. Pecíolo-raquis Una disposición similar de los tejidos se observó en las especies de los otros grupos y géneros estudiados. Los cordones vasculares presentan un pro- ceso de reducción en número y diferenciación otros taxa estudiados, se distinguen en la base del pecíolo dos CVG, excepto en Pleopeltis fallax (Schltdl. & Cham.) Mickel & Beitel, que presenta uno. Los CVP están ausentes en algu- nas especies del grupo Polypodium vulgare, P. munchii Christ y Pleopeltis fallax. A partir de la base del pecíolo y en dirección acrópeta existe una tendencia a la fusión de los CVP entre sí, de los CVP con los CVG y finalmente de los dos CVG, hasta formar un cordón vascular de forma más o menos triangular en la parte distal del raquis (Fig. 3) a excepción del tipo XI donde se conservan los dos CVG y uno de los CVP. A su vez, este cordón vascular cambia gradual- mente la configuración del tejido xilemático en dirección acrópeta; al formarse presenta forma de “X” luego pasa sucesivamente a formas tipo “mariposa”, “T” o “yunque” y, finalmente al tipo “hongo” (Fig. 2D). La vena primaria de los segmentos de la lámina tiene un cordón vascular como el descrito en la última fase de las fusiones del eje pecíolo-raquis y se encuentra protegida por capas de células esclerosadas en la parte superior e inferior (Figs. 5A, B). 2C), el pecíolo es redondeado, con tendencia a aplanarse o acanalarse una o dos veces en la superficie adaxial. En la superficie del canal se observan tricomas simples, cateniformes o capitados, sobre todo en dirección acrópeta, en P. colpodes, P. hispidulum y P. plesiosorum. Se reconocen 11 tipos de configuracio- nes de los cordones vasculares con base en la fusión/división de los mismos (Fig. 3). Tipo I, Pleopeltis fallax. Tipo II, Polypodium munchii, también presente en las especies europeas del grupo Polypodium vulgare (P. vulgare y P. someyae Yatabe) según Rödl-Linder (1990). Tipo IIIa, Polypodium californicum Kaulf. y Synammia feuillei. Tipo IIIb, Polypodium eatonii y P. colpodes. Tipo IIIc, Polypodium polypodioides y Pleopeltis angusta Humb. & Bonpl. ex Willd. Tipo IV, típico de la mayor parte de las especies de los grupos de Polypo- dium dulce y P. plesiosorum (Polypodium arcanum, P. conterminans, P. flagellare, P. hispidulum, P. martensii, P. plesiosorum y P. rhodopleuron). Pecíolo-raquis 2C) y el periciclo (p) de una capa celular excepto en la región externa del meristele. En la porción central del cordón vas- cular se dispone el tejido xilemático que tiene forma de arco (“C”) con el protoxilema exarco (Fig. 2C). Entre el periciclo y los elementos del xilema se encuentra el tejido floemático, dis- puesto en dos bandas alargadas a cada lado del xilema. Una disposición similar de los tejidos se observó en las especies de los otros grupos y géneros estudiados. L d l 2C), el pecíolo es redondeado, con tendencia a aplanarse o acanalarse una o dos veces en la superficie adaxial. En la superficie del canal se observan tricomas simples, cateniformes o capitados, sobre todo en dirección acrópeta, en P. colpodes, P. hispidulum y P. plesiosorum. La epidermis es simple y presenta una cutícula gruesa. Debajo de la epidermis se diferencia una hipodermis (hd) constituida por tres a seis hile- ras de fibras con engrosamiento de la pared de mediano a muy notable, que en ocasiones llega a obliterar el lumen. Esta banda de esclerénquima forma un cilindro continuo a lo largo del pecío- lo excepto en la parte distal y en el raquis, que suele interrumpirse a la altura de las alas foliares (Fig. 2D). Bajo la hipodermis, existen de cinco a ocho hileras de células parenquimáticas con forma isodiamétrica que ocasionalmente contie- nen granos de almidón. En Polypodium kunze- anum se observan cristales prismáticos y fibras dispersas en el parénquima. En el tejido paren- quimático se encuentran embebidos (uno) dos cordones vasculares grandes (CVG) en posición adaxial y uno a 13 pequeños (CVP), formando un arco en la posición abaxial; cada uno de los cordones está rodeado por una hilera de células, cuya pared tangencial interna presenta un engro- samiento sobresaliente de celulosa con taninos, denominada vaina circumendodermal (vc) (V. Hernández, T. Terrazas, K. Mehltreter & G. Ángeles, en prep.). En el interior de cada vaina se desarrolla la endodermis (e) formada por una capa de células (Fig. 2C) y el periciclo (p) de una capa celular excepto en la región externa del meristele. En la porción central del cordón vas- cular se dispone el tejido xilemático que tiene forma de arco (“C”) con el protoxilema exarco (Fig. 2C). Entre el periciclo y los elementos del xilema se encuentra el tejido floemático, dis- puesto en dos bandas alargadas a cada lado del xilema. Pecíolo-raquis C. Corte transversal en la base del pecíolo en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (130 X y recuadro a 400 X). D. Corte transversal en la base del eje raquis en Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (200 X). CVG=cordones vasculares grandes; CVP=cordones vasculares pequeños; e=endodermis; hd=hipodermis; mx=metaxilema; p=periciclo; px=protoxilema; tf=trazas foliares; vc=vaina circumendodermal. Flechas=zona de abscisión. Fig. 2. Phyllopodium-petiole and rachis. A, B. Morphology and anatomy of phyllopodium. C. Transverse section of petiole base in Polypodium colpodes (D. Tejero Díez 4270 IZTA) (130 X and square 400 X). D. Transverse section in the base of the rachis axis in Polypodium colpodes. (D. Tejero Díez 4270 IZTA) (200 X). CVG=large vascular strands; CVP=small vascular strands; hd=hypodermis; vc=circumendodermal sheath; e=endodermis; p=pericycle; mx=metaxylem; px=protoxylem; tf=leaf trace. Arrows=abscission zone. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 961 2C), el pecíolo es redondeado, con tendencia a aplanarse o acanalarse una o dos veces en la superficie adaxial. En la superficie del canal se observan tricomas simples, cateniformes o capitados, sobre todo en dirección acrópeta, en P. colpodes, P. hispidulum y P. plesiosorum. La epidermis es simple y presenta una cutícula gruesa. Debajo de la epidermis se diferencia una hipodermis (hd) constituida por tres a seis hile- ras de fibras con engrosamiento de la pared de mediano a muy notable, que en ocasiones llega a obliterar el lumen. Esta banda de esclerénquima forma un cilindro continuo a lo largo del pecío- lo excepto en la parte distal y en el raquis, que suele interrumpirse a la altura de las alas foliares (Fig. 2D). Bajo la hipodermis, existen de cinco a ocho hileras de células parenquimáticas con forma isodiamétrica que ocasionalmente contie- nen granos de almidón. En Polypodium kunze- anum se observan cristales prismáticos y fibras dispersas en el parénquima. En el tejido paren- quimático se encuentran embebidos (uno) dos cordones vasculares grandes (CVG) en posición adaxial y uno a 13 pequeños (CVP), formando un arco en la posición abaxial; cada uno de los cordones está rodeado por una hilera de células, cuya pared tangencial interna presenta un engro- samiento sobresaliente de celulosa con taninos, denominada vaina circumendodermal (vc) (V. Hernández, T. Terrazas, K. Mehltreter & G. Ángeles, en prep.). En el interior de cada vaina se desarrolla la endodermis (e) formada por una capa de células (Fig. Pecíolo-raquis Tipo V, especies de del grupo de Polypodium polypodioides (hoja escamosa) con dos variantes: Tipo Va, con fusión de los CVG en la parte distal del pecíolo (Polypodium alansmithii R.C. Moran, P. fraternum, P. gutta- tum Maxon, P. lepidotrichum (Fée) Maxon, P. pyrrholepis (Fée) Maxon, P. plebeium Schltdl. & Cham. y P. remotum Desv.) y Tipo Vb, sin Los cordones vasculares presentan un pro- ceso de reducción en número y diferenciación en tamaño al pasar del rizoma-filopodio al eje pecíolo-raquis (Fig. 3). En las especies del grupo de Polypodium plesiosorum y en los Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 962 PECÍOLO COSTA / RAQUIS parte basal parte media parte alta parte basal parte media parte alta TIPOS I II IIIa IIIb IIIc IV Va Vb VI VII VIII IX X XI Fig. 3. Patrones de fusión/división de cordones vasculares a lo largo del pecíolo-costa/raquis. Cordones vasculares grandes (CVG)=círculos grandes, cordones vasculares pequeños (CVP)= círculos pequeños y trazas foliares=(•). Fig. 3. Fusion/fission patterns in vascular strands along petiole-rachis. Large vascular stands (CVG) are represented with large circles, small vascular stands (CVP) with small circles and leaf trace with (•). PECÍOLO COSTA / RAQUIS parte basal parte media parte alta parte basal parte media parte alta TIPOS I II IIIa IIIb IIIc IV Va Vb VI VII VIII IX X XI Fig. 3. Patrones de fusión/división de cordones vasculares a lo largo del pecíolo-costa/raquis. Cordones vasculares grandes (CVG)=círculos grandes, cordones vasculares pequeños (CVP)= círculos pequeños y trazas foliares=(•). Fig. 3. Fusion/fission patterns in vascular strands along petiole-rachis. Large vascular stands (CVG) are represented with large circles, small vascular stands (CVP) with small circles and leaf trace with (•). PECÍOLO COSTA / RAQUIS parte basal parte media parte alta parte basal parte media parte alta TIPOS I II IIIa IIIb IIIc IV Va Vb VI VII VIII IX X XI e fusión/división de cordones vasculares a lo largo del pecíolo-costa/raquis. Cordone grandes, cordones vasculares pequeños (CVP)= círculos pequeños y trazas foliares=(• ones de fusión/división de cordones vasculares a lo largo del pecíolo-costa/raquis. Cordones vasculares grandes culos grandes, cordones vasculares pequeños (CVP)= círculos pequeños y trazas foliares=(•). ones de fusión/división de cordones vasculares a lo largo del pecíolo-costa/raquis. Cordones vasculares grandes culos grandes, cordones vasculares pequeños (CVP)= círculos pequeños y trazas foliares=(•). Fig. 3. Fusion/fission patterns in vascular strands along petiole-rachis. Pecíolo-raquis Large vascular stands (CVG) are represented with large circles, small vascular stands (CVP) with small circles and leaf trace with (•). Fig. 3. Fusion/fission patterns in vascular strands along petiole-rachis. Large vascular stands (CVG) are represented with large circles, small vascular stands (CVP) with small circles and leaf trace with (•). Fig. 3. Fusion/fission patterns in vascular strands along petiole-rachis. Large vascular stands (CVG) are represented with large circles, small vascular stands (CVP) with small circles and leaf trace with (•). fusión de los CVG y desaparición de los CVP en la parte distal del pecíolo (Polypodium furfuraceum y Pleopeltis mexicana (Fée) Mic- kel & Beitel). Tipo intermedio entre los tipos IV y V, Polypodium rhachipterygium. Tipo VI, Pecluma dispersa. Tipo VII, Polypodium echinolepis Fée y P. subpetiolatum Hook. Tipo VIII, Phlebodium pseudoaureum y Polypodium pleurosorum. Tipo IX, Serpocaulon triseriale, Polypodium kunzeanum. Tipo X, definido por Rödl-Linder (1990) quien lo observó en Gonio- phlebium percussum (Cav.) Wagner & Grether y G. demersum (Brause) Rodl-Linder. Tipo XI, Serpocaulon falcaria (Kunze) A. R. Sm. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 963 Lámina formado por células parenquimáticas dispuestas en forma de columna, “Y” o “X” (Fig. 5A), con cámaras subestomáticas rodeadas por células parenquimáticas globosas en la parte abaxial; lo presentaron Polypodium arcanum, P. californi- cum, P. colpodes, P. hispidulum, P. plesiosorum, P. rhachipterygium, P. rhodopleuron, Phlebo- dium pseudoaureum y Serpocaulon falcaria. En el mesofilo bifacial (dorsiventral) (Fig. 5B), se observan dos a cuatro capas de células isodiamétricas a cuadrangulares de parénqui- ma denso (en empalizada) hacia la superficie adaxial y un parénquima laxo (esponjoso) hacia la abaxial; lo presentaron Polypodium echinolepis, P. fraternum, P. subpetiolatum, Serpocaulon triseriale y las especies del grupo Polypodium polypodioides y Pleopeltis. En vista superficial, las células epidérmi- cas típicas de ambas superficies tienen tama- ños variables y muestran una gradación desde paredes fuertemente lobuladas (Polypodium rhachipterygium) hasta sinuosas (Polypodium guttatum y Phlebodium pseudoaureum). En la superficie abaxial se observan estomas general- mente distribuidos y orientados en dirección al eje longitudinal de la hoja. Se reconocieron cinco tipos estomáticos: anomocítico, copolocítico en dos versiones: copolocítico “a” donde la célula anexa está conectada con dos células opuestas y copolocítico “b” donde la célula anexa está conectada con una célula opuesta; diacítico y polocítico (Figs. 4A-H). DISCUSIÓN Los tipos de anastomosis de las venas secundarias y superiores y la forma en que se fusionan los cordones vasculares a lo largo del eje pecíolo-raquis de la hoja, fueron los caracteres más relevantes para reconocer dos grupos de especies homogéneas en el complejo de Polypodium plesiosorum. Otros caracteres como la división de la lámina, los tipos esto- máticos o la anatomía del mesofilo laminar ayudaron a definir similitudes con otros grupos fuera del complejo bajo estudio. La mayoría de las especies pubescentes de los grupos de Polypodium plesiosorum y P. dulce estudiados presentan tricomas catenifor- mes, hialinos, con una a cuatro células de lon- gitud (Fig. 4J), con excepción de P. echinolepis cuyos tricomas son uni a bicelulares capitados (Fig. 4I). Pecíolo-raquis En cada una de las 21 especies estudiadas para este carácter, se obser- vó al menos tres de estos tipos en porcentajes variables, con excepción de Synammia feuillei que presentó exclusivamente estomas de tipo polocítico (Cuadro 3). Los tipos anomocítico, copolocítico b y polocítico se presentan en mayor proporción en las especies revisadas (Cuadro 3). En la zona de la vena primaria se distin- guen por debajo de la epidermis varias capas de células con paredes engrosadas que pueden o no estar esclerosadas. La vena primaria presen- ta una estructura semejante a la descrita para la porción distal del raquis (Fig. 5). En la epidermis adaxial, cerca de la parte marginal de la lámina, las células epidérmicas están interrumpidas por los hidátodos. Éstos están en conexión con las últimas ramifica- ciones de las venas foliares y se aprecian, en vista superficial, como una glándula de forma ovada, compuesta de células cuadrangulares dispuestas en círculos concéntricos, unida a una hilera de células alargadas rectangulares (el pie) que corresponde a la parte terminal de la vena (Fig. 4K). Forma y división de la lámina. Entre las especies estudiadas del grupo de Polypodium plesiosorum (sensu Moran 1995), la división de la lámina resultó muy variable al encontrarse láminas pinnatisectas (7 especies), pinnatipartidas (1) y pinnadas (3). Los estados intermedios basipinnados que se presentan en En vista transversal, las epidermis adaxial y abaxial son de un estrato celular, presentan una cutícula muy delgada y células con forma de domos cortos (Fig. 5A). En las especies estudiadas se observaron dos tipos de mesofilo. El unifacial, que está Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 964 1 1 1 1 1 1 1 1 1 1 1 1’ 1’ 1’ 1” 2 2 2 2 2 2 2 2 2 3 3 3 3 4 A B C F E D G H J K I 50 µM 100 µM E 50 µM Fig. 4. Estructura epidérmica. Tipos estomáticos A-H. A=anomocítico. B, F=polocítico. C=copolocítico. D, E=diacítico. G, H=copolocíticos [G es la variante donde existen una o dos células anexas opuestas y H es la variante donde existen hasta tres célula anexas opuestas (1’). Tricomas I-J. I=tricoma glandular en P. echinolepis. J=tricoma cateniforme común en las especies estudiadas. Hidatodo K. p Los números corresponden a la posición de las células subsidiarias que rodean a las de guarda. Fig. 4. Epidermal structure. Stomatal types A-H. A=anomocytic. B, F=polocytic. C=copolocytic. D, E=diacytic. G, H=copolocytics [G is a variant where two subsidiary cells are opposited y H is a variant where three subsidiary cells are opposited. Trichomes I-J. I=gland trichome in P. echinolepis. J=catenate trichome, common in the studied species. Hydathode K. Numbers correspond to subsidiary cells position that surround guard ones. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 965 Cuadro 3 Taxones Anomocítico Polocítico Diacítico Copolocítico A Copolocítico B Grupo P. plesiosorum P. arcanum 16 44 2 0 38 P. colpodes 10 48 4 4 34 P. eatonii 8 40 10 4 38 P. flagellare 2 56 14 0 28 P. fraternum 12 54 0 0 34 P. hispidulum 26 42 0 0 26 P. kunzeanum 6 28 18 4 44 P. plesiosorum 20 52 0 0 28 P. pleurosorum 2 26 2 20 50 P. rhachipterygium 28 48 2 0 22 P. rhodopleuron 4 44 4 0 48 Grupo P. dulce P. echinolepis 30 58 0 0 12 P. Forma y división de la lámina. martensii 16 22 8 12 42 Grupo P. polypodioides P. alan-smithii 44 34 6 0 16 P. guttatum 50 38 0 0 12 P. plebeium 34 46 4 0 16 Grupo P. vulgare P. californicum 12 38 0 18 32 PHLEBODIUM P. pseudoaureum 28 32 18 2 10 SERPOCAULON S. falcaria 4 36 2 8 50 S. triseriale 2 34 0 30 34 SYNAMMIA S. feuillei 0 100 0 0 0 P. castaneum, P. colpodes y semipinnada en P. flagellare son aparentemente casos derivados de la condición pinnatisecta, dado que sus seg- mentos basales son adnatos, sursumcurrentes y la porción apical de la lámina es subconforme. En Polypodium eatonii la condición dual del grado de división de la lámina aparentemente es un caso de estrés hídrico (Tejero & Pacheco 2004). Por otra parte, Polypodium fraternum, P. kunzeanum y P. pleurosorum, son disímiles al resto de las especies del grupo dado que la lámina es pinnada, las pinnas proximales tienen la base cuneada o redondeado-sésil y el seg- mento apical es conforme o levemente subcon- forme. En términos generales el contorno de la lámina en las especies del grupo es deltado a ovado lanceolado en diferentes proporciones, dado que no se reducen demasiado los segmen- tos basales como en Pecluma (ej. P. dispersa). Si acaso, en algunos ejemplares de Polypodium conterminans, P. hispidulm, P. plesiosorum y P. rhodopleuron, la base de la lámina se ve afectada porque los segmentos tienden a estar reflejos. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 966 A B ep ep pes cs es vs x pem pes cs vc 1 mm 1 mm en+pe A 1 mm Fig. 5. Corte transversal de lámina en Polypodium. A. Mesofilo unifacial en P. rhodopleuron (Tejero-Díez 4193 IZTA). B. Mesofilo bifacial en P. fraternum (D. Tejero-Díez 4245 IZTA). cs=cámara subestomática; ep=epidermis; en=endodermis; es=esclerénquima; pe=periciclo; pem= parénquima en empalizada; pes=parénquima esponjoso; vc=vaina circumendodermal; vs=vena secundaria; x=xilema. Fig. 5. Corte transversal de lámina en Polypodium. A. Mesofilo unifacial en P. rhodopleuron (Tejero-Díez 4193 IZTA). B. Mesofilo bifacial en P. fraternum (D. Tejero-Díez 4245 IZTA). cs=cámara subestomática; ep=epidermis; en=endodermis; es=esclerénquima; pe=periciclo; pem= parénquima en empalizada; pes=parénquima esponjoso; vc=vaina circumendodermal; vs=vena secundaria; x=xilema. Fig. 5. Corte transversal de lámina en Polypodium. A. Mesofilo unifacial en P. rhodopleuron (Tejero-Díez 4193 IZTA). B. Mesofilo bifacial en P. fraternum (D. Tejero-Díez 4245 IZTA). Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 Forma y división de la lámina. cs=cámara subestomática; ep=epidermis; en=endodermis; es=esclerénquima; pe=periciclo; pem= parénquima en empalizada; pes=parénquima esponjoso; vc=vaina circumendodermal; vs=vena secundaria; x=xilema. f j cs=cámara subestomática; ep=epidermis; en=endodermis; es=esclerénquima; pe=periciclo; pem= parén empalizada; pes=parénquima esponjoso; vc=vaina circumendodermal; vs=vena secundaria; x=xilema. Fig. 5. Transverse section of blade in Polypodium. A. Unifacial mesophyll in P. rhodopleuron (Tejero-Díez 4193 IZTA). B. Bifacial mesophyll in P. fraternum (D. Tejero-Díez 4245 IZTA). cs=substomatal chamber; ep=epidermis en=endodermis; es=sclerenchyma; pe=pericycle; pem=palisade parenchyma; Fig. 5. Transverse section of blade in Polypodium. A. Unifacial mesophyll in P. rhodopleuron (Tejero-Díez 4193 IZTA). B. Bifacial mesophyll in P. fraternum (D. Tejero-Díez 4245 IZTA). Bifacial mesophyll in P. fraternum (D. Tejero-Díez 4245 IZTA). cs=substomatal chamber; ep=epidermis en=endodermis; es=sclerenchyma; pe=pericycle; pem=palisade parenchyma; pes=spongy parenchyma; vc=circumendodermal sheath; vs=secondary vein; x=xylem. cs=substomatal chamber; ep=epidermis en=endodermis; es=sclerenchyma; pe=pericycle; pem=palisade pa pes=spongy parenchyma; vc=circumendodermal sheath; vs=secondary vein; x=xylem. Polypodium rhachipterygium es la única especie del complejo P. plesiosorum con hoja pinnatipartida que además presenta un ala lar- gamente decurrente sobre el pecíolo derivado del par de segmentos proximales; este último carácter lo comparte con varias especies del grupo de P. polypodioides, con Pleopeltis y con unas cuantas especies del grupo de Polypodium dulce (sensu Moran 1995). de tener características celulares como la pre- sencia de cristales prismáticos y fibras dispersas en el parénquima del peciolo-raquis, es la única especie con las pinnas articuladas al raquis, carácter que comparte con el género asiático Goniophlebium (Rödl-Linder 1990). El margen crenulado a crenado es la con- dición generalizada en las especies estudiadas. Este carácter se puede presentar indistinta- mente en todo el segmento o sólo en la mitad distal. En ocasiones es poco evidente o bien puede estar muy pronunciado e incluso formar Polypodium kunzeanum es un caso espe- cial dentro del complejo P. plesiosorum y entre los polipodiaceos de América, ya que además Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 967 lóbulos en el margen acroscópico de los seg- mentos proximales, como en ciertos individuos de Polypodium plesiosorum var. plesiosorum. La condición de margen crenado aparentemen- te es común en los polipodios americanos, ya que también aparece en algunas especies del grupo P. vulgare, en la mayoría de las especies del grupo P. polypodioides y en varias especies del grupo P. dulce y de Serpocaulon. Por el contrario, el margen entero a repando entre los polipodiáceos es raro, observándose sólo en Polypodium rhachipterygium y P. Forma y división de la lámina. eatonii del complejo bajo estudio. Este tipo de margen es propio de los géneros Phlebodium y Pecluma, también se presenta en algunas especies de Pleopeltis, del grupo de Polypodium dulce y Serpocaulon. poblaciones con venas libres (Tejero-Díez 4359 IZTA) o anastomosadas (Tejero-Díez 4245 IZTA) y, en menor proporción, también P. rhachipterygium. Con respecto al tipo de anastomosis de las venas terciarias o de orden superior, las uniones más sencillas aparentemente están representadas por el tipo III. Polypodium conterminans según Moran (1995) y Mickel & Smith (2004) pertene- ce al grupo de P. dulce; sin embargo, el ejemplar tipo de P. conterminans y 80% de los ejempla- res estudiados presentan aréolas pentagonales semejantes a P. plesiosorum. Considerando este carácter, junto con la similitud de las escamas del rizoma (Tejero-Díez & Pacheco en prep.), se propone que P. conterminans forme parte del grupo de P. plesiosorum (sensu stricto). Uniones de venas más complejas están representadas por el tipo IV registradas en Polypodium arcanum, P. castaneum, P. col- podes, P. eatonii y P. flagellare. Este carácter apoya la propuesta de Tejero-Díez & Pacheco (2004) de definir a estas especies como un grupo homogéneo. Anatomía Sen & Hennipman (1981) reconocieron en la familia Polypodiaceae un total de nueve tipos estomáticos y más de tres tipos se pueden encontrar con diferente predominio en una misma especie. A partir del análisis llevado a cabo, se observó que Synammia feuillei fue la única especie que presentó un único tipo esto- mático, el polocítico; en el resto de las especies existe combinación de tipos estomáticos seme- jante a lo citado por Sen & Hennipman (1981) y ninguno de ellos permite caracterizar algún grupo taxonómico de los aquí considerados. De acuerdo con Crabbe et al. (1975) y Pichi- Sermolli (1977), en Polypodium la presencia del tipo polocítico aunado a los tipos copolocí- tico y anomocítico es una condición primitiva, característica de Polypodiaceae (incluyendo a los helechos grammitidáceos, Smith et al. 2006). En este sentido, llama la atención la pre- sencia de un sólo tipo estomático en Synammia feuillei; especie de distribución austral que según Schneider et al. (2006) constituye una línea filogenética independiente de Polypodia- ceae en Sudamérica, lejanamente relacionada con el grupo Serpocaulon tal como había suge- rido Tryon & Tryon (1982). Disposición de las venas Wagner (1952) y de la Sota (1973) men- cionaron que las formas primitivas en los polipodiáceos poseían venas reticuladas y, por tanto, las especies con venas libres, general- mente distribuidas en las zona templadas de ambos hemisferios, pueden considerarse como derivadas. Sin embargo, en el Nuevo Mundo, donde los polipodiáceos han tenido patrones secundarios de evolución (Hennipman et al. 1990, Lellinger 1993), ambos tipos de venas pueden apreciarse; aparentemente en los gru- pos que se originan en el neotrópico del hemis- ferio norte surgen venas anastomosadas a partir del complejo de Polypodium vulgare y P. dulce de venación abierta (Haufler & Ranker 1995, Haufler et al. 1995), mientras que los taxa del hemisferio sur las presentan exclusivamente anastomosadas tal como se observa en los géneros Serpocaulon y Synammia. El tipo V (de aréolas poliangulares), es un carácter que exhibieron Polypodium pleuroso- rum y P. kunzeanum. La venación parecida de P. pleurosorum con Phlebodium y, la tendencia a dividir la vena fértil como en el grupo VI, han sido argumentos para apoyar la reclasifi- cación de P. pleurosorum como un miembro de Phlebodium: Ph. inaequale T. Moore (Tejero- Díez et al. 2009). Por otra parte, Polypodium kunzeanum, debido a las particularidades his- tológicas y anatómicas anotadas, tendrá que reconsiderarse taxonómicamente en el futuro. Los tipos I a V de organización en las venas estériles se encuentran en las distintas especies del grupo de Polypodium polypodioi- des y Pleopeltis dando la impresión de un con- tinuum evolutivo, carácter que se complementa con la forma peculiar de división de la vena fértil tipo VII. Esta última forma de ramifica- ción de la vena fértil no se observó en ninguna especie del grupo P. plesiosorum. De la Sota (1966) hace un análisis más detallado de los tipos V y VII de aréolas fértiles en el grupo Polypodium squamatum L. e indica su parecido En los complejos de Polypodium dulce y P. vulgare, definidos en parte por su venación abierta, P. californicum presenta ocasional- mente aréolas pentagonales; en P. echinolepis existen poblaciones con venas anastomo- sadas y, en P. conterminans, generalmen- te son anastomosadas con proporción de venas libres semejante a P. plesiosorum. En el grupo bajo estudio, P. fraternum presenta Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 968 Tryon 1982); grupos que Windham (1993) y Otto (2007) consideran como Pleopeltis. Disposición de las venas Por otra parte el tipo VI observado en Pecluma, está relacionado con el peciolo terete, carácter que es distintivo de este género (Price 1983). con las del género Pleopeltis. Este estudio y otros similares (Windham 1993, Hooper 1994), junto con los de índole molecular (Schneider et al. 2004, Otto 2007) apoyan la idea de reubicar a las especies del grupo de Polypodium polypo- dioides dentro de Pleopeltis. AGRADECIMIENTOS Finalmente, la presencia de aréolas pentago- nales y la forma y tipo de división de la lámina, permitieron agrupar a Polypodium conterminans (originalmente considerado en el grupo de P. dulce de venación libre sensu Moran 1995) con P. hispidulum, P. plesiosorum y P. rhodopleu- ron. Estas cuatro especies comparten caracteres similares, además de las mencionadas, en las escamas del rizoma y por ello pueden constituir el grupo P. plesiosorum típico. El presente estudio se realizó gracias a la beca de doctorado 160253 otorgada al primer autor por el Consejo Nacional de Ciencia y Tecnología. Las laminillas histológicas fueron realizadas por Alexander Rojas Alvarado, bajo la supervisión de Estela Sandoval Zapotitla, en el laboratorio de histología vegetal del Jardín Botánico del Instituto de Biología de la UNAM. A Magali Santillán se deben los dibu- jos y a Alin Torres Díaz la configuración de las figuras. La crítica al manuscrito original (tesis) la realizaron Fernando Chiang Cabrera del Instituto Biología de la UNAM, Adolfo Espejo Serna y Blanca Pérez García del Depto. de Bio- logía de la UAM-Iztapalapa y Francisco Lorea Hernández del Instituto de Ecología A. C. y, para esta versión, colaboraron Blanca Pérez García y Mónica Ponce del Instituto de Botá- nica Darwinion, Argentina. Los comentarios de cinco dictaminadores anónimos ayudaron a mejorar sustancialmente este manuscrito. Con base en la forma en que se anasto- mosan las venas (tipo IV) y a la distribución de los cordones vasculares a lo largo del eje pecíolo-raquis, este trabajo apoya la propuesta de Tejero-Díez & Pacheco (2004) con relación a que Polypodium arcanum, P. castaneum, P. col- podes, P. eatonii y P. fagellare forman un grupo homogéneo, denominado grupo P. colpodes. g , g p p Polypodium fraternum, P. kunzeanum, P. pleurosorum y P. rhachipterygium no tienen el patrón de caracteres de los grupos de P. ple- siosorum o P. colpodes descritos. Polypodium fraternum por su hoja pinnada, venación tipo I y III, cordones vasculares Tipo V, aunado a la presencia de escamas comosas en el rizoma (Tejero-Díez 2005), es posible que pertenezca a una sección cercana al grupo de Polypodium dulce de Moran (1995). La especie sudame- ricana Polypodium kunzeanum presenta las pinnas articuladas al raquis, carácter único en las especies de la familia Polypodiaceae en América. Pecíolo-raquis Rödl-Linder (1990) describió en su estu- dio para Goniophlebium cuatro series de dis- posiciones de cordones vasculares a lo largo del eje pecíolo-raquis. Además de éstas, en este trabajo se describen otras siete, consi- deradas como “tipos”, a partir del análisis correspondiente en algunas especies de Peclu- ma, Phlebodium, Pleopeltis y Polypodium. En su conjunto, estos tipos aportan información relevante sobre las relaciones taxonómicas y filogenéticas de Polypodiaceae (Tejero-Díez 2005). Polypodium eatonii y P. colpodes pre- sentan el tipo II junto con P. californicum perteneciente al grupo de P. vulgare, con el cual según Haufler & Ranker (1995) algunas especies del complejo P. plesiosorum tiene una relación filogenética. El resto de las especies del grupo de P. plesiosorum y P. dulce (Moran, 1995) muestran el tipo III, con excepción de P. subpetiolatum y P. echinolepis (especies de hoja pinnada) que presentan mayor canti- dad de CVP. Ello hace que estas dos últimas especies puedan formar un grupo propio, más relacionadas a Phlebodium que con las propias del grupo de Polypodium dulce, dentro del cual Moran (1995) las reconoce. Las especies estudiadas de Serpocaulon y Polypodium kun- zeanum presentan formas de fusión/división que corresponden a algunas de las series reconocidas por Rödl-Linder (1990) para Goniophlebium, como son los taxa americanos actualmente reconocidos como Serpocaulon y el grupo Goniophlebium argutum (Wall. ex Hook.) J. Sm. ex Hook. respectivamente. De acuerdo con de la Sota (1966), el grupo de Polypodium squamatum presenta el tipo IV, el mismo que se observó en P. guttatum y P. plebeium (grupo P. pycnocarpum de Tryon & Con respecto al mesofilo, los grupos Polypodium plesiosorum y P. vulgare presentan una lámina con estructura unifacial, con excep- ción de P. fraternum que es bifacial, mientras que las especies estudiadas de los grupos P. dulce y P. polypodioides y en Pleopeltis y Serpocaulon la tienen bifacial. Aparentemente, entre los helechos polipodiáceos la condición bifacial podría ser una situación derivada, adquirida como consecuencia de la adaptación a medios ecológicos xerófilos, pero es intere- sante señalar que esta condición de la lámina es el común denominador en los grupos y géneros señalados. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 969 AGRADECIMIENTOS El carácter anterior y el patrón de fusión y división de cordones vasculares en el pecíolo-raquis tipo IX, la hace semejante a Goniophlebium propio de Asia. Polypodium pleurosorum, por su sistema de venas anasto- mosado tipo V y sistema de cordones vascu- lares del pecíolo-raquis tipo VIII, se relaciona con el género Phlebodium; por ello Tejero-Díez et al. (2009) sugieren volver a utilizar el nom- bre Phlebodium inaequale T. Moore que Moore (1855) le asignó. Polypodium rhachipterygium, que presenta venación tipo I y III, lámina pin- natipartida y ala peciolar, aunado a caracteres propios del rizoma y soro (Tejero-Díez 2005), permite vislumbrar su relación con el grupo de P. dulce. REFERENCIAS Aguilar Rodríguez, S. 1998. Apéndice I: Técnicas de laboratorio para el estudio de los embriófitos, p. 247- 272. In D. Tejero-Díez & M.P. Granillo-Velázquez (eds.). Plantae; Introducción al estudio de las plantas con embrión. Universidad Nacional Autónoma de México-Facultad de Estudios Superiores Iztacala, Tlalnepantla, Edo. México, México. Gómez, L.D. 1975. Contribuciones a la pteridología costa- rricense VIII: La hibridización en el trópico; Micro- gramma x Polypodium y Polypodium aspidiolepis Baker. Brenesia 6: 49-57. Haufler, C.H. & T.A. Ranker. 1995. Rbcl sequences provi- de phylogenetic insights among sister species of the fern genus Polypodium. Amer. Fern J. 85: 361-374. Anthony, N.C. & E.A.C.L.E. Schlepe. 1985. xPleopodium- a putative intergeneric fern hybrid from Africa. Bothalia 15: 555-559. Haufler, C.H., D.E. Soltis & P.S. Soltis. 1995. Phylogeny of the Polypodium vulgare complex: insights from chloroplast DNA restriction site data. Syst. Bot. 20: 110-119. Beck, C.B., R. Schmid & G.W. Rothwell. 1982. Stelar morphology and the primary vascular systems of seed plants. Bot. Rev. 48: 691-815. Hennipman, E.P. 1986. The systematics of the Polypo- diaceae (Filicales): posing the problems. Acta Bot. Neerl. 35: 115. Bosman, M.T.M. 1991. A monograph of the fern genus Microsorium. Leiden Bot. Ser. 14: 1-161. Copeland, E.B. 1947. Genera filicum; the genera of ferns. Chronica Botanica. Waltham, Massachusetts, EEUU. Hennipman, E., P. Veldhoen & K.U. Kramer. 1990. Poly- podiaceae, p. 203-230. In K. Kubitzki (ed.). The families and genera of vascular plants. Springer, Berlín, Alemania. Crabbe, J.A., A.C. Jermy & J.T. Mickel. 1975. A new generic sequence for pteridophyte herbarium. Fern Gaz. 11: 141-162. Hensen, R.V. 1990. Revision of the Polypodium loriceum complex (Filicales, Polypodiaceae). Nova Hedwigia 50: 279-336. Ching, R.C. 1933. The study of Chinese ferns. XI. Polypo- dium. Contr. Inst. Bot. Acad. Peiping 2: 31-52. Hetterscheid, W.L.A. & E.P. Hennipman. 1984. Venation patterns, soral characteristics and shape of the fronds of the microsorioid Polypodiaceae. Bot. Jahrb. Syst. 105: 11-47. Ching, R.C. 1940. On natural classification of the family “Polypodiaceae”. Sunyatsenia 5: 201-268. Christensen, C. 1928. On the systematic position of Polypodium vulgare. Dansk Bot. Ark. 5: 1-10. Hooper, E.A. 1994. Biosystematic analysis of the Pleopel- tis macrocarpa complex in the Neotropics. Ph.D. Thesis, University of Kansas, Kansas, EEUU. De la Sota, E.R. 1963. Conceptos sobre la evolución de la venación en las filices leptosporangiadas superiores. Ameghiniana 3: 37-42. Iwatsuki, K. 1962. On the reticulate venation of the The- lypteroid ferns. Acta Phytotaxonom. Geobot. 20: 219-227. RESUMEN El complejo Polypodium plesisorum es un grupo de morfología heterogénea. Se estudió la arquitectura y anatomía foliar de 12 especies que constituyen el complejo de Polypodium plesiosorum reconocidas por Moran, y otras 23 especies de Polypodium (complejos P. dulce, P. polypodioides y P. vulgare sunsu Moran) y géneros relacionados taxonómicamente (Goniophlebium, Phlebodium, Pleopeltis, Serpocaulon, Synammia) con fines comparativos. Dos grupos de especies homogéneas fueron establecidos en el complejo P. plesisorum. Uno de ellos se reconoce por el tipo de anastomosis de las venas (tipo III): Polypodium conterminans (originalmente considerado en el grupo de P. dulce de venación libre según Moran), P. hispidulum, P. plesiosorum y P. rhodopleuron. El segundo grupo formado por Polypodium arcanum, P. castaneum, P. colpodes, P. eatonii y P. flagellare, se reconoce por el tipo de anastomosis de las venas (tipo IV) y la forma en que se fusionan los cordones vasculares a lo largo del eje pecíolo- raquis de la hoja. Los caracteres estudiados ayudaron a relacionar las especies restantes con otros grupos fuera del complejo estudiado. Palabras clave: Polypodiaceae, complejo Polypodium ple- siosorum, arquitectura foliar, anatomía foliar, pecíolo. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 970 Fryns-Claessens, E. & W. van Cotthem. 1973. A new clas- sification of the ontogenetic types of stomata. Bot. Rev. 39: 71-137. REFERENCIAS De la Sota, E.R. 1965. Las especies escamosas del género Polypodium (s. str.) en Brasil. Rev. Mus. La Plata, Secc. Bot. 9: 243-271. Johansen, D.A. 1940. Plant microtechnique. McGraw-Hill, New York, EEUU. De la Sota, E.R. 1966. Revisión de las especies americanas del grupo Polypodium squamatum L. Rev. Mus. La Plata, Secc. Bot. 10: 69-186. Kaur, S. 1978. Contributions to the morphology of Tec- taria: venation pattern of sterile and fertile leaves. Phytomorphology 28: 14-19. De la Sota, E.R. 1973. On the classification and phylogeny of the Polypodiaceae, p. 229-244. In A.C. Jermy, J.A. Crabbe & B.A. Thomas (eds.). The phylogeny and classification of the ferns. J. Linn. Soc. Bot. 67(1). Lellinger, D.B. 1993. Application of the name Gonio- phlebium and new subgeneric name in Polypodium. Amer. Fern J. 83: 37-38. Lloyd, R.M. & F.A. Lang. 1964. The Polypodium vulgare complex in North America. Brit. Fern Gaz. 9: 168- 177. Evans, A.M. 1969. Interspecific relationships in the Polypo- dium pectinatum-plumula complex. Ann. Missouri Bot. Gard. 55: 193-293. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 971 Presl, C.B. 1836. Tentamen pteridographiae. T. Haase, Praga, Rep. Checa. Lucansky, T.W. & R.A. White. 1974. Comparative studies of the nodal and vascular anatomy in the neotropical Cyatheaceae. III. Nodal and petiole patterns; sum- mary and conclusions. Amer. J. Bot. 61: 472-480. Price , M.G. 1983. Pecluma, a new tropical American fern genus. Amer. Fern J. 73: 109-116. Maxon, W.R. 1903. Studies of certain Mexican and Guate- malan species of Polypodium. Contr. U.S. Natl. Herb. 8: 271-280. Ranker, T.A., A.R. Smith, B.S. Parris, J.M. Geiger, C.H. Haufler, S.C.K. Starub & H. Schneider. 2004. Phylo- geny and evolution of grammitid ferns (Grammitida- ceae): a case of rampant morphological homoplasy. Taxon 53: 415-428. Maxon, W.R. 1916a. Polypodium furfuraceum and related species. Contr. U.S. Natl. Herb. 17: 557-579. Rödl-Linder, G. 1990. A monograph of the fern genus Goniophlebium (Polypodiaceae). Blumea 34: 277- 423. Maxon, W.R. 1916b. Polypodium squamatum and its allies. Contr. U.S. Natl. Herb. 17: 579-596. Maxon, W.R. 1916c. Studies of tropical American ferns: No 6. Contr. U.S. Natl. Herb. 17: 541-608. Saiki, Y., M. Matsumoto & Y. Mitsuda. 1988. Vascular patterns in the petioles of Aspleniaceae. Proc. ISSP 273-278. Mickel, J.T. & J.M. Beitel. 1987. Notes on xPleopodium and Pleopeltis in tropical America. Amer. Fern J. 77: 16-27. Salgado A.E. 1982. Venation pattern in Philippine Tectar- oid fern. REFERENCIAS De la Salle University, Integrated Research Center. Monograph Series Number 3. Manila, Filipinas. Mickel, J.T. & J.M. Beitel. 1988. Pteridophyte flora of Oaxaca, Mexico. Mem. New York Bot. Gard. 46: 1-568. Salino, A., T.E. Almeida, A.R. Smith, A. Navarro-Gómez, H.P. Kreier & H. Schneider. 2008. A new species of Microgramma (Polypodiaceae) from Brazil and recircumscription of the genus based on phylogenetic evidence. Syst. Bot. 33: 630-635. Mickel, J.T. & A.R. Smith. 2004. The pteridophytes of Mexico. Mem. New York Bot. Gard. 88: 1-1054. Mitsuta, S. 1984. Studies in the venation of Polypodia- ceae II. Venation of Polypodiaceae (2) and some other ferns. Mem. Fac. Sci. Kyoto Univ. Ser. Biol. 9: 57-85. Schneider, H., A.R. Smith, R. Cranfill, T.E. Hildebrand, C.H. Haufler & T.A. Ranker. 2004. Unraveling the phylogeny of polygrammoid ferns (Polypodiaceae & Grammitidaceae): exploring aspects of the diversifi- cation of epiphytic plants. Molec. Phylogenet. Evol. 31: 1041-1063. Moore, T. 1855. New garden ferns. No. IV. 8. Phlebodium inaequale. Gard. Chron. 1855: 660. Moran, R.C. 1995. Polypodium L., p. 349-365. In R.C. Moran & R. Riba (eds.). Psilotaceae a Salviniceae. Flora Meso- americana. Vol 1. Universidad Nacional Autónoma de México, Distrito Federal, México; Missouri Botanical Garden y The Natural History Museum. Schneider, H., H.P. Kreier, R. Willson & A.R. Smith. 2006. The Synammia enigma: evidence for a tempe- rate lineage of polygrammoid ferns (Polypodiaceae, Polypodiidae) in southern South America. Syst. Bot. 31: 31-41. Neuroth, R. 1996. Biosystematik und Evolution des Polypodium vulgare-Komplexes (Polypodiaceae, Pteridophyta). Dissert. Bot. 256: 1-209. Schuettpelz, E. & K.M. Pryer. 2007. Fern phylogeny inferred from 400 leptosporangiate species and three plastid genes. Taxon 56: 1037–1050. Otto, E.M. 2007. Untersuchungen zur Phylogenie der epiphytischen Farngattung Pleopeltis. Diplomarbeit. Albrecht-von-Haller-Institut für Pflanzenwissenschaf- ten an der Biologischen Fakultät der Georg-August- Universität zu Göttingen, Göttingen, Alemania. Sen, U. & E. Hennipman. 1981. Structure and ontogeny of stomata on Polypodiaceae. Blumea 27: 175-201. Smith, A.R. 1981. Pteridophytes, p. 1-130. In D.E. Bree- dlove (ed.). Flora of Chiapas Vol. 2. California Aca- demy of Sciences, San Francisco, EEUU. Payne, W.W. 1968. The use of cellulose acetate film for the production of epidermal cast. Ward´s Bull. 7: 6-7. Smith, A.R., H. Schneider, B.S. Parris, C.H. Haufler, T.A. Ranker, J.P. Terrien & J.M.O. Geiger. 2000. Phylo- geny of Grammitidaceae and Polypodiaceae inferred Pichi-Sermolli, R.E.G. 1977. Tentamen Pteridophytorum genera in taxonomicum ordinem redigendi. Webbia 31: 313-512. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. REFERENCIAS 58 (3): 955-976, September 2010 972 from two large data sets: rbcL and morphology. Amer. J. Bot. 87: 97. Tryon, R.M. & A.F. Tryon. 1982. Ferns and allied plants, with special reference to tropical America. Springer, New York, EEUU. Smith, A.R., H.P. Kreier, C.H. Haufler, T.A. Ranker & H. Schneider. 2006. Serpocaulon, a new genus segrega- ted from Polypodium. Taxon 55: 919-930. Wagner, W.H. 1952. Types of foliar dichotomy in living ferns. Amer. J. Bot. 39: 578-592. Steeves, T.A. 1988. Phylogenetic and ontogenetic inter- pretation of the fern vascular system. Proc. ISSP 279-284. Wagner, W.H. 1979. Reticulate veins in the systematics of modern ferns. Taxon 28: 87-95. Wagner, W.H. 1986. The new world fern genus Margina- riopsis (Polypodiaceae): an example of leaf dimorphy and coenosory in generic delimitation. Bull. Torrey Bot. Club 113: 159-167. Stewart, W.N. & G.W. Rothwell. 1993. Paleobotany and the evolution of plants. Cambridge University, Cam- bridge, Gran Bretaña. Stolze, R.G. 1981. Ferns and fern allies of Guatemala. Part II: Polypodiaceae. Fieldiana, Bot. n.s. 6: 1-522. Wagner, W.H. & F.S. Wagner. 1975. A hybrid polypody from the new world tropics. Fern Gaz. 11: 125-135. Tejero-Díez, J.D. 2005. Revisión taxonómica del com- plejo Polypodium plesiosorum Kunze (Polypodia- ceae, Polypodiophyta). Tesis Doctorado, Universidad Autónoma Metropolitana, Distrito Federal, México. Wardlaw, C.W. 1944. Experimental and analytical studies of pteridophytes. IV. Stelar morphology: experimen- tal observations on the relation between leaf develop- ment and stelar morphology in species of Dryopteris and Onoclea. Ann. Bot. 8: 387-399. Tejero-Díez, J.D. & L. Pacheco. 2004. Taxa nuevos, nomenclatura, redefiniciones y distribución en las especies relacionadas a Polypodium colpodes Kunze (Polypodiaceae, Pteridophyta). Acta Bot. Mex. 67: 75-115. Weatherby, C.A. 1922. The group of Polypodium lanceola- tum in North America. Contr. Gray Herb. 65: 3-14. Weatherby, C.A. 1939. The group of Polypodium polypodioides. Contr. Gray Herb. 124: 22-35. Tejero-Díez, J.D., J.T. Mickel & A.R. Smith. 2009. A hybrid Phlebodium (Polypodiaceae, Polypodiophyta) and its influence on the circumscription of the genus. Amer. Fern J. 99: 109-116. Weatherby, C.A. 1947. Polypodium lepidopteris and its relatives in Brazil. Contr. Gray Herb. 165: 76-82. White, R.A. 1984. Comparative development of vascular tissue patterns in the shoot apex of ferns, p. 53-107. In R.A. White y W.C. Dickison (eds.). Contemporary problems in plant anatomy. Academic, New York, EEUU. Tindale, M.D. 1960. Vein patterns in Microsorium scan- dens and its allies. Amer. Fern J. 50: 241-245. Tryon, R.M. 1964. APPENDIX I Vouchers used for anatomy study. Polypodium species are arranged in groups agreement with Moran (1995) except Serpocaulon (=P. loriceum). Collect of D. Tejero-Díez. Morphologic units of study: 1=histology of petiole and blade. 2=stomatal apparatus. 3=petiole rachis series. 4=leaf architecture Taxones Referencia Unidades morfológicas de estudio POLYPODIUM 1 2 3 4 Grupo P. plesiosorum P. arcanum var. arcanum 4218 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, ca. Zacapoaxtla X 4230 (IZTA, UAMIZ); Méx., Hgo., km 22 entre Metepec y Tenango de Doria X X 4240 (IZTA, UAMIZ); Méx., Hgo., Agua Blanca, salto San Pedrito X X X 4250 (IZTA, UAMIZ); Méx., Hgo., Molango, Laguna Atezca X 4269 (IZTA, UAMIZ); Méx., Oax., Ixtlán de Juárez, km 142 -carretera fed. 175 X P. arcanum var. septentrionale Mickel 4160 (IZTA, UAMIZ); Méx., Tamps., Ocampo (km 30 entre Ocampo y Tula) X X X X P. arcanum var. bakeri Tejero & Mickel 4364 (IZTA, UAMIZ); Méx., Gro., Heliodoro Castillo, ca. El Carrizal X 4365 (IZTA, UAMIZ); ibídem, entre Filo de Caballo y Pto. El Gallo X P. castaneum T. B. Croat 42213 (MO); El Salvador, Sonsonate, ca. cúspide Cerro Verde X X X X P. colpodes var. colpodes 2074 (IZTA); Méx., Méx., Almoloya de Alquisiras, entronque a Ixtapan Los Sabinos X 3194 (IZTA); Méx., Méx., Sto. Tomás de los Plátanos, base de la presa 4302-4303 (CR, IZTA, UAMIZ); Costa Rica, Guanacaste, P. Nal. Palo Verde X 4313 (CR, IZTA, UAMIZ); Costa Rica, San José, Dota, entre Sta. María y COPEI X X 4360 (IZTA, UAMIZ); Méx., Chis., Ixtapa, Pte. Río Lajas X 4368 (IZTA); Méx., Gro., Eleodoro Castillo, camino a Puerto El Gallo X P. colpodes var. oligomeron (H. Christ)Tejero 4311 (CR, INB, IZTA, UAMIZ); Costa Rica, San José, Dota, ca. Copei. X P. eatonii 4315 (IZTA, UAMIZ); Méx., Ver., Banderilla, a 3 km al E de Banderilla X X X X P. flagellare 4306 (CR, INB, IZTA, UAMIZ); Costa Rica, Puntarenas, Garabito, P. Nal. Carará X X Referencia Unidades morfológicas de estudio 1 2 3 4 4218 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, ca. Zacapoaxtla X 4230 (IZTA, UAMIZ); Méx., Hgo., km 22 entre Metepec y Tenango de Doria X X 4240 (IZTA, UAMIZ); Méx., Hgo., Agua Blanca, salto San Pedrito X X X 4250 (IZTA, UAMIZ); Méx., Hgo., Molango, Laguna Atezca X 4269 (IZTA, UAMIZ); Méx., Oax., Ixtlán de Juárez, km 142 -carretera fed. REFERENCIAS Evolution in the leaf of living ferns. Mem. Torrey Bot. Club 21: 73-82. Windham, M.D. 1993. New taxa and nomenclatural chan- ges in the North American fern flora. Contr. Univ. Michigan Herb. 19: 31-61. Tryon, R.M. 1970. The classification of the Cyatheaceae. Contr. Gray Herb. 200: 3-53. Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 973 APÉNDICE I Especímenes de referencia (estudio anatómico). Especies de Polypodium en grupos de acuerdo con Moran (1995) excepto Serpocaulon (=P. loriceum). Recolectas de D. Tejero-Díez. Unidades morfológicas de estudio: 1=histología del pecíolo y lámina. 2=aparato estomático. 3=series pecíolo raquis. 4=arquitectura foliar APPENDIX I 175 X e 4160 (IZTA, UAMIZ); Méx., Tamps., Ocampo (km 30 entre Ocampo y Tula) X X X X 4364 (IZTA, UAMIZ); Méx., Gro., Heliodoro Castillo, ca. El Carrizal X 4365 (IZTA, UAMIZ); ibídem, entre Filo de Caballo y Pto. El Gallo X T. B. Croat 42213 (MO); El Salvador, Sonsonate, ca. cúspide Cerro Verde X X X X 2074 (IZTA); Méx., Méx., Almoloya de Alquisiras, entronque a Ixtapan Los Sabinos X 3194 (IZTA); Méx., Méx., Sto. Tomás de los Plátanos, base de la presa 4302-4303 (CR, IZTA, UAMIZ); Costa Rica, Guanacaste, P. Nal. Palo Verde X 4313 (CR, IZTA, UAMIZ); Costa Rica, San José, Dota, entre Sta. María y COPEI X X 4360 (IZTA, UAMIZ); Méx., Chis., Ixtapa, Pte. Río Lajas X 4368 (IZTA); Méx., Gro., Eleodoro Castillo, camino a Puerto El Gallo X 4311 (CR, INB, IZTA, UAMIZ); Costa Rica, San José, Dota, ca. Copei. X 4315 (IZTA, UAMIZ); Méx., Ver., Banderilla, a 3 km al E de Banderilla X X X X 4306 (CR, INB, IZTA, UAMIZ); Costa Rica, Puntarenas, Garabito, P. Nal. Carará X X Grupo P. plesiosorum P. arcanum var. arcanum 4218 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, ca. Zacapoaxtla X 4230 (IZTA, UAMIZ); Méx., Hgo., km 22 entre Metepec y Tenango de Doria X X 4240 (IZTA, UAMIZ); Méx., Hgo., Agua Blanca, salto San Pedrito X X X 4250 (IZTA, UAMIZ); Méx., Hgo., Molango, Laguna Atezca X 4269 (IZTA, UAMIZ); Méx., Oax., Ixtlán de Juárez, km 142 -carretera fed. 175 X P. arcanum var. septentrionale Mickel 4160 (IZTA, UAMIZ); Méx., Tamps., Ocampo (km 30 entre Ocampo y Tula) X X X X P. arcanum var. bakeri Tejero & Mickel 4364 (IZTA, UAMIZ); Méx., Gro., Heliodoro Castillo, ca. El Carrizal X 4365 (IZTA, UAMIZ); ibídem, entre Filo de Caballo y Pto. El Gallo X P. castaneum T. B. Croat 42213 (MO); El Salvador, Sonsonate, ca. cúspide Cerro Verde X X X X P. colpodes var. colpodes 2074 (IZTA); Méx., Méx., Almoloya de Alquisiras, entronque a Ixtapan Los Sabinos X 3194 (IZTA); Méx., Méx., Sto. Tomás de los Plátanos, base de la presa 4302-4303 (CR, IZTA, UAMIZ); Costa Rica, Guanacaste, P. Nal. Palo Verde X 4313 (CR, IZTA, UAMIZ); Costa Rica, San José, Dota, entre Sta. María y COPEI X X 4360 (IZTA, UAMIZ); Méx., Chis., Ixtapa, Pte. Río Lajas X 4368 (IZTA); Méx., Gro., Eleodoro Castillo, camino a Puerto El Gallo X P. colpodes var. APPENDIX I oligomeron (H. Christ)Tejero 4311 (CR, INB, IZTA, UAMIZ); Costa Rica, San José, Dota, ca. Copei. X P. eatonii 4315 (IZTA, UAMIZ); Méx., Ver., Banderilla, a 3 km al E de Banderilla X X X X P. flagellare 4306 (CR, INB, IZTA, UAMIZ); Costa Rica, Puntarenas, Garabito, P. Nal. Carará X X P. colpodes var. oligomeron (H. Christ)Tejero P. eatonii P. flagellare Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 974 Rev. Biol. Trop. (Int. J. Trop. Biol. Taxones P. fraternum P. hispidulum P. kunzeanum P. plesiosorum var. plesiosorum P. plesiosorum var. rubicundum Tejero P. pleurosorum P. rhachipterygium P. rhodopleuron Grupo P. dulce P. conterminans P. echinolepis P. martensii Referencia Unidades morfológicas de estudio 4314 (CR, INB, IZTA, UAMIZ); Costa Rica, San José, Puriscal, ca. San Martín X X C. W. Dodge 16567 (MO); Panamá, Chiriquí, Alajuela, ca. Madden Daan X L. D. Gómez 467 (CR); Costa Rica, San Pablo Turrubares X 4202 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X X 4245 (IZTA, UAMIZ); Méx., Hgo., Zacualtipán, ca. Carpinteros X X X X E. Martínez & G. Aguilar 10007 (MEXU); Méx., Chis., Ocosingo, ca. Bonampak X H. Türckheim 8053 (P); Guatemala, Alta Verapaz, Cubilguitz X 4328 (IZTA, UAMIZ); Méx., Chis., Palenque, zona arqueológica X X 4329 (IZTA, UAMIZ); Méx., Chis., salto Misol-Ha X M. Grayum 8602 (INB); Costa Rica, San José, Puriscal, Cerro Bola X C. Linderman 2599 (BM); Brazil, Matto Grosso, cascada Cupin X H. H. Smith 2564 (L, U); Colombia, sin loc. X X 4129 (IZTA, UAMIZ); Méx., Qro., Pinal de Amoles, entre Cadereyta y Pinal A. X X X 4132 (IZTA, UAMIZ); Méx., Qro., Pinal de Amoles, entre Pinal Amoles. y Jalpan X X 4137-4138 (IZTA, UAMIZ); Méx., Tamps., Gómez Farías, reserva de la biosfera. El Cielo. X X 4205-4206 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X 4220 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, ca. Zacapoaxtla X 4330 (IZTA, UAMIZ); Méx., Ver., Maltrata, km 235-carretera fed. 150 X 4308 (INB, IZTA, UAMIZ); Costa Rica, Cartago, Paraíso, ca. Orosí X X X X 4366 (IZTA, UAMIZ); Méx., Gro., Heliodoro Castillo, ca. Filo de Caballo X X M. Pérez Farrera 246 (UAMIZ); Méx., Chis, Villaflores, cerro Tres Picos X 4185 (IZTA, UAMIZ); Méx., Ver., Catemaco, ca. Coyame X X X X Ramamoorthy 3348 (IZTA, MEXU); Méx., Ver., San Andrés Tuxtla, estación biológca UNAM X X Ramamoorthy 4157 (IZTA, MEXU); Ibídem X 4078 (IZTA, UAMIZ) ; Méx., Hgo., Tlalchinol, entrada a Lontla 5 km a Huejutla X X X 4193 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X 4228 (IZTA, UAMIZ) ; Méx., Hgo., Tenango de Doria, km 19 carr. estatal 126 X X 4243 (IZTA, UAMIZ); Méx., Hgo., Zacualtipán, ca. Carpinteros X 4358 (IZTA, UAMIZ); Méx., Chis., Jitotol, ca. Pueblo Nuevo X X X X 4213 (IZTA); Méx., Pue., Cuetzalán, salto San Andrés X X X X 4088 (ENCB, IZTA), Méx., Hgo., Zacualtipán, ca. Zoquizoquiapán X X X Referencia Taxones Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. Rev. Biol. Trop. (Int. J. Trop. Biol. Taxones P. fraternum P. hispidulum P. kunzeanum P. plesiosorum var. plesiosorum P. plesiosorum var. rubicundum Tejero P. pleurosorum P. rhachipterygium P. rhodopleuron Grupo P. dulce P. conterminans P. echinolepis P. martensii 58 (3): 955-976, September 2010 975 Referencia morfológicas de estudio 4180 (IZTA, UAMIZ); Méx., Méx., Chapa de Mota, ca. Chapa de Mota. X X X 4363 (IZTA, UAMIZ); Méx., Chis., P. Nal. Cañón Sumidero X 4074 (IZTA); Méx., Hgo., Tlalchinol, ca. Zacualtipán X X X X 4355 (IZTA, UAMIZ); Méx., Oax., Ixtlán de Juárez, ca. Calpulalpan X X 4192 (IZTA, UAMIZ); Méx., Pue, Acatzingo, km 27 rumbo a San Salvador. El Seco X X X X 4332 (IZTA, UAMIZ); Méx., Hgo, Zacualtipán, a 2 km desv. a Tianguistengo X X 4340 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X de la Sota (1966), fig. 12.D X 4353 (IZTA, UAMIZ); Méx., Oax., km 110-carretera fed. 175 X X 4194 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X 4346 (IZTA, UAMIZ); Méx., Pue., Cuetzalán, salto San Andrés X X 4356 (IZTA, UAMIZ); Méx., Oax., Ixtlán de Juárez, ca. Calpulalpan X X 4333 (IZTA, UAMIZ); Méx., Hgo, Zacualtipán, a 2 km desv. a Tianguistengo X Sánchez Glez. 10 V 2001 (CHAPA, IZTA); Méx., Méx., Texcoco, cerro Tláloc X X X X Rödl-Linder (1990), fig. 3.a X X X X 4341 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X 4195 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X X 4362 (IZTA, UAMIZ); Méx., Chis., ca. San Cristóbal X X 4352 (IZTA, UAMIZ); Méx., Oax., Valle Nal., ca. Esperanza X X 4347 (IZTA, UAMIZ); Méx., Pue., Cuetzalán, salto San Andrés X X 4334 (IZTA, UAMIZ); Méx., Hgo, Zacualtipán, a 2 km desv. a Tianguistengo X X 4196 (IZTA, UAMIZ); Méx., Pue., Zacapoaxtla, Apulco, salto La Gloria X X X X 4351 (IZTA, UMIZ); Méx., Oax., Valle Nacional, ca. Esperanza X 4186 (IZTA, UAMIZ); Méx., Ver., Catemaco, ca. Coyame X X X X Harvey s.n. (K); Chile, Talcuhano X Levin 509 (P); Chile, sin localidad X Taxones Grupo P. polypodioides P. alan-smithii P. furfuraceum P. guttatum P. lepidotrichum Rev. Biol. Trop. (Int. J. Trop. Biol. ISSN-0034-7744) Vol. 58 (3): 955-976, September 2010 976
https://openalex.org/W3204720739	https://www.mdpi.com/2076-3417/11/19/8804/pdf	English	null	Assessment of Ultraviolet Impact on Main Pigment Content in Purple Basil (Ocimum basilicum L.) by the Spectrometric Method and Hyperspectral Images Analysis	Applied sciences	2,021	cc-by	8,221	Citation: Proshkin, Y.A.; Smirnov, A.A.; Semenova, N.A.; Dorokhov, A.S.; Burynin, D.A.; Ivanitskikh, A.S.; Panchenko, V.A. Assessment of Ultraviolet Impact on Main Pigment Content in Purple Basil (Ocimum basilicum L.) by the Spectrometric Method and Hyperspectral Images Analysis. Appl. Sci. 2021, 11, 8804. https://doi.org/10.3390/ app11198804 Keywords: basil; ultraviolet radiation; photosynthetic pigments; hyperspectral images; vegetative indices Academic Editor: Ana M. Cavaco Received: 11 August 2021 Accepted: 18 September 2021 Published: 22 September 2021 Received: 11 August 2021 Accepted: 18 September 2021 Published: 22 September 2021 Assessment of Ultraviolet Impact on Main Pigment Content in Purple Basil (Ocimum basilicum L.) by the Spectrometric Method and Hyperspectral Images Analysis Yuri A. Proshkin 1 , Alexandr A. Smirnov 1,* , Natalya A. Semenova 1 , Alexey S. Dorokhov 1, Dmitry A. Burynin 1, Alina S. Ivanitskikh 1 and Vladimir A. Panchenko 1,2 1 Federal Scientiﬁc Agroengineering Center VIM, (FSAC VIM), Federal State Budgetary Scientiﬁc Institution, 109428 Moscow, Russia; yproshkin@gmail.com (Y.A.P.); natalia.86@inbox.ru (N.A.S.); dorokhov.vim@yandex.ru (A.S.D.); burynin@gmail.com (D.A.B.); alinena@yandex.ru (A.S.I.); pancheska@mail.ru (V.A.P.) 1 Federal Scientiﬁc Agroengineering Center VIM, (FSAC VIM), Federal State Budgetary Scientiﬁc Institution, 109428 Moscow, Russia; yproshkin@gmail.com (Y.A.P.); natalia.86@inbox.ru (N.A.S.); dorokhov.vim@yandex.ru (A.S.D.); burynin@gmail.com (D.A.B.); alinena@yandex.ru (A.S.I.); pancheska@mail.ru (V.A.P.) 2 Department of Theoretical and Applied Mechanics, Russian University of Transport, 127994 Moscow, Russia * Correspondence: as984788@gmail.com; Tel.: +7-9057-236-219 Abstract: This research is aimed at the assessing the impact of the ultraviolet radiation in the A, B, and C ranges (as additives to the main light) on general plan condition, the stress experienced by them, the pigment concentration in the leaves and leaf reﬂective characteristics. Under studying, there were the photo-protective reactions of the purple variety basil plants. The plants were grown in plastic pots in a phyto-chamber equipped with an automatic microclimate system. The phyto-chamber was divided into four compartments where, in addition to the main lighting, there were installed the additional LEDs emitting their radiation in the ranges UV-A, UV-B, and UV-C. Plant reactions were evaluated by the contents of the main pigments as detected by the spectrometric method. Then correlations were revealed between those values and the vegetative indices obtained based on the hyperspectral images. A strong correlation (R2 > 0.83) was observed between the values of the vegetative indices ARI and mARI and the anthocyanins concentration in basil leaves. A weak correlation (R2 = 0.0479) was found between the ARI and mARI values and the carotenoids index CRI700, which is attributed to the shielding effect of the anthocyanins. Deviations in the results are inﬂuenced by leaf surface unevenness, its thickness and density. Additional research is needed including developing reﬂection indices taking into account the shielding effect of the purple pigments. applied sciences applied sciences applied sciences applied sciences 1. Introduction Many papers are devoted to studying of the inﬂuence of the sulfhydryl compounds (thiourea and dithiotreitol) and other active additives on the ﬂavonoids concentration and accumulation rates as far as the ﬂavonoids reduce signiﬁcantly the stress caused by UV radiation [12–15]. First of all, the results of the studies of UV radiation’s effect on plants are important for enterprises growing agricultural products in an artiﬁcial climate. After all, the addition of UV spectrum rays to the main illumination will allow obtaining green and vegetable crops with the speciﬁed characteristics in terms of the contents of acids, sugars, anthocyanins and antioxidants and thereby increasing the nutritional value and taste qualities of the products. Pigment concentration measurement by the chemical method takes quite a long time, limiting the ability to analyze a large number of samples, so sub-sampling is used instead. Modern digital technologies allow estimating the pigment concentration in a leaf by the analysis of its reﬂection spectra (analysis of the hyperspectral images) [16–19]. Based on such data, plant stress states can be determined [20–22] as well as the content of substances of interest, for example, shikimic acid [23]. Such measurements are carried out in a non- invasive way and require much less labor. However, in cases of the analysis of the purple- leaved crop, difﬁculties are encountered with assessment of the chlorophyll concentration. Analysis of the published research on the UV radiation inﬂuences on green crops has revealed that only a few works are devoted to the effects of UV-A and UV-C on the purple- leaved crops. Additionally, the non-invasive methods are developed quite poorly for the pigment concentration measuring [24,25]. Taking the above into account, it was decided to study the photo-protective reactions of the red-leaved green crops under exposure to UV-A, UV-B, and UV-C radiation in addition to the main light. The plant reactions were evaluated by the change in the pigment concentration (by chemical method) and by the correlations presence between the obtained values and the vegetative indices calculated with aid of the hyperspectral images. 1. Introduction The negative anthropogenic impact on the Earth biosphere results in such serious climatic changes as the growth of atmosphere and ocean annual average temperatures, the formation of greenhouse gases in large amounts, the ozone layer depletion and conse- quently rises the UV radiation intensity. The impacts studying of the UV radiation shows that it affects the ecosystems’ biodiversity negatively and results in migrations and/or extinction of large species of animals and plants [1–4]. The ozone layer depletion and the inevitable rise of the levels of UV-A, UV-B and UV-C radiation affects negatively the organisms’ physiology, increases the frequency of cancer diseases in animals and inhibits plants’ vegetation and productivity [5,6]. In many publications devoted to this problem, the inﬂuence was studied of the light spectral composition on the plants’ growth and de- velopment; and the mechanisms of plant protection from the UV radiation were described including pigment synthesis and accumulation processes, the features of accumulation of the secondary metabolites and the functions of those substances [7–10]. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional afﬁl- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). It is well-known that in small doses, UV radiation can stimulate metabolic processes, thereby inducing a strengthening effect in plants [11]. When the plants are exposed to https://www.mdpi.com/journal/applsci Appl. Sci. 2021, 11, 8804. https://doi.org/10.3390/app11198804 Appl. Sci. 2021, 11, 8804 2 of 13 small doses of UV radiation, some insigniﬁcant changes occur in their morphological parameters, such as the plants’ fresh and dry mass, the average leaf area and the plant height. That is why the plants’ mechanisms of adaptation and photoprotection are studied mainly by the following metabolic parameters: the concentration of pigments and micro- and macro-nutrients, the content of nitrates, sugars and acids in the leaves, and the fruits of the plants. Mostly, the studies on this topic are carried out in an artiﬁcial environment, where conditions are created enabling the control of UV radiation exposure doses and du- ration. 1. Introduction yp p g The aims of the conducted research were: to evaluate the effects of different UV radiation ranges (UV-A, UV-B, and UV-C) on the pigment concentration in plant leaves and the reﬂective characteristics of leaves; to give a general assessment of a stress state for the studied plants; to make a comparison between the results of the pigment concentration measurements obtained by the invasive biochemical method and the same obtained by the non-invasive research method (the optical method based on spectral analysis of the digital images made by a hyperspectral camera). 2.1. Vegetable Material The experiment was carried out on the following purple-leaved varieties of the sweet basil (Ocimum basilicum L.): “Red Ruby”, “Gastronome” and “Markus” (“Prestige seeds” Ltd., Moscow, Russian Federation). Sweet basil is a herbaceous plant with a high content of essential oils that signiﬁcantly affect both the taste and the ﬂavor characteristics of the plant leaves. These essential oils are the secondary metabolites with pronounced antioxidant, antifungal, and antibacterial properties. “Red Ruby” is an early-middle variety of basil. The plants are compact, not large. This variety is intended for growing both in protected soil (regardless in seedling or seedless culture) and in open ground. It has a strong clove aroma. “Markus” is an early-middle variety of basil, too. It is recommended for growing both in open and protected ground. The variety adapts easily to various substrates, has a Appl. Sci. 2021, 11, 8804 3 of 13 high cold resistance and is resistant to edge burn. It is rich in essential oils that give it the clove aroma. high cold resistance and is resistant to edge burn. It is rich in essential oils that give it the clove aroma. “Gastronome” is a midseason-ripening, cold-resistant and leveled basil variety with stable color and the speciﬁc clove aroma. 2.2. Cultivation Conditions The plants were grown in 1-L volume plastic pots. Seven seeds were sown in each pot; after germination, only three plants were left in each. As a substrate, the neutralized high-moor peat Agrobalt-C (“Rostorﬁninvest” CJSC, Moscow, Russian Federation) was used. Four groups of the basil plants of the three varieties (“Markus”, “Gastronome”, “Red Ruby”) were planted, ﬁve pots in each group; in total, sixty plants of each variety. For 60 days, the plants were cultivated in the phyto-chamber equipped with the automatic microclimate system under the following conditions: at the temperature range in day/night of 26/20 ± 1.0 ◦C and at the relative air humidity of 60 ± 10%. Watering was carried out by the drip method. The nutrient solutions were prepared on the basis of de-ionized water. The chemical composition of the nutrient solution was as follows: N-NO3 9.64 mM/L; N-NH4 1.07 mM/L; P-PO4 1.00 mM/L; K 5.77 mM/L; Ca 2.00 mM/L; Mg 1.65 mM/L; S-SO4 1.75 mM/L; Fe 15.00 µM/L; B 20.00 µM/L; Cu 1.00 µM/L; Zn 5.00 µM/L; Mn 10.00 µM/L; Mo 1.00 µM/L. 2.3. Irradiation Conditions The volume of the phyto-chamber was divided by opaque panels into four equal compartments. In three of the four zones, in addition to the main lighting, there were installed additional LEDs emitting their radiation in the UV-A, UV-B, and UV-C ranges. (The fourth zone was for control). The spectral characteristics of the emitters in each zone are shown in Table 1. The UV irradiators consisted of an aluminum radiator with printed circuit boards with LEDs attached to them. The technical speciﬁcations of the installed LEDs are presented in Table 2. As a percentage of visible light, the intensity of the used UV emitters, comparable to that of sunlight, are for the Moscow summer season for 2019 and 2020 (with the exception of UV-C, which is absent under normal conditions, but it can be present in places with ozone holes in earth′s atmosphere). Table 1. Spectral properties of radiation sources. No. Experiment Variant Main Lighting, W m−2 UV-Radiation, W m−2 B 400–500 nm G 500–600 nm R 600–700 nm FR 700–800 nm UV-A 315–400 nm UV-B 280–315 nm UV-C 100–280 nm 1 UV-A 5 ± 0.5 9.5 ±0.3 15.8 ± 0.4 1.5 ± 0.1 1.5 ± 0.2 - - 2 UV-B - 0.015 ± 0.005 - 3 UV-C - - 0.010 ± 0.002 4 Control - - - Table 2. Characteristics of UV-LEDs. Characteristics Type of UV-LEDs UV-A Raytron Solution UV-B Szyunju UV-C Epistar Peak current, mA 750 40 40 Output power, W 2.4 ÷ 2.6 2 ÷ 4 4 ÷ 6 Direct voltage, V 3.2 ÷ 3.8 5 ÷ 8 5 ÷ 8 Viewing angle, ◦ 60 120 120 Peak wavelength, nm 365 310 275 Table 1. Spectral properties of radiation sources. Table 2. Characteristics of UV-LEDs. In the photosynthetically active radiation zone, the exposure was provided by the combined LED-based irradiators of various spectral compositions. The light period dura- Appl. Sci. 2021, 11, 8804 4 of 13 tion was 16 h. The average irradiation intensity was 30 ± 1,2 Wm−2, which corresponds to the lower recommended level required for the normal basil cultivation. The UV irradiators worked simultaneously with the main light. The UV irradiance was monitored with the portable UV radiometer TKA-PCM (STP “TKA”, Saint-Petersburg, Russian Federation). 2.4. Analysis of Pigment Content The quantitative analysis of the pigment content was carried out on the 30th, 45th and 60th days of the cultivation. Five plants of each variety were chosen, with the 3rd leaf from the top taken for analysis. In order to determine the quantitative content of the chlorophylls and the carotenoids, samples weighing 0.1 g were selected. The pigments were extracted from the plant tissues using 100% acetone. Weighing was performed on the laboratory electronic scales Sartorius LA230S laboratory scale (error rate 0.0001 g). After the ﬁltration, the obtained extract was analyzed in the spectrophotometer SPECS SSP-705 (manufactured by “Spectroscopic systems” CJSC, Russia). The optical density was determined at the following wavelengths: 644 nm (peak absorption of the chlorophyll b), 662 nm (peak absorption of the chlorophyll a) and 440 nm (peak absorption of the carotenoids) as described by Semenova et al. [26]. The absorbing layer thickness in the cuvettes was 10 mm. The quantitative concentration of pigments was calculated by Holm–Wettstein method (for 100% acetone). In order to determine the quantity of the anthocyanins, samples weighing 0.3 g were taken. Then the pigments were extracted from the plant tissues with use of the 1% hydrochloric acid solution; the tissues were kept in a water bath for 15 min at the temperature of 40–45 ◦C. After ﬁltration, the samples remaining on the ﬁlter were extracted again using the 1% hydrochloric acid solution. The obtained extract was analyzed in the spectrophotometer SPECX SSP-705. The optical density was determined at the wavelengths of 510 nm and 657 nm (adjusted for the green pigment content accordingly) [27]. The thickness of the absorbing layer in the cuvette was 10 mm. The total content of the anthocyanins was calculated by cyanidin-3,5-diglycoside. 2.5. Vegetative Indices Calculation The spectral characteristics of the leaves’ reﬂectivity were measured with aid of the portable hyperspectral camera Specim IQ (Spectral Imaging Ltd., Oulu, Finland). The reﬂection spectra were obtained for each individual pixel of the image in the range of 400–1000 nm (spectral resolution 7 nm, 204 spectral bands, spatial sampling of 512 pixels). Based on them, the vegetative indices were calculated. As the light source for Specim IQ, four halogen lamps were used. They were installed symmetrically relative to the camera position. p The spectral characteristics of the leaves’ reﬂectivity were measured on the 30th day of the cultivation. After the obtained hyperspectral images processing, the following well-known indices for the plants state assessment were calculated: the photochemical reﬂection index PRI (often used as the stress indicator), the indices ARI and mARI (used to assess the anthocyanins content in plant leaves), the chlorophyll indices MND705 and MND750/700 and the carotenoid index CRI700 [28,29]. The indices were calculated using the following formulas: PRI = ((R531 −R570))/((R531 + R570)) (1) ARI = (1/R550) −(1/R700) (2) mARI = R850 ((1/R550) −(1/R700)) (3) MND705 = (R750 −R705)/(R750 + R705 −2·R445) (4) R750/700 = R750/R700 (5) CRI700 = (1/R510) −(1/R700) (6) PRI = ((R531 −R570))/((R531 + R570)) (1) ARI = (1/R550) −(1/R700) (2) mARI = R850 ((1/R550) −(1/R700)) (3) MND705 = (R750 −R705)/(R750 + R705 −2·R445) (4) R750/700 = R750/R700 (5) CRI700 = (1/R510) −(1/R700) (6) (6) Appl. Sci. 2021, 11, 8804 5 of 13 The peak sensitivity of the reﬂection index to the anthocyanins concentration is found in the band of 550 nm. The 700 nm band is the chlorophyll absorption band; it is used in connection with the bands overlap of the light absorption by the chlorophyll and the anthocyanins. The band of 850 nm for the mARI index was taken as the closest one sensitive to reﬂection from the near-infrared range. 3. Results and Discussion 3.1. Analysis of Basil Plants Habitus 2.6. Statistical Analysis The biochemical parameters were processed by applying ANOVA. To estimate the statistical signiﬁcance of the considered parameters, the F-test and the least signiﬁcant difference test were applied in the RStudio software. 3.1. Analysis of Basil Plants Habitus The ratio of total chlorophyll to carotenoids showed that the plants actively photosynthesized; there were no signs of a physiological aging. EVIEW 7 of 15 (a) (b) (c) Figure 2. Average concentration of chlorophylls (a+b) in leaves, mg/g. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Figure 2. Average concentration of chlorophylls (a+b) in leaves, mg/g. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). The “Gastronome” variety showed the greatest increase in the concentration of chloro- phylls (by 17%) and carotenoids (by 25%) in the group with UV-A radiation on the 45th day of the cultivation. In the other variants of the UV irradiation, there was no signiﬁcant differences from the control group. The “Markus” variety showed a noticeable UV radiation effect on the average con (a) (b) (b) (a) (c) (c) Figure 2. Average concentration of chlorophylls (a+b) in leaves, mg/g. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and k Figure 2. Average concentration of chlorophylls (a+b) in leaves, mg/g. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). “Markus” (c). The “Gastronome” variety showed the greatest increase in the concentration of chloro- phylls (by 17%) and carotenoids (by 25%) in the group with UV-A radiation on the 45th day of the cultivation. In the other variants of the UV irradiation, there was no signiﬁcant differences from the control group. “Markus” (c). The “Gastronome” variety showed the greatest increase in the concentration of chloro- phylls (by 17%) and carotenoids (by 25%) in the group with UV-A radiation on the 45th day of the cultivation. In the other variants of the UV irradiation, there was no signiﬁcant differences from the control group. The “Markus” variety showed a noticeable UV radiation effect on the average con- centration of leaf total chlorophyll, mainly, on the 30th day of the cultivation. All the UV irradiation sources reduced the total chlorophyll concentration; along with this, the pattern changes in the carotenoid concentration differed from that of the chlorophyll. Carotenoids reduce the stress effect and protect the chlorophyll molecules from destruction caused by the light during the photo-oxidation [30]. 3.1. Analysis of Basil Plants Habitus The exposure to UV-A, UV-B, and UV-C had a signiﬁcant impact on the basil growth and development. It was found that the UV-A and UV-B irradiation did stimulate the basil plants growth and development, although there was observed a variety-speciﬁc reaction of the “Red Ruby” variety to the UV-A and UV-B irradiation. In contrast to above, the UV-C irradiation caused substantial growth inhibition and burns on the leaves of all basil varieties (Figure 1). VIEW 6 of 15 Figure 1. Basil plants of varieties “Red Ruby”, “Gastronome” and “Markus” on 45th and 60th days of cultivation; from left to right: control, UV-A, UV-B, and UV-C. Figure 1. Basil plants of varieties “Red Ruby”, “Gastronome” and “Markus” on 45th and 60th days of cultivation; from left to right: control, UV-A, UV-B, and UV-C. Figure 1. Basil plants of varieties “Red Ruby”, “Gastronome” and “Markus” on 45th and 60th days of cultivation; from left to right: control, UV-A, UV-B, and UV-C. Figure 1. Basil plants of varieties “Red Ruby”, “Gastronome” and “Markus” on 45th and 60th days of cultivation; from left to right: control, UV-A, UV-B, and UV-C. 3 2 Analysis of Pigment Content 3.2. Analysis of Pigment Content 3.2. Analysis of Pigment Content In the case of the “Red Ruby” variety: on the 30th and 45th days of the cultivation, the UV radiation had no significant effect on the chlorophyll concentration in leaves; the average concentration in all groups was 4.8–5.5 mg/g (Figure 2). In all the observed groups, the average carotenoids concentration in the leaves was approximately 1 mg/g (Fi u e 3) No i ifi a t effe t of the UV adiatio o the a e a e a ote oid o e t a In the case of the “Red Ruby” variety: on the 30th and 45th days of the cultivation, the UV radiation had no signiﬁcant effect on the chlorophyll concentration in leaves; the average concentration in all groups was 4.8–5.5 mg/g (Figure 2). In all the observed groups, the average carotenoids concentration in the leaves was approximately 1 mg/g (Figure 3). No signiﬁcant effect of the UV radiation on the average carotenoids concentration in the Appl. Sci. 2021, 11, 8804 6 of 13 leaves was revealed. The ratio of total chlorophyll to carotenoids showed that the plants actively photosynthesized; there were no signs of a physiological aging. EVIEW 7 of 15 leaves was revealed. 3.1. Analysis of Basil Plants Habitus Judging on the change in the ratio of carotenoids and chlorophylls, it can be concluded that the UV-C irradiation caused the greater stress and provoked the protective reaction emergence of the photosynthetic system of the basil plants of the “Markus” variety during the active growth period (on 30th day). (a) (b) p y g g p y Analyzing the anthocyanin concentration in the plant leaves, it is found that in the “Red Ruby” variety, the average anthocyanin concentration in the leaves depends strongly on the exposure of the UV-A, UV-B, and UV-C irradiation. The UV-C radiation had the greatest effect on the 60th day of the cultivation (Figure 4). All groups differed with the UV radiation range, and there was observed a signiﬁcant increase in anthocyanin concentration in leaves between the 45th and 60th day of cultivation by about 20–30%. In the case of the “Gastronome” variety, the average anthocyanin concentration in the leaves on the 60th day of the cultivation under the UV irradiation of different ranges has no signiﬁcant difference from the control group. In the experiment variants of the “Markus” variety, the average anthocyanin concentrations obtained on the 30th, 45th and 60th days of the cultivation are the highest in the variants with the UV-B and UV-C radiation. 7 of 13 Appl. Sci. 2021, 11, 8804 (a) (b) 8 of 15 (c) entration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). h i i i h l l i i f d h i h Figure 3. Average concentration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). (c) Figure 3. Average concentration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Analyzing the anthocyanin concentration in the plant leaves, it is found that in the “Red Ruby” variety, the average anthocyanin concentration in the leaves depends strongly on the exposure of the UV-A, UV-B, and UV-C irradiation. The UV-C radiation had the greatest effect on the 60th day of the cultivation (Figure 4). 3.1. Analysis of Basil Plants Habitus All groups differed with the UV radiation range, and there was observed a significant increase in anthocyanin concentration in leaves between the 45th and 60th day of cultivation by about 20–30%. In the case of the “Gastronome” variety, the average anthocyanin concentration in the leaves on the 60th day of the cultivation under the UV irradiation of different ranges has no sig- nificant difference from the control group. In the experiment variants of the “Markus” variety, the average anthocyanin concentrations obtained on the 30th, 45th and 60th days of the cultivation are the highest in the variants with the UV-B and UV-C radiation. (a) (b) 8 of 15 (c) Figure 3. Average concentration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Analyzing the anthocyanin concentration in the plant leaves, it is found that in the (c) ntration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Figure 3. Average concentration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). y g y p “Red Ruby” variety, the average anthocyanin concentration in the leaves depends strongly on the exposure of the UV-A, UV-B, and UV-C irradiation. The UV-C radiation had the greatest effect on the 60th day of the cultivation (Figure 4). All groups differed with the UV radiation range, and there was observed a significant increase in anthocyanin concentration in leaves between the 45th and 60th day of cultivation by about 20–30%. In the case of the “Gastronome” variety, the average anthocyanin concentration in the leaves on the 60th day of the cultivation under the UV irradiation of different ranges has no sig- nificant difference from the control group. In the experiment variants of the “Markus” variety, the average anthocyanin concentrations obtained on the 30th, 45th and 60th days of the cultivation are the highest in the variants with the UV-B and UV-C radiation. (c) conc the tration of carotenoids in leaves. Obtained on 30th, 45th and 60th days of asil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Figure 3. Average concentration of carotenoids in leaves. 3.1. Analysis of Basil Plants Habitus Obtained on 30th, 45th and 60th days of cultivation of following basil varieties: “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). of the cultivation are the highest in the variants with the UV B and UV C radiation. anthocyanin concentration in the plant leaves, it is found that in the y, the average anthocyanin concentration in the leaves depends osure of the UV-A, UV-B, and UV-C irradiation. The UV-C radiation ect on the 60th day of the cultivation (Figure 4). All groups differed n range, and there was observed a significant increase in anthocyanin ves between the 45th and 60th day of cultivation by about 20–30%. In ronome” variety, the average anthocyanin concentration in the leaves he cultivation under the UV irradiation of different ranges has no sig- rom the control group. In the experiment variants of the “Markus” anthocyanin concentrations obtained on the 30th, 45th and 60th days e the highest in the variants with the UV-B and UV-C radiation. (a) (b) (a) (b) (c) Figure 4. Average anthocyanins concentration on 30th, 45th and 60th day of cultivation of basil “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). y p the average anthocyanin concentration in the leaves depends ure of the UV-A, UV-B, and UV-C irradiation. The UV-C radiation t on the 60th day of the cultivation (Figure 4). All groups differed range, and there was observed a significant increase in anthocyanin s between the 45th and 60th day of cultivation by about 20–30%. In nome” variety, the average anthocyanin concentration in the leaves cultivation under the UV irradiation of different ranges has no sig- m the control group. In the experiment variants of the “Markus” nthocyanin concentrations obtained on the 30th, 45th and 60th days he highest in the variants with the UV-B and UV-C radiation. (a) (b) (b) (c) (c) Figure 4. Average anthocyanins concentration on 30th, 45th and 60th day of cultivation of basil “Red Ruby” (a), “Gastronome” (b) and “Markus” (c). Appl. Sci. 2021, 11, 8804 8 of 13 e varie h The high anthocyanins concentration in the variety “Red Ruby” on the 60th day of the cultivation indicates a greater response to the UV stress caused by the high-intensity lighting and the UV irradiation during budding and ﬂowering than in the varieties “Gas- tronome” and “Markus”. they take on the photo-protective function. 3.3. 3.3. Analysis of Vegetative Indices thocyanins are concentrat In the case of the hyperspectral analysis conducted with the purple-leaved crops, some difﬁculties arise in the indices’ calculation based on the light reﬂection in the spectrum green region. It is well-known that in the basil varieties with purple leaves, the anthocyanins are concentrated in the adaxial and abaxial epidermal cells in form of vacuolar solutions [31]. The anthocyanins can screen more than 90% of the green light giving protection to the mesophyll of the basil leaves from the photo-oxidative stress [32]. Figure 5 shows the reﬂection spectra of the purple and green leaves of the “Markus” basil variety. At a high anthocyanins concentration, the light absorption in the green part of the spectrum increases and so does the reﬂection in the near-infrared range. That is why it is impossible to identify the relationship between the PRI index and the stress states of the purple basil. So, in the future in order to assess the stress conditions, we used the vegetative indices associated with the photosynthetic pigment concentration. solutions [31]. The anthocyanins can screen more than 90% of the gre tection to the mesophyll of the basil leaves from the photo-oxidative shows the reflection spectra of the purple and green leaves of the “Ma At a high anthocyanins concentration, the light absorption in the gre trum increases and so does the reflection in the near-infrared range. T possible to identify the relationship between the PRI index and the purple basil. So, in the future in order to assess the stress conditions, w tive indices associated with the photosynthetic pigment concentration Figure 5 Reflection spectra (A) and appearance (B) of purple and green basil le Figure 5. Reﬂection spectra (A) and appearance (B) of purple and green basil leaves of “Markus” on 30th day of cultivation. Fi 5 R fl ti t (A) d (B) f l d b il l Figure 5. Reﬂection spectra (A) and appearance (B) of purple and green basil leaves of “Markus” on 30th day of cultivation. Figure 5. Reflection spectra (A) and appearance (B) of purple and green basil l 30th day of cultivation. 3.1. Analysis of Basil Plants Habitus Analysis of vegetative indices I th f th h t l l i d t d ith th Most probably, the higher carotenoids concentration in the varieties “Gastronome” and “Markus” is attributed to the fact that due to the lower anthocyanins concentration, they take on the photo-protective function. In the case of the hyperspectral analysis conducted with the pu some difficulties arise in the indices' calculation based on the light ref trum green region. It is well-known that in the basil varieties with pu 3.3. Analysis of Vegetative Indices thocyanins are concentrat Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation. Figure 7. Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation. The analysis revealed the uneven distribution of the indices of R750/R700 for chloro phylls and mARI for anthocyanins over the leaf area (Figures 6 and 7). Similar results were obtained for other indices under study (the data are not provided). This can be attributed Table 3. Average values of vegetative indices for basil varieties “Red Ruby”, “Gastronome” and “Markus” on 30th day of cultivation. The analysis revealed the uneven distribution of the indices of R750/R700 for chlor phylls and mARI for anthocyanins over the leaf area (Figures 6 and 7). Similar results we alues of vegetative indices for basil varieties “Red Ruby”, “Gastronome” and “Markus” on 30th day The analysis revealed the uneven distribution of the indices of R750/R700 for chloro phylls and mARI for anthocyanins over the leaf area (Figures 6 and 7). Similar results were obtained for other indices under study (the data are not provided). This can be attributed Table 3. Average values of vegetative indices for basil varieties “Red Ruby”, “Gastronome” and “Markus” on 30th day of cultivation. to the varietal characteristics of the purple basil, which, depending on a variety, is charac- terized by instability of the purple pigments (anthocyanins) in both individual leaves and the whole plant [33]. Additionally, the uneven distribution of the index is affected by shadows due to uneven lighting and the glare of reflected light from the glossy surface of the leaf. To process the obtained hyperspectral images, the software “Envi 5.2” was used, the vegetative indices average values were obtained for whole plants on the 30th day of cultivation. For analyzing, from images of individual leaves, several areas were selected not containing any glare and shadows; then the average index value for the selected areas was determined (Table 3). 3.3. Analysis of Vegetative Indices thocyanins are concentrat The main advantage of the spectral method based on the reflecti The main advantage of the spectral method based on the reﬂection spectra analysis for the assessment of the pigment content consists in the fact that it is possible to determine their concentrations both for the whole plant and for an individual leaf (Figures 6 and 7). Appl. Sci. 2021, 11, 8804 Appl. Sci. 2021, 11, x FOR 9 of 13 10 of 15 Figure 6. Graphic image of mARI index of basil leaves in false colors on 30th day of cultivation Figure 6. Graphic image of mARI index of basil leaves in false colors on 30th day of cultivation. Figure 6. Graphic image of mARI index of basil leaves in false colors on 30th day of cultivation Figure 6. Graphic image of mARI index of basil leaves in false colors on 30th day of cultivation. The analysis revealed the uneven distribution of the indices of R750/R700 for chloro- phylls and mARI for anthocyanins over the leaf area (Figures 6 and 7). Similar results were obtained for other indices under study (the data are not provided). This can be attributed to the varietal characteristics of the purple basil, which, depending on a variety, is char- acterized by instability of the purple pigments (anthocyanins) in both individual leaves and the whole plant [33]. Additionally, the uneven distribution of the index is affected by shadows due to uneven lighting and the glare of reﬂected light from the glossy surface of the leaf. To process the obtained hyperspectral images, the software “Envi 5.2” was used, the vegetative indices average values were obtained for whole plants on the 30th day of cultivation. For analyzing, from images of individual leaves, several areas were selected not containing any glare and shadows; then the average index value for the selected areas was determined (Table 3). 10 of 13 11 of 15 Appl. Sci. 2021, 11, 8804 Appl. Sci. 2021, 11, x FOR Figure 7. Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation. Figure 7. Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation. Figure 7. Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation. Figure 7. Graphic image of R750/R700 index of basil leaves in false colors on 30th day of cultivation Figure 7. 3.3. Analysis of Vegetative Indices thocyanins are concentrat Plant Variety Sample Group Index ARI Index mARI Index R750/R700 Index MND705 Index CRI700 “Red Ruby” Control 9.07 9.15 3.14 0.52 6.77 UV-A 10.28 10.21 3.22 0.5 9.07 UV-B 12.14 11.36 3.15 0.50 9,47 UV-C 10.73 10.52 3.65 0.59 8.00 “Gastronome” Control 9.81 10.01 2.78 0.49 6.91 UV-A 12.13 11.71 3.59 0.55 8.97 UV-B 9.21 8.73 4.16 0.59 7.94 UV-C 11.87 11.84 3.39 0.55 9.13 “Markus” Control 9.32 9.30 4.30 0.65 6.74 UV-A 9.01 8.99 3.29 0.52 7.79 UV-B 9.48 9.32 3.81 0.57 6.84 UV-C 10.16 10.23 4.48 0.61 6.92 Appl. Sci. 2021, 11, 8804 11 of 13 In order to test the effectiveness of the proposed method, the correlation was estimated between the obtained values of the vegetative indices ARI and mARI and the anthocyanins concentration value in the basil leaves obtained by the chemical method (Figure 8). Figure 8. Evaluation of correlation between pigment concentration and vegetation indices obtained on 30th day of cultivation. (a) Correlation between CRI700 index and carotenoids concentration (b) correlation between the R750/R500 index and a+b chlorophylls concentration; (c) correlation between MND705 index and a+b chlorophylls concentration; (d) correlation between ARI index and anthocyanins concentration; (e) correlation between mARI index and anthocyanins concentration. Figure 8. Evaluation of correlation between pigment concentration and vegetation indices obtained on 30th day of cultivation. (a) Correlation between CRI700 index and carotenoids concentration; (b) correlation between the R750/R500 index and a+b chlorophylls concentration; (c) correlation between MND705 index and a+b chlorophylls concentration; (d) correlation between ARI index and anthocyanins concentration; (e) correlation between mARI index and anthocyanins concentration. The determination coefﬁcients R2 = 0.82 for ARI and R2 = 0.85 for mARI indicate a high degree of correlation, which indicates the acceptability of this method for the anthocyanins content assessment in plant leaves. For the carotenoids, a weak correlation was observed (R2 = 0.0479), which is explained by the shielding effect of the anthocyanins as far as the CRI700 index is determined at the reﬂection band of 510 nm, which coincides with the maximum absorption of the anthocyanins [32]. The weak correlations of the R750/R700 and MND705 indices obtained for the chlorophylls can be explained by the large unevenness of the reﬂection spectra over the leaf area (Figure 6), which is typical for the purple basil. After all, on the crops with the green leaves, the strong correlations of these indices with the chlorophyll content were revealed [34]. 3.3. Analysis of Vegetative Indices thocyanins are concentrat Additionally, the deviations present in the results are inﬂuenced by such factors as the geometry of a leaf, the uneven lighting, the morphological features (thickness and density of the sheet), etc. Appl. Sci. 2021, 11, 8804 12 of 13 For a more accurate assessment of the stress factors’ impact on the concentrations of the chlorophylls and the carotenoids in the purple-leaved crops and on the overall productivity, it is necessary to conduct some additional research and develop such reﬂection indices that take into account the shielding effect of the purple pigments. Data Availability Statement: No additional data available. Conﬂicts of Interest: The authors declare no conﬂict of interest. The funders had no role in the design of the study, in the collection, analyses, or interpretation of the data, in the writing of the manuscript, or in the decision to publish the results. 8. Skarupova, D.; Vostalova, J.; Svobodova, A.R. Ultraviolet A protective potential of plant extracts and phytochemicals. Biomed. Pap. Med. Fac. Univ. Palacky Olomouc Czech Repub. 2020, 164, 1–22. [CrossRef] [PubMed] 6. Bornman, J.F.; Barnes, P.W.; Robinson, S.A.; Ballaré, C.L.; Flinte, S.D.; Caldwell, M.M. Solar ultraviolet radiation and ozone depletion-driven climate change: Effects on terrestrial ecosystems. Photochem. Photobiol. Sci. 2015, 14, 88–107. [CrossRef] 7. Takshak, S.; Agrawal, S. Defense potential of secondary metabolites in medicinal plants under UV-B stress. J. Photochem. Photobiol. B Biol. 2019, 193, 51–88. [CrossRef] p 5. Rafanelli, C.; Damiani, A.; Benedetti, E.; De Simone, S.; Anav, A.; Ciattaglia, L.; Di Menno, I. UV Solar Radiation in Polar Regions: Consequences for the Environment and Human Health. UV Radiat. Glob. Clim. Chang. 2010, 73–105. [CrossRef] 4. Neale, R.E.; Barnes, P.W.; Robson, T.M.; Neale, P.J.; Williamson, C.E.; Zepp, R.G.; Wilson, S.R.; Madronich, S.; Andrady, A.L.; Heikkilä, A.M.; et al. Environmental effects of stratospheric ozone depletion, UV radiation, and interactions with climate change: UNEP Environmental Effects Assessment Panel, Update 2020. Photochem. Photobiol. Sci. 2021, 20, 1–67. [CrossRef] [PubMed] 4. Conclusions The highest concentration of anthocyanins in leaves was recorded for the “Red Ruby” variety under continuous UV-C irradiation (30% higher than in controls); it was relatively lower in the samples under UV-B and UV-A (20–25% higher than in controls). Most probably, the higher anthocyanins concentration in the case of the “Red Ruby” variety compared to the “Gastronome” and “Markus” varieties under the UV irradiation of the studied ranges indicates a greater adaptability of the “Red Ruby” variety to the stresses caused by the UV radiation. The strong correlation between the obtained values of the vegetative indices ARI and mARI and the anthocyanins concentration in the basil leaves shows the possibility of using hyperspectral cameras to study the biochemical processes occurring during the plants’ vegetation. These results can be used for the development of light growing quantitative recommendations, for productivity increasing, and for food products obtaining with a high content of the anthocyanins. However, for the indices of chlorophyll and carotenoids, the correlations are found to be low, which is primarily due to the shielding effect of the anthocyanins and the uneven color of the leaves. Author Contributions: Conceptualization and methodology, A.A.S. and Y.A.P.; validation, N.A.S.; formal analysis, Y.A.P. and N.A.S.; investigation and resources, A.S.I.; writing—original draft prepa- ration, Y.A.P. and A.A.S.; writing—review and editing, N.A.S. and A.A.S., visualization, D.A.B.; supervision and funding acquisition, A.S.D. and V.A.P.; project administration, A.S.D. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by a grant of the Ministry of Science and Higher Education of the Russian Federation for large scientiﬁc projects in priority areas of scientiﬁc and technological development (subsidy identiﬁer 075-15-2020-774). Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: No additional data available. Data Availability Statement: No additional data available. q g 6. Bornman, J.F.; Barnes, P.W.; Robinson, S.A.; Ballaré, C.L.; Flinte, S.D.; Caldwell, M.M. Solar ultraviolet radiation and ozone depletion-driven climate change: Effects on terrestrial ecosystems. Photochem. Photobiol. Sci. 2015, 14, 88–107. [CrossRef] 7 Takshak S ; Agrawal S Defense potential of secondary metabolites in medicinal plants under UV B stress J Photochem Photobiol References The potential use of the UV-A and UV-B to improve tomato quality and preference for consumers. Sci. Hortic. 2019, 246, 777–784. [CrossRef] p q y p 12. Pandey, M.; Srivastava, A.K.; Suprasanna, P.; D’Souza, S.F. Thiourea mediates alleviation of UV-B stres Indian mustard (Brassica juncea L.). J. Plant Interact. 2012, 7, 143–150. [CrossRef] 13. Boghdady, M.S.; Desoky, E.M.; Azoz, S.N.; Nassar, D.M.A. Effect of Selenium on Growth, Physiological Aspects and Productivity of Faba Bean (Vicia faba L.). Egypt. J. Agron. 2017, 39, 83–97. [CrossRef] f gyp g 14. Germ, M.; Stibilj, V.; Šircelj, H.; Jerše, A.; Kroﬂiˇc, A.; Golob, A.; Marši´c, N.K. Biofortiﬁcation of common buckwheat microgreens and seeds with different forms of selenium and iodine. J. Sci. Food Agric. 2019, 99, 4353–4362. [CrossRef] [PubMed] 15. Golob, A.; Novak, T.; Marsic, N.K.; Sircelj, H.; Stibilj, V.; Jerse, A.; Kroﬂic, A.; Germ, M. Biofortiﬁcation with selenium and iodine changes morphological properties of Brassica oleracea L. var. gongylodes) and increases their contents in tubers. Plant Physiol. Biochem. 2020, 150, 234–243. [CrossRef] 16. Sytar, O.; Zivcak, M.; Neugart, S.; Brestic, M. Assessment of hyperspectral indicators related to the content of phenolic compounds and multispectral ﬂuorescence records in chicory leaves exposed to various light environments. Plant Physiol. Biochem. 2020, 154, 429–438. [CrossRef] 17. Alt, V.V.; Gurova, T.A.; Elkin, O.V.; Klimenko, D.N.; Maximov, L.V.; Pestunov, I.A.; Dubrovskaya, O.A.; Genaev, M.A.; Erst, T.V.; Genaev, K.A.; et al. The use of Specim IQ, a hyperspectral camera, for plant analysis. Vavilov J. Genet. Breed. 2020, 24, 259–266. [CrossRef] 18. An, G.Q.; Xing, M.F.; He, B.B.; Liao, C.H.; Huang, X.D.; Shang, J.L.; Kang, H.Q. Using Machine Learning for Estimating Rice Chlorophyll Content from In Situ Hyperspectral Data. Remote Sens. 2020, 12, 3104. [CrossRef] 19. Zhang, J.; Tian, H.Q.; Wang, D.; Li, H.; Mouazen, A.M. A novel spectral index for estimation of relativ sugar beet. Comput. Electron. Agric. 2021, 184, 106088. [CrossRef] g p g 20. Li, L.; Zhang, Q.; Huang, D.F. A Review of Imaging Techniques for Plant Phenotyping. Sensors 2014, 14, 20078–20111. [CrossRef] 20. Li, L.; Zhang, Q.; Huang, D.F. A Review of Imaging Techniques for Plant Phenotyping. Sensors 2014, 14, 20078–20111. [CrossRef] 21. Behmann, J.; Steinrucken, J.; Plumer, L. Detection of early plant stress responses in hyperspectral images. ISPRS J. Photogramm. 20. Li, L.; Zhang, Q.; Huang, D.F. A Review of Imaging Techniques for Plant Phenotyping. Sensors 2014, 14, 20078–20111. [CrossRef] 21. References 1. Ganzeveld, L.; Bouwman, L.; Stehfest, E.; van Vuuren, D.P.; Eickhout, B.; Lelieveld, J. Impact of future land use and land cover changes on atmospheric chemistry-climate interactions. J. Geophys. Res. Atmos. 2010, 115, D23301. [CrossRef] 2. Parrish, D.D.; Derwent, R.G.; Staehelin, J. Long-term changes in northern mid-latitude tropospheric ozone concentrations: 1. Ganzeveld, L.; Bouwman, L.; Stehfest, E.; van Vuuren, D.P.; Eickhout, B.; Lelieveld, J. Impact of future land use and land cover changes on atmospheric chemistry-climate interactions. J. Geophys. Res. Atmos. 2010, 115, D23301. [CrossRef] 1. Ganzeveld, L.; Bouwman, L.; Stehfest, E.; van Vuuren, D.P.; Eickhout, B.; Lelieveld, J. Impact of future land use and land cover changes on atmospheric chemistry-climate interactions. J. Geophys. Res. Atmos. 2010, 115, D23301. [CrossRef] , ; , ; , ; , ; , ; , J p changes on atmospheric chemistry-climate interactions. J. Geophys. Res. Atmos. 2010, 115, D23301. [CrossRef] 2. Parrish, D.D.; Derwent, R.G.; Staehelin, J. Long-term changes in northern mid-latitude tropospheric ozone concentrations: Synthesis of two recent analyses. Atmos. Environ. 2021, 248, 118227. [CrossRef] g p y p y 2. Parrish, D.D.; Derwent, R.G.; Staehelin, J. Long-term changes in northern mid-latitude tropospheric ozone concentrations: Synthesis of two recent analyses. Atmos. Environ. 2021, 248, 118227. [CrossRef] 2. Parrish, D.D.; Derwent, R.G.; Staehelin, J. Long-term changes in northern mid Synthesis of two recent analyses. Atmos. Environ. 2021, 248, 118227. [CrossRef] 3. Bodeker, G.E.; Kremser, S. Indicators of Antarctic ozone depletion: 1979 to 2019. Atmos. Chem. Phys. 2021, 21, 5289–5300. [CrossRef] 4. Neale, R.E.; Barnes, P.W.; Robson, T.M.; Neale, P.J.; Williamson, C.E.; Zepp, R.G.; Wilson, S.R.; Madronich, S.; Andrady, A.L.; Heikkilä, A.M.; et al. Environmental effects of stratospheric ozone depletion, UV radiation, and interactions with climate change: UNEP Environmental Effects Assessment Panel, Update 2020. Photochem. Photobiol. Sci. 2021, 20, 1–67. [CrossRef] [PubMed] Appl. Sci. 2021, 11, 8804 13 of 13 13 of 13 9. Emiliani, J.; D’Andrea, L.; Falcone Ferreyra, M.L.; Maulión, E.; Rodríguez, E.J.; Rodriguez-Concepcion, M.; Casati, P. A role for beta, beta-xanthophylls in Arabidopsis UV-B photoprotection. J. Exp. Bot. 2018, 69, 4921–4933. [CrossRef] [PubMed] 10. Pandey, A.; Jaiswal, D.; Agrawal, S.B. Ultraviolet-B mediated biochemical and metabolic responses of a medicinal plant Adhatoda vasica Nees at different growth stages. J. Photochem. Photobiol. B Biol. 2021, 216, 112142. [CrossRef] 11. Mariz-Ponte, N.; Martins, S.; Gonçalves, A.; Correia, C.M.; Ribeiro, C.; Celeste Dias, M.; Santos, C. References Behmann, J.; Steinrucken, J.; Plumer, L. Detection of early plant stress responses in hyperspectral images. ISPRS J. Photogramm. Remote Sens. 2014, 93, 98–111. [CrossRef] 21. Behmann, J.; Steinrucken, J.; Plumer, L. Detection of early plant stress responses in hyperspectral images. ISPRS J. Photogramm. Remote Sens. 2014, 93, 98–111. [CrossRef] 22. Proctor, C.; Dao, P.D.; He, Y.H. Close-range, heavy-duty hyperspectral imaging for tracking drought im SINE model. J. Quant. Spectrosc. Radiat. Transf. 2021, 263, 107528. [CrossRef] 23. Feng, X.P.; Yu, C.L.; Chen, Y.; Peng, J.Y.; Ye, L.H.; Shen, T.T.; Wen, Y.; He, Y. Non-destructive Determination of Shikimic Acid Concentration in Transgenic Maize Exhibiting Glyphosate Tolerance Using Chlorophyll Fluorescence and Hyperspectral Imaging. Front. Plant Sci. 2018, 9, 468. [CrossRef] [PubMed] 24. Blackburn, G.A. Hyperspectral remote sensing of plant pigments. J. Exp. Bot. 2007, 58, 855–867. [C 25. Gitelson, A.A.; Chivkunova, O.B.; Merzlyak, M.N. Nondestructive estimation of anthocyanins and chlorophylls in anthocyanic leaves. Am. J. Bot. 2009, 96, 1861–1868. [CrossRef] 26. Semenova, N.A.; Smirnov, A.; Grishin, A.; Pishchalnikov, R.Y.; Chesalin, D.D.; Gudkov, S.V.; Chilingaryan, N.O.; Skorokhodova, A.N.; Dorokhov, A.S.; Izmailov, A.Y. The Effect of Plant Growth Compensation by Adding Silicon-Containing Fertilizer under Light Stress Conditions. Plants 2021, 10, 1287. [CrossRef] g [ ] 27. Maslennikov, P.; Chupakhina, G.; Skrypnick, L.; Fedurayev, P.; Poltavskaya, R. Content of anthocyanin and carotenoid pigments in medicinal plants. Bull. Mosc. Reg. State Univ. 2013, 1, 6. p g 28. Gitelson, A.A.; Merzlyak, M.N.; Chivkunova, O.B. Optical properties and nondestructive estimation of anthocyanin content in plant leaves. Photochem. Photobiol. 2001, 74, 38–45. [CrossRef] 29. Ferri, C.P.; Formaggio, A.R.; Schiavinato, M.A. Narrow band spectral indexes for chlorophyll determination in soybean canopies [Glycine max (L.) Merril]. Braz. J. Plant Physiol. 2004, 16, 131–136. [CrossRef] ashevsky, N.D. The structure of the photosynthetic apparatus of the plant genues Tamarix which grows in the Nat. Sci. 2013, 2, 11–13. 30. Afanasyev, I.A.; Smashevsky, N.D. The structure of the photosynthetic apparatus of the plant genues Ta Astrakhan region. Nat. Sci. 2013, 2, 11–13. g 31. Landi, M.; Guidi, L.; Pardossi, A.; Tattini, M.; Gould, K.S. Photoprotection by foliar anthocyanins mitigates effects of boron toxicity in sweet basil (Ocimum basilicum). Planta 2014, 240, 941–953. [CrossRef] y 32. Nichelmann, L.; Bilger, W. Quantiﬁcation of light screening by anthocyanins in leaves of Berberis thunbergia. Planta 2017, 246, 1069–1082. [CrossRef] [PubMed] 32. Nichelmann, L.; Bilger, W. Quantiﬁcation of light screening by anthocyanins in leaves of Berberis thunbergia. References Planta 2017, 246, 1069–1082. [CrossRef] [PubMed] 33. Phippen, W.B.; Simon, J.E. Anthocyanin inheritance and instability in purple basil (Ocimum basilicum L.). J. Hered. 2000, 91, 289–296. [CrossRef] [PubMed] 33. Phippen, W.B.; Simon, J.E. Anthocyanin inheritance and instability in purple basil (Ocimum basilicum L.). J. Hered. 2000, 91, 289–296. [CrossRef] [PubMed] 34. Sims, D.A.; Gamon, J.A. Relationships between leaf pigment content and spectral reﬂectance across a wide range of species, leaf structures and developmental stages. Remote Sens. Environ. 2002, 81, 337–354. [CrossRef] 34. Sims, D.A.; Gamon, J.A. Relationships between leaf pigment content and spectral reﬂectance across a wide range of species, leaf structures and developmental stages. Remote Sens. Environ. 2002, 81, 337–354. [CrossRef]
https://openalex.org/W1969764409	https://www.scielo.br/j/eagri/a/8dCxShnwjjpntKtKQWq9HSM/?lang=pt&format=pdf	Portuguese	null	Decomposição de culturas de cobertura no sistema plantio direto, manejadas mecânica e quimicamente	Engenharia agrícola	2,007	cc-by	3,708	1 Pós-Graduando, UNESP, Jaboticabal - SP, Bolsista Capes, gn_reis@yahoo.com.br 2 Prof. Adjunto, Departamento de Engenharia Rural, UNESP, Jaboticabal - SP. Bolsista de Produtividade CNPq. 3 Prof. Dr., Departamento de Engenharia Rural, UNESP, Jaboticabal - SP. Bolsista de Produtividade CNPq. 4Engenheiro Agrônomo. Recebido pelo Conselho Editorial em: 18-4-2006 Aprovado pelo Conselho Editorial em: 16-10-2006 E A í J b i b l 27 1 194 2 DECOMPOSIÇÃO DE CULTURAS DE COBERTURA NO SISTEMA PLANTIO DIRETO, MANEJADAS MECÂNICA E QUIMICAMENTE GUSTAVO N. DOS REIS1, CARLOS E. A. FURLANI2, ROUVERSON P. DA SILVA3, JULIANO R. GERLACH4, JORGE W. CORTEZ1, DANILO C. C. GROTTA1 RESUMO: O sistema plantio direto preconiza o uso de culturas de cobertura para produção de massa seca sobre o solo. O presente trabalho teve por objetivo estudar a produção e a taxa de decomposição de duas coberturas vegetais - crotalária juncea (Crotalária juncea L.) e mucuna- cinza (Stilozobium niveum L) - em três sistemas de manejo (rolo-faca, triturador de palhas e herbicida), e a capacidade de campo efetiva dos manejos. O experimento foi realizado na área do Laboratório de Máquinas e Mecanização Agrícola (LAMMA), do Departamento de Engenharia Rural, UNESP, Jaboticabal - SP, com delineamento em blocos ao acaso, em esquema fatorial 3x2, originando seis tratamentos, com quatro repetições. Foram analisadas as produções de massa seca das culturas de cobertura antes e após o manejo, a velocidade de deslocamento e a capacidade de campo efetiva para cada conjunto (trator-equipamento). As análises dos valores obtidos permitiram verificar que o fator manejo não interfere na decomposição da massa seca das coberturas vegetais e que as duas culturas de cobertura apresentaram massas semelhantes aos 30; 70 e 125 dias após a semeadura, diferindo aos 97 dias, época na qual a crotalária apresentou maior quantidade de massa seca. Aos 30; 51 e 71 dias após o manejo, as massas secas das culturas foram semelhantes. O manejo com herbicida apresentou maior capacidade de campo efetiva. PALAVRAS-CHAVE: mucuna-cinza, Crotalária juncea, capacidade de campo efetiv KEYWORDS: mucuna-cinza, crotalária juncea, effective field capacity. 1 Pós-Graduando, UNESP, Jaboticabal - SP, Bolsista Capes, gn_reis@yahoo.com.br 2 Prof. Adjunto, Departamento de Engenharia Rural, UNESP, Jaboticabal - SP. Bolsista de Produtividade CNPq. 3 Prof. Dr., Departamento de Engenharia Rural, UNESP, Jaboticabal - SP. Bolsista de Produtividade CNPq. 4Engenheiro Agrônomo. INTRODUÇÃO O Sistema Plantio Direto (SPD) é caracterizado pela permanência de palha, rotação de culturas e restos vegetais na superfície do solo, que posteriormente são revolvidos apenas na fileira de deposição de sementes e fertilizantes, com o controle de plantas daninhas realizado com a aplicação de herbicidas e pelos efeitos alelopáticos de cada planta de cobertura. A manutenção de resíduos culturais na superfície do solo no SPD promove condições distintas às propriedades do solo em relação à incorporação promovida pelo sistema de preparo convencional. O revolvimento mínimo do solo leva à decomposição mais lenta e gradual da matéria orgânica, apresentando como conseqüência alterações físicas, químicas e biológicas no solo, repercutidas na sua fertilidade e na produtividade das culturas, características já observadas por MOODY et al. (1961). A diversificação de plantas com sistemas radiculares capazes de explorar diferentes profundidades do solo, com potencial diferenciado de reciclagem, proporciona melhor equilíbrio dos nutrientes e incremento na qualidade e na atividade biológica do solo (EMBRAPA, 1998). Segundo SILVA et al. (1999), devem ser escolhidas, para cobertura, espécies que produzam grande quantidade de matéria seca, resistentes ao ataque de pragas e moléstias, com sementes uniformes de bom poder germinativo e exigência relativamente baixa quanto ao preparo e fertilidade do solo, rápido crescimento, fácil manejo e sistema radicular profundo, dispensando tratos culturais. Os meios mecânicos de corte das culturas de cobertura podem variar em função do equipamento disponível na propriedade e das adaptações e regulagens efetuadas nesses equipamentos. Para tanto, poderão ser utilizados equipamentos, como alfanje, foice ou gadanho; triturador de palhas; roçadora ou rolo-faca. No caso de plantas perenes, que rebrotam o ano todo, os cortes poderão ocorrer com maior freqüência durante o ano agrícola (SILVEIRA et al., 1985). De acordo com BENEZ (2002), para aperfeiçoar o uso da máquina, o agricultor deve, inicialmente, realizar a análise operacional para verificar quais são as operações que representam as reais condições da propriedade. O desempenho econômico do sistema tratorizado é diretamente proporcional a sua capacidade operacional, ou seja, quanto maior seu valor, menores serão os custos, melhorando economicamente o trabalho do sistema (MIALHE, 1974). Pressupõe-se que os manejos de diferentes culturas de coberturas influenciam na taxa de decomposição da massa seca sobre a superfície do solo, refletindo na velocidade de deslocamento e, conseqüentemente, na capacidade de campo efetiva do conjunto utilizados nesses manejos. INTRODUÇÃO O presente trabalho teve por objetivo estudar a produção e a taxa de decomposição de duas coberturas vegetais (crotalária juncea e mucuna-cinza) em três sistemas de manejo (rolo-faca, triturador de palhas e herbicida), e verificar qual equipamento proporciona melhor capacidade de campo efetiva. ACTING OF COVERING CULTURES IN NO TILLAGE SYSTEM, HANDLING MECHANICS AND CHEMISTRY ABSTRACT: The no tillage using covering cultures about the production dry mass surface in the soil. The present work had for objective to study the production and acting of two vegetable coverings crotalária juncea (Crotalaria juncea L.) and mucuna-cinza (Stilozobium niveum), in three systems of handlings (roll-knife, crusher of straws and herbicide), and and the effective field capacity. The experiment was accomplished in area of Laboratório de Máquinas e Mecanização Agrícola (LAMMA), of the Departamento de Engenharia Rural, UNESP; Jaboticabal - SP, with experimental design consisted of a factorial array 3x2 with completely randomized blocks, origin six treatments with four repetitions. They were measuring and analyzed parameters of production of dry mass of the coverings vegetable after the sowing and handling, displacement speed and the effective field capacity execute for each group (tractor-equipment). The analysis of the obtained values allowed verifying that the factor handling didn’t interfere in the mass it evaporates of covering vegetable. The cultures present similar masses to the 30; 71 and 125 days after the sowing, differing to the 97 days, was the Crotalaria juncea L. present larger mass earnings. In the powders handling the two cultures have masses similar to the 30; 51 and 71 days after the handling, in other words, during whole the experiment. The handling with herbicide present larger effective field capacity. KEYWORDS: mucuna-cinza, crotalária juncea, effective field capacity. Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 195 Decomposição de culturas de cobertura no sistema plantio direto Decomposição de culturas de cobertura no sistema plantio direto TABELA 1. Descrição das culturas de cobertura utilizadas. TABELA 1. Descrição das culturas de cobertura utilizadas. Pureza Germinação Valor Cultural Espécie Quantidade (kg ha-1) (%) Crotalária (Crotalaria juncea L.) 25 98,5 76 74,6 Mucuna (Stilozobium niveum L.) 60 98,3 81 79,6 TABELA 1. Descrição das culturas de cobertura utilizadas. ELA 1. Descrição das culturas de cobertura utilizadas. Utilizou-se fertilizante da fórmula 4-20-20 na dose de 150 kg ha-1 na semeadura de crotalária e mucuna, no dia 6-4-2004, e herbicida gliphosate em pré-emergência, na dose de 3 e 6 L ha-1 para o manejo químico das culturas de cobertura. Os manejos foram realizados após a ultima coleta da massa seca (125 dias após a semeadura). Durante a instalação e a condução deste trabalho, foram utilizadas as máquinas e implementos descritos na Tabela 2. TABELA 2. Equipamentos utilizados. Tratores Marca/ Modelo Potência Características Ford 6600 63 kW (86 cv) 4x2 MF 620 77 kW (105 cv) 4x2 TDA Tratores MF 660 110 kW (150 cv) 4x2 TDA Equipamentos Marca/ Modelo Características Semeadora Jumil 2650 Exacta Quatro linhas de plantio, espaçadas de 0,5 m. Pulverizador Jacto Condor Tanque de 600 L, 24 bicos tipo leque, largura de 12 m. Triturador Jumil Trimax 2500 Montado, largura de corte de 2,3 m. Rolo-faca LAMMA Montado, largura de corte de 1,6 m, cilindro de 1 m de diâmetro, massa de 1.360 kg e distância entre facas de 0,2 m. TABELA 2. Equipamentos utilizados. Para a determinação da massa seca, utilizou-se de um quadrado de ferro de 0,25 m2, tesoura de poda, estufa elétrica e balança de precisão de 0,01 g. As avaliações aos 30; 60; 97 e 125 dias após a semeadura (DAS) e aos 30; 51 e 71 dias após o manejo (DAM) foram feitas coletando-se uma amostra por parcela, secada em estufa elétrica a 70 o C até atingir massa constante. A largura efetiva de trabalho dos equipamentos foi medida diretamente sobre as passadas dos equipamentos nas parcelas e suas médias foram: rolo-faca (1,60 m), triturador de palhas (2,15 m) e herbicida (12,00 m). A partir do espaço percorrido e do tempo de percurso medido por meio de cronômetro digital em todas as parcelas experimentais, determinou-se a velocidade real de deslocamento. MATERIAL E MÉTODOS O experimento foi conduzido na área experimental do Laboratório de Máquinas e Mecanização Agrícola (LAMMA) da UNESP - Jaboticabal - SP, localizada nas coordenadas geodésicas: latitude 21º14’S e longitude 48º16’W, com altitude média de 560 m, declividade média de 4% e clima Cwa (subtropical), de acordo com a classificação de Köeppen. O solo da área experimental foi classificado como Latossolo Vermelho Eutroférrico, conforme o Sistema Brasileiro de Classificação de Solos (EMBRAPA, 1999). O delineamento experimental foi de blocos casualizados, em esquema fatorial 3x2, originando seis tratamentos, com quatro repetições, totalizando 24 observações. Os tratamentos Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 Gustavo N. dos Reis, Carlos E. A. Furlani, Rouverson P. da Silva et al. 196 foram três métodos de manejo (rolo-faca, triturador de palhas e herbicida), em duas coberturas vegetais - crotalária juncea (Crotalária juncea L.) e mucuna-cinza (Stilozobium niveum L). Cada parcela experimental ocupou área de 300 m2 (25 x 12 m), distribuídas em quatro blocos, e entre as parcelas, no sentido longitudinal, foi reservado espaço de 15 m destinado à realização de manobras, tráfego de máquinas e estabilização dos conjuntos. Na semeadura, foram utilizadas as espécies e suas respectivas quantidades de sementes por unidade de área, conforme Tabela 1. TABELA 1. Descrição das culturas de cobertura utilizadas. No intuito de estabilizar as determinações, o movimento começava antes da baliza de início da parcela, e o tempo era cronometrado quando o referencial no trator coincidia sobre esta, e interrompido depois de decorridos os 25 m da parcela, ou seja, quando o referencial no trator coincidia com a baliza final. Para cada conjunto, ao longo do trabalho, foram selecionadas marchas distintas, de acordo com a recomendação do fabricante (Tabela 3), no intuito de obter o melhor desempenho dos órgãos ativos dos equipamentos. Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 Decomposição de culturas de cobertura no sistema plantio direto Decomposição de culturas de cobertura no sistema plantio direto Decomposição de culturas de cobertura no sistema plantio direto 197 TABELA 3. Relação de tratores e velocidades de trabalho. Trator Manejo Velocidade (km h-1) FORD 6600 (4x2) Herbicida 7,0 MF 660 (4x2 TDA) Rolo-faca 7,5 MF 620 (4x2 TDA) Triturador de palhas 7,0 Manejo Velocidade (km h-1) Herbicida 7,0 Rolo-faca 7,5 iturador de palhas 7,0 A capacidade de campo efetiva foi determinada segundo MIALHE (1974), com base na largura de trabalho real do equipamento e velocidade real de deslocamento proporcionada pela marcha selecionada em cada manejo. Para o cálculo da capacidade de campo efetiva, utilizou-se da seguinte equação: 10 Vrt Lrt Cce = (1) 10 Vrt Lrt Cce = (1) 10 Vrt Lrt Cce = (1) em que, em que, q Cce - capacidade de campo efetiva, ha h-1; 1 Vrt - velocidade real de trabalho, km h-1; Lrt - largura real de trabalho do equipamento, m, e 10 - fator de conversão. édias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tuke obabilidade; significativo (p≤0,01); C.V. - coeficiente de variação; NS - não-significativo. Médias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de probabilidade; significativo (p≤0,01); C.V. - coeficiente de variação; NS - não-significativo. RESULTADOS E DISCUSSÃO As produções de matéria seca foram de 981 e 415 kg ha-1 aos 30 dias, e 8.175 e 5.092 kg ha-1 aos 60 dias para crotalária e mucuna, respectivamente. Aos 90 dias após a semeadura, as culturas de coberturas não diferiram entre si. Resultado similar foi encontrado por MACIEL (2002) para as culturas de coberturas crotalária e mucuna-cinza, com a primeira apresentando maior produção aos 30 e 60 dias após a semeadura. As produções de massa seca de crotalária foram de 6.065 e 5.628 kg ha-1 aos 30 e 60 DAS e da mucuna de 2.537 e 2.921 kg ha-1 aos 30 e 60 DAS, respectivamente. Porém, aos 90 DAS, a crotalária e a mucuna-cinza não diferiram quanto à produção da massa seca. O autor afirma que isso se deve ao rápido crescimento inicial da crotalária, mas que, ao final do ciclo, a mucuna consegue igualar a produção de massa seca. De acordo com a Tabela 5, observa-se semelhança estatística entre os fatores estudados aos 30; 51 e 71 DAM (dias após manejo). A massa seca das culturas de cobertura não teve influência dos tratamentos até mais de dois meses após os manejos, apesar de as parcelas de triturador e rolo- faca terem apresentado maior produção de massa ao final do ciclo. Esses dados possibilitam a utilização de qualquer equipamento de manejo apresentado neste trabalho, ou seja, pode-se utilizar o pulverizador (herbicida) que confere alta capacidade de campo efetiva (Tabela 6) ou, então, os equipamentos de manejo mecânico que apresentam menor capacidade de campo efetiva, porém independem das condições de umidade que o herbicida necessita para ser aplicado e translocado na planta. Para BARROS (2002), as culturas de cobertura crotalária e mucuna-cinza diferiram após os manejos, apresentando 10.000 e 7.300 kg ha-1 de massa seca no solo, respectivamente. MACIEL (2002), avaliando diferentes sistemas de manejo (triturador de palhas, ceifadora de tambor e pulverizador), não encontrou diferença na quantidade de massa seca aos 30; 60; 90 e 120 dias após a semeadura. Observa-se que ocorreu aumento de massa seca após 30 DAM em relação aos 125 DAS, devido à presença de massa seca no solo antes do manejo, fazendo com que se acarretasse aumento aproximado de 700 kg. TABELA 5. Síntese de análise de variância e do teste de médias para massa seca da parte aérea das coberturas vegetais aos 30; 51 e 71 DAM (dias após manejo). RESULTADOS E DISCUSSÃO Pela Tabela 4, observa-se que as parcelas nas quais seriam montados os manejos e as culturas de coberturas não afetaram a produção de massa seca da parte aérea aos 30 e 70 DAS (dias após semeadura). Para a massa seca coletada aos 97 DAS, houve diferença significativa entre as culturas de cobertura, com a crotalária apresentando maior quantidade de massa em relação à mucuna, fato esse relacionado à escassez de chuvas no período e ao rápido crescimento inicial da crotalária. Para as parcelas que receberiam os manejos, nota-se diferença significativa para o triturador de palhas em relação aos demais, sendo esse o de menor acúmulo de massa seca. Aos 125 DAS, houve semelhança estatística entre as parcelas nas quais seriam montados os manejos com triturador de palhas e rolo-faca, diferenciando-se da parcela que receberia o manejo com herbicida, que proporcionou maior quantidade de massa seca. TABELA 4. Síntese de análise de variância e do teste de médias para massa seca da parte aérea das coberturas vegetais aos 30; 70; 97 e 125 DAS (dias após a semeadura). Massa seca (kg ha-1) Fatores 30 DAS 70 DAS 97 DAS 125 DAS Manejos Triturador 425 a 3.265 a 2.854 b 3.222 b Rolo-faca 487 a 3.280 a 3.815 a 3.182 b Herbicida 482 a 3.655 a 3.902 a 3.512 a Culturas Mucuna 459 a 3.318 a 3.107 b 3.283 a Crotalária 471 a 3.481 a 3.940 a 3.328 a Teste F Manejos (M) 0,52NS 1,27NS 6,84 7,63 Culturas (C) 0,05NS 0,52NS 10,52 0,36NS M x C 0,05NS 1,46NS 2,14NS 2,92NS C.V. (%) 28,9 16,3 17,8 5,5 Médias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tukey, a 5% de probabilidade; significativo (p≤0,01); C.V. - coeficiente de variação; NS - não-significativo. TABELA 4. Síntese de análise de variância e do teste de médias para massa seca da pa coberturas vegetais aos 30; 70; 97 e 125 DAS (dias após a semeadura). Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 198 Gustavo N. dos Reis, Carlos E. A. Furlani, Rouverson P. da Silva et al. BARROS (2002), ao analisar a produção de matéria seca da parte aérea da crotalária e da mucuna-cinza aos 30; 60 e 90 dias após a semeadura, verificou que, nas duas primeiras avaliações, a crotalária foi superior estatisticamente à mucuna. édias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tuke obabilidade; C.V. - coeficiente de variação; NS - não-significativo. édias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tuke obabilidade; C V coeficiente de variação; NS não significativo seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tukey, a 5% dade; C.V. - coeficiente de variação; NS - não-significativo. RESULTADOS E DISCUSSÃO Massa Seca (kg ha-1) Fatores 30 DAM 51 DAM 71 DAM Manejos Triturador 4.145 a 3.052 a 2.371 a Rolo-faca 4.002 a 3.697 a 2.707 a Herbicida 4.015 a 3.987 a 2.542 a Culturas Mucuna 4.048 a 3.663 a 2.629 a Crotalária 4.046 a 3.495 a 2.451 a Teste F Manejos (M) 0,43NS 3,08NS 1,11NS Culturas (C) 0,0002NS 0,29NS 0,94NS M x C 2,37NS 2,41NS 0,50NS C.V. (%) 7,2 21,5 17,7 Médias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tukey, a 5% de probabilidade; C.V. - coeficiente de variação; NS - não-significativo. TABELA 5. Síntese de análise de variância e do teste de médias para massa seca da parte aérea das coberturas vegetais aos 30; 51 e 71 DAM (dias após manejo). Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 Decomposição de culturas de cobertura no sistema plantio direto 199 Na Figura 1, é apresentada a taxa de decomposição da massa seca após o manejo e observa- se que, logo após o manejo até os 51 dias, a decomposição é de aproximadamente 18 kg ha-1 dia-1, enquanto, a partir de 51 dias, a decomposição se acelera, ocorrendo taxa acima de 50 kg ha-1. BARROS (2002) observou que a decomposição foi aos 55 dias após o manejo para a crotalária e mucuna-cinza foi semelhante, ou seja, 43,3% e 39,1% da massa seca original, respectivamente. Mucuna: y = -34,474x + 5193,4 R2 = 0,9277 Crotalária: y = -38,797x + 5296,4 R2 = 0,9641 2000 2500 3000 3500 4000 4500 20 30 40 50 60 70 80 Dias após manejo Massa (kg ha-1) Mucuna Crotalária FIGURA 1. Taxa de decomposição da massa seca após manejo. FIGURA 1. Taxa de decomposição da massa seca após manejo. Pela Tabela 6, nota-se semelhança entre as culturas de cobertura para as variáveis velocidade real de deslocamento e capacidade de campo efetiva. Para o fator manejo, observa-se diferença estatística entre as velocidades, mostrando que o manejo com rolo-faca permitiu desenvolver maior velocidade, seguido do herbicida e triturador de palhas, respectivamente. A variável capacidade de campo efetiva diferiu estatisticamente, sendo o manejo com herbicida o de maior valor, devido sua maior largura real de trabalho, seguido pelo triturador de palhas e rolo-faca. MACIEL (2002) encontrou maior velocidade de deslocamento para o triturador de palhas e o pulverizador, discordando deste experimento. RESULTADOS E DISCUSSÃO Entretanto, o autor obteve maior capacidade de campo efetiva para o pulverizador (herbicida), devido a sua maior largura de trabalho, o que também ficou evidenciado neste experimento. TABELA 6. Síntese de análise de variância e do teste de médias para velocidade real de deslocamento (km h-1) e capacidade de campo efetiva (ha h-1). Fatores Velocidade Cce Manejo Triturador 6,6 c 1,4 b Rolo-faca 7,1 a 1,1 c Herbicida 6,8 b 8,4 a Cultura Mucuna 6,8 a 3,6 a Crotalária 6,8 a 3,6 a Teste F Manejo (M) 65,28 35.954,33 Cultura (C) 0,79NS 1,00NS M x C 2,42NS 2,33NS C.V. (%) 1,34 1,69 Médias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tukey, a 5% de probabilidade; significativo (p≤0,01); C.V. coeficiente de variação; NS não-significativo. TABELA 6. Síntese de análise de variância e do teste de médias para velocid deslocamento (km h-1) e capacidade de campo efetiva (ha h-1). Médias seguidas de mesmas letras minúsculas nas colunas não diferem entre si, pelo teste de Tukey, a 5% de probabilidade; significativo (p≤0,01); C.V. coeficiente de variação; NS não-significativo. Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 200 Gustavo N. dos Reis, Carlos E. A. Furlani, Rouverson P. da Silva et al. CONCLUSÕES A produção de massa seca foi igual nas duas culturas, e as taxas de decomposição da crotalária e da mucuna foram semelhantes para os três manejos estudados. O manejo das culturas de cobertura com pulverizador (químico) proporcionou maior capacidade de campo efetiva. Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007 REFERÊNCIAS BARROS, F.L. Semeadura direta da cultura do milho (Zea mays L.) em função de diferentes manejos da cobertura vegetal. 2002. 63 f. Monografia (Graduação em Agronomia) - Faculdade de Ciências Agrárias e Veterinárias, Universidade Estadual Paulista, Jaboticabal, 2002. BENEZ, S.H. Racionalização do uso de máquinas em sistema de plantio direto. In: ENCONTRO NACIONAL DE PLANTIO DIRETO NA PALHA, 1., 2002, Águas de Lindóia. Anais... Águas de Lindóia: Nobel, 2002. p.8. EMBRAPA. Serviço Nacional de Produção de Informações. Sistema plantio direto. Brasília, 1998. 248 p. EMBRAPA. Centro Nacional de Pesquisas de Solos. Sistema brasileiro de classificação de solos. Brasília, 1999. 412 p. MACIEL, R.F. Manejo das culturas de crotalária (Crotalária juncea L.) e mucuna-cinza (Stizolobuim niveum) visando ao plantio direto. 2002. 53 f. Monografia (Graduação em Agronomia) - Faculdade de Ciências Agrárias e Veterinárias, Universidade Estadual Paulista, Jaboticabal, 2002. MIALHE, L. G. Manual de mecanização agrícola. São Paulo: Agronômica Ceres, 1974. 301 p. MOODY, J.E.; SHEAR, G.M.; JONES JUNIOR, J.N. Growing corn without tillage. Soil Science Society of American Procedure, Madison, v.6, n.7, p.516, 1961. SILVA, J.A.A.; DONADIO, L.C.; CARLOS, J.A.D. Adubação verde em cítrus. Jaboticabal: Funep, 1999. 37 p. SILVEIRA, G.M.; KURACHI, S.A.; FUJIWARA, M. Métodos mecânicos e químicos de controle de ervas daninhas em cafezal. Bragantia, Campinas, v.44, n.1, p.173-8, 1985. Eng. Agríc., Jaboticabal, v.27, n.1, p.194-200, jan./abr. 2007
https://openalex.org/W2207394670	https://escholarship.org/content/qt0tb6003t/qt0tb6003t.pdf?t=qaezxj	English	null	A four gene signature predicts benefit from anthracyclines: evidence from the BR9601 and MA.5 clinical trials	Oncotarget	2,015	cc-by	5,165	A four gene signature predicts benefit from anthracyclines: evidence from the BR9601 and MA.5 clinical trials Bartlett1,5 1 Transformative Pathology, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 2 Informatics and Bio-computing, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 1 Transformative Pathology, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 2 Informatics and Bio-computing, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 3 Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada 4 Department of Pharmacology and Toxicology, University of Toronto, Toronto, ON, Canada 5 Edinburgh Cancer Research UK Centre, MRC IGMM, University of Edinburgh, Crewe Road South, Edinburgh, UK 6 Leeds Institute of Cancer and Pathology and Cancer Research UK Centre, St James’ University Hospital, Leeds, UK 7 Sunnybrook Odette Cancer Centre, Toronto, ON, Canada 10 NCIC Clinical Trials Group (NCIC CTG] and Queen’s University, Kingston, ON, Canada Correspondence to: Melanie Spears, email: Melanie.Spears@oicr.on.ca Keywords: breast cancer, anthracycline, chromosome instability, predictive biomarker Received: July 20, 2015 Accepted: August 10, 2015 Published: September 10, 2015 Published: September 10, 2015 This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Powered by the California Digital Library University of California eScholarship.org Oncotarget, Vol. 6, No. 31 www.impactjournals.com/oncotarget/ ABSTRACT Chromosome instability (CIN) in solid tumours results in multiple numerical and structural chromosomal aberrations and is associated with poor prognosis in multiple tumour types. Recent evidence demonstrated CEP17 duplication, a CIN marker, is a predictive marker of anthracycline benefit. An analysis of the BR9601 and MA.5 clinical trials was performed to test the role of existing CIN gene expression signatures as predictive markers of anthracycline sensitivity in breast cancer. Univariate analysis demonstrated, high CIN25 expression score was associated with improved distant relapse free survival (DRFS) (HR: 0.74, 95% CI 0.54-0.99, p = 0.046). High tumour CIN70 and CIN25 scores were associated with aggressive clinicopathological phenotype and increased sensitivity to anthracycline therapy compared to low CIN scores. However, in a prospectively planned multivariate analysis only pathological grade, nodal status and tumour size were significant predictors of outcome for CIN25/CIN70. A limited gene signature was generated, patients with low tumour CIN4 scores benefited from anthracycline treatment significantly more than those with high CIN4 scores (HR 0.37, 95% CI 0.20-0.56, p = 0.001). In multivariate analyses the treatment by marker interaction for CIN4/anthracyclines demonstrated hazard ratio of 0.35 (95% CI 0.15-0.80, p = 0.012) for DRFS. This data shows CIN4 is independent predictor of anthracycline benefit for DRFS in breast cancer. A four gene signature predicts benefit from anthracyclines: evidence from the BR9601 and MA.5 clinical trials A four gene signature predicts benefit from anthracyclines: evidence from the BR9601 and MA.5 clinical trials Melanie Spears1, Fouad Yousif2, Nicola Lyttle1, Paul C. Boutros2,3,4, Alison F Munro5, Chris Twelves6, Kathleen I. Pritchard7,8, Mark N. Levine9, Lois Shepherd10 and John MS. Bartlett1,5 1 Transformative Pathology, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 2 Informatics and Bio-computing, Ontario Institute for Cancer Research, MaRS Centre, Toronto, ON, Canada 3 Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada 4 Department of Pharmacology and Toxicology, University of Toronto, Toronto, ON, Canada 5 Edinburgh Cancer Research UK Centre, MRC IGMM, University of Edinburgh, Crewe Road South, Edinburgh, UK 6 Leeds Institute of Cancer and Pathology and Cancer Research UK Centre, St James’ University Hospital, Leeds, UK 7 Sunnybrook Odette Cancer Centre, Toronto, ON, Canada 8 University of Toronto, Toronto, ON, Canada 9 McMaster University and Hamilton Health Sciences, Hamilton, ON, Canada 10 NCIC Clinical Trials Group (NCIC CTG] and Queen’s University, Kingston, ON, Canada Correspondence to: Melanie Spears, email: Melanie.Spears@oicr.on.ca Keywords: breast cancer, anthracycline, chromosome instability, predictive biomarker Received: July 20, 2015 Accepted: August 10, 2015 Published: September 10, 2015 This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Melanie Spears1, Fouad Yousif2, Nicola Lyttle1, Paul C. Boutros2,3,4, Alison F. Munro5, Chris Twelves6, Kathleen I. Pritchard7,8, Mark N. Levine9, Lois Shepherd10 and John MS. Bartlett1,5 Melanie Spears1, Fouad Yousif2, Nicola Lyttle1, Paul C. Boutros2,3,4, Alison F. Munro5, Chris Twelves6, Kathleen I. Pritchard7,8, Mark N. Levine9, Lois Shepherd10 and John MS. Correlation of CIN25 and 70 and clinicopathological parameters with clinical outcomes Correlation of CIN25 and 70 and clinicopathological parameters with clinical outcomes We successfully analysed 282 of 321 (87.9%) and 421 of 440 (95.7%) tumours from BR9601 and MA.5, respectively. High CIN25 and CIN70 scores were defined as above the median as previously described. In univariate analysis using continuous clinicopathological biomarkers, high CIN25 and CIN70 scores were associated with younger age (p < 0.0001), high tumour grade (p < 0.0001), PgR negativity (p < 0.0001) and ER negativity (p < 0.0001) but not with tumour size, nodal status or HER2 status. p CIN is a phenotype description of genomic instability at the karyotypic level that results in multiple alterations in chromosomal number or structure. Multiple mechanisms drive CIN, including compromised spindle assembly checkpoint (SAC), sister chromatid cohesion defects, additional centrosomes and abnormal spindle kinetochore attachments. Pre-mitotic mechanisms may also include defects in DNA repair and replication pathways. CIN is associated with poor prognosis in many patients with solid tumours [11-12]. Critically for the current study, cell lines with high CIN phenotype enter mitotic catastrophe if challenged with anthracyclines, possibly as a result of defective SAC and other G2/M checkpoints [13]. In silico analysis has identified two mRNA signatures associated with CIN, “CIN25” and “CIN70”, as predictive of prognosis in a number of cancer datasets [14]. Furthermore high CIN70 signature expression was associated with paclitaxel resistance in ovarian cancers [15]. The CIN70 signature incorporates many genes whose mRNA expression levels correlate with proliferation, and have a role in the cell cycle [11]. CIN signatures as prognostic markers for OS and DRFS In a preplanned analysis the prognostic significance of CIN25 and CIN70 was tested on the entire patient cohort, irrespective of allocated adjuvant chemotherapy or trial. No significant association between CIN70 expression and DRFS (HR: 1.14, 95% CI 0.91-1.43, p = 0.273) or OS (HR: 1.14, 95% CI 0.88-1.45, p = 0.278) was evident. By contrast, tumours with high CIN25 scores were associated with reduced DRFS (HR: 1.43, 95%CI 1.11- 1.67, p = 0.004, Figure 2A) and OS (HR: 1.45, 95%CI 1.14-1.85, p = 0.003, Figure 2B). After multivariate analysis and adjustment for nodal status, grade, size, age, HER2, ER and PgR status, a high CIN25 score was not an independent predictor for DRFS or OS. Our work and that of others support a link between SAC dysregulation, a potential cause of CIN, CIN itself [10] and markers of CIN [7-9] and benefit from anthracycline containing polychemotherapy in vivo. Preclinical evidence linking CIN to anthracycline sensitivity support our hypothesis that CIN provides a potential clinically useful and relevant means of selecting those patients who are likely to benefit from anthracycline containing chemotherapy. Successful validation of such an approach would further support the selective use of anthracycline based chemotherapy and provide a viable diagnostic approach to support such selective use. RESULTS breast cancer. Significant resources have been devoted to developing markers that may predict anthracycline benefit (HER2, TOP2A, Ch17CEP and TIMP1) with limited success [5-9]. To date Ch17CEP (duplication of the peri-centromeric α-satellite region of chromosome 17) and TOP2A are the only markers that have shown consistent results across a number of clinical trials [7-9]. The functional pathways related to anthracycline benefit in Ch17CEP/TOP2A altered tumours remain unknown. Research from our group has linked the predictive effect of CEP17 in vivo to chromosome instability (CIN), which itself is predictive of anthracycline benefit in the BR9601 trial [10] and in preclinical models. INTRODUCTION to polychemotherapy [1]. Anthracyclines, however, cause significant toxicities including life threatening cardiotoxicity and leukaemia [2-4]. Selecting the subset of patients who will benefit from adjuvant anthracycline whilst sparing other patients unnecessary toxicities remains a clinically highly relevant approach for early Meta-analyses performed by the Early Breast Cancer Trialists Collaborative Group (EBCTCG) confirm a significant increase in disease free and overall survival (OS) following the addition of anthracyclines www.impactjournals.com/oncotarget Oncotarget 31693 CIN signatures as predictive markers of anthracycline benefit In a preplanned analysis no significant differential benefit in either DRFS or OS from E-CMF treatment was demonstrated in univariate analysis between patients whose tumours had high or low CIN70 expression (Table 1). In univariate analysis, patients whose tumours had high CIN25 gene expression scores had an increased DRFS (HR: 0.74, 95%CI 0.54-0.99, p = 0.046) when treated with E-CMF compared with patients treated with CMF alone (Table 1); a similar association was seen with respect to OS, although this did not reach statistical significance (HR: 0.76, 95%CI 0.56-1.05, p = 0.095). Conversely, there was no apparent differential benefit of E-CMF vs. CMF in patients with low CIN25 scores for DRFS (HR: 0.85, 95%CI 0.61-1.21, p = 0.374) or OS (HR: 0.87, 95%CI 0.61-1.29, p = 0.535). A multivariate analysis To further validate the role of existing CIN gene expression signatures as markers of anthracycline sensitivity we assessed these gene expression signatures in a prospectively planned and powered retrospective analysis of two pivotal clinical trials (BR9601 and MA.5). In addition we identified a new minimal gene set encapsulating the predictive value of these assay and validated its ability to stratify patients according to anthracycline benefit using clinical outcome. www.impactjournals.com/oncotarget Oncotarget 31694 Oncotarg 31695 ww.impactjournals.com/oncotarget Table 1: Hazard ratios for overall survival and distant relapse free survival comparing epirubicin plus cyclophosphamid methotrexate and fluorouracil [E-CMF] with CMF alone by biomarker status. Figure 1: Schematic REMARK diagram representing the BR9601 and MA.5 cohorts. Table 1: Hazard ratios for overall survival and distant relapse free survival comparing epirubicin plus cyclophosphamide, methotrexate and fluorouracil [E-CMF] with CMF alone by biomarker status. Figure 1: Schematic REMARK diagram representing the BR9601 and MA.5 cohorts. Figure 1: Schematic REMARK diagram representing the BR9601 and MA.5 cohorts. www.impactjournals.com/oncotarget Oncotarget 31695 Table 2: Multivariate analysis for DRFS and OS. no significant benefit from E-CMF treatment versus CMF treatment was demonstrated in patients whose tumours were grade 3 and had low CIN25 gene expression (Table 1). No significant benefit from adjuvant E-CMF versus CMF was demonstrated in patients with grade I and II tumours irrespective of CIN25 gene expression scores (Table 1). with adjustment for size, nodal status, ER, pathological grade, HER2, CIN25, treatment and CIN25 by treatment interaction showed only pathological grade, nodal status, tumour size and CEP17 to be significant predictors of outcome (Table 1). CIN signatures as predictive markers of anthracycline benefit The hazard ratio for the treatment by marker effect of CIN25 was 0.86 (95% CI 0.54-1.36, p = 0.519) for DRFS and 0.86 (95% CI 0.53-1.40, p = 0.549) for OS (Table 1). The uncorrected hazard ratio for treatment by marker effect of CIN25 in grade III tumours was 0.81 (95% CI 0.45-1.46, p = 0.479) for DRFS and 0.78 (95% CI 0.42-1.43, p = 0.413) for OS (Table 1). CIN4 as a predictor for anthracycline benefit Previous research identified a significant association between CIN gene expression and grade III tumours [9]; therefore, an exploratory analysis was performed in patients with grade III tumours. In univariate analysis, patients with grade III tumours that had high CIN25 gene expression scores had longer DRFS (HR: 0.66, 96%CI 0.46-0.94, p = 0.021) and OS (HR: 0.70, 95%CI 0.49- 1.00, p = 0.05) when treated with E-CMF than those treated with CMF alone (Figure 3, Table 1). By contrast, In order to select a minimal set of genes that reflects CIN we used the merged clinical cohort (containing both BR9601 and MA.5) and applied a bootstrap aggregation approach to allow a training and validation approach to be tested (see methods). A CIN4 score was generated using the expression values of 4 genes, HDGF, KIAA0286, RFC4 and MSH6, weighted by their regression coefficients. Patients with www.impactjournals.com/oncotarget Oncotarget 31696 Figure 2: Kaplan-Meier survival curves for low CIN25 score [red line] and high CIN25 score (blue line) for distant relapse free survival (A) and overall survival (B). Figure 2: Kaplan-Meier survival curves for low CIN25 score [red line] and high CIN25 score (blue line) for distant relapse free survival (A) and overall survival (B). Figure 3: Kaplan-Meier survival curves for epirubicin plus cyclophosphamide, methotrexate and fluorouracil (E-CMF) treated low CIN25 (orange), E-CMF high CIN25 (red), CMF treated low CIN25 (light blue), and CMF high CIN25 (dark blue) for distant relapse free survival (A) overall survival (B) Figure 3: Kaplan-Meier survival curves for epirubicin plus cyclophosphamide, methotrexate and fluorouracil (E-CMF) treated low CIN25 (orange), E-CMF high CIN25 (red), CMF treated low CIN25 (light blue), and CMF high CIN25 (dark blue) for distant relapse free survival (A) overall survival (B) Figure 4: Kaplan-Meier survival curve for low CIN4 score treated with epirubicin plus cyclophosphamide, methotrexate and fluorouracil (E-CMF) (orange line), high CIN4 score treated with E-CMF (red line), low CIN4 score treated with CMF (light blue line) and high CIN4 score treated with CMF (dark blue line) for distant relapse free survival for overall survival (A) and distant relapse free survival (B). DISCUSSION Chromosome segregation is a tightly orchestrated process and when this process goes awry CIN and aneuploidy results [16]. The original CIN signature produced by Carter and colleagues used aneuploidy as a surrogate marker for CIN and both this and mRNA CIN signatures were associated with poor prognosis in multiple cancers [11-12]. Furthermore, a study performed by Szasz and colleagues identified a four gene signature, AURKA, FOXM1, TOP2A and TPX2 from the CIN70 signature based on the high level of correlation with histological grade and in silico expression [17]. Their four gene score was able to stratify grade 2 breast cancer patients into good and poor prognostics cohorts even better than Ki67 and the mitotic index [17]. CIN has also been linked to sensitivity to specific chemotherapy agents, including anthracyclines [10]. In this study we examined the role of CIN gene signatures as predictive biomarkers for anthracycline benefit in breast cancer. Our CIN4 gene signature differs from existing candidates in being an independent predictor of anthracycline benefit. Interestingly, three of the four genes in our CIN4 signature are involved in DNA repair/ DNA-binding activity. Anthracyclines are thought to exert their actions by intercalation with DNA, generation of free radicals and crosslinking DNA to proteins [25]. Therefore, dysregulation of genes involved in DNA repair may plausibly lead to anthracycline sensitivity. To validate this gene signature in an independent dataset with 80% power at the p = 0.05 level would require only 300 patients treated in an adjuvant anthracycline trial. In conclusion we have demonstrated that a 4 gene signature, related to CIN, represents an independent predictor of anthracycline sensitivity. yi In a prospectively planned molecular analysis of two retrospective anthracycline trials univariate analyses and an analysis including stratification by grade, data for the CIN25 signature further supported our hypothesis that CIN is associated with anthracycline sensitivity in early breast cancer. However we were unable to demonstrate statistically significant predictive value for the predefined CIN25 and CIN70 signatures in multivariate regression analyses, when correcting for conventional prognostic markers and including the previously validated CEP17 predictive biomarker [10-7]. Several studies have demonstrated correlations between grade and chromosomal instability [10-18-19]. Moreover, low and intermediate grade tumours have fewer structural genomic aberrations and numerical aberrations in whole chromosomes [18-20], perhaps reflecting a reduced incidence of CIN. MATERIALS AND METHODS MATERIALS AND METHODS CIN4 as a predictor for anthracycline benefit Figure 4: Kaplan-Meier survival curve for low CIN4 score treated with epirubicin plus cyclophosphamide, methotrexate and fluorouracil (E-CMF) (orange line), high CIN4 score treated with E-CMF (red line), low CIN4 score treated with CMF (light blue line) and high CIN4 score treated with CMF (dark blue line) for distant relapse free survival for overall survival (A) and distant relapse free survival (B). www.impactjournals.com/oncotarget Oncotarget 31697 cancers that have a low CIN4 score had superior DRFS (HR 2.72, 95%CI 1.48-5.02, p = 0.001) and OS (HR 2.00, 95%CI 1.09-3.69, p = 0.03) when treated with the anthracycline compared to CMF alone (Figure 4). In multivariate analysis, the hazard ratio for treatment marker effect of CIN4 was 0.35 (95% CI 0.15-0.79, p = 0.01) for DRFS and 0.35 (95%CI 0.15-0.80, p = 0.01) (Table 2). gene panel and a combination of expression clustering, bootstrap aggregation and Cox regression modelling, we were able to identify a set of 4 genes, HDGF, KIAA0286, RFC4 and MSH6 which were significantly predictive of anthracycline benefit. Our study demonstrated that high CIN4 score was associated with improved DRFS in patients treated with anthracycline therapy. There are a number of multigene signatures associated with prognosis available commercially [21], some of which are being tested in prospective randomized trials such as “Microarray in Node-Negative Disease May Avoid Chemotherapy Trial” (MINDACT) [22] and “Trial Assigning Individualised Options for Treatment (Rx)” (TAILORx) [23]; both studies are assessing whether it is possible to avoid the use of chemotherapy in patients predicted not to benefit. However, no signatures that are specifically associated with benefit from specific chemotherapy agents have been validated to level I evidence as described by Simon et al [24]. At this point it still remains largely unclear which subpopulation of patients will benefit from adjuvant anthracycline therapy and which patients are being treated unnecessarily. Therefore a validated diagnostic test is urgently required to identify which subgroup of patients should be treated with anthracycline and which should be offered alternative therapies. The CIN4 signature identified in this study is a credible candidate to address this specific challenge. DISCUSSION This may explain why, in our study there appeared to be benefit from the addition of an anthracycline in patients with high grade tumours and high CIN25 signature but not those with grade 1 and 2 tumours. To validate the observed effect of CIN25 as a predictor of anthracycline sensitivity in an independent dataset with 80% power at the p = 0.05 level we would require a 2470 patients. BR9601 trial The BR9601 trial recruited 374 pre- and post-menopausal women with completely excised, histologically confirmed breast cancer and a clear indication for adjuvant chemotherapy. Patients p We also used a training and validation approach to derive a novel, minimum 4 gene signature “CIN4” (US patent number 62/024,729) as a predictive biomarker for anthracycline benefit. Using genes within the CIN70 www.impactjournals.com/oncotarget Oncotarget 31698 were randomized between 8 cycles of CMF (i.v. cyclophosphamide 750 mg/m2, methotrexate 50 mg/ m2 and 5-fluorouracil 600 mg/m2) every 21 days, and E-CMF (4 cycles of epirubicin 100 mg/m2 every 21 days followed by 4 cycles of the same CMF regimen ); patient characteristics are shown in supplementary data (Figure 1, Supplementary Table 1). The protocol was approved by central and local ethics committees, and each patient provided written informed consent prior to randomization. For the current analysis, tissue blocks were retrieved and RNA was extracted. The primary outcomes of the BR9601 study were RFS and OS, however distant relapse free survival was also reported . reading, counts for each RNA species were extracted and analyzed. The nCounter CodeSet contains two types of built-in controls: positive controls (spiked RNA at various concentrations to assess the overall assay performance) and negative controls (probes for background calculation). The raw data were normalized to the standard curve generated via the nCounter system spike-in controls present in all reactions. Statistics The SPSS (v20) statistical package was used for statistical analysis. Kaplan-Meier estimates of survival were used for analysis of distant relapse free (DRFS) and overall survival (OS). The Cox’s proportional hazard model was used to obtain hazard ratios for relapse or death. When comparing outcomes between the treatment arms within the groups of patients identified by biomarker expression, p-values were not calculated for sub-groups to avoid multiple testing and bias where one group was much smaller than the other. The Cox model was instead used to identify statistically significant interactions (p < 0.05) between biomarkers and outcome on the different treatments (treatment by marker effect), in models that also included biomarker status (marker effect) and treatment, as covariates. For the CIN25 and CIN70 signatures values were dichotomised around the median as previously described and a combined, preplanned treatment by marker analysis was performed using results from both the BR9601 and MA5 clinical trial cohorts. MA.5 trial The MA.5 trial randomized 716 premenopausal women with node-positive breast cancer to receive either adjuvant CEF or CMF (Figure 1, Supplementary Table 1). The CEF regimen consisted of 6 cycles of epirubicin 60 mg/m2 and 5-fluorouracil (5-FU) 500 mg/ m2, both delivered intravenously on days 1 and 8, and oral cyclophosphamide 75 mg/m2 daily on days 1 through 14. Patients randomized to the CEF regimen also received antibiotic prophylaxis throughout. The CMF regimen consisted of 6 cycles of methotrexate 40 mg/m2 and 5-FU 600 mg/m2, both delivered intravenously on days 1 and 8, and oral cyclophosphamide 100 mg/m2 daily on days 1 through 14. The MA.5 protocol was approved by the institutional review board at each participating center and registered as NCI-V90-0027 on cancer.gov. Written informed consent was obtained from each woman. RNA extraction The combined BR9601/MA5 cohort was split to two groups according to the randomized treatment. Using Affinity propagation clustering (R package apcluster1), the 70 genes were clustered into 9 groups according to their expression profiles (Supplementary Figure 2). A multivariate Cox model was fitted for each gene, adjusting for clinical variables including HER2, ER, PgR, tumour size, grade and nodal status. The top genes from each expression cluster, with the most significant p-value in the anthracycline treated cohort and a non-significant CMF cohort, were selected to make a list of 21 genes (Supplementary Figure 1). The number of genes selected from each cluster was weighted by the size of the cluster; as a result, more genes were selected for large clusters compared to small clusters (Supplementary Table 4). Each cluster had to be represented at least once even if the gene in that cluster was not significant. The gene breakdown for clusters 1-9 was as follows: CDC2; KIF20A; HDGF; MDUFAB1, CDC3A; CDC6, MAD2L1, NXT1, TOPK; FEN1, CCT5; DKC1, ECT2; KIAA0286, MCM2, RFC4, MSH6; ch.TOG, CNAP1, TOP2A, RRM1 (Supplementary Table 4). From this list, all possible combinations of 2, Total RNA from FFPE tissue samples (2 x 10µM full sections) were extracted using the RecoverAll Total Nucleic Acid Isolation kit (Life Technologies) according to the manufacturers protocol and concentrations were determined using the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies). ACKNOWLEDGMENTS AND FUNDING 9. Bartlett, J. M., McConkey, C. C., Munro, A. F., Desmedt, C., Dunn, J. A., Larsimont, D. P., O’Malley, F. P., Cameron, D. A., Earl, H. M., Poole, C. J., Shepherd, L. E., Cardoso, F., Jensen, M. B., Caldas, C., Twelves, C. J., Rea, D. W., Ejlertsen, B., Di, Leo A., and Pritchard, K. I.Predicting Anthracycline Benefit: TOP2A and CEP17-Not Only but Also. J. Clin. Oncol. 2015; 33:1680-7. This work was supported by the funding from OICR. We thank the Government of Ontario for funding, which is provided through the Ontario Ministry of Research and Innovation. PCB was supported by a TFRI New Investigator Award and a CIHR New Investigator Award. 10. Munro, A. F., Twelves, C., Thomas, J. S., Cameron, D. A., and Bartlett, J. M.Chromosome instability and benefit from adjuvant anthracyclines in breast cancer. Br J Cancer. 2012; 107: 71-74. CONFLICTS OF INTEREST All authors state that they have no financial conflicts of interest. 11. Carter, S. L., Eklund, A. C., Kohane, I. S., Harris, L. N., and Szallasi, Z.A signature of chromosomal instability inferred from gene expression profiles predicts clinical outcome in multiple human cancers. Nat Genet. 2006; 38: 1043-1048. Gene expression analysis RNA (400ng) was used with the nCounter system, according to the manufacturer’s protocol (NanoString ® Technologies, Seattle, WA, USA). In brief, 5µl of RNA was hybridized at 96oC overnight with the NanoString Codeset. Probes for the analysis of test genes and controls were synthesized by NanoString technologies, including probes for the 70 genes of interest and 6 normalising genes (Supplementary Table 2). All 76 genes and controls were assayed simultaneously in multiplexed reactions. After probe hybridizations and NanoString nCounter digital www.impactjournals.com/oncotarget Oncotarget 31699 Jordan, S., Twelves, C. J., Cameron, D. A., Thomas, J., Campbell, F. M., Rea, D. W., Provenzano, E., Caldas, C., Pharoah, P., Hiller, L., Earl, H., and Poole, C. J.Predictive markers of anthracycline benefit: a prospectively planned analysis of the UK National Epirubicin Adjuvant Trial (NEAT/BR9601). Lancet Oncol. 2010; 11: 266-274. 3, 4 and 5 gene signatures were examined (210, 1330, 5985, and 20349 combinations respectively). Using bootstrapped aggregation (bagging), each combination was bootstrapped 100 times, with the median area under the curve (AUC) noted as the bagged variable. In each bootstrap, the treatment cohort was split into 60% training and 40% test sets [26-27-28-29]. The AUC was calculated from the test sets (R package survivalROC2) (Supplementary Table 5). The gene signature selected had the greatest AUC for a treatment by marker effect in both the training and validation approaches and included four genes, HDGF, KIAA0286, RFC4 and MSH6, termed the CIN4 signature. A multivariate Cox regression was fitted using the four genes, adjusting for the same clinical variables mentioned above (Supplementary Table 6). A CIN4 score was generated using the expression values of the four genes, weighted by their regression coefficients. 7. 7. Pritchard, K. I., Munro, A., O’Malley, F. P., Tu, D., Li, X., Levine, M. N., Shepherd, L., Chia, S., and Bartlett, J. M.Chromosome 17 centromere (CEP17) duplication as a predictor of anthracycline response: evidence from the NCIC Clinical Trials Group (NCIC CTG) MA.5 Trial. Breast Cancer Res Treat. 2012; 131: 541-551. 8. Bartlett, J. M. S., Desmedt, C., Munro, A., O’Malley, F. P., Larsimont, D., Di Leo, A., Cameron, D. A., Isola, J., Shepherd, L., Twelves, C. J., Pritchard, K. I., and TIIa, Metaanal Grp.Chromosome 17 polysomy: a unifying hypothesis underlying benefit from adjuvant anthracyclines? Cancer Research. 2009; 69: 364S-364S. REFERENCES A., McConkey, C., www.impactjournals.com/oncotarget Oncotarget 31700 24: 123-140. 24: 123-140. 16. Bakhoum, S. F. and Compton, D. A.Chromosomal instability and cancer: a complex relationship with therapeutic potential. J Clin Invest. 2012; 122: 1138-1143. 17. Szasz, A. M., Li, Q., Eklund, A. C., Sztupinszki, Z., Rowan, A., Tokes, A. M., Szekely, B., Kiss, A., Szendroi, M., Gyorffy, B., Szallasi, Z., Swanton, C., and Kulka, J.The CIN4 chromosomal instability qPCR classifier defines tumor aneuploidy and stratifies outcome in grade 2 breast cancer. PLoS One. 2013; 8: e56707- 18. A’Hern, R. P., Jamal-Hanjani, M., Szasz, A. M., Johnston, S. R., Reis-Filho, J. S., Roylance, R., and Swanton, C.Taxane benefit in breast cancer—a role for grade and chromosomal stability. Nat Rev Clin Oncol. 2013; 10: 357- 364. 19. Sauer, T., Beraki, K., Jebsen, P. W., Ormerod, E., and Naess, O.Ploidy analysis by in situ hybridization of interphase cell nuclei in fine-needle aspirates from breast carcinomas: correlation with cytologic grading. Diagn. Cytopathol. 1997; 17: 267-271. 20. Dellas, A., Torhorst, J., Schultheiss, E., Mihatsch, M. J., and Moch, H.DNA sequence losses on chromosomes 11p and 18q are associated with clinical outcome in lymph node- negative ductal breast cancer. Clin Cancer Res. 2002; 8: 1210-1216. 21. Harbeck, N., Sotlar, K., Wuerstlein, R., and Doisneau- Sixou, S.Molecular and protein markers for clinical decision making in breast cancer: today and tomorrow. Cancer Treat Rev. 2014; 40: 434-444. 22. Rutgers, E., Piccart-Gebhart, M. J., Bogaerts, J., Delaloge, S., Veer, L. V., Rubio, I. T., Viale, G., Thompson, A. M., Passalacqua, R., Nitz, U., Vindevoghel, A., Pierga, J. Y., Ravdin, P. M., Werutsky, G., and Cardoso, F.The EORTC 10041/BIG 03-04 MINDACT trial is feasible: results of the pilot phase. Eur J Cancer. 2011; 47: 2742-2749. 23. Zujewski, J. A. and Kamin, L.Trial assessing individualized options for treatment for breast cancer: the TAILORx trial. Future. Oncol. 2008; 4: 603-610. 24. Simon, R. M., Paik, S., and Hayes, D. F.Use of archived specimens in evaluation of prognostic and predictive biomarkers. J Natl Cancer Inst. 2009; 101: 1446-1452. 25. Minotti, G., Recalcati, S., Menna, P., Salvatorelli, E., Corna, G., and Cairo, G.Doxorubicin cardiotoxicity and the control of iron metabolism: quinone-dependent and independent mechanisms. Methods Enzymol. 2004; 378: 340-361. 26. Govindan, G. and Nair, A. S.Bagging with CTD—a novel signature for the hierarchical prediction of secreted protein trafficking in eukaryotes. Genomics Proteomics Bioinformatics. 2013; 11: 385-390. 27. Tan, A. C. REFERENCES 1. Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. 2005; 365: 1687-1717. 12. Habermann, J. K., Doering, J., Hautaniemi, S., Roblick, U. J., Bundgen, N. K., Nicorici, D., Kronenwett, U., Rathnagiriswaran, S., Mettu, R. K., Ma, Y., Kruger, S., Bruch, H. P., Auer, G., Guo, N. L., and Ried, T.The gene expression signature of genomic instability in breast cancer is an independent predictor of clinical outcome. Int J Cancer. 2009; 124: 1552-1564. 2. Swain, S. M., Whaley, F. S., and Ewer, M. S.Congestive heart failure in patients treated with doxorubicin: a retrospective analysis of three trials. Cancer. 2003; 97: 2869-2879. 3. Mercuro, G., Cadeddu, C., Piras, A., Dessi, M., Madeddu, C., Deidda, M., Serpe, R., Massa, E., and Mantovani, G.Early epirubicin-induced myocardial dysfunction revealed by serial tissue Doppler echocardiography: correlation with inflammatory and oxidative stress markers. Oncologist. 2007; 12: 1124-1133. 3. Mercuro, G., Cadeddu, C., Piras, A., Dessi, M., Madeddu, C., Deidda, M., Serpe, R., Massa, E., and Mantovani, G.Early epirubicin-induced myocardial dysfunction revealed by serial tissue Doppler echocardiography: correlation with inflammatory and oxidative stress markers. Oncologist. 2007; 12: 1124-1133. 13. Munro, A. F., Cameron, D. A., and Bartlett, J. M.Targeting anthracyclines in early breast cancer: new candidate predictive biomarkers emerge. Oncogene. 2010; 29: 5231- 5240. 14. Carter, S. L., Eklund, A. C., Kohane, I. S., Harris, L. N., and Szallasi, Z.A signature of chromosomal instability inferred from gene expression profiles predicts clinical outcome in multiple human cancers. Nat Genet. 2006; 38: 1043-1048. 4. Jones, L. W., Haykowsky, M. J., Swartz, J. J., Douglas, P. S., and Mackey, J. R.Early breast cancer therapy and cardiovascular injury. J. Am. Coll. Cardiol. 2007; 50: 1435- 1441. 15. Swanton, C., Nicke, B., Schuett, M., Eklund, A. C., Ng, C., Li, Q., Hardcastle, T., Lee, A., Roy, R., East, P., Kschischo, M., Endesfelder, D., Wylie, P., Kim, S. N., Chen, J. G., Howell, M., Ried, T., Habermann, J. K., Auer, G., Brenton, J. D., Szallasi, Z., and Downward, J.Chromosomal instability determines taxane response. Proc Natl Acad Sci U S A. 2009; 106: 8671-8676. 5. Pritchard, K. I., Shepherd, L. E., O’Malley, F. P., Andrulis, I. L., Tu, D., Bramwell, V. H., and Levine, M. N.HER2 and responsiveness of breast cancer to adjuvant chemotherapy. 2006; 354: 2103-2111. 6. Bartlett, J. M., Munro, A. F., Dunn, J. REFERENCES and Gilbert, D.Ensemble machine learning on gene expression data for cancer classification. Appl. Bioinformatics. 2003; 2: S75-S83. 28. Zhang, H and Singer, BH.Recursive partitioning and applications. 2010; Second: 29. Breiman L.Bagging Predictors. Machine Learning. 1996; www.impactjournals.com/oncotarget Oncotarget 31701
https://openalex.org/W2735040917	https://www.scielo.br/j/mr/a/wMmf8F54vPBPKWQvqv73x3b/?lang=en&format=pdf	English	null	Experimental Study of the Compressive Strength of Multi-Coat Renders	Materials research	2,017	cc-by	6,787	1. Introduction in terms of in-service loads; mechanical incompatibility of render and its support, among others. This property is a standard requirement in EN 998-13 and under EN 1015-114, the device used to test the compressive strength of mortars should have 40 mm square plates so as to obtain the strength of 40 mm cubic specimens. But these samples do not represent the render traditionally used on façades since, in accordance with EN 13914-15 and other national technical specifications suggested by the National Laboratory of Civil Engineering6, conventional render should be applied in at least two coats (base and top coats) and, sometimes by using a spatterdash coat for adherence purposes. The study of in-service compressive strength is widely used in the assessment of concrete structures under actual service conditions. In these in-situ evaluations, several cores are drilled and collected for compression strength testing in the lab, according to EN 137911. Also the concrete core strength results are usually correlated with other indirect methods, such as rebound hammer index and ultrasonic pulse velocity. The concept of estimating actual concrete in-situ strength relative to cubic specimens is employed when the values of the strength of cores are used for structural calculation or to give the strength of an existing structure2. It is therefore hard to directly evaluate the compressive strength of multi-coat systems, and so indirect methods have been used, namely a pendulum rebound hammer or ultrasound on the walls surface. When comparing standard specimens produced in the lab with the same concrete as that used in construction works, their strength will differ from that determined in situ, even if we consider that it is possible to obtain perfect cores of the same size as the test pieces. This is because of differences in compacting and curing2. Therefore, the study of in-service compressive strength should be performed on cores collected onsite. Several authors have tried to discuss the direct assessment of compressive strength of lime-based renders applied on old buildings walls by carrying out tests on cores extracted from existing render systems. Válek and Veiga7 used irregular samples of historic mortars collected in-situ to study a simple adaptation of the method to determine the compressive strength recommended by the European Standard EN 1015- 114, to enable the compressive testing of irregular samples collected from coatings on old buildings. Experimental Study of the Compressive Strength of Multi-Coat Renders António Soaresa, Inês Flores-Colena, Jorge de Britoa* aDepartamento de Engenharia Civil, Arquitectura e Georrecursos, Investigação e Inovação em Engenharia Civil para a Sustentabilidade, Universidade Técnica de Lisboa, Av. Rovisco Pais, 1049-001, Lisboa, Portugal * e-mail: jb@civil.ist.utl.pt © 2017 © 2017 DOI: http://dx.doi.org/10.1590/1980-5373-MR-2016-0557 Materials Research. 2017; 20(5): 1254-1265 DOI: http://dx.doi.org/10.1590/1980-5373-MR-2016-0557 ; ( ) Received: July 25, 2016; Revised: May 09, 2017; Accepted: June 20, 2017 Mortar compressive strength is a standard requirement and should be tested on stand-ardized specimens with a 40 mm square cross section. However, these specimens with only one type of mortar (base-coat layer) do not represent the render traditionally ap-plied to façades since that must consist of at least two layers (base coat and top coat, with different formulations and thicknesses). The in-service compressive strength of the render is therefore currently evaluated on-site by means of indirect methods (e.g. pendulum rebound hammer or ultrasound). Addi-tionally, studies on evaluating the compressive strength in multilayer coating systems, based on the laboratory analysis of the mortar samples collected in situ, are few and far between. This paper intends to discuss the method of assessing in-service compressive strength of applied multi-coat renders systems, by testing cores collected after pull-off tests. Therefore, 422 compressive strength tests were carried out to establish a relation-ship between the compressive strength of single- and multi-coat renders and identify which render layers influence the system’s compressive strength. The experimental cam-paign results have shown the great influence of the top coat on multi-coat renders’ per-formance and the feasibility of studying the in-service compressive strength after a pull-off test; and also allowed identifying relevant parameters to evaluate the strength of samples collected in-situ. 1. Introduction Flores-Colen10 studied the in-service mechanical performance of cement-based traditional and pre-dosed renders, through in-situ and lab testing on collected samples. The research concluded that the pull-off technique, despite being destructive, is a probing method that can help the study of render mechanical parameters. The authors identify the benefits of studying the render cores after a pull-off test. Good correlations were obtained between adhesion results and compression tests (R2 > 0.80). The compressive strength was studied in standard specimens and cores. Also, acceptable results allowed estimating the cores compressive strength in terms of ten times the stress in cohesive failure from the pull-off test (R2 > 0.70). However, core compressive strength results obtained in the lab showed high variation coefficients (CV > 25%) and needed further investigation. Figure 1. Steel moulds for reference specimens (left) and with base coat in fresh state and hardened spatterdash (right). Figure 2. Final aspect of the circular reference specimens, with several layers. Due to the thickness of the in-situ renders, it is also necessary to take into account the limitation this causes in the geometry of the sample (test piece) for the evaluation of the compressive strength of the render, since the compressive strength of products such as concrete and mortar is affected by the size of the test piece11-12 (cited by13), and it decreases in concrete through a power trend the more slender the test piece14-15. A number of authors16-18, have recently studied the compressive strength of mortars and found that, as with concrete it increases as slenderness decreases through a power trend. Soares et al.18 further obtained for industrial and traditional renders a power trend relationship between compressive strength (as a function of the strength index) and slenderness (h/d) of the test piece (0.60 < R2 < 0.80), except for traditional multilayer render (spatterdash, base coat, top coat), making it difficult to directly compare the compressive strength of a multi-coat render with that of standard specimens. Figure 2. Final aspect of the circular reference specimens, with several layers. The expected compressive strength of the render was characterized by standard specimens testing at 7, 14, 8 and 90 days. The results are given in Table 1. The cores were extracted from render applied to brick and tested, along with test pieces moulded in the laboratory (reference specimens), using the same constituents and having the same geometric characteristics as the cores. 1. Introduction The adaptation involved applying a confining mortar with a 1:3 volumetric In the case of renders, the study of in-service compressive strength can help the diagnosis of the causes of several render anomalies related with high susceptibility to cracks; low resistance to the impact loads; inadequate specification Experimental Study of the Compressive Strength of Multi-Coat Renders 1255 ratio (CEM II, 32.5: sand), on the two parallel sides of the irregular sample. compressive strength from multilayer cores, it is important to assess the relationship between the compressive strength of test pieces and cores from a single layer and a multilayer and to identify the coating layers that affect the compressive strength results. Magalhães and Veiga8 used reference samples measuring 20 x 40 x 80 [mm3], cut from pieces measuring 40 x 40 x 160 [mm3], to assess the method. Confining mortar was applied for the compression test. The compressive strength values obtained with these samples were compared with those of the standard test specimens, resulting in a good linear relationship (R2 > 0.96); with this correlation the values of compressive strength obtained by the adapted method were higher than the ones obtained by the standard method. The higher compressive strength of the reference specimens (RSs) relative to the standard specimens (SSs) was explained by the greater carbonation of the RSs, which were tested over a 1-3 year period after the SSs were tested8. The authors concluded that the proposed method can make a valid contribution, albeit without furnishing rigorous results in absolute terms, to finding the mechanical strength of old buildings9. Soares et al.19 studied (in the laboratory) the possibility of producing reference specimens with more than one layer (Figures 1 and 2), to represent conventional multi-coat cores of render, with the purpose of evaluating the compressive strength of cementitious mortars used in walls rendering. The author obtained good results, finding a linear relationship between compressive strength values for cores and reference specimens of traditional multilayer render close to y = x (with a slope of 1 ± 0.22). The same author found that this relationship does not change significantly with age (R2 = 0.86) when grouping the results of the all ages (7, 14, 28 and 90 days). Figure 1. Steel moulds for reference specimens (left) and with base coat in fresh state and hardened spatterdash (right). 2.2 Application of renders prepared in the laboratory The initial curing of the different types of render studies was harmonized by keeping the single-coat traditional render samples in polyethylene bags for 11 days, except for the ones tested at 7 days, which were only kept in the bags for 4 days, since it would not be possible to carry out the whole pull-off adhesion test procedure and the subsequent compressive strength test. The mortar for the traditional render was applied according to EN 13914-15. This standard does not allow for the waiting time required between applying each coat of render since this varies depending on the climatic conditions. We followed the suggestions in Gomes et al.6 for the 3-coat traditional render, leaving an interval of 3 days between applying the spatterdash and the base coat, and 8 days between that and the top coat. 2.4 Storage and curing of the specimens The specimens were stored and cured as described in EN 1015-114 by being placed in a conditioned chamber at a temperature of 20 ± 2 ºC and relative humidity of 65 ± 5%. The samples were wrapped in polyethylene bags for the initial curing. The water content of each formulation was determined based on the flow table test so as to arrive at a flow value of 175 ± 10 mm, per EN 1015-221, apart for the spatterdash that had a fluid consistency. Table 2 shows the composition of each coat of the traditional render. Because the top coat was applied 8 days after the base coat the samples were kept in the polyethylene bags for a further 3 days to ensure that this last layer was wet cured.f Table 1. Compressive strength values measured in standardized specimens (SS Single-coat traditional render/base coat of multi-coat traditional render Age [days] Compressive strength of standardized specimens (SSs) [N/mm2] for the evaluation of its influence on the compressive strength of the coating system. The authors intend to identify the layers with more relevance to assess compressive strength of multilayer render cores and evaluate the feasibility of this method to evaluate directly the in-service compressive strength of existing renders, taking advantage of pull-off cores without resorting to complex methods such as finite elements analysis or computational models. The surface of the intermediate layers of the traditional multi-coat render was roughened to encourage better bonding between coats. 2.1 Preparation of traditional render In the reference specimens of multi-coat render the mould of the next layer was placed around the one already produced so that the bond between the coats would be continuous (Figure 1, right). The traditional mortar was produced as prescribed in EN 196-120, using river sand as aggregate and CEM II/B-L 32.5 N cement as binder. The volumetric ratios used for each traditional render coat were as established in LNEC6, within the standards of good practice, as follows: spatterdash at 1:2; base coat 1:4, and top coat 1:4.5. 2. Experimental Procedure The reference specimens were prepared19, based on procedures for standard size specimens, using steel moulds (Figure 1, left, with inner measurements similar to those of the cores (Cs) obtained with the pull-off adhesion test (render applied on a brick). Reference specimens (RSs) were produced and conventional multi-coat renders were applied to bricks (spatterdash, base coat and top coat). Control conventional renders were also produced (one coat with the same composition as the base coat of a multi-coat render). The relevant thicknesses are given in Table 2. The mortar was placed in the moulds and compacted by tapping from above with a trowel to achieve a uniform distribution of fresh mortar in the mould, and finally pressed to give a smooth finish. Figure 2 shows the final look of the circular cross section reference specimens. 1. Introduction This paper presents part of the research conducted and discusses the relationship between the compressive strength of conventional single-layer and multilayer cement-based render. The aim is to study the possibility of evaluating the mechanical strength of a core or specimen of render with several coats, using a single-coat specimen. In summary, there are very few studies on evaluating the compressive strength of multilayer coating systems through standard procedures. Given the possibility of evaluating Soares et al. 1256 Materials Research Table 1. Compressive strength values measured in standardized specimens (SSs) Single-coat traditional render/base coat of multi-coat traditional render Age [days] 7 14 28 90 Compressive strength of standardized specimens (SSs) [N/mm2] 8,8 10,6 10,8 13,1 1256 Soares et al. Materials Research 2.6 Preparation of cores for the compressive strength test Before the cores could be used for the compressive strength test the steel disc and glue required for the pull-off adhesion test had to be removed without damaging the core. The procedure described by Flores-Colen10 was used, whereby the glue is melted on an electric stove and the core is then detached from the disc by means of a spatula (Figure 3). The irregularities on the cores’ surface, caused by contact with the brick, were smoothed with a file in preparation for contact with the compression instrument’s plates (Figure 4). As the cross sectional area of the specimen was sometimes larger than that of the testing machine’s plates, two steel plates with an area greater than the sample’s surface area were used (Figure 5). They were placed on the ends of the core such that the compression force would be distributed over the entire area. These plates were used on all the cores to keep the test conditions identical. A steel wedge was used with the cores of reduced thickness (Figure 5) to adjust their height so that they would fit between the compression machine’s plates. Before the cores could be used for the compressive strength test the steel disc and glue required for the pull-off adhesion test had to be removed without damaging the core. The procedure described by Flores-Colen10 was used, whereby the glue is melted on an electric stove and the core is then detached from the disc by means of a spatula (Figure 3). The traditional multi-coat render tested at 7 days was produced without a top coat, i.e. it comprised only spatterdash and base coat. This aspect of cores tested at age 7 days, plus the fact that a slope closest to y = x was obtained at this age, with a good correlation coefficient (R2 = 0.76 for cores and R2 = 0.99 for reference specimens), may indicate that the spatterdash coat (the render coat with the strongest cement ratio) does not exert a significant influence on compressive strength. The irregularities on the cores’ surface, caused by contact with the brick, were smoothed with a file in preparation for contact with the compression instrument’s plates (Figure 4). As the cross sectional area of the specimen was sometimes larger than that of the testing machine’s plates, two steel plates with an area greater than the sample’s surface area were used (Figure 5). 3.1 Compressive strength relationship between multi-coat and single-coat cores Analysis of the relationship between the strengths found for single- and multi-coat render cores showed that the relationship at 7 days has a slope of 0.87 (which is close to y = x), with a correlation coefficient R2 = 0.76 (Figure 6), while at other ages the slope increases, with correlation coefficients of the same order of magnitude. The slope increase is most marked at 28 days. However, the slope obtained with the 28 day results has a correlation coefficient R2 ≈ 0.38, which can be ascribed to the 3 highest compressive strength values at this age for the single-coat render coming from specimens whose slenderness (h/d) is between 0.33 and 0.37, which are low slenderness values. According to Soares et al.18, these slenderness values result in high compressive strength values of considerable variability, and therefore not representative. 2.6 Preparation of cores for the compressive strength test They were placed on the ends of the core such that the compression force would be distributed over the entire area. These plates were used on all the cores to keep the test conditions identical. A steel wedge was used with the cores of reduced thickness (Figure 5) to adjust their height so that they would fit between the compression machine’s plates. The relationship obtained by grouping all the ages shows a trend line in Figure 6 with a slope of 0.94, which is close to y = x. However, this has R2 ≈ 0.18, so it cannot be representative of the various ages treated separately, which was expected since three-coat examples (aged 14, 28 and 90 days) are being combined with two-coat examples (aged 7 days, without top coat). When only three-coat examples are grouped together we find a considerable increase in the correlation coefficient up to R2 ≈ 0.58, which, even though it is not a high value, can be explained by the deviation from the trend line in the 28 day results with a low R2 value. 2.5 Compressive strength test It should also be remembered that according to EN 13914-15 the spatterdash should not be regarded as a coat and should be ignored when measuring the total thickness of the coat. But according to LNEC18 it can contribute to compressive strength, and so for the purpose of this work it was applied evenly to the surface of the substrate and is deemed a layer The procedure established in EN 1015-114 was followed, using a Seidner Form+Test instrument (model 505/200/10 DM1), with a 200 kN load cell, at a rate between 50 N/s and 500 N/s. Cores from the pull-off adhesion test in accordance with EN 1015-1222 were subjected to compression testing. In Experimental Study of the Compressive Strength of Multi-Coat Renders 1257 Table 2. Characteristics of the traditional render Traditional render Spatterdash Base coat Top coat Thickness (mm) Type 1 3 15 6 Type 2 3 25 6 Type 5 - 15 - Type 6 - 25 - Volumetric ratio 1:2 1:4 1:4,5 Water/cement ratio 0.70 1.00 1.06 Caption: Type 1 and Type 2 - multi-coat traditional renders Type 5 and Type 6 - control traditional renders with one coat. Table 2. Characteristics of the traditional render compressive strength of traditional multi-coat render through the value determined for a single-layer traditional render (made with the same mixture and being the same size as the base coat of the multi-coat render), i.e. a simple method without resorting to complex techniques or computational models. addition to the 50 mm diameter circular cross section cores, 40 mm and 50 mm square cross section cores were also prepared, as suggested by RILEM23 for in-situ tests which recommends using square cross section cores with an area the same as the circular cross section ones. Therefore, three 50 mm diameter circular cross section cores, two 40 mm square cross section cores and one 50 mm square cross section core were taken from each brick. RSs with similar characteristics to the Cs were also tested in compression, making a total of 422 valid tests, as detailed in Tables 3 to 5. Some of the tests at 28 days were conducted using a Controls test device (model L11D2) with a 250 kN load cell, at the same rate of loading, which may explain the lower values of compressive strength at 28 days of some specimens (relative to 14 days) 3. Presentation and Discussion of the Results The compressive strength values obtained for both the Cs and RSs were compared to see if is possible to evaluate the 1258 Soares et al. Materials Research Table 3. Number of compressive strength tests Table 3. Number of compressive strength tests Type of render Type of specimen Multi-coat traditional render Single-coat traditional render Core diameter/cross section D50 L40 L50 D50 L40 L50 Tests C 62 47 23 24 15 7 RS 47 55 48 31 34 29 Total C 132 46 RS 150 94 C 178 RS 244 422 Caption: C - cores taken from render applied to brick; RS - reference specimens; D50 - 50 mm diameter circular cross section specimens; L40 - 40 mm square cross section specimens; L50 - 50 mm square cross section specimens. Caption: C - cores taken from render applied to brick; RS - reference specimens; D50 - 50 mm diameter circular cross section specimens; L40 - 40 mm square cross section specimens; L50 - 50 mm square cross section specimens. Table 4. 3. Presentation and Discussion of the Results Average results and standard deviations of the compressive strength tests of cores Type of render Core diameter/cross section 7 days 14 days 28 days 90 days Multi-coat traditional render (base coat with 15 mm thickness) D50 15.38 (± 1.04) 11.59 (± 0.91) 10.05 (± 2.95) 12.46 (± 2.25) L40 14.42 (± 4.08) 9.15 (± 0.91) 10.81 (± 4.79) 14.76 (± 0.97) L50 22.38 (± 0.00) 15.06 (± 1.70) 10.63 (± 4.10) - Multi-coat traditional render (base coat with 25 mm thickness) D50 10.55 (± 1.08) 8.01 (± 1.90) 9.54 (± 1.71) 9.34 (± 0.87) L40 8.57 (± 0.77) 8.00 (± 0.93) 8.00 (± 2.68) 10.68 (± 1.55) L50 12.00 (± 0.25) 11.83 (± 3.46) 10.48 (± 1.59) 12.28 (± 2.06) Single-coat traditional render (with 15 mm thickness) D50 18.09 (± 2.70) 24.34 (± 1.09) 22.61 (± 1.64) 26.09 (± 1.67) L40 15.49 (± 0.17) 20.53 (± 0.26) 20.20 (± 0.25) 27.06 (± 0.69) L50 18.48 (± 0.00) 36.06 (± 0.00) 27.17 (± 0.00) - Single-coat traditional render (with 25 mm thickness) D50 8.41 (± 0.67) 12.31 (± 3.21) 10.97 (± 0.37) 11.72 (± 1.74) L40 8.28 (± 0.50) 11.04 (± 0.37) 10.25 (± 0.00) 14.15 (± 040) L50 9.66 (± 0.00) 14.85 (± 0.00) 11.49 (± 0.00) 16.38 (± 0.00) Type of render Core diameter/cross section 7 days 14 days 28 days 90 days Multi-coat traditional render (base coat with 15 mm thickness) D50 15.38 (± 1.04) 11.59 (± 0.91) 10.05 (± 2.95) 12.46 (± 2.25) L40 14.42 (± 4.08) 9.15 (± 0.91) 10.81 (± 4.79) 14.76 (± 0.97) L50 22.38 (± 0.00) 15.06 (± 1.70) 10.63 (± 4.10) - Multi-coat traditional render (base coat with 25 mm thickness) D50 10.55 (± 1.08) 8.01 (± 1.90) 9.54 (± 1.71) 9.34 (± 0.87) L40 8.57 (± 0.77) 8.00 (± 0.93) 8.00 (± 2.68) 10.68 (± 1.55) L50 12.00 (± 0.25) 11.83 (± 3.46) 10.48 (± 1.59) 12.28 (± 2.06) Single-coat traditional render (with 15 mm thickness) D50 18.09 (± 2.70) 24.34 (± 1.09) 22.61 (± 1.64) 26.09 (± 1.67) L40 15.49 (± 0.17) 20.53 (± 0.26) 20.20 (± 0.25) 27.06 (± 0.69) L50 18.48 (± 0.00) 36.06 (± 0.00) 27.17 (± 0.00) - Single-coat traditional render (with 25 mm thickness) D50 8.41 (± 0.67) 12.31 (± 3.21) 10.97 (± 0.37) 11.72 (± 1.74) L40 8.28 (± 0.50) 11.04 (± 0.37) 10.25 (± 0.00) 14.15 (± 040) L50 9.66 (± 0.00) 14.85 (± 0.00) 11.49 (± 0.00) 16.38 (± 0.00) Caption: D50 - 50 mm diameter circular cross section specimens; L40 - 40 mm square cross section specimens; L50 - 50 mm square ti i Table 4. 3. Presentation and Discussion of the Results Average results and standard deviations of the compressive strength tests of cores Experimental Study of the Compressive Strength of Multi-Coat Renders 1259 Table 5. Average results and standard deviations of the compressive strength tests of reference specimens Type of render Reference specimen diameter/ cross section 7 days 14 days 28 days 90 days Multi-coat traditional render (base coat with 15 mm thickness) D50 19.78 (± 1.51) 15.27 (± 1.20) 13.42 (± 1.65) 18.69 (± 1.01) L40 12.50 (± 0.50) 9.13 (± 0.51) 9.17 (± 1.16) 12.08 (± 1.52) L50 17.63 (± 2.06) 13.61 (± 0.46) 12.16 (± 1.18) 16.97 (± 0.85) Multi-coat traditional render (base coat with 25 mm thickness) D50 8.50 (± 0.81) 7.93 (± 1.06) 8.93 (± 1.60) 12.57 (± 0.61) L40 7.03 (± 0.37) 7.28 (± 1.09) 7.05 (± 1.40) 7.51 (± 1.19) L50 8.96 (± 0.50) 9.90 (± 0.88) 9.89 (± 0.93) 12.49 (± 0.78) Single-coat traditional render (with 15 mm thickness) D50 24.18 (± 0.77) 28.86 (± 1.93) 20.28 (± 1.14) 25.96 (± 3.23) L40 15.07 (± 0.93) 15.64 (± 0.96) 11.68 (± 1.86) 16.79 (± 3.07) L50 20.50 (± 2.11) 24.97 (± 0.00) 18.04 (± 0.63) 26.17 (± 2.61) Single-coat traditional render (with 25 mm thickness) D50 11.65 (± 0.59) 15.67 (± 1.25) 11.63 (± 0.44) 12.32 (± 1.37) L40 9.22 (± 0.77) 10.29 (± 1.42) 9.71 (± 0.88) 8.69 (± 1.10) L50 10.93 (± 0.47) 12.13 (± 0.23) 10.82 (± 0.56) 10.69 (± 1.23) Caption: D50 - 50 mm diameter circular cross section specimens; L40 - 40 mm square cross section specimens; L50 - 50 mm square cross section specimens. Average results and standard deviations of the compressive strength tests of reference specimens Caption: D50 - 50 mm diameter circular cross section specimens; L40 - 40 mm square cross section specimens; L50 - 50 mm square cross section specimens. Figure 3. Heating the epoxy glue and removal of steel disc from the core. the single-coat traditional render (produced with the same product as the base coat of the multi-coat traditional render. This can be explained by the higher slenderness (h/d) of the multi-coat traditional render, or by it having a top coat that is produced with a lower cement ratio. 3. Presentation and Discussion of the Results Another reason for slope y > x can be related to water suction by the substrate that can modify the internal structure and the density of mortars after application, affecting the coats’ performance (e.g. spatterdash and base coat). Figure 3. Heating the epoxy glue and removal of steel disc from the core. Figure 4. Rectification of surfaces of cores and reference specimens. 3.2 Compressive strength relationship between multi-coat and single-coat reference specimens The relationship between reference specimens of traditional single-coat and multi-coat renders (Figure 7) shows good correlation coefficients (R2) at each test age analysed individually, i.e. they are all above 0.8, and the slopes are in most cases lower than those from the cores’ analysis. The higher R2 values in the reference specimens may be related with the greater size precision that is achieved in moulding these specimens by comparison with the cores that result from the pull-off test, which may cause some variation in dimensions and non-visible damage, such as microcraking.f Figure 4. Rectification of surfaces of cores and reference specimens. Figure 5. Compressive strength test using two steel plates and a steel wedge. To better appreciate the difference in the ratios of each layer of the multi-coat traditional render, Figure 8 (left) shows a reference specimen (cross section) after the compressive strength test, where it is possible to see the different layers of the render. Note that the spatterdash coat is the most compact and the others are successively less so. Figure 5. Compressive strength test using two steel plates and a steel wedge. Thus, bearing in mind the slope y > x observed in the comparison of single-coat traditional and multi-coat (spatterdash, base coat and top coat) render, it can be said that the multi-coat traditional render has lower strength than When comparing the cross section after the compression test on the spatterdash layer (Figure 8) with that of the top coat 1260 Soares et al. Materials Research Figure 6. Relationship between the strength of multi-coat and single-coat cores of tradi-tional render. Figure 6. Relationship between the strength of multi-coat and single-coat cores of tradi-tional render. It was noted that the strength of the two-coat (base coat and top coat) specimens is slightly higher than that of the three-coat ones, as shown in Figure 9, with a change in the trend line slopes of about 0.62 (0.62 = 2.18 - 1.56), which results in a difference of 2.8 N/mm2 in comparison with higher value strengths (worse case). Where strengths are lower, the difference between them in the specimens with and without spatterdash is about 1 N/mm2, which supports the provision in EN 13914-15 whereby spatterdash is not regarded as a coating layer. 3.2 Compressive strength relationship between multi-coat and single-coat reference specimens (Figure 8, right) we can see that the spatterdash layer does not suffer significant degradation under the compressive strength test, noting that this is a square cross section, while the top coat suffers greater degradation and loss of its cross section area, which indicates that it was predominant in the failure mechanism during the test. (Figure 8, right) we can see that the spatterdash layer does not suffer significant degradation under the compressive strength test, noting that this is a square cross section, while the top coat suffers greater degradation and loss of its cross section area, which indicates that it was predominant in the failure mechanism during the test. A small comparative analysis of the multilayer specimens, with and without spatterdash was possible in the compressive strength test at 90 days (Figure 9). 1261 Experimental Study of the Compressive Strength of Multi-Coat Renders Figure 7. Relationship between the strength of multi-coat and single-coat reference specimens of traditional render Figure 7. Relationship between the strength of multi-coat and single-coat reference specimens of traditional render Figure 8. Square cross section reference specimen of multi-coat traditional render after compressive strength test (profile view, left; base of spatterdash layer, centre; surface of top coat, right). Figure 7. Relationship between the strength of multi-coat and single-coat reference specimens of traditional render Figure 7. Relationship between the strength of multi-coat and single-coat reference specimens of traditional render Figure 8. Square cross section reference specimen of multi-coat traditional render after compressive strength test (profile view, left; base of spatterdash layer, centre; surface of top coat, right). Figure 8. Square cross section reference specimen of multi-coat traditional render after compressive strength test (profile view, left; base of spatterdash layer, centre; surface of top coat, right). 1262 Soares et al. Materials Research Figure 9 Relationship between strengths of reference specimens of multi coat and sin gle coat traditional render at 90 days of age Figure 9 Relationship between strengths of reference specimens of multi-coat and sin-gle-coat traditional render at 90 days of age 3.4 The contribution of each coat to the multi- coat render 5. References This paper discussed alternative methods to better characterize the compressive strength of multi-coat renders on in-service building façades, instead of using standard lab specimens with a 40 mm thick single-coat mortar. The experimental results have shown that the global compressive strength of a multi-coat render depends on the number and type of coats: 1. CSN EN 13791 - Assessment of in-situ compressive strength in structures and pre-cast concrete components, 2007. Brussels, Belgium, 2005. 2. Concrete Society - Concrete core testing for strength. Technical report 11, London, the Concrete Society, 1976. 3. CEN EN 998-1: 2010 - Specification for mortar. Rendering and plastering mortar. Brussels, Belgium, 2010. • The spatterdash (the coat with the highest cement content) does not significantly influence the compressive strength results and has a non-uniform thickness;f 4. CEN EN 1015-11:1999 (Ed.1) - Methods of test for mortar for masonry. Part 11: Determination of flexural and compressive strength of hardened mortar. Brussels, Bel-gium, 1999. strength results and has a non-uniform thickness;f • The base coat thickness seems not to affect the relationship between multi-coat and single-coat specimens (as long as the thickness of the base coat is similar in the two types of specimens); 5. CEN EN 13914-1:2005 (Ed. 1) - Design, preparation and application of external rendering and plastering - External rendering. Brussels, Belgium, 2005. • The top coat (the coat with lowest cement content) has a high influence on the multi-coat mortar compressive strength, showing less compressive strength than the respective single-coat render (slope > 1). Therefore the contribution of this coat should be taken into account in the interpretation of in-service compressive strength results. 6. Gomes R, Veiga M, Brito J. Influence of execution procedures on current renders (in Portuguese). 1st National Congress on Construction Mortars : APFAC, 2005. http://www.apfac.pt/ congresso2005/comunicacoes/Paper%2021.pdf 7. Válek, J., Veiga, M. - Characterisation of mechanical properties of historic mortars - testing of irregular samples. Structural Studies, Repairs and Maintenance of Heritage Architecture IX. WIT Press 83, 2005, pp. 365 - 374. In this context, two alternative methods can be used to assess the in-service mechanical performance of multi-coat renders: 1) to produce lab specimens with the same layers, that can be difficult to represent exactly the applied render, without the substrate interaction; or 2) collect cores from on-site testing, for example after a pull-off test. Figure 9. Relationship between strengths of reference specimens of multi-coat and sin-gle-coat traditional render at 90 Figure 9. Relationship between strengths of reference specimens of multi-coat and sin-gle-coat traditional render at 90 days of age. 3.3 Compressive strength relationship between multi-coat cores and single-coat reference specimens So it has been an indication that there is a direct relationship between the strength determined with multi-coat traditional render and that found with single-coat traditional render (produced with the same mix and layer size as the multi-coat traditional base coat render). But it has only been possible to establish a relationship between the strengths obtained with single-coat render specimens and multi-coat render cores aged 7 days (where the traditional render cores do not have a top coat). Despite the acceptable correlations found from the analysis of the relationship between multi-coat traditional render and single-coat for Cs and RSs, treated separately, it was not possible to relate multi-coat traditional render cores with single-coat traditional reference specimens, as we can see in Figure 10, where the best relationship occurs for the 7-day results (no top coat) with a slope of 0.94 (close to y = x), with a satisfactory correlation coefficient (R2 ≈ 0.59). So a more detailed study on this issue is needed to establish a relationship associated with having a top coat, since this layer has a strong impact on compressive strength. For this 1263 Experimental Study of the Compressive Strength of Multi-Coat Renders Figure 10 Relationship between the strength of multi coat cores and reference speci mens of a single coat traditional render The results enable the assessment of the contribution of the different layers of multi-coat renders to overall compressive strength.i purpose, the performance of individual tests for each layer is necessary, varying cement contents and thicknesses. In addition it is important to carry out auxiliary tests, such as: the elasticity modulus and Poisson measurements. Then, the comparison of the previous experimental results with a finite element analysis should be followed to study in depth the effect of the top coat on the compressive strength value. It is concluded that the first coat, the spatterdash does not significantly influence compressive strength. The results listed below enable this last conclusion to be drawn: • At 7 days of age (multi-coat system consisting of spatterdash and base coat) - compressive strength values are very close to those obtained with single- coat traditional render (same constitution as the multilayer render’s base coat); 1264 Soares et al. Figure 9. Relationship between strengths of reference specimens of multi-coat and sin-gle-coat traditional render at 90 Materials Research • Comparison of a multi-coat render (spatterdash, base coat and top coat) with a two-layer one (base coat and top coat) at 90 days of age did not find any significant differences in the compressive strength values, a certain façade (including damage and non-damage areas). The core experimental results also showed a high variability with lower correlations than the ones observed in reference specimens, due to several factors, such as: the difference on sizes and slenderness of the cores, the possible microcracking damage during cores extraction; or the water suction of the substrate. However, cores represent the real in-service conditions and minimize the use of complex numerical methods; their combination with other in-situ techniques (to measure indirectly mechanical characteristics and pull- off as a probe technique) reduces the results variability and increase the accuracy of the multi-coat diagnosis based on cores assessment. • In practical application onsite the spatterdash’s thickness is not uniform and its contribution to the analysis of compressive strength can be disregarded. Meanwhile, the third layer, the top coat, had a strong influence on the compressive test value for multi-coat renders. It was found that this layer suffered greater degradation in the compressive strength test (with a greater loss of area than occurred with the spatterdash) and therefore further investigation is needed. Future experimental work should try to establish a relationship between the compressive strength of a multi-coat render and various formulations of the top coat. Also, the influence of each coat, with various thickness and formulations, should be detailed by further development. These experimental studies can support numerical relationships between the contributions of each coat to the overall of multilayer render compressive strength. The results at 90 days presented an R2 ≥ 0.77 and showed a slope near to those obtained at 14 and 28 days, indicating a stable trend with the age of the render, which is useful since in-service performance tests are commonly performed on existing buildings. 5. References The use of cores takes into account the real in-service conditions (the influence of the substrate; the degradation agents’ action; the application process; among others) and is useful in comparative assessments, for example: to compare the mechanical behaviour of a multi-coat rendering mortar on 8. Magalhães A, Veiga M.. Physical and mechanical characterisation of historic mortars. Application to the evaluation of the state of conservation. Materiales de Construcción. 2009;59(295):61-77. 9. Indelicato, F. - Estimate of concrete cube strength by means of different diameter cores: A statistical approach. Materials and Structures, Volume 33, Issue 3, 1997, pp. 131-138. 10. Flores-Colen, I. - Evaluation methodology of the in-service performance of façades from the predictive maintenance standpoint (in Portuguese), PhD Thesis on Civil Engi-neering, Technical University of Lisbon, Lisbon, Portugal, 2009, 487 p. Experimental Study of the Compressive Strength of Multi-Coat Renders 1265 11. Bazant, Z. P. - Size effect in blunt fracture, concrete, rock, metal. Journal of Engi-neering Mechanics, Volume 110, Issue 4, 1984, pp. 518-535. 18. Soares, A., Flores-Colen, I., de Brito, J. - Influence of slenderness on the compres-sive strength evaluation of cores of renders. Materials and Structures, Volume 48, Issue 5, 2015, pp. 1449-1460. 12. Bazant, Z., Xi, Y., Reid, S. G. - Statistical size effect in quasi- brittle structures: I. Is Weibull theory applicable? Journal of Engineering Mechanics, Volume 117, Issue 11, 1991, pp. 2609-2622. 19. Soares, A., Flores-Colen, I., de Brito, J. - The use of non-standard specimens to study the compressive strength of multi-coat renders. Journal of Testing and Evaluation, Volume 43, Issue 5, 2015, pp. 1055-1065. 13. Majeed, S. - Effect of specimen size on compressive, modulus of rupture and split-ting strength of cement mortar. Journal of Applied Sciences, Volume 11, Issue 3, 2011, pp. 584-588. 20. CEN EN 196-1 - Methods of testing cement. Part 1: Determination of strength. Brussels, Belgium, 2005. 21. CEN EN 1015-2:1998 (Ed.1) - Methods of test for mortar for masonry. Part 2: Bulk sampling of mortars and preparation of test mortars”. Brussels, Belgium, 1998. 14. ASTM C42 - Standard method of obtaining and testing drilled cores and sawn beams of concrete, PA, USA, 1992. 22. European Committee for Standardization. European Standard - EN 1015-12:2000 (Ed.1) - Methods of test for mortar for masonry. Part 12: Determination of adhesive strength of hardened rendering and plastering mortars on substrates. Brussels: European Committee for Standardization; 2000. 15. Gonnerman, H. 5. References - Effect of size and shape of test specimen on compressive strength of concrete. Philadelphia: ASTM Proc. 25, Part. II, 1925, pp. 237-250. 16. Drdácký, M., Masín, D., Mekonome, M.D., Slízková, Z. - Compression testes on non-standard historic mortar specimens, HMC08 - Historical Mortars Conference, LNEC, Lisbon, 2008, CD. 23. RILEM - MDT.D.3 Determination “in situ” of the adhesive strength of rendering and plastering mortars to their substrate. RILEM Recommendation, Materials and Struc-tures, Volume 37, Issue 7, 2004, pp. 488-490. 17. Benedetti, A., Luca, P. - Experimental characterization of mortar by testing on small specimens. 15th International Brick and Block Masonry Conference. Florianópolis, Brazil, 2012, CD.
https://openalex.org/W2752211585	https://europepmc.org/articles/pmc6168974?pdf=render	English	null	Albumin Counteracts Immune-Suppressive Effects of Lipid Mediators in Patients With Advanced Liver Disease	Clinical gastroenterology and hepatology	2,018	cc-by	10,575	Albumin Counteracts Immune-Suppressive Effects of Lipid Mediators in Patients With Advanced Liver Disease Louise China,* Alexander Maini,* Simon S. Skene,‡ Zainib Shabir,‡ Yvonne Sylvestre,‡ Romain A. Colas,§ Lucy Ly,§ Natalia Becares Salles,* Vittorio Belloti,* Jesmond Dalli,§ Derek W. Gilroy,* and Alastair O’Brien* *Division of Medicine, University College London, London, United Kingdom; ‡University of London Comprehensive Clinical Trials Unit, London, United Kingdom; and §Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom BACKGROUND & AIMS: Patients with acute decompensation and acute-on-chronic liver failure (AD/ACLF) have immune dysfunction, which increases their risk for infections; however, there are no effective treatments to restore their immune function. We investigated whether the potentially immune-restorative effects of albumin are mediated by its effects on prostaglandin E2 (PGE2) and other lipids. We analyzed bloods samples from 45 of 79 patients with AD/ACLF and serum levels of albumin less than 30 g/L for whom infusion of 20% human albumin solution (HAS) increased serum levels of albumin 30 g/L or more in a feasibility study of effects of 20% HAS. Immune function was determined by comparison of macrophage function following addition of plasma samples. We also used samples from 12 healthy individuals. We measured binding of plasma proteins to PGE2 and serum levels of endotoxin (lipopolysaccharide) and cytokines; using 10 patients’ samples, we investigated the effects of PGE2 inhibitors. We performed a comprehensive lipid metabolomic analysis using samples from 10 different patients, before and after HAS administration. At baseline, AD/ACLF patient plasma induced signiﬁcantly lower production of tumor necrosis factor by healthy macrophages than plasma from healthy individuals (P < .0001). Plasma from patients after HAS infusion induced signiﬁcantly higher levels of tumor necrosis factor produc- tion by macrophages (19.5 – 4.8 ng/mL) compared with plasma collected before treatment (17.7 – 4.5 ng/mL; P [ .0013). There was a signiﬁcantly lower proportion of plasma protein (albumin) binding to PGE2 from patients with AD/ACLF plasma (mean, 61.9%) compared with plasma from control subjects (77.1%; P [ .0012). AD/ACLF plasma protein binding to PGE2 increased following HAS treatment compared with baseline (mean increase, 8.7%; P < .0001). Circulating levels of PGE2, lipopolysaccharide, and inﬂammatory or anti-inﬂammatory cytokines were higher in patients with AD/ACLF than healthy volunteers. Unexpectedly, HAS infusion had no effect on mediator levels. Principal component analysis of baseline levels of lipids that induce or resolve inﬂammation identiﬁed 2 distinct groups of patients that differed according to baseline plasma level of lipopolysaccharide. Abbreviations used in this paper: ACLF, acute-on-chronic liver failure; AD, acute decompensation; CI, conﬁdence interval; CRP, C-reactive protein; HAS, human albumin solution; IL, interleukin; LM, lipid mediator; LPS, lipopolysaccharide; MDM, monocyte-derived macrophages; PGE2, prostaglandin E2; SD, standard deviation; TNF, tumor necrosis factor; WCC, white cell count. Clinical Gastroenterology and Hepatology 2018;16:738–747 Clinical Gastroenterology and Hepatology 2018;16:738–747 Plasma Protein Binding Capacity Paired plasma samples pretreatment/post-treatment with 20% HAS were obtained from 52 of 79 patients in the ATTIRE feasibility trial. In 45 of 52 patients, the post- treatment sample corresponded to restoration of serum albumin 30 g/L (the primary endpoint) on mean treatment day 3.29 (standard deviation [SD], 1.27). These patients had mean pretreatment serum albumin 23.98 g/L (range, 12–29 g/L). In the other 7 of 52 patients the post-treatment sample was when patient had reached highest serum albumin level, and a sample had been taken that day. Plasma PGE2 binding was assessed in these samples with healthy volunteer samples (n ¼ 12) as comparator. Speciﬁcally, we aimed to conﬁrm that elevated circulating PGE2 levels contributed to immune suppres- sion; examine whether exogenous albumin improved PGE2-albumin binding and/or increased catalysis; compare PGE2 binding in commercial albumin prepara- tions; determine whether any improvement in immune dysfunction observed following 20% HAS infusion was via a PGE2 effect; and examine the potential interaction of infused albumin with other plasma lipids (including proresolving mediators, molecules with host protective actions10), proinﬂammatory and anti-inﬂammatory cytokines, and endotoxin. Finally, we correlated laboratory ﬁndings with patient clinical characteristics. The amount of PGE2 bound by plasma was deter- mined using equilibrium dialysis (Thermo Scientiﬁc Single-Use RED [rapid equilibrium dialysis] Plate, IL), which enabled quantiﬁcation of bound versus free PGE2 via postdialysis sample scintillation counting (Supplementary Methods). The 20% HAS from commercial suppliers Zenalb (BPL Herts, UK), Albu- norm (Octapharm, Manchester, UK), and Alburex (CSL Behring, West Sussex, UK) including 2 different batches of Zenalb and Alburex were assessed and compared with fatty acid free albumin from human serum (Sigma-Aldrich, UK). Albumin concentration was diluted to 20 g/L (300 mM) and checked using bromocresol green. Albumin Counteracts Immune-Suppression Albumin Counteracts Immune-Suppression May 2018 739 by London-Brent research ethics committee (ref:15/LO/ 0104). All authors had access to study data and reviewed and approved the ﬁnal manuscript. Studies were per- formed as follows with laboratory researcher blinded to whether the sample was pre- or post-HAS infusion. See editorial on page 633, and related article on page 748. A defective immune response in patients with acute decompensation (AD) or acute-on-chronic liver fail- ure (ACLF) is widely considered to underlie susceptibility to bacterial infection.1–3 However, despite multiple studies the mechanisms underlying immune dysfunction in AD/ACLF remain unclear. We developed a model in which healthy volunteers’ monocyte-derived macrophages (MDMs) were incubated with plasma from patients with AD/ACLF and measured tumor necrosis factor (TNF) pro- duction, a validated marker of monocyte function in critical illness.4 Using this model, we demonstrated elevated plasma prostaglandin E2 (PGE2) and its potential role in immune suppression in patients with AD/ACLF. We also proposed a beneﬁcial effect of transfusing 20% human albumin solution (HAS) to antagonize PGE2’s effects.5 A Laboratory Outcomes The key secondary endpoint for ATTIRE feasibility study was change in immune function determined by patient plasma-induced healthy volunteer MDM dysfunction, as measured by endotoxin-stimulated TNF production (lipopolysaccharide [LPS]; Salmonella abortus equi S-form [TLRgrade, Enzo Life Science], NY), for 4 hours in presence of 25% patient plasma pre- and post-HAS treatment. TNF was measured with enzyme- linked immunosorbent assay (R&D systems, MN) as previously.5 Plasma samples analyzed were from admission (pre-HAS infusion) compared with samples once serum albumin had reached 30 g/L. The same assay was repeated using a monocyte cell line (mono- mac 6) for validation. Experiments were in a single centralized laboratory. Laboratory and matching clin- ical data were exchanged simultaneously between researcher and statisticians at the Comprehensive Clinical Trials Unit at University College London (Supplementary Methods). Albumin has been reported to bind and catalyze PGE2 inactivation.6 Albumin is synthesized in the liver, therefore levels decrease in AD/ACLF, and so PGE2 should be more bioavailable. Defective functional binding capacity of albu- min has been described in cirrhosis,7 again theoretically further enhancing bioavailability of PGE2. However the actual PGE2-albumin binding relationship in liver disease has never been explored. Studies have shown other poten- tial immunomodulatory roles for albumin8,9 but these have used samples from single center observational cohorts. We performed immune function analysis of patients with AD using samples collected from a feasibility trial in preparation of the ATTIRE trial (Albumin To prevenT Infection in chronic liveR failure). Our feasibility trial included79patientswithAD/ACLFwho received20% HAS. An accompanying manuscript in this issue details the clin- ical outcomes of these patients. The current article provides mechanistic insights into the potential immune restorative effect of targeted 20% HAS infusions in AD/ACLF. Albumin Counteracts Immune-Suppressive Effects of Lipid Mediators in Patients With Advanced Liver Disease Sample analyses after HAS treatment indicated that albumin regulates circulating levels of lipid mediators, but this effect was distinct in each group. CONCLUSIONS: CONCLUSIONS: Analysis of bloodsamples from patients with AD/ACLF participating in a feasibilitystudy of 20% HAS infusions has shown that infusions to raise serum albumin above 30 g/L reversed plasma-mediated immune dysfunction by binding and inactivating PGE2. We also describe a method to classify the inﬂammatory response in AD/ACLF, based on lipid proﬁle, which could improve identiﬁcation of patients most likely to respond to HAS treatment. A randomized controlled trial is needed to determine whether these effects of HAS reduce infections in AD/ACLF. Trial registered with European Medicines Agency (EudraCT 2014-002300-24) and adopted by NIHR (ISRCTN14174793). Keywords: ATTIRE Trial; Resolution Phase Lipid Mediators; Immune Regulation; ALD. Most current article © 2018 by the AGA Institute. Published by Elsevier, Inc. This is an open access article under the CC BY license (http://creativecommons.org/ licenses/by/4.0/). 1542-3565/$36.00 https://doi.org/10.1016/j.cgh.2017.08.027 Plasma-Mediated Immune Dysfunction Pre– and Post–20% Human Albumin Solution Infusions We selected (while blinded) aliquots from 10 sam- ple pairs that had showed at least a 15% difference in MDM TNF production following HAS infusion. A total of 15% was considered representative because 20% HAS infusions produced a mean >14% increase in MDM TNF (see results section). Experiments were performed with healthy volunteer plasma as compar- ator. PGE2 receptor antagonists AH6809 50 mM (EP1-3 antagonist) and MF498 10 mM (EP4 antagonist) (ie, pan PGE2 receptor blockade) were added to samples before LPS stimulation and TNF measured. Samples from Days 4, 5, and 10 of HAS treatment were also used. Patient plasma treatment signiﬁcantly reduced endotoxin (LPS)-stimulated production of TNF from healthy MDMs compared with healthy volunteer plasma (P < .0001) (Figure 1A). There was a signiﬁcant increase in MDM TNF production of 14.3% (95% con- ﬁdence interval [CI], 5.1%–23.5%; 17.7 4.5 ng/mL to 19.5 4.8 ng/mL; P ¼ .0013) (Figure 1B) following addition of post-HAS treatment plasma compared with pretreatment from 45 paired samples. In total, 30 of 45 (78%) had improved MDM TNF production post-treatment. A differentiated monocyte cell line showed similar ﬁndings of 10.2% (95% CI, 2.5%–17.9%; P ¼ .014) (Supplementary Figure 1A). There was a trend toward increased TNF production from patients not incrementing 30 g/L (n ¼ 7) compared with pretreatment (Supplementary Figure 1B). There was no change in mean white cell count (WCC) or C-reactive protein (CRP) between pre and post samples but serum bilirubin was reduced by a mean 25% (Supplementary Table 2). Lipid Mediator Metabolomic Data Samples from 10 patients pre- and post-HAS infusion from a top recruiting sites were chosen for analysis in view of the complexity of processing required and need for standardized collection and storage (Supplementary Methods). Plasma PGE2 concentrations pretreatment were highly variable but substantially elevated with a mean 52.5 pg/mL (SD, 44.6; n ¼ 10) compared with published healthy volunteer concentrations using this technique (mean, 4.1; SD, 0.2 pg/mL).12 Albumin infusion had no overall effect on total plasma PGE2 concentrations, which measure free and albumin-bound PGE2 (pretreatment, 52.5 [13.4] pg/mL vs post-treatment, 49.9 [8.1] pg/mL) (Supplementary Tables 3 and 4). Lipopolysaccharide Detection and Cytokine Measurement LPS and proinﬂammatory and anti-inﬂammatory cytokines were assayed in the 45 paired patient plasma samples using HEK293 cells and BD Cytometric Bead Array Human Soluble Protein Kit (BD Biosciences, Oxford, UK) (Supplementary Methods). Targeted 20% Human Albumin Solution Infusions Improved Acute Decompensation and Acute-on-Chronic Liver Failure Plasma Ability to Bind Prostaglandin E2 Both by Increasing Albumin Concentration and Functional Capacity With No Effect on Overall Prostaglandin E2 Concentration Targeted 20% Human Albumin Solution Infusions Improved Acute Decompensation and Acute-on-Chronic Liver Failure Plasma Ability to Bind Prostaglandin E2 Both by Increasing Albumin Concentration and Functional Capacity With No Effect on Overall Prostaglandin E2 Concentration Statistical Analysis For plasma analysis (a-d) a paired Student t test compared pre- and post-HAS treatment groups (Prism 7, CA). For lipid metabololipidomic analyses, Simca 14.1 (Malmo, Sweden) was used as below. Albumin bound PGE2 with very low afﬁnity and calculated dissociation constant was approximately 280 mM. This low binding afﬁnity suggests that decreasing serum albumin to AD/ACLF patient levels combined with observed increases in PGE2 concen- tration could result in increases in free circulating PGE2 to pathophysiological levels. To illustrate this we performed theoretical calculations (Supplementary Table 5). Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Monocyte-Derived Macrophage Functional Studies Results Recruitment and Baseline Characteristics Recruitment and Baseline Characteristics Plasma-Mediated Immune Dysfunction Pre– and Post–20% Human Albumin Solution Infusions Clinical Gastroenterology and Hepatology Vol. 16, No. 5 740 China et al 740 China et al 740 Study Design and Patients ATTIRE’s protocol paper was published11 and the full protocol is available online. Ethical approval was granted Plasma-Mediated Immune Dysfunction Pre– and Post–20% Human Albumin Solution Infusions Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Recruitment and Baseline Characteristics LPS MDMs TNF production in presence of healthy volunteer plasma was 6.88 ng/mL more in presence of AD plasma (CI, 4.85–8.91 ng/mL; P < 0001) (B) LPS MDM TNF production in presence of plasma pre and post HAS treatment (n 45 patients Figure 1. Targeted 20% HAS infusions reverse immune dysfunction in AD/ACLF by improving ability of patients with AD/ ACLF plasma to bind PGE2. (A) Endotoxin (LPS) stimulated MDM TNF production in presence of patient plasma pretreatment with 20% HAS (n ¼ 45 patients) compared with nonautologous healthy volunteer plasma. LPS MDMs TNF production in presence of healthy volunteer plasma was 6.88 ng/mL more in presence of AD plasma (CI, 4.85–8.91 ng/mL; P < .0001). (B) LPS MDM TNF production in presence of plasma pre- and post-HAS treatment (n ¼ 45 patients incremented serum albumin >30 g/L). Mean post-treatment TNF increase 1.75 ng/mL (0.72–2.77; P ¼ .0013), 14.5% (5.1%–23.5%). (C) Percentage PGE2/3H-PGE2 bound to healthy volunteer and AD/ACLF plasma protein using equilibrium dialysis. Post-HAS treatment plasma binds more PGE2 than pre-HAS (mean increase, 8.7%; CI, 5.2%–12.1%; P < .0001; n ¼ 45). (D) Percentage PGE2/3H-PGE2 bound in different HAS products or Sigma albumin diluted to 20 g/L albumin in phosphate-buffered saline (n ¼ 3). (E) LPS MDM TNF production in presence of pretreatment patient plasma (n ¼ 10) in presence/absence of EP2 (AH6890-50 mM) and EP4 (MF498-10 mM) receptor antagonists compared with post-treatment effect. (F) LPS MDM TNF production in presence of AD/ACLF plasma Day 5 and 10 post-treatment with 20% HAS. (Supplementary Figure 1C) suggesting functional improvement in binding capacity. post-HAS treatment with mean increase of 8.7% (95% CI, 5.2%–12.1%; P < .0001) (Figure 1C). The binding may have improved because of increased plasma albumin concentration following treatment. To investigate functional alterations in binding we selected 23 patient samples with a greater or equal improvement in binding compared with overall mean 8.7% PGE2 bound (mean, 16.1%; 95% CI, 6.0%–15.0%; P < .0001). Pretreatment and post-treatment plasma was diluted to 18 g/L albumin and post-treatment plasma bound signiﬁcantly more PGE2 than pretreatment (mean increase, 10.9%; 95% CI, 5.2%–16.7%; P ¼ .0007) Recruitment and Baseline Characteristics Baseline characteristics were as follows: mean age, 53.4 years; male, 66%; and alcohol primary cause of cirrhosis, 96% (Supplementary Table 1). Mean Model for End-Stage Liver Disease score was 20.9 (SD, 6.62); 17 of 79 patients had 1 extra hepatic organ dysfunction at baseline and 21 (27%) ACLF grade 1–3. Baseline albumin levels were <25 g/L in 67%. AD/ACLF plasma bound a mean of 15.2% less PGE2 compared with healthy volunteer plasma (n ¼ 12; 77.1% vs 61.9%; P ¼ .0012). The 45 paired patient samples showed an improved ability to bind PGE2 Albumin Counteracts Immune-Suppression 741 741 May 2018 Figure 1. Targeted 20% HAS infusions reverse immune dysfunction in AD/ACLF by improving ability of patients with AD/ ACLF plasma to bind PGE2. (A) Endotoxin (LPS) stimulated MDM TNF production in presence of patient plasma pretreatment with 20% HAS (n ¼ 45 patients) compared with nonautologous healthy volunteer plasma. LPS MDMs TNF production in presence of healthy volunteer plasma was 6.88 ng/mL more in presence of AD plasma (CI, 4.85–8.91 ng/mL; P < .0001). (B) LPS MDM TNF production in presence of plasma pre- and post-HAS treatment (n ¼ 45 patients incremented serum albumin >30 g/L). Mean post-treatment TNF increase 1.75 ng/mL (0.72–2.77; P ¼ .0013), 14.5% (5.1%–23.5%). (C) Percentage PGE2/3H-PGE2 bound to healthy volunteer and AD/ACLF plasma protein using equilibrium dialysis. Post-HAS treatment plasma binds more PGE2 than pre-HAS (mean increase, 8.7%; CI, 5.2%–12.1%; P < .0001; n ¼ 45). (D) Percentage PGE2/3H-PGE2 bound in different HAS products or Sigma albumin diluted to 20 g/L albumin in phosphate-buffered saline (n ¼ 3). (E) LPS MDM TNF production in presence of pretreatment patient plasma (n ¼ 10) in presence/absence of EP2 (AH6890-50 mM) and EP4 (MF498-10 mM) receptor antagonists compared with post-treatment effect. (F) LPS MDM TNF production in presence of AD/ACLF plasma Day 5 and 10 post-treatment with 20% HAS. Figure 1. Targeted 20% HAS infusions reverse immune dysfunction in AD/ACLF by improving ability of patients with AD/ ACLF plasma to bind PGE2. (A) Endotoxin (LPS) stimulated MDM TNF production in presence of patient plasma pretreatment with 20% HAS (n ¼ 45 patients) compared with nonautologous healthy volunteer plasma. 20% Human Albumin Solution Infusions Seem to Improve Immune Function in Patients With Acute Decompensation and Acute-on-Chronic Liver Failure by Reversing the Immune Suppressive Effect of Prostaglandin E2 With Effect Maintained to at Least Day 10 of Treatment 20% Human Albumin Solution Infusions Seem to Improve Immune Function in Patients With Acute Decompensation and Acute-on-Chronic Liver Failure by Reversing the Immune Suppressive Effect of Prostaglandin E2 With Effect Maintained to at Least Day 10 of Treatment was conducted in accordance with published criteria including matching retention time and at least 6 diag- nostic ions in tandem mass spectrum12 (Figure 2, Supplementary Tables 3 and 4). We identiﬁed mediators from each major essential fatty acid metabolome including 13 series resolvins 1, protectins 1, and lipoxins A4. Using multivariate analysis of plasma LM proﬁles pre and post albumin treatment we found that each of these groups segregated into 2 distinct clusters (Figure 3A and B). These data indicate that albumin treatment regulates circulating LM levels. However, overall LM levels pre and post albumin treat- ment did not demonstrate statistically signiﬁcant changes (Supplementary Tables 3 and 4). Therefore, we questioned whether responses in plasma LM proﬁles following albumin were dependent on pretreatment LM levels. LPS-induced TNF production from MDMs pretreated with pan-PGE2 receptor blockade (EP1-3-AH6890 and EP4-MF698) before addition of pre-HAS treatment plasma was increased to a similar level as when post-HAS plasma was added (without PGE2 antagonists). Mean increase was 3.51 ng/mL (P ¼ .0013; 95% CI, 1.78–5.24) (Figure 1E). However pan-PGE2 receptor blockade had no signiﬁcant effect on MDMs treated with post–20% HAS plasma (P ¼ .0945). These antagonists had no effect on MDM TNF production when added to healthy plasma samples (Supplementary Figure 1D). The increased MDM TNF production between pre- and post-HAS treatment was maintained but not increased up to Day 10 of treatment in 10 samples analyzed (Figure 1F, Supplementary Figure 1E). Principal component analysis of pretreatment LM proﬁles identiﬁed 2 distinct patient groups characterized by 5 patients per group (Figure 3B). Assessment of LM proﬁles in each group pre albumin treatment demonstrated a hypoactivated proﬁle (Group 1) with reduced concentrations of several proresolving and inﬂammation-initiating mediators including n-3 docosa- pentaenoic acid–derived protectins, AA-derived lipoxins and prostaglandins, thromboxane B2, and leukotriene B4 (Figure 3C). The second hyperactivated group demon- strated overall elevated LM concentrations (Group 2; Figure 3B and C). Patients in the hyperactivated group had elevated WCC, temperature, cytokine, and CRP levels and statistically signiﬁcant increases in plasma endotoxin concentration (Table 1). Investigation of peripheral blood LM levels pre and post albumin administration demonstrated a re-equilibration of several mediator families. 20% Human Albumin Solution Infusions Seem to Improve Immune Function in Patients With Acute Decompensation and Acute-on-Chronic Liver Failure by Reversing the Immune Suppressive Effect of Prostaglandin E2 With Effect Maintained to at Least Day 10 of Treatment LM concentrations for several of the families identiﬁed in the hyperactivated group were found to be decreased, whereas in the hypoactivated group mediator concentrations increased post albumin treatment (Figure 4A–F). These results demonstrate that plasma LM proﬁles identify 2 distinct patients groups, hypoactivated and hyperactivated, and regulation of plasma LM proﬁles by albumin is distinct in each. The endotoxin and cytokine levels did not change signiﬁ- cantly following HAS in either group and there was no difference in clinical outcomes. Targeted20% HumanAlbumin SolutionInfusions had No Signiﬁcant Effect on Elevated Plasma Concentrations of Lipopolysaccharide and Proinﬂammatory/Anti-Inﬂammatory Cytokines Seen in Patients With Acute Decompensation and Acute-on-Chronic Liver Failure There was a trend toward reduction but no signiﬁ- cant differences in total plasma proinﬂammatory and anti-inﬂammatory cytokine levels assayed (TNF, interleukin [IL] 1b, IL6, IL10, and IL8) and LPS concen- trations in 45 paired samples (Supplementary Table 6). Principal Component Analysis of Baseline (Pretreatment) Inﬂammation Initiating and Proresolving Plasma Lipid Mediators Identiﬁed 2 Distinct Acute Decompensation and Acute-on-Chronic Liver Failure Patient Groups and Targeted 20% Human Albumin Solution Infusions Demonstrated Distinct Responses Between These Groups Commercially Available 20% Human Albumin Solution Tested Bound Prostaglandin E2 to a Similar Degree There were no signiﬁcant differences in PGE2 binding among samples tested (Figure 1D) and values were less than healthy volunteer plasma binding. Clinical Gastroenterology and Hepatology Vol. 16, No. 5 742 China et al Clinical Gastroenterology and Hepatology Vol. 16, No. 5 China et al 742 20% Human Albumin Solution Infusions Seem to Improve Immune Function in Patients With Acute Decompensation and Acute-on-Chronic Liver Failure by Reversing the Immune Suppressive Effect of Prostaglandin E2 With Effect Maintained to at Least Day 10 of Treatment Discussion We investigated plasma lipid mediator (LM) proﬁles for essential fatty acid–derived (docosahexaenoic acid, n-3 docosapentaenoid acid, eicosapentaenoic acid, and arachidonic acid), proresolving mediators: resolvins, protectins, maresins, and lipoxins in 10 plasma samples pre–20% HAS infusion and once serum albumin had reached 30 g/L following treatment. We quantiﬁed the classic inﬂammation-initiating eicosanoids (prostaglan- dins, thromboxane B2, and leukotrienes). Identiﬁcation This represents the ﬁrst demonstration of a potential pharmacological immune restorative role for 20% HAS infusions in patients with AD/ACLF through its ability to bind PGE2 using samples from a multicenter inter- ventional trial. We show that 20% HAS infusions seemed to reverse AD/ACLF patient plasma-induced macrophage dysfunction restoring TNF production to- wards levels seen when macrophages were incubated with healthy plasma. We had no control arm and this May 2018 May 2018 Albumin Counteracts Immune-Suppression 743 743 Figure 2. AD/ACLF plasma proresolving and inﬂammation-initiating mediator proﬁles. Plasma collected pre and post albumin administration and lipid mediators proﬁled using liquid chromatography–tandem mass spectrometry LM metabololipidomics. (A) Representative multiple reaction chromatograms for identiﬁed lipid media- tors and (B) tandem mass spectrometry spectrum used in identiﬁcation of AT-lipoxins A4 and MaR1. Representative of 10 patients. AA, arachidonic acid; DHA, docosahexae- noic acid; DPA, docosa- pentaenoic acid; EPA, eicosapentaenoic acid; m/z, mass to charge ratio. endotoxin or exerts a beneﬁcial effect on proin- ﬂammatory cytokines.16–18 Immune function is an extremely complex process and we designed a prag- matic assay to investigate samples from multiple sites that we have validated by showing similar effects in freshly isolated monocytes from patients with ACLF.19 We show for the ﬁrst time in a prospective trial that 20% HAS infusion reversed the immune suppressive effects of PGE2 in AD/ACLF by improving plasma binding to this molecule, thereby inactivating it. This effect persisted in samples tested to Day 10 but did not improve once serum albumin was >30 g/L. PGE2 binds albumin but no other plasma proteins20 and we found albumin to have a very low binding afﬁnity for PGE2 supporting the presence of free unbound PGE2 within plasma at pathophysiological concentrations. Albumin infusion improved plasma protein binding to PGE2. Studies using PGE2 (E-prostanoid) receptor antagonists demonstrated a similar immune restorative effect to 20% HAS infusion and had no effect in samples post- HAS infusion supporting this immune restorative effect of albumin occurring via PGE2 inhibition. Discussion LM proﬁles interrogated using partial least square discriminant analysis. (A) 2D score plot with baseline and post treatment plasma LM proﬁles interrogated using principle component analysis. 2D score plot (top), 2D loading plot (bottom). (B) 2D loading plot with pre-treatment plasma pro-resolving and inﬂammation- initiating mediators identiﬁes 2 distinct groups: group 1 (red) and group 2 (blue). (C) Assessment of lipid mediator proﬁles in each of these groups pre albumin treatment demonstrated a hypo-activated proﬁle (Group 1) with reduced concentrations of several pro- resolving and inﬂammation initiating mediators. The second hyper-activated group was characterized by overall elevated lipid mediator concentrations (Group 2). Results mean standard error of the mean, representative of 10 patients. Figure 3. Albumin administrationshifts AD/ACLF plasma LM proﬁles and prealbumin LM proﬁles identiﬁes 2distinct AD/ACLF patient groups with a hyperactivated (group 2) and hypoactivated (group 1) LM phenotype. LM proﬁles interrogated using partial least square discriminant analysis. (A) 2D score plot with baseline and post treatment plasma LM proﬁles interrogated using principle component analysis. 2D score plot (top), 2D loading plot (bottom). (B) 2D loading plot with pre-treatment plasma pro-resolving and inﬂammation- initiating mediators identiﬁes 2 distinct groups: group 1 (red) and group 2 (blue). (C) Assessment of lipid mediator proﬁles in each of these groups pre albumin treatment demonstrated a hypo-activated proﬁle (Group 1) with reduced concentrations of several pro- resolving and inﬂammation initiating mediators. The second hyper-activated group was characterized by overall elevated lipid mediator concentrations (Group 2). Results mean standard error of the mean, representative of 10 patients. tionshifts AD/ACLF plasma LM proﬁles and prealbumin LM proﬁles identiﬁes 2distinct AD/ACLF patient Table 1. Discussion Unex- pectedly plasma endotoxin and proinﬂammatory/ anti-inﬂammatory cytokine concentrations were unaf- fected by albumin infusions, therefore effects observed were not via direct modulation of these. We found no may be caused by other patient care aspects because median time between samples was 4 days, with an overall 25% improvement in bilirubin observed, which cannot be excluded as a confounder. We again show signiﬁcantly elevated prostaglandins and demonstrate for the ﬁrst time proresolving mediators resolvins, protectins, maresins, and lipoxins in patients with AD/ ACLF. These autacoids stimulate key cellular resolution events, enhancing macrophage apoptotic cell clear- ance.10 Finally principal component analysis of LMs divided patients into a hyperinﬂammatory and hypoin- ﬂammatory proﬁle that could be differentiated by plasma endotoxin concentrations from 10 patients analyzed with nonsigniﬁcantly elevated WCC, tempera- ture, cytokine, and CRP levels. Importantly, 20% HAS infusions seemed to equilibrate the inﬂammatory bal- ance of inﬂammation initiating eicosanoids and pro- resolving mediators between these groups without affecting endotoxin or cytokine levels. These data sug- gest a further novel immune restorative effect for albumin. may be caused by other patient care aspects because median time between samples was 4 days, with an overall 25% improvement in bilirubin observed, which cannot be excluded as a confounder. We again show signiﬁcantly elevated prostaglandins and demonstrate for the ﬁrst time proresolving mediators resolvins, protectins, maresins, and lipoxins in patients with AD/ ACLF. These autacoids stimulate key cellular resolution events, enhancing macrophage apoptotic cell clear- ance.10 Finally principal component analysis of LMs divided patients into a hyperinﬂammatory and hypoin- ﬂammatory proﬁle that could be differentiated by plasma endotoxin concentrations from 10 patients analyzed with nonsigniﬁcantly elevated WCC, tempera- ture, cytokine, and CRP levels. Importantly, 20% HAS infusions seemed to equilibrate the inﬂammatory bal- ance of inﬂammation initiating eicosanoids and pro- resolving mediators between these groups without affecting endotoxin or cytokine levels. These data sug- gest a further novel immune restorative effect for albumin. Albumin is considered to have immune modulatory effects in AD/ACLF13–15 but no prospective trial has identiﬁed mechanistic action beyond volume expansion. We previously demonstrated its potential to antagonize the effects of PGE2 and others have suggested it binds Clinical Gastroenterology and Hepatology Vol. 16, No. 5 744 China et al Figure 3. Albumin administrationshifts AD/ACLF plasma LM proﬁles and prealbumin LM proﬁles identiﬁes 2distinct AD/ACLF patient groups with a hyperactivated (group 2) and hypoactivated (group 1) LM phenotype. Discussion Clinical Characteristics of Group 1 (Hypoinﬂammatory) and Group 2 (Hyperinﬂammatory) Patients as Deﬁned by Lipid Mediator Principal Component Analysis at Baseline (n ¼ 5 Patients per Group) Group 1: hypoinﬂammatory lipid mediator proﬁle, mean (SD) Group 2: hyperinﬂammatory lipid mediator proﬁle, mean (SD) MELD 18.9 (5.4) 20.1 (7.9) Age 45.66 (13.52) 48.63 (13.85) Pre 20% HAS Post 20% HAS Pre 20% HAS Post 20% HAS Serum albumin (g/L) 22.4 (6.1) 28.2 (5.2) 20.8 (3) 29.6 (4.5) Temperature (C) 36.84 (1.0) 37.96 (1.1) White cell count (x109/L) 11.24 (4.5) 9.18 (5.83) 18.16 (14.2) 4.6 (16.4) CRP (mg/mL) 70.8 (76) 107.0 (46.8) 92.6 (93.3) 46.0 (45.4) Heart rate (bpm) 104.4 (14.5) Not available 106.8 (23.7) Not available Endotoxin (pg/mL) 3.7 (2.3) 7.2 (3.5) 23.44 (12.3) 19.0 (13.3) TNF (pg/mL) 0.75 (0.5) 3.0 (1.7) 2.27 (1.7) 5.4 (7.5) IL1b (pg/mL) 0.67 (1.5) 1.6 (1.7) 4.13 (6.8) 2.3 (2.8) IL6 (pg/mL) 227.7 (407.2) 7135 (6649) 898.1 (1949.4) 14,121 (31,356) IL8 (pg/mL) 372.9 (96.9) 576.7 (285.1) 442.3 (173.2) 828.2 (1271.9) IL10 (pg/mL) 0.6 (3.2) 120.0 (110.5) 12.9 (12.6) 239.9 (13.3) CRP, C-reactive protein; HAS, human albumin solution; IL, interleukin; MELD, model for end-stage liver disease; SD, standard deviation; TNF, tumor necrosis factor. Table 1. Clinical Characteristics of Group 1 (Hypoinﬂammatory) and Group 2 (Hyperinﬂammatory) Patients as Deﬁned by Lipid Mediator Principal Component Analysis at Baseline (n ¼ 5 Patients per Group) CRP, C-reactive protein; HAS, human albumin solution; IL, interleukin; MELD, model for end-stage liver disease; SD, standard deviation; TNF, tumor necrosis factor. Albumin Counteracts Immune-Suppression May 2018 745 Figure 4. Albumin differentially regulates plasma LM proﬁles in patients with AD/ACLF with hyperactivated and hypoactivated LM phenotypes. Plasma pre and post albumin administration was LM proﬁled using liquid chromatography–tandem mass spectrometry based LM metabololipidomics and quantiﬁed using multiple reaction monitoring. (A–F) Results represent cumulative levels of proresolving mediator and inﬂammation-initiating eicosanoids found to be regulated by albumin admin- istration. Five patients per group per interval. Figure 4. Albumin differentially regulates plasma LM proﬁles in patients with AD/ACLF with hyperactivated and hypoactivated LM phenotypes. Plasma pre and post albumin administration was LM proﬁled using liquid chromatography–tandem mass spectrometry based LM metabololipidomics and quantiﬁed using multiple reaction monitoring. (A–F) Results represent cumulative levels of proresolving mediator and inﬂammation-initiating eicosanoids found to be regulated by albumin admin- istration. Five patients per group per interval. Discussion HAS rectiﬁes this LM imbalance leading to a normalized, appropriate inﬂammatory response with potential improved outcome. IV, intravenous 746 China et al Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Figure 5. Schematic version of our hypothesis that prophylactic human albumin infusions improve immune response in ACLF via 2 mechanisms. (A) HAS increases circulating albumin concentration and function improving binding of immunosuppressive PGE2, reducing free/bioactive PGE2 and restoring monocyte/macrophage function. (B) Patients with ACLF can be divided into hypoinﬂammatory and hyperinﬂammatory responses deﬁned by LM metabolomics, both with potential adverse outcomes. HAS rectiﬁes this LM imbalance leading to a normalized, appropriate inﬂammatory response with potential improved outcome. IV, intravenous samples, which might have been expected to fall if the patients were substantially better. immunophenotyping has never previously been described in liver disease. Although the hyperactivated group had elevated concentrations of endotoxin and cytokine pre- treatment, these levels were unaffected by HAS infusion. These data therefore offer a completely novel opportunity to study the effect of albumin on the immune system. Figure 5 provides a schematic version of our hypothesis. Plasma LM proﬁling of the 4 major essential fatty acid bioactive metabolomes demonstrated albumin adminis- tration caused a shift in peripheral blood LM proﬁles. Post hoc analysis of prealbumin LM proﬁles identiﬁed 2 distinct groups, a hyperactivated proﬁle with elevated levels of inﬂammation-initiating eicosanoids and proresolving me- diators and a hypoactivated proﬁle with reduced LMs. Al- bumin administration led to distinct regulation of LM proﬁles in each group suggesting that it may activate different protective mechanisms in these groups. Immu- nophenotyping sepsis studies have shown both hyper- activated and hypoactivated proﬁles can lead to a negative outcome23,24; indeed recent evidence has shown that a hyperactivated plasma lipid signature predicts death in sepsis.25 Albumin may therefore have further beneﬁcial immune effects. The potential role of these LMs in inﬂammation and infection and possible utility of LM In summary using samples from our multicenter feasibility study we present novel evidence that targeted albumin infusions seem to exert a beneﬁcial immune effect in patients with AD/ACLF via its ability to bind PGE2, but do not reduce overall total circulating levels. We identify for the ﬁrst time proresolving LMs in advanced liver disease and propose that LM metabolic analysis could immunophenotype these patients. Finally, a second novel potential immune restorative role in which albumin infusions rectify both hyperinﬂammatory and hypoinﬂammatory LM proﬁles was demonstrated. Discussion Its weak binding of PGE2 and lack of effect on absolute levels may explain the absence of renal or gastrointes- tinal side effects in contrast to nonsteroidal anti-inﬂammatory drugs, which alter eicosanoid proﬁles at these sites. Studies have identiﬁed structural and functional alterations in commercial HAS compared with healthy albumin22 and we showed reduced PGE2 binding compared with healthy volunteer plasma. It may be the immune effects of albumin could be enhanced with further research. Again we observed heterogeneity in immune dysfunction and response to albumin using our immune assay may identify patients most likely to beneﬁt from this approach. difference in the ability of commercial albumins tested to bind PGE2. Patients recruited to our randomized controlled trial will be given HAS customarily used at that site and therefore the absence of any differences between manufacturers was important. Overall PGE2 concentrations were unaffected by HAS infusion, which challenges previous data demonstrating PGE2 catalysis by albumin.6 This observation may be related to our assay measuring total PGE2, both free and albumin bound. We hope to develop techniques to differentiate between the 2 to determine whether free PGE2 is catalyzed by albumin. Albumin is present in low concentrations in AD/ACLF and has decreased functional efﬁcacy7 caused by post- transcriptional modiﬁcation.21 Administration of 20% HAS not only improved albumin concentration but also functional capacity to bind immunosuppressive PGE2. Taken together these data suggest that 20% HAS infusions act pharmacologically to improve immune function in AD/ACLF through albumin’s ability to bind elevated circulating levels of immunosuppressive PGE2. A weakness was the lack of a control (untreated arm) and therefore immune function may have improved because of patients recovering; indeed serum bilirubin fell by 25% between samples. However the CRP and WCC were unchanged, and we previously showed immune dysfunction in AD persisted throughout hospital admission.10 Furthermore, plasma cytokine and endo- toxin levels did not alter between pre- and post-HAS 746 China et al China et al Clinical Gastroenterology and Hepatology Vol. 16, No. 5 746 Figure 5. Schematic version of our hypothesis that prophylactic human albumin infusions improve immune response in ACLF via 2 mechanisms. (A) HAS increases circulating albumin concentration and function improving binding of immunosuppressive PGE2, reducing free/bioactive PGE2 and restoring monocyte/macrophage function. (B) Patients with ACLF can be divided into hypoinﬂammatory and hyperinﬂammatory responses deﬁned by LM metabolomics, both with potential adverse outcomes. References 1. Fernandez J, Navasa M, Gomez J, et al. Bacterial infections in cirrhosis: epidemiological changes with invasive procedures and norﬂoxacin prophylaxis. Hepatology 2002;35:140–148. 18. Simon-Talero M, Garcia-Martinez R, Torrens M, et al. Effects of intravenous albumin in patients with cirrhosis and episodic he- patic encephalopathy: a randomized double-blind study. J Hepatol 2013;59:1184–1192. 2. Bajaj JS, O’Leary JG, Reddy KR, et al. Second infections independently increase mortality in hospitalized patients with cirrhosis: the North American consortium for the study of end- stage liver disease (NACSELD) experience. Hepatology 2012; 56:2328–2335. 19. Maini A, China L, Gilroy D, et al. Progression of cirrhotic liver disease towards acute-on-chronic liver failure triggers changes in innate immune cell phenotype and their response to pro-inﬂammatory stimuli. J Hepatol 2017;66. 3. Rajkovic I, Williams R. Abnormalities of neutrophil phagocytosis, intracellular killing and metabolic activity in alcoholic cirrhosis and hepatitis. Hepatology 1986;6:252–262. 20. Raz A. Interaction of prostaglandins with blood plasma proteins. Biochem J 1972;130:631–636. 4. Ploder M, Pelinka L, Roth E, et al. Lipopolysaccharide induced TNF production and monocyte human leucocyte antigen dr expression is correlated with survival in septic trauma patients. Shock 2006;25:129–134. 21. Domenicali M, Baldassarre M, Giannone FA, et al. Post- transcriptional changes of serum albumin: clinical and prog- nostic signiﬁcance in hospitalized patients with cirrhosis. Hepatology 2014;60:1851–1860. 5. O’Brien AJ, Fullerton JN, Massey KA, et al. Immunosuppression in acutely decompensated cirrhosis is mediated by prosta- glandin E2. Nat Med 2014;20:518–523. 5. O’Brien AJ, Fullerton JN, Massey KA, et al. Immunosuppression in acutely decompensated cirrhosis is mediated by prosta- glandin E2. Nat Med 2014;20:518–523. 22. Bar-Or D, Bar-Or R, Rael LT, et al. Heterogeneity and oxidation status of commercial human albumin preparations in clinical use. Crit Care Med 2005;33:1638–1641. 6. Yang J, Petersen CE, Ha CE, et al. Structural insights into human serum albumin-mediated prostaglandin catalysis. Protein Sci 2002;11:538–545. 6. Yang J, Petersen CE, Ha CE, et al. Structural insights into human serum albumin-mediated prostaglandin catalysis. Protein Sci 2002;11:538–545. 23. Hotchkiss RS, Opal S. Immunotherapy for sepsis: a new approach against an ancient foe. N Engl J Med 2010;363:87–89. 7. Jalan R, Schnurr K, Mookerjee RP, et al. Alterations in the functional capacity of albumin in patients with decompensated cirrhosis is associated with increased mortality. Hepatology 2009;50:555–564. 7. Jalan R, Schnurr K, Mookerjee RP, et al. Alterations in the functional capacity of albumin in patients with decompensated cirrhosis is associated with increased mortality. Hepatology 2009;50:555–564. Discussion We believe our study provides the ﬁrst evidence for an May 2018 747 immune-based mechanism of 20% HAS in AD/ACLF. However, a control arm was not included in the study design. Sample collection from ATTIRE stage 2, our randomized controlled trial to assess whether our 20% HAS infusion regimen leads to a reduction in infection, renal dysfunction, and death in patients with AD/ACLF compared with standard care, will provide further opportunity to investigate the role of these LMs in AD/ ACLF. 12. Colas RA, Shinohara M, Dalli J, et al. Identiﬁcation and signature proﬁles for pro-resolving and inﬂammatory lipid mediators in human tissue. Am J Physiol Cell Physiol 2014; 307:C39–C54. 13. Arroyo V, Garcia-Martinez R, Salvatella X. Human serum albumin, systemic inﬂammation, and cirrhosis. J Hepatol 2014; 61:396–407. 14. Bernardi M, Ricci CS, Zaccherini G. Role of human albumin in the management of complications of liver cirrhosis. J Clin Exp Hepatol 2014;4:302–311. 15. Garcia-Martinez R, Caraceni P, Bernardi M, et al. Albumin: pathophysiologic basis of its role in the treatment of cirrhosis and its complications. Hepatology 2013;58:1836–1846. Supplementary Material Note: To access the supplementary material accom- panying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at https://doi.org/10.1016/j.cgh.2017.08.027. 16. Garcia-Martinez R, Andreola F, Mehta G, et al. Immunomodu- latory and antioxidant function of albumin stabilises the endo- thelium and improves survival in a rodent model of chronic liver failure. J Hepatol 2015;62:799–806. 17. Lee KC, Baker LA, Stanzani G, et al. Extracorporeal liver assist device to exchange albumin and remove endotoxin in acute liver failure: results of a pivotal pre-clinical study. J Hepatol 2015; 63:634–642. Reprint requests Address requests for reprints to: Louise China, Institute of Liver and Digestive Health, Division of Medicine, University College London, The Rayne Building, 5 University Street, London, WC1E 6JF, United Kingdom. e-mail: l.china@ucl.ac.uk; fax: þ44 (0)20 3108 8519. References 24. Said EA, Dupuy FP, Trautmann L, et al. Programmed death-1- induced interleukin-10 production by monocytes impairs CD4þ T cell activation during HIV infection. Nat Med 2010; 16:452–459. 8. Chen TA, Tsao YC, Chen A, et al. Effect of intravenous albumin on endotoxin removal, cytokines, and nitric oxide production in patients with cirrhosis and spontaneous bacterial peritonitis. Scand J Gastroenterol 2009;44:619–625. 25. Dalli J, Colas RA, Quintana C, et al. Human sepsis eicosanoid and proresolving lipid mediator temporal proﬁles: correlations with survival and clinical outcomes. Crit Care Med 2017; 45:58–68. 9. Giannone F, Domenicali M, Baldassarre M, et al. Ischaemia- modiﬁed albumin: a marker of bacterial infection in hospitalized patients with cirrhosis. Liver Int 2015;35:2425–2432. Reprint requests Address requests for reprints to: Louise China, Institute of Liver and Digestive Health, Division of Medicine, University College London, The Rayne Building, 5 University Street, London, WC1E 6JF, United Kingdom. e-mail: l.china@ucl.ac.uk; fax: þ44 (0)20 3108 8519. 10. Serhan CN, Chiang N, Dalli J. The resolution code of acute inﬂammation: novel pro-resolving lipid mediators in resolution. Semin Immunol 2015;27:200–215. 11. China L, Muirhead N, Skene SS, et al. ATTIRE: Albumin To prevenT Infection in chronic liveR failurE: study protocol for a single-arm feasibility trial. BMJ Open 2016;6:e010132. 11. China L, Muirhead N, Skene SS, et al. ATTIRE: Albumin To prevenT Infection in chronic liveR failurE: study protocol for a single-arm feasibility trial. BMJ Open 2016;6:e010132. Supplementary Methods incubated with the detection reagent for 4 hours before being read for absorbance at 640 nm on a FLUOStar Omega Plate reader (BMG Labtech, Ortenberg, Germany). Plasma-Induced Monocyte-Derived Macrophage Dysfunction Cytokine Measurement Brieﬂy, cultured human MDMs were stimulated with 1 ng/mL of LPS (Salmonella abortus equi S-form [TLRgrade], Enzo Life Sciences) for 4 hours in the pres- ence of 25% patient plasma pretreatment and post- treatment, and TNF production was measured (enzyme-linked immunosorbent assay, R&D Systems) as shown previously. Improvement in macrophage function was predeﬁned as a signiﬁcant increase in LPS-induced TNF production. Experiments were performed in the presence of plasma from healthy control subjects as a comparator. Pairs of patient plasma samples were iden- tiﬁed that contained a pretreatment sample (serum albumin <30 g/L) and a subsequent post-treatment sample when serum albumin was ﬁrst measured at >30 g/L. Beads with the appropriate cytokines (IL1b, IL6, IL8, IL10, TNF-a) were mixed with standards as provided to produce a standard curve. Samples were diluted in sample diluent. Assay was then performed as per the instructions. Beads were read on a BD FACSVerse ﬂow cytometer (3 lasers: 405 nm, 488 nm, and 640 nm; 10-parameter analysis; BD Biosciences). Data were acquired using BD FACSuite (BD Biosciences). Data were analyzed using FCAP Array software v3.0 (Soft Flow Inc, Hungary). Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Plasma Protein Binding Capacity To determine the amount of PGE2 bound, equilibrium dialysis using a Thermo Scientiﬁc Single-Use Rapid Equilibrium Dialysis Plate was used, which enabled quantiﬁcation of bound versus free PGE2 via postdialysis sample scintillation counting. Patient plasma was incu- bated with H3-PGE2 mixed with unlabeled PGE2 (molar ratio, 2727:1; ﬁnal concentration, 2.73 mM). This was then dialyzed against phosphate-buffered saline in the Rapid Equilibrium Dialysis Plate for 4 hours at 37C. Counts in sample chamber and buffer chamber were measured enabling percentage or total concentration of PGE2 bound to be calculated (% bound ¼ 100 – [cpm buffer chamber/cpm plasma chamber] 100). Lipid Mediator Metabolomic Data Plasma was placed in 4 volumes of ice cold methanol containing deuterium-labelled internal standards: d4-PGE2, d5-LXA4, d5-RvD2, d4-LTB4, d5-LTC4, d5-LTD4, d5-LTE4, and d8-5S-HETE (500 pg each; Cayman Chem- icals). These were then kept at 20C for 45 minutes to allow for protein precipitation and lipid mediators were extracted using C-18 based Solid Phase Extraction as in Colas et al12 (PubMed identiﬁer: 24696140). Methyl formate fractions were brought to dryness using a TurboVap LP (Biotage) and products suspended in water-methanol (50:50 vol/vol) for liquid chromatogra- phy–tandem mass spectrometry based proﬁling. Here a Shimadzu LC-20AD HPLC and a Shimadzu SIL-20AC autoinjector (Shimadzu, Kyoto, Japan), paired with a QTrap 5500 (ABSciex, Warrington, UK) were used and operated as described in Colas et al12 (PubMed identiﬁer: 24696140). To monitor each lipid mediator and deuterium-labelled internal standard, a multiple reaction monitoring method was developed using parent ions and characteristic diagnostic ion fragments as in Colas et al12 (PubMed identiﬁer: 24696140). This was coupled to an information-dependent acquisition and an enhanced product ion scan. Identiﬁcation criteria included match- ing retention time to synthetic standards and at least 6 diagnostic ions in the tandem mass spectrometry spec- trum for each molecule. Calibration curves were obtained for each molecule using authentic and synthetic compound mixtures and deuterium-labelled lipid medi- ator at 0.78, 1.56, 3.12, 6.25, 12.5, 25, 50, 100, and 200 pg. Standards for liquid chromatography–tandem mass spectrometry proﬁling were produced biogenically, purchased from Cayman Chemicals, or provided by Dr Charles N. Serhan (supported by National Institutes of Health funded P01GM095467 to CNS). Linear calibration curves were obtained for each lipid mediator, which gave r2 values of 0.98–0.99. Conﬂicts of interest The authors disclose no conﬂicts 747.e1 China et al Trial Funding and Sponsor (PIs). The Trial Management Group is responsible for developing the design, coordination and strategic man- agement of the trial. The work is supported by the Health Innovation Challenge fund (Wellcome Trust and Department of Health) award number 164699. This publication pre- sents independent research commissioned by the Health Innovation Challenge Fund, a parallel funding partner- ship between the Department of Health and Wellcome Trust. The views expressed in this publication are those of the authors and not necessarily those of the Depart- ment of Health or Wellcome Trust. The trial sponsor is University College London with trial management activ- ities conducted by the University College London Comprehensive Clinical Trials Unit. Research ethical approval was granted by the London-Brent research ethics committee (ref: 15/LO/0104). Acknowledgments The authors have had the support of the following individuals via trial oversight committees: Professor Graeme Alexander, Professor Stephen Brett, Professor Mauro Bernardi, Professor Dominique Valla, Dr Vipul Jaraith, Mr Tim Clayton, Mr Brennan Kahan, Mr John Crookenden, and Ms Susan Tebbs. A list of trial sites and PIs is obtained by contacting attire@ucl.ac.uk. Participating Hospitals and Personnel The Research Steering Group operates on behalf of the funders to ensure that appropriate milestones have been met in the delivery of the trial. It consists of the chief investigator, an independent expert, and repre- sentatives of the Welcome Trust and Department of Health. Basildon University Hospital (Site PI Dr G. Wright), Bristol Royal Inﬁrmary (Site PI Dr J. Portal), Singleton Hospital Swansea (Site PI Dr C. Lye Ch’ng), Newcastle Freeman Hospital (Site PI Dr S. McPherson), Royal London Hospital (Site PI Dr Y. Kallis), Hull Royal Inﬁrmary (Site PI Dr Lynsey Corless), Royal Free Hospital (Site PI Professor R. Jalan), North Tees and Hartlepool NHS Foundation Trust (Site PI Professor Jane Metcalf), Royal Liverpool University Hospital (Site PI Dr P. Richardson), and University Hospitals Birmingham NHS Foundation Trust (Site PI Dr A. Elsharkawy). Role of the Funding Source The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. This publication presents inde- pendent research commissioned by the Health Innova- tion Challenge Fund (Wellcome Trust and Department of Health) award number 164699, a parallel funding part- nership between the Department of Health and Well- come Trust. The views expressed in this publication are those of the authors and not necessarily those of the Department of Health or Wellcome Trust. Trial Steering Committee The Trial Steering Committee is the independent group responsible for oversight of the trial to safeguard the interests of trial patients. The Trial Steering Com- mittee provides advice to the chief investigator, Clinical Trials Unit, funder, and sponsor on all aspects of the trial through its independent chair. Independent Data Monitoring Committee The Independent Data Monitoring Committee is responsible for safeguarding the interests of trial patients, monitoring the accumulating data, and making recommendations to the Trial Steering Committee on whether the trial should continue as planned. It com- prises a clinical chair (independent hepatologist), inde- pendent gastroenterologist, and an independent statistician all with expertise in clinical trials. Lipopolysaccharide Detection HEK293 cells are transfected to stably express TLR4 and a nuclear factor-kB-inducible secreted embryonic alkaline phosphatase reporter gene. QUANTI-Blue detection medium changes color in the presence of secreted embryonic alkaline phosphatase in the spec- trum of 620–655 nm. Because the absorbance is in direct proportion to the amount of endotoxin present, the concentration of endotoxin can be calculated from a standard curve obtained using serial dilutions of the HEK-Blue Endotoxin Standard (a preparation of Escher- ichia coli 055:B5 LPS standardized against Food and Drug Administration–approved control standard endo- toxin). Samples were diluted in endotoxin-free water (Sigma, UK) and then incubated with the HEK293 cells for 24 hours. The supernatant from these cells was then Trial Management Group The Trial Management Group comprises the chief investigator, clinical research fellow, clinical project manager, trial statistician, trial manager, data manager, health economist and P5 trial site principle investigator Clinical Gastroenterology and Hepatology Vol. 16, No. 5 747.e3 China et al 747.e3 China et al Supplementary Figure 1. Targeted 20% HAS infusions reverse immune dysfunction in AD/ACLF by improving AD/ACLF plasma PGE2 binding. (A) TNF production from MonoMac-6 cell line (MM6) stimulated with LPS in the presence of patient plasma pretreatment and post-treatment with 20% HAS (n ¼ 45 patients as above). Mean increase in MM6 TNF-a production was 0.211 ng/mL (10.2%; CI, 0.0517–0.369 ng/mL; P ¼ .014). (B) TNF production from healthy volunteer MDMs stimulated with LPS in the presence of patient plasma pretreatment and post-treatment with 20% HAS (n ¼ 7 patients; patients that did not increment serum albumin to 30 g/L and had a pretreatment and post-treatment sample available for analysis). Sample numbers too small for statistical analysis. (C) Percentage of PGE2/3H-PGE2 bound to patient plasma protein using equilibrium dialysis, comparing patient plasma pretreatment and post-treatment with 20% HAS. Data shown with undiluted samples and when all samples had been diluted to the same albumin concentration (18 g/L). (D) TNF production from healthy volunteer MDMs stimulated with LPS in presence of healthy plasma (n ¼ 4) and presence/absence of AH6890 (50 mM) and MF498 (10 mM) and 1 ng/mL PGE2. (E) TNF production from healthy volunteer MDMs stimulated with LPS in the presence of patient plasma post-treatment with 20% HAS, samples Day 2 and 4 post-treatment. HV, healthy volunteer; Tx, treatment. Supplementary Figure 1. Targeted 20% HAS infusions reverse immune dysfunction in AD/ACLF by improving AD/ACLF plasma PGE2 binding. (A) TNF production from MonoMac-6 cell line (MM6) stimulated with LPS in the presence of patient plasma pretreatment and post-treatment with 20% HAS (n ¼ 45 patients as above). Mean increase in MM6 TNF-a production was 0.211 ng/mL (10.2%; CI, 0.0517–0.369 ng/mL; P ¼ .014). (B) TNF production from healthy volunteer MDMs stimulated with LPS in the presence of patient plasma pretreatment and post-treatment with 20% HAS (n ¼ 7 patients; patients that did not increment serum albumin to 30 g/L and had a pretreatment and post-treatment sample available for analysis). Sample numbers too small for statistical analysis. (C) Percentage of PGE2/3H-PGE2 bound to patient plasma protein using equilibrium dialysis, comparing patient plasma pretreatment and post-treatment with 20% HAS. Trial Management Group Data shown with undiluted samples and when all samples had been diluted to the same albumin concentration (18 g/L). (D) TNF production from healthy volunteer MDMs stimulated with LPS in presence of healthy plasma (n ¼ 4) and presence/absence of AH6890 (50 mM) and MF498 (10 mM) and 1 ng/mL PGE2. (E) TNF production from healthy volunteer MDMs stimulated with LPS in the presence of patient plasma post-treatment with 20% HAS, samples Day 2 and 4 post-treatment. HV, healthy volunteer; Tx, treatment. Albumin Counteracts Immune-Suppression 747.e4 Albumin Counteracts Immune-Suppression 747.e4 May 2018 May 2018 Supplementary Table 1. Baseline Demographics and Clinical Characteristics of the Analysis Population (n ¼ 79) Characteristic Mean (SD) Age, y 53.41 (11.63) Serum albumin, g/L 23.95 (3.51) Days since admission 1.81 (0.88) MELD 20.90 (6.62) Creatinine 91.2 (78.2) n (%) Male 52 (66) Admitted to ICU 2 (3) Prescribed antibiotics 41 (52) Diagnosis of infection 27 (34) Etiology of cirrhosisa n (%) Alcohol 76 (96) Hepatitis B 1 (1) Hepatitis C 11 (14) NAFLD 4 (5) Other etiologies 2 (3) Organ failure according to proposed deﬁnitions n (%) Renal 8 (10) Respiratory 9 (11) Circulatory 13 (16) Brain 3 (4) ACLF gradeb n (%) Grade 0 58 (73) Grade 1 11 (14) Grade 2 6 (8) Grade 3 4 (5) ACLF, acute-on-chronic liver failure; ICU, intensive care unit; MELD, Model for End-Stage Liver Disease; NAFLD, nonalcoholic fatty liver disease; SD, standard deviation. aSome patients have more than 1 liver cirrhosis etiology. bAccording to European foundation for the study of chronic liver failure criteria. Supplementary Table 2. Differences in Blood Test Values Pre-HAS Infusion and After Albumin Restored 30 g/L in Samples Used for Immune Function Analysis Mean (SD) Pre–20% HAS infusion, serum albumin <30 g/L Mean (SD) serum albumin restored 30 g/L P value (paired Student t test) White cell count 9.3 (6.4) 8.6 (7.5) .23 CRP 44 (52) 30 (31) .39 Bilirubin 166 (154) 145 (116) .02 CRP, C-reactive protein; HAS, human albumin solution; SD, standard deviation. fferences in Blood Test Values Pre-HAS Infusion and After Albumin Restored 30 g/L in Samples sed for Immune Function Analysis CRP, C-reactive protein; HAS, human albumin solution; SD, standard deviation. CRP, C-reactive protein; HAS, human albumin solution; SD, standard deviation. Clinical Gastroenterology and Hepatology Vol. 16, No. 5 747.e5 China et al Supplementary Table 3. May 2018 Plasma LM Proﬁles in Prealbumin Administration LM Levels Were Assessed Using LM Metabololipidomics DHA bioactive metabolome Q1 Q3 Group 1 Group 2 2-2529 AOB 3-2445 AOB 8-1822 AOB 9-2417 AOB 10-2200 AOB 5-1026 AOB 4-1486 AOB 1-2320 AOB 6-2419 AOB 7-1702 AOB RvD1 375 141 — — 2.2 0.8 2.7 3.3 1.7 19.1 1.0 1.8 RvD2 375 141 2.9 1.7 2.1 0.9 3.8 — — — 4.9 11.2 RvD3 375 147 — — 2.8 5.1 0.7 3.1 3.4 1.1 0.0 0.6 RvD4 375 101 — — 0.2 0.7 0.8 1.0 0.5 1.5 1.3 0.2 RvD5 359 199 7.2 6.8 — 1.3 — 12.4 3.3 13.8 4.6 8.6 RvD6 359 101 — — 0.8 — 0.6 23.1 115.0 — 5.0 — 17R-RvD1 375 141 139.6 0.2 — — — — — — — — 17R-RvD3 375 147 — — 15.1 0.3 — — — — — 2.6 PD1 359 153 0.3 3.8 — — — — 10.7 7.4 3.3 1.0 17R-PD1 359 153 — 1.1 7.8 3.9 3.6 — — 1.6 3.5 — 10S,17S-diHDHA 359 153 — 1.1 — — 11.1 — — — — — 22-OH-PD1 375 153 1.7 — — — — 0.9 — — — — MaR1 359 221 — — — — — 1.1 — — — — 7S,14S-diHDHA 359 221 — — — 11.5 — — — 6.6 — — 4S,14S-diHDHA 359 101 10.3 — 4.9 4.9 — 1.8 — — — 4.7 n-3 DPA bioactive metabolome RvT1 377 239 3.0 1.4 1.2 2.8 0.3 3.5 3.5 4.4 2.1 0.3 RvT2 377 197 0.6 — — — — 1.4 — — 0.3 0.3 RvT3 377 197 2.3 0.7 3.3 — 1.5 — 2.0 20.6 0.3 1.0 RvT4 377 211 9.5 3.0 33.3 — 8.4 — 7.9 77.7 1.5 5.1 RvD1n-3 DPA 377 143 47.4 2.3 12.7 5.0 — 2.9 3.7 35.5 — 6.8 RvD2n-3 DPA 377 215 2.2 2.2 — 0.8 — 1.7 2.2 — — 2.3 RvD5n-3 DPA 361 199 — — 26.0 — — 9.5 — 266.9 3.4 — PD1n-3 DPA 361 183 10.9 34.2 413.2 13.8 — 259.8 265.9 — 192.7 327.9 10S,17S-diHDPA 361 183 — — — — — — — — — — MaR1n-3 DPA 361 223 154.1 7.4 — 12.1 — — — 30.4 13.6 — 7S, 14S-diHDPA 361 223 — 8.3 — 6.2 — 15.0 41.4 18.4 5.5 28.7 EPA bioactive metabolome RvE1 349 195 1.6 — 1.1 — 20.7 1.5 0.8 7.6 — 0.6 RvE2 333 199 6.6 — 11.1 — 118.1 9.1 3.2 29.1 — 3.0 RvE3 333 201 2.5 26.2 37.9 2.4 5.0 3.1 18.8 7.1 23.7 13.8 AA bioactive metabolome LXA4 351 217 0.8 — — — 1.0 — 1.2 — 1.7 8.0 LXB4 351 221 16.0 — 147.4 — — — 64.0 169.6 44.0 167.8 5S,15S-diHETE 335 235 10.8 — 148.3 35.0 29.4 486.1 118.4 110.1 29.6 83.1 AT-LXA4 351 217 — — 17.7 — — 12.0 6.1 3.5 — 3.5 AT-LXB4 351 221 — — — — 6.6 — — — — 21.9 LTB4 335 195 40.6 33.4 9.1 15.9 93.3 27.5 40.2 7.0 281.3 596.6 5S,12S-diHETE 335 195 4.5 0.4 2.4 11.1 3.3 5.3 2.4 1.3 9.7 13.0 20-OH-LTB4 351 195 3.6 43.0 3.8 15.3 50.4 11.6 87.1 4.0 486.7 1106.0 LTC4 626 189 12.2 35.8 38.7 139.5 28.9 88.6 — 11.7 — 23.0 LTD4 497 189 — 12.4 17.7 17.2 7.8 — — — — — LTE4 440 189 0.8 59.3 144.1 62.9 61.1 240.2 34.9 10.7 97.9 0.1 PGD2 351 189 8.1 13.1 4.1 30.1 14.9 55.5 36.9 6.0 34.0 34.7 PGE2 351 189 14.0 23.4 32.4 11.4 24.8 111.9 105.2 7.5 112.5 81.5 PGF2a 353 193 25.5 26.9 60.4 55.1 63.2 167.5 180.4 62.8 48.2 71.2 TxB2 369 169 1142.4 623.8 2566.2 780.2 1781.9 1939.9 6946.8 408.4 15370.6 9075.9 NOTE. (daughter ion). AA, arachidonic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid. NOTE. Results are expressed as pg/mL. Em dash ¼ below limit; limit z 0.1 pg. With Q1, M-H (parent ion); and Q3, diagnostic ion in the tandem mass spectrometry (daughter ion) May 2018 Results are expressed as pg/mL. Em dash ¼ below limit; limit z 0.1 pg. With Q1, M-H (parent ion); and Q3, diagnostic ion in the tandem mass spectrometry (daughter ion) e 3. Plasma LM Proﬁles in Prealbumin Administration LM Levels Were Assessed Using LM Metabololipidomics Supplementary Table 3. Plasma LM Proﬁles in Prealbumin Administration LM Levels Were Assessed Using LM Metabololipidomics Albumin Counteracts Immune-Suppression 747.e6 Albumin Counteracts Immune-Suppression 747.e6 May 2018 Supplementary Table 4. Plasma LM Proﬁles Post Albumin Administration Supplementary Table 4. May 2018 Plasma LM Proﬁles Post Albumin Administration DHA bioactive metabolome Q1 Q3 Group 1 Group 2 2-2529 AOB 3-2445 AOB 8-1822 AOB 9-2417 AOB 10-2200 AOB 5-1026 AOB 4-1486 AOB 1-2320 AOB 6-2419 AOB 7-1702 AOB RvD1 375 141 0.3 — 0.4 2.1 1.1 — 2.5 9.4 0.4 3.4 RvD2 375 141 — 0.6 2.5 — 3.6 — 4.6 — 2.7 7.4 RvD3 375 147 — 0.7 — 6.7 — 3.4 3.1 0.8 0.4 — RvD4 375 101 0.1 1.4 2.2 1.9 3.6 0.4 0.6 2.4 0.9 0.9 RvD5 359 199 3.6 3.6 5.6 14.1 18.2 7.4 7.8 1.6 7.5 17.5 RvD6 359 101 4.5 2.8 3.3 0.6 6.2 24.7 31.7 2.7 10.2 1.8 17R-RvD1 375 141 — 1.7 — 1.0 — — — — 1.2 1.2 17R-RvD3 375 147 0.2 — — 0.5 — 0.3 — 1.0 — 2.8 PD1 359 153 1.3 2.1 0.3 1.1 2.1 — — 2.8 4.9 8.8 17R-PD1 359 153 5.1 3.9 — 1.5 3.4 — — — — 1.5 10S,17S-diHDHA 359 153 — — — — 0.9 — — — 0.6 — 22-OH-PD1 375 153 — — — — 0.2 — 1.2 2.7 — 1.2 MaR1 359 221 — — — — 1.7 2.9 — — — — 7S,14S-diHDHA 359 221 3.1 — 1.9 20.0 3.2 — — 8.6 3.2 — 4S,14S-diHDHA 359 101 — — 1.6 6.6 — — — 19.5 — 3.8 n-3 DPA bioactive metabolome RvT1 377 239 0.5 0.5 0.6 0.5 0.5 7.4 1.7 0.7 0.8 8.8 RvT2 377 197 0.2 1.0 1.1 — 1.1 1.4 0.8 2.5 1.5 — RvT3 377 197 0.6 0.3 0.6 — 0.8 — 0.6 3.8 0.3 2.7 RvT4 377 211 2.8 1.5 2.0 — 1.9 — 4.2 15.7 1.2 10.7 RvD1n-3 DPA 377 143 16.7 20.3 7.5 15.3 9.1 4.8 4.5 18.9 9.2 3.2 RvD2n-3 DPA 377 215 — 0.2 — 1.1 2.7 1.0 1.1 3.2 — 1.8 RvD5n-3 DPA 361 199 3.7 9.3 — 3.3 0.8 8.5 — 3.2 2.5 4.6 PD1n-3 DPA 361 183 24.4 47.3 299.5 108.5 41.5 574.0 422.6 — — 367.1 10S,17S-diHDPA 361 183 — — — — — — — — — — MaR1n-3 DPA 361 223 — 4.7 10.3 78.1 2.7 — 13.0 — — 109.7 7S, 14S-diHDPA 361 223 — 6.5 9.7 81.7 — — 84.7 — 1.8 39.2 EPA bioactive metabolome RvE1 349 195 0.9 — 0.5 0.5 0.4 25.4 5.4 1.0 8.7 0.6 RvE2 333 199 3.9 — 1.9 1.2 1.0 207.1 39.1 4.3 32.3 2.5 RvE3 333 201 0.8 4.0 13.3 — — — 3.5 — 2.9 181.5 AA bioactive metabolome LXA4 351 217 0.7 — — 0.4 — 1.5 0.4 — 1.5 0.7 LXB4 351 221 13.5 6.8 6.8 11.0 189.1 0.0 11.0 189.1 0.0 13.5 5S,15S-diHETE 335 235 58.8 182.7 47.9 71.2 21.4 282.4 126.7 58.6 11.0 45.2 AT-LXA4 351 217 — — — 3.5 — 17.6 3.5 — 17.6 — AT-LXB4 351 221 — — — — — 129.3 — — 129.3 — LTB4 335 195 8.4 12.4 4.6 11.2 52.4 20.5 15.2 6.7 84.2 85.5 5S,12S-diHETE 335 195 13.2 4.3 2.0 11.2 52.4 5.7 10.3 8.1 56.8 4.2 20-OH-LTB4 351 195 15.6 4.2 — — — 19.5 5.2 26.5 129.3 — LTC4 626 189 310.6 143.6 104.0 108.4 29.1 67.5 187.3 71.6 86.1 158.3 LTD4 497 189 — — 181.7 — — — — — 11.8 — LTE4 440 189 16.3 15.4 111.6 115.8 0.4 560.8 29.6 12.3 54.9 — PGD2 351 189 21.2 34.8 26.8 13.9 47.3 36.9 13.5 14.5 27.4 32.8 PGE2 351 189 58.9 55.4 71.8 8.2 69.4 69.5 25.1 3.8 72.0 65.1 PGF2a 353 193 28.0 92.6 114.2 41.4 107.8 65.6 58.1 37.6 31.6 84.8 TxB2 369 169 3167.6 938.9 9291.8 943.0 4320.1 1790.9 3309.4 782.5 8298.5 5740.6 NOTE. and Q3, diagnostic ion in the tandem mass spectrometry (daughter ion). AA, arachidonic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; LM, lipid mediator; PG, prostaglandin. AA, arachidonic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; LM, lipid me os c o e a de ass spec o e y (daug e o ) c acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; EPA, eicosapentaenoic acid; LM, lipid mediator; PG, prost he tandem mass spectrometry (daughter ion). ACLF, acute-on-chronic liver failure; PGE2, prostaglandin E2. May 2018 LM levels were assessed using LM-metabololipidomics. Results are expressed as pg/mL. Em dash ¼ below limit; limit z 0.1 pg. With Q1, M-H (parent ion); and Q3, diagnostic ion in the tandem mass spectrometry (daughter ion). AA arachidonic acid; DHA docosahexaenoic acid; DPA docosapentaenoic acid; EPA eicosapentaenoic acid; LM lipid mediator; PG prostaglandin Clinical Gastroenterology and Hepatology Vol. 16, No. 5 747.e7 China et al 747.e7 China et al Supplementary Table 5. Theoretical Calculations Comparing the Effects of Low and High Binding Afﬁnity Albumin on Free Circulating Levels of PGE2 Based on Concentrations of PGE2 and Serum Albumin Found in Healthy Patients and Patients With ACLF Albumin at 40 g/L (w600 mM) (Healthy) PGE2, pg/mL PGE2, pM High binding afﬁnity, Kd mM Free PGE2, pg/mL Low binding afﬁnity, Kd mM Free PGE2, pg/mL 2 5.7 0.00007 1.4 10 28.4 0.02 0.0003 200 2.5 20 56.7 0.0007 5 Albumin at 20 g/L (w300 mM) (ACLF) PGE2, pg/mL PGE2, pM Kd mM Free PGE2, pg/mL Kd mM Free PGE2, pg/mL 2 5.7 0.0001 0.80 10 28.4 0.02 0.0007 200 4.0 20 56.7 0.0013 7.99 ACLF, acute-on-chronic liver failure; PGE2, prostaglandin E2. 747.e7 China et al Clinical Gastroenterology and Hepatology Vol. 16, No. 5 Supplementary Table 5. Theoretical Calculations Comparing the Effects of Low and High Binding Afﬁnity Albumin on Free Circulating Levels of PGE2 Based on Concentrations of PGE2 and Serum Albumin Found in Healthy Patients and Patients With ACLF Supplementary Table 6. Plasma Cytokine and LPS Concentrations at Baseline and Following 20%-HAS Treatment Once Serum Albumin >30 g/L Mean plasma level pretreatment, pg/mL (n ¼ 45) Mean plasma level post-treatment, pg/mL ( n ¼ 45) Mean change post-treatment, pg/mL CI, pg/mL TNF-a 1.32 1.30 -0.01 -0.42 to 0.40 IL6 100.88 85.10 -17.46 -49.05 to 14.13 IL8 708.76 458.61 -252.80 -555.70 to 50.21 IL10 2.78 3.24 þ0.45 -0.64 to 1.53 IL1b 1.28 1.14 -0.16 -0.99 to 0.68 LPS 17.71 15.69 -2.022 -4.792 to 0.7477 CI, conﬁdence interval; HAS, human albumin solution; IL, interleukin; LPS, lipopolysaccharide; TNF, tumor necrosis factor. Supplementary Table 6. Plasma Cytokine and LPS Concentrations at Baseline and Following 20%-HAS Treatment Once Serum Albumin >30 g/L . Plasma Cytokine and LPS Concentrations at Baseline and Following 20%-HAS Treatment Once Serum Albumin >30 g/L
https://openalex.org/W4211029864	https://boris.unibe.ch/165825/1/Clinical_Exp_Dental_Res_-_2022_-_Tentolouri_-_Masseter_muscle_thickness_and_vertical_cephalometric_characteristics_in.pdf	English	null	Masseter muscle thickness and vertical cephalometric characteristics in children with Class II malocclusion	Clinical and experimental dental research	2,022	cc-by	6,537	O R I G I N A L A R T I C L E O R I G I N A L A R T I C L E Eirini Tentolouri1 \| Gregory S. Antonarakis1 \| Ioanna Georgiakaki2 \| Stavros Kiliaridis1,3 Eirini Tentolouri1 \| Gregory S. Antonarakis1 \| Ioanna Georgiakaki2 \| Stavros Kiliaridis1,3 Stavros Kiliaridis1,3 1Division of Orthodontics, University Clinics of Dental Medicine, University of Geneva, Geneva, Switzerland 2Private Practice, Thessaloniki, Greece 3Department of Orthodontics and Dentofacial Orthopedics, University of Bern, Bern, Switzerland 1Division of Orthodontics, University Clinics of Dental Medicine, University of Geneva, Geneva, Switzerland 2Private Practice, Thessaloniki, Greece 3Department of Orthodontics and Dentofacial Orthopedics, University of Bern, Bern, Switzerland Correspondence Methods: The current study design was retrospective and cross‐sectional, looking at a sample of 211 growing children with Class II malocclusion between the ages of 6 and 15 derived from two centers. Ultrasonographic masseter muscle thickness measurements and vertical cephalometric variables, including the gonial angle, were evaluated before any orthodontic treatment had been carried out. Multiple linear regression analysis was used to examine the association between masseter muscle thickness and vertical cephalometric measurements, including age and patient origin as independent variables in the analysis. Results: In the present sample, masseter muscle thickness was found to be in- dependent of sex, but correlated with age, with older children presenting thicker masseter muscles. In the total patient sample, using multiple regression analyses, children with thicker masseter muscles had significantly smaller intermaxillary and gonial angles. No other cephalometric vertical characteristics showed associations with masseter muscle thickness. Conclusion: In growing children with Class II malocclusion, those with thicker masseter muscles are more likely to display smaller intermaxillary and gonial angles respectively. Abstract Background: Masseter muscle thickness and its relationship with vertical craniofacial morphology have been extensively studied in adults, but data on children are lacking. Objective: To examine the association between masseter muscle thickness and vertical cephalometric parameters in a group of Class II malocclusion growing children. Correspondence Eirini Tentolouri, Division of Orthodontics, University Clinics of Dental Medicine, University of Geneva, 1 rue Michel‐Servet, 1206 Geneva, Switzerland. Email: eirini.tentolouri@gmail.com Funding information None Correspondence Eirini Tentolouri, Division of Orthodontics, University Clinics of Dental Medicine, University of Geneva, 1 rue Michel‐Servet, 1206 Geneva, Switzerland. Email: eirini.tentolouri@gmail.com Clin Exp Dent Res. 2022;1–8. wileyonlinelibrary.com/journal/cre2 \| 1 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Clinical and Experimental Dental Research published by John Wiley & Sons Ltd. Received: 6 October 2021 \| Revised: 23 December 2021 \| Accepted: 29 December 2021 Received: 6 October 2021 \| Revised: 23 December 2021 \| Accepted: 29 December 2021 DOI: 10.1002/cre2.528 1 \| INTRODUCTION Various cephalometric characteristics have been proposed to correlate to a favorable response to functional appliance treatment such as a low mandibular plane angle, low basal‐plane angle, high Jarabak ratio, short mandibular corpus and ramus height, short cranial base, and small anterior and posterior lower face heights (Kumar et al., 2013; Patel et al., 2002). Although of fundamental interest in helping us understand var- iation in the response to functional appliance treatment, the direct influence of the masticatory muscles on treatment outcome does not have everyday clinical applicability as orthodontic offices are very seldomly equipped with an ultrasound machine or a bite force mea- suring gauge. Finding a clinically useful variable, such as the gonial angle, that can act as a proxy for the masticatory muscle functional capacity would therefore be very useful in a clinical setting, as this may aid in treatment planning or more accurate prediction of treat- ment outcome. Masticatory muscles, which are directly involved in functional appliance treatment, have also been thought to play an important role in defining the degree of success of this treatment. In fact, recent studies show that the initial condition of the masticatory muscles evaluated either through ultrasonographic masseter muscle thickness measurements or through maximal molar bite force mea- surements may partly determine functional appliance treatment outcomes (Antonarakis & Kiliaridis, 2015; Antonarakis et al., 2012; Kiliaridis et al., 2010). Moreover, functional appliance treatment seems to also have an effect on the masticatory muscles, as witnessed by temporary atrophy of masseter muscles following functional treatment (Kiliaridis et al., 2010). The aim of our study was thus to investigate the association between masseter muscle thickness and vertical cephalometric vari- ables, including the gonial angle, in a homogeneous group of growing children with Class II malocclusion. In an important study that tried to ascertain what pretreatment cephalometric variables may be predictive of individual mandibular outcomes following functional appliance treatment in children with Class II malocclusion, the discriminant analysis identified a single predictive variable, namely the gonial angle (Franchi & Baccetti, 2006). The authors go on to suggest that children with a pretreat- ment gonial angle of smaller than 125.5° are expected to respond favorably (a greater skeletal mandibular effect) to functional appli- ance treatment while those with a gonial angle greater than 125.5° will be expected to respond unfavorably. 1 \| INTRODUCTION Based on these data, it has thus been suggested that the gonial angle may be able to be used as a proxy for the functional capacity of the masticatory muscles, however, the samples in these previous studies were relatively small. Different therapeutic approaches for the correction of Class II mal- occlusion in growing individuals exist, with functional appliances being a widely used means of treatment with potentially successful yet varying results. Functional appliances act on the sagittal and vertical position of the mandible, bringing about both skeletal and dentoalveolar effects. Their skeletal effect is through the displace- ment of the mandible downwards and forwards which causes soft tissue and muscle stretching and myotatic reflexes (Bishara & Ziaja, 1989; Carels & van der Linden, 1987; Graber & Neumann, 1984). The dentoalveolar effect arises from the shift in the dental arches towards a Class I molar relationship and incisor compensation (Macey‐Dare & Nixon, 1999; Vargervik & Harvold, 1985). Masticatory muscle function is related to increased loading of the jaws and bone apposition which can have an effect on the gonial angle (Kiliaridis, 2006). The gonial angle is a site of muscle attachment (masseter and medial pterygoid muscles) and thus the size and activity of these muscles may affect the morphology of the gonial angle, and perhaps the dentofacial morphology more generally. Wolff's law explains that bone morphology is affected by muscle thickness (Wolff, 1870). This means that there is an association be- tween the function of muscles and the internal structure and shape of the bone (Dibbets, 1992). Putting this hypothesis into practice, re- search suggests that there is a negative association between mass- eter muscle thickness and vertical facial morphology (Hannam & Wood, 1989; Kiliaridis & Kalebo, 1991; Raadsheer et al., 1996; Weijs & Hillen, 1986). All of these cited studies however investigate adult samples, besides one (Raadsheer et al., 1996). That one study how- ever looked at a heterogeneous group of individuals, including both children and adults, from 7 to 22 years of age and evaluated an- thropometric instead of cephalometric data. The large variation in interindividual response to functional appliances may be due to several factors including compliance, appliance choice, growth potential, facial type, treatment timing, and skeletal maturity (Barton & Cook, 1997; Bishara & Ziaja, 1989; Carels & van der Linden, 1987; Celli et al., 2010; Tulloch et al., 1990; Woodside, 1998). 1 \| INTRODUCTION More recent studies (Antonarakis & Kiliaridis, 2015; Kiliaridis et al., 2010) also find asso- ciations between the gonial angle and treatment outcome following functional appliance treatment, which agree with the results pre- viously mentioned. A study on children treated with twin blocks found that the only predictive variable for successful treatment was the condylion–gonion–menton angle (Cretella Lombardo et al., 2020). 2 \| 2 TENTOLOURI ET AL. K E Y W O R D S children, Class II malocclusion, masseter muscle, vertical characteristics This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. nder the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any mediu perly cited. © 2022 The Authors. Clinical and Experimental Dental Research published by John Wiley & Sons Ltd. wileyonlinelibrary.com/journal/cre2 \| 1 1 \| INTRODUCTION 2 \| 2 \| 2.2 \| Methods The two principal outcomes evaluated were pretreatment masseter muscle thickness evaluated by ultrasonographic measurements, and pretreatment cephalometric analysis focusing on the vertical dimension. FIGURE 1 Cephalometric measurements used: ANB angle—the angle formed between points A, nasion (N), and B; intermaxillary angle—the angle formed between the maxillary and mandibular planes, with the maxillary plane being defined as a plane between the anterior nasal spine (ANS) and posterior nasal spines (PNS), and the mandibular plane being defined as a plane between menton (Me) and gonion (Go); mandibular plane angle—the angle formed between the Sella‐Nasion (SN) plane, and the MeGo plane; gonial angle—the angle formed between articulare (Ar), Go, and Me; mandibular ramus height— the distance between Ar and Go; facial height index—ratio of posterior facial height (PFH, defined by S‐Go) to anterior face height (AFH, defined by N‐Me) 2.2.1 \| Masseter muscle thickness measurements The thickness of the masseter muscles was measured using ultra- sonography, as per the method described by Kiliaridis and Kälebo (Kiliaridis & Kalebo, 1991), and modified by Raadsheer et al. (1994). Children involved in our study were examined by one of two op- erators (G. S. A. or I. G.), in either of the two centers, using a real‐time scanner (Pie Medical Scanner 480) with a 7.5 MHz linear array transducer. The two operators had been calibrated to the senior author having originally developed the method (S. K.). Imaging and measurements were performed bilaterally with the subjects seated in an upright position, with their heads in a natural head position without a headrest. Measurements were taken with the muscles in contraction. Children were asked to clench maximally in the inter- cuspal position so that the masseter muscles were contracted. Scanning of the masseter muscle was then performed on a level halfway between the zygomatic arch and the gonial angle. To avoid erroneous measurements, the scan plane was orientated perpendi- cular to the anterior border of the muscle and to the surface of the underlying ramus, and light pressure was applied to the muscles. The registrations were performed twice for each muscle, and the final muscle thickness was calculated as a mean of the duplicate mea- surements of the contracted muscles. The measurements were taken directly from the image at the time of scanning with a read‐out dis- tance to the nearest 0.1 mm. FIGURE 1 Cephalometric measurements used: ANB angle—the angle formed between points A, nasion (N), and B; intermaxillary angle—the angle formed between the maxillary and mandibular planes with the maxillary plane being defined as a plane between the FIGURE 1 Cephalometric measurements used: ANB angle—the angle formed between points A, nasion (N), and B; intermaxillary angle—the angle formed between the maxillary and mandibular planes, with the maxillary plane being defined as a plane between the anterior nasal spine (ANS) and posterior nasal spines (PNS), and the mandibular plane being defined as a plane between menton (Me) and gonion (Go); mandibular plane angle—the angle formed between the Sella‐Nasion (SN) plane, and the MeGo plane; gonial angle—the angle formed between articulare (Ar), Go, and Me; mandibular ramus height— the distance between Ar and Go; facial height index—ratio of posterior facial height (PFH, defined by S‐Go) to anterior face height (AFH, defined by N‐Me) 2.1 \| Subjects The present retrospective cross‐sectional study was performed on a sample of 211 growing children with Class II malocclusion (115 females and 96 males) between the ages of 6 and 15, from two different centers. The patients were seen either at our University Orthodontic Clinic (n = 82) or in a private practice limited to ortho- dontics (n = 129). The study was approved by the Cantonal Research Ethics Commission (no. 2016‐00292), and informed consent was not applicable since the data were collected retrospectively. The inclusion criteria were the following: children with Class II malocclusion presenting for orthodontic treatment, where pretreat- ment records had been taken; the presence of a pretreatment TENTOLOURI ET AL. 3 cephalometric radiograph; ultrasonographic masseter muscle thick- ness measurements having been undertaken. cephalometric analysis was carried out; the measurements that were used are shown in Figure 1. cephalometric analysis was carried out; the measurements that were used are shown in Figure 1. The exclusion criteria were the following: patients without the presence of the first permanent molars; patients with any craniofacial anomaly or syndrome; juvenile idiopathic arthritis or signs of condylar lesions or temporomandibular dysfunction or disorders; lateral ce- phalometric radiographs of insufficient diagnostic quality; insufficient data with regard to the masseter muscle thickness measurements (not done bilaterally, or not done in contraction). 2.3 \| Statistics The statistical analysis of our data was performed using the Statistical Package for Social Sciences version 25.0 (SPSS Inc). First, sex, age, and the ANB angle were tested for possible associations with masseter muscle thickness. Linear regression analysis was then used to study the association between masseter muscle thickness and the vertical cephalometric measurements, for the total patient sample. Finally, multiple regression analyses were performed including age, sex, patient origin (from which center), and/or the ANB angle as 2.4 \| Error of the method Paired t‐tests were used to calculate the systematic error and Dahlberg's formula to measure random error (Houston, 1983), for both ultrasonographic masseter muscle thickness measurements and the cephalometric analysis. The error of the method for the masseter muscle thickness measurements was carried out by the two in- vestigators having carried out these measurements. Each investigator performed repeated measurements on twenty patients on two se- parate occasions, 2–4 weeks apart. No systematic error was found, and random error was found to be 0.3 mm for one of the in- vestigators and 0.4 mm for the other investigator. The error of the method for the cephalometric analysis was calculated by repeating cephalometric tracings on 20 radiographs, with a 2‐week interval separating the two measurements. No significant systematic error was found, and random error was found not to exceed 1° for angular measurements and 0.9 mm for linear measurements. Masseter muscle thickness in the present sample was independent of sex but correlated with age (R = .419, p < .001), with older children having thicker masseter muscles. The scatter plot in Figure 2 shows the correlation between masseter muscle thickness and age. No as- sociation was found between the ANB angle and masseter muscle thickness. When performing a regression model with masseter muscle thickness as the dependent variable and age, sex, and patient origin as independent variables, the model is statistically significant (R = .437, p < .001) with only age being significant within the model (β coefficient = .403, p < .001). 3.3 \| Masseter muscle thickness and vertical cephalometric variables Using simple bivariate regression analysis, several of the cephalo- metric variables measured showed statistically significant correlations with masseter muscle thickness. There were significant, although weak, correlations of masseter muscle thickness with the inter- maxillary angle (R = .166, p = .026), the mandibular plane angle (R = .198, p = .016), the gonial angle (R = .185, p = .008), and the facial height ratio (R = .167, p = .043). Thicker masseter muscles were seen in children with a smaller intermaxillary and mandibular plane angle, a smaller gonial angle, and a smaller facial height ratio. additional independent variables and masseter muscle thickness as the dependent variable. All correlations were considered significant at p < .05. Baseline data for the sample are presented in Table 1. No sta- tistically significant differences were found between females and males for any of the variables examined. 2.2.2 \| Cephalometric analysis All lateral cephalometric radiographs were carried out in a standar- dized way, with the head fixed in a cephalostat and with the teeth in occlusion. One operator (E. T.) performed the cephalometric analysis of all radiographs using OnyxCeph3TM (Image Instruments). The magnification of all radiographs was adjusted to zero. A limited 4 TENTOLOURI ET AL. 3.1 \| Baseline data BLE 2 Multiple linear regression analysis results with the cephalometric variable as the dependent variable Note: With regard to patient origin, Center 1 represents the private practice while Center 2 represents the University Orthodontic Clinic. Satiroğlu et al., 2005; Weijs & Hillen, 1984) finding a negative asso- ciation between masseter muscle thickness and facial morphology, and vertical facial height. Cephalometric variables such as the man- dibular plane angle and the gonial angle have been negatively asso- ciated with masseter muscle thickness, and mandibular ramus height and posterior facial height have been positively associated with masseter muscle thickness (Kubota et al., 1998; Rohila et al., 2012). 3.1 \| Baseline data Cephalometric variables such as the man- TABLE 2 Multiple linear regression analysis results with the cephalometric variable as the dependent variable R Significance of model β Coefficient Significance of variable Model 1—gonial angle .362 p < .001 Constant 133.8 p < .001 Masseter muscle thickness (mm) −.8 p = .003 Age (years) −.2 p = .491 Sex (0 = female and 1 = male) −.2 p = .785 Patient origin (Center 1 or 2) 2.8 p = .001 Model 2—intermaxillary angle .298 p = .001 Constant 39.9 p < .001 Masseter muscle thickness (mm) −0.8 p = .001 Age (years) −0.2 p = .401 Sex (0 = female and 1 = male) −0.5 p = .481 Patient origin (Center 1 or 2) 0.1 p = .892 Note: With regard to patient origin, Center 1 represents the private practice while Center 2 represents the University Orthodontic Clinic. FIGURE 2 Scatter plot showing the correlation between masseter muscle thickness and age (R = .419, p < .001) TENTOLOURI ET AL. \| 5 FIGURE 2 Scatter plot showing the correlation between masseter muscle thickness and age (R = .419, p < .001) TENTOLOURI ET AL. \| 5 FIGURE 2 Scatter plot showing the correlation between masseter muscle thickness and age (R = .419, p < .001) TENTOLOURI ET AL. \| 5 TENTOLOURI ET AL. 5 FIGURE 2 Scatter plot showing the correlation between masseter muscle thickness and age (R = .419, p < .001) FIGURE 2 Scatter plot showing the correlation between masseter muscle thickness and age (R = .419, p < .001) TABLE 2 Multiple linear regression analysis results with the cephalometric variable as the dependent variable R Significance of model β Coefficient Significance of variable Model 1—gonial angle .362 p < .001 Constant 133.8 p < .001 Masseter muscle thickness (mm) −.8 p = .003 Age (years) −.2 p = .491 Sex (0 = female and 1 = male) −.2 p = .785 Patient origin (Center 1 or 2) 2.8 p = .001 Model 2—intermaxillary angle .298 p = .001 Constant 39.9 p < .001 Masseter muscle thickness (mm) −0.8 p = .001 Age (years) −0.2 p = .401 Sex (0 = female and 1 = male) −0.5 p = .481 Patient origin (Center 1 or 2) 0.1 p = .892 Note: With regard to patient origin, Center 1 represents the private practice while Center 2 represents the University Orthodontic Clinic. 3.1 \| Baseline data The subjects involved in this study consisted of 211 children (115 females and 96 males) with Class II division 1 malocclusion, between the age of 6 and 15 (mean age: 10.3, standard deviation [SD]: 1.9 years). A total of 129 of the patients originated from one center, and 82 from the other center. The sample was collected based on a dental Class II malocclusion, and the mean skeletal anteroposterior relationships (as measured with the ANB angle) was 5.3 (SD: 1.8) degrees. The mean masseter muscle thickness, measured under contraction, was found to be 11.7 (SD: 1.7) mm in the present sample. When performing multiple regression analyses, age, sex, and center were added as additional independent variables to the re- gression models. Analyses show significant correlations with the gonial angle and the intermaxillary angle, both of which showed a negative correlation with masseter muscle thickness. The results of these multiple regression analyses are shown in Table 2. TABLE 1 Baseline data for the sample of Class II malocclusion children Total sample (n = 211) Females (n = 115) Males (n = 96) Mean SD Mean SD Mean SD Age (years) 10.3 1.9 10.3 1.9 10.4 1.9 Masseter muscle thickness (mm) 11.7 1.7 11.5 1.7 11.9 1.6 ANB (degrees) 5.3 1.8 5.2 1.9 5.3 1.6 Intermaxillary angle (degrees) 28.1 5.4 28.4 5.4 27.8 5.3 Mandibular plane angle (degrees) 34.9 5.4 35.5 5.5 34.2 5.2 Gonial angle (degrees) 126.3 5.9 126.6 6.0 125.9 5.9 Mandibular ramus height (mm) 42.9 8.7 42.2 8.7 43.8 8.8 Facial height ratio (%) 63.5 4.6 63.2 4.6 64.0 4.8 Total sample (n = 211) Females (n = 115) Males (n = 96) Mean SD Mean SD Mean SD Age (years) 10.3 1.9 10.3 1.9 10.4 1.9 Masseter muscle thickness (mm) 11.7 1.7 11.5 1.7 11.9 1.6 ANB (degrees) 5.3 1.8 5.2 1.9 5.3 1.6 Intermaxillary angle (degrees) 28.1 5.4 28.4 5.4 27.8 5.3 Mandibular plane angle (degrees) 34.9 5.4 35.5 5.5 34.2 5.2 Gonial angle (degrees) 126.3 5.9 126.6 6.0 125.9 5.9 Mandibular ramus height (mm) 42.9 8.7 42.2 8.7 43.8 8.8 Facial height ratio (%) 63.5 4.6 63.2 4.6 64.0 4.8 4 \| DISCUSSION The results of the present study show associations between the Satiroğlu et al., 2005; Weijs & Hillen, 1984) finding a negative asso- ciation between masseter muscle thickness and facial morphology, and vertical facial height. 4 \| DISCUSSION Satiroğlu et al., 2005; Weijs & Hillen, 1984) finding a negative asso- ciation between masseter muscle thickness and facial morphology, and vertical facial height. Cephalometric variables such as the man- dibular plane angle and the gonial angle have been negatively asso- ciated with masseter muscle thickness, and mandibular ramus height and posterior facial height have been positively associated with masseter muscle thickness (Kubota et al., 1998; Rohila et al., 2012). The results of the present study show associations between the thickness of the masseter muscle and the intermaxillary and gonial angles in a sample of growing children with Class II malocclusion. More specifically, in children with thicker masseter muscles, a smaller intermaxillary and gonial angle were found. Other methods to measure the functional capacity of the mas- ticatory muscles have also been looked at in relation to the vertical facial morphology, often arriving at similar conclusions. Vertical facial dimensions have been found to be negatively correlated with Previous studies, although having been carried out in adults, have provided results which point in the same direction, with numerous studies (Bakke et al., 1992; Benington et al., 1999; Farella et al., 2003; Kiliaridis & Kalebo, 1991; Raadsheer et al., 1996; Rani & Ravi, 2010; 6 6 TENTOLOURI ET AL. maximal bite force and the electromyographic activity of the masti- catory muscles (Custodio et al., 2011; Takeuchi‐Sato et al., 2019). Computer tomography (CT) studies have also found comparable re- sults with masseter muscle thickness and length showing a negative correlation with the mandibular plane angle (Azaroual et al., 2014). were found in the present sample probably because children were less than 15 years of age, and sex differences normally appear after puberty and into adulthood. maximal bite force and the electromyographic activity of the masti- catory muscles (Custodio et al., 2011; Takeuchi‐Sato et al., 2019). Computer tomography (CT) studies have also found comparable re- sults with masseter muscle thickness and length showing a negative correlation with the mandibular plane angle (Azaroual et al., 2014). Data resulting from the present study can provide clinical re- levance with regard to the treatment of Class II malocclusion in growing children with the use of functional appliances. The gonial angle is a site of muscle attachment and could therefore be used as a representative variable defining the masticatory system. 4 \| DISCUSSION A large go- nial angle may indicate that the masticatory system is weaker and a smaller gonial angle that there is a better developed masticatory system. As this is easily measurable in any lateral cephalometric radiograph, it could be looked at in all children to evaluate the masticatory system before commencing functional appliance treat- ment to be able to partly predict treatment outcomes. The majority of the data available however to date are derived from adult samples. Few studies have looked at associations between the functional capacity of the masticatory muscles and vertical cra- niofacial morphology in growing children. In a sample of 60 pre- pubertal children, Lione et al. (2013) found that the masseter muscles presented a significantly decreased ultrasonographic volume in dolichofacial subjects compared with brachyfacial or normofacial subjects by looking at the mandibular plane angle. In a study on 61 growing children, Biondi et al. (2016) showed that masseter muscle thickness was progressively decreased in low‐angle, normal‐ angle, and high‐angle subjects, with the vertical skeletal pattern being evaluated using the Frankfurt‐mandibular plane angle. Another study carried out on 72 children between 8.5 and 9.5 years of age found that the importance of the masseter muscle is more evident in the vertical facial morphology of females, whereby there was a negative association between masseter muscle thickness and the inter- maxillary angle (Charalampidou et al., 2008). Muscular factors in relation to vertical craniofacial morphology are one of many factors that can play an important role in dictating this morphology. Particularly in children who are in their mixed dentition, the number of interocclusal contacts during the period of the transition of the dentition may be responsible for some of the variation in muscle thickness measurements, which may attenuate any association between muscles and craniofacial morphology. Occlusal contacts have been found to be associated with muscle thickness and function (Bakke et al., 1992; Ferrario et al., 2002). There are of course other variables such as function, genetics, or ethnic origin which may be equally or more important for craniofacial growth. A complex interaction between genetics and environmental influences can determine the development of craniofacial morphol- ogy during growth. Using other methods to evaluate the functional capacity of the masticatory muscles, comparable results have also been obtained. Ingervall and Thilander (1974) showed that children with greater muscle activity during maximal contraction presented a more rec- tangular facial type. 4 \| DISCUSSION However, one study in growing children looking at maximal bite force found correlations between bite force and craniofacial dimensions only for boys (Sonnesen & Bakke, 2005). Using three‐dimensional CT imaging in growing children, it has been found that in dolichofacial subjects with an increased man- dibular plane angle, a decreased posterior facial height, and a de- creased facial height ratio, the angle between the anterior border of the masseter and the Frankfurt horizontal plane is likely to be con- siderably less than in brachyfacial subjects (Chan et al., 2008; Wong et al., 2016). CONFLICT OF INTERESTS Celli, D., Gasperoni, E., Oliva, B., & Deli, R. (2010). Assessment of mandibular growth and response to functional appliance treatment in prepubertal patients with different auxologic categories. Progress in Orthodontics, 11(1), 20–26. https://doi.org/10.1016/j.pio.2010. 04.010 The authors declare that there are no conflicts of interests. The authors declare that there are no conflicts of interests. nevertheless calibrated to the senior author. Moreover, the center of origin was included as an independent variable in multiple regression analyses in an effort to account for these differences and possible systematic error. Benington, P. C., Gardener, J. E., & Hunt, N. P. (1999). Masseter muscle volume measured using ultrasonography and its relationship with facial morphology. European Journal of Orthodontics, 21(6), 659–670. https://doi.org/10.1093/ejo/21.6.659 https://doi.org/10.1093/ejo/21.6.659 Biondi, K., Lorusso, P., Fastuca, R., Mangano, A., Zecca, P. A., Bosco, M., Caprioglio, A., & Levrini, L. (2016). Evaluation of masseter muscle in different vertical skeletal patterns in growing patients. European Journal of Paediatric Dentistry, 17(1), 47–52. AUTHOR CONTRIBUTIONS Gregory S. Antonarakis and Stavros Kiliaridis conceived, designed, and supervised the study. Eirini Tentolouri and Ioanna Georgiakaki conducted sample collection and analysis. Gregory S. Antonarakis carried out the statistical elaboration of the project. Eirini Tentolouri, Gregory S. Antonarakis, and Stavros Kiliaridis contributed to the in- terpretation of the data and writing of the manuscript. All authors read and approved the final manuscript. Chan, H. J., Woods, M., & Stella, D. (2008). Mandibular muscle morphology in children with different vertical facial patterns: A 3‐ dimensional computed tomography study. American Journal of Orthodontics and Dentofacial Orthopedics, 133(1), 10.e1‐13. https://doi.org/10.1016/j.ajodo.2007.05.013 Charalampidou, M., Kjellberg, H., Georgiakaki, I., & Kiliaridis, S. (2008). Masseter muscle thickness and mechanical advantage in relation to vertical craniofacial morphology in children. Acta Odontologica Scandinavica, 66(1), 23–30. https://doi.org/10.1080/0001635070 1884604 5 \| CONCLUSIONS Bishara, S. E., & Ziaja, R. R. (1989). Functional appliances: A review. American Journal of Orthodontics and Dentofacial Orthopedics, 95(3), 250–258. https://doi.org/10.1016/0889-5406(89)90055-3 • Masseter muscle thickness in growing children with Class II mal- occlusion is independent of sex but correlated with age, with older patients presenting thicker masseter muscles. • Masseter muscle thickness in growing children with Class II mal- occlusion is independent of sex but correlated with age, with older patients presenting thicker masseter muscles. Bresin, A. (2001). Effects of masticatory muscle function and bite‐raising on mandibular morphology in the growing rat. Swedish Dental Journal, (150), 1–49. • In children with thicker masseter muscles, smaller intermaxillary and gonial angles respectively were found. • In children with thicker masseter muscles, smaller intermaxillary and gonial angles respectively were found. Carels, C., & van der Linden, F. P. (1987). Concepts on functional appliances' mode of action. American Journal of Orthodontics and Dentofacial Orthopedics, 92(2), 162–168. https://doi.org/10.1016/ 0889-5406(87)90371-4 ORCID Eirini Tentolouri http://orcid.org/0000-0003-0954-2079 Eirini Tentolouri http://orcid.org/0000-0003-0954-2079 Custodio, W., Gomes, S. G., Faot, F., Garcia, R. C., & Del Bel Cury, A. A. (2011). Occlusal force, electromyographic activity of masticatory muscles and mandibular flexure of subjects with different facial types. Journal of Applied Oral Science, 19(4), 343–349. https://doi. Gregory S. Antonarakis http://orcid.org/0000-0001-6741-6922 Gregory S. Antonarakis http://orcid.org/0000-0001-6741-6922 4.1 \| Limitations The present study evaluated a large sample of children, more than 200, and it is thus to date the largest study attempting to look at associations between masseter muscle thickness and vertical cra- niofacial morphology in growing children. Limitations however in- clude that its design is retrospective and cross‐sectional in nature. Ideally, a prospective longitudinal study would be able to give more concrete answers with regard to the development of the vertical craniofacial morphology in children in relation to their masticatory functional capacity. It is unclear at what point during growth and development the associations between the masticatory muscles and vertical facial morphology become more evident, or what the exact cause and ef- fect is. Proffit and Fields (1983) suggested that low bite force in subjects with a hyperdivergent facial pattern might allow excessive eruption of posterior teeth and backward rotation of the mandible although they were unable to find any significant differences be- tween normal‐ and long‐face children with regard to bite force. This explanation is also offered by other authors, although based on a rat experimental model, agreeing that the hypofunction of the muscles in a hyperdivergent pattern produces weaker forces, and this results in a greater eruption of posterior molars and more vertical growth (Bresin, 2001). A further possible limitation is the fact that the study was carried out in a homogeneous sample, looking specifically at children with Class II malocclusion, which raises questions about generalizability to other malocclusions or populations. The advantage of including only children with Class II malocclusion however is to ensure a certain homogeneity and clinical relevance in respect to treatment outcomes in this group of children such as with functional appliances. Finally, the patients are derived from two different centers with differences in the populations, and the operator carrying out the ultrasono- graphic masseter muscle thickness measurements. Interoperator re- liability was not carried out since measurements were not carried out on the same patients by the two examiners, but both examiners were The originality in the present study is that associations between masseter muscle thickness and vertical craniofacial parameters are investigated in a large homogeneous sample of children only with Class II malocclusion. One must keep in mind that no sex differences TENTOLOURI ET AL. 7 DATA AVAILABILITY STATEMENT Cretella Lombardo, E., Franchi, L., Gastaldi, G., Giuntini, V., Lione, R., Cozza, P., & Pavoni, C. (2020). Development of a prediction model for short‐term success of functional treatment of Class II malocclusion. International Journal of Environmental Research and Public Health, 17(12), 4473. https://doi.org/10.3390/ijerph1712 4473 The data that support the findings of this study are available from the corresponding author upon reasonable request. 8 \| Dental Research, 21(3), 402–407. https://doi.org/10.4103/0970- 9290.70812 Journal of Physical Anthropology, 80(4), 429–445. https://doi.org/10. 1002/ajpa.1330800404 Journal of Physical Anthropology, 80(4), 429–445. https://doi.org/10. 1002/ajpa.1330800404 Houston, W. J. (1983). The analysis of errors in orthodontic measurements. American Journal of Orthodontics, 83(5), 382–390. https://doi.org/10.1016/0002-9416(83)90322-6 Rohila, A. K., Sharma, V. P., Shrivastav, P. K., Nagar, A., & Singh, G. P. (2012). An ultrasonographic evaluation of masseter muscle thickness in different dentofacial patterns. Indian Journal of Dental Research, 23(6), 726–731. https://doi.org/10.4103/0970-9290.111247 Ingervall, B., & Thilander, B. (1974). Relation between facial morphology and activity of the masticatory muscles. Journal of Oral Rehabilitation, 1(2), 131–147. https://doi.org/10.1111/j.1365- 2842.1974.tb00771.x Satiroğlu, F., Arun, T., & Işik, F. (2005). Comparative data on facial morphology and muscle thickness using ultrasonography. European Journal of Orthodontics, 27(6), 562–567. https://doi.org/10.1093/ ejo/cji052 Kiliaridis, S. (2006). The importance of masticatory muscle function in dentofacial growth. Seminars in Orthodontics, 12(2), 110–119. https://doi.org/10.1053/j.sodo.2006.01.004 Sonnesen, L., & Bakke, M. (2005). Molar bite force in relation to occlusion, craniofacial dimensions, and head posture in pre‐orthodontic children. European Journal of Orthodontics, 27(1), 58–63. https:// doi.org/10.1093/ejo/cjh069 Kiliaridis, S., & Kalebo, P. (1991). Masseter muscle thickness measured by ultrasonography and its relation to facial morphology. Journal of Dental Research, 70(9), 1262–1265. https://doi.org/10.1177/0022 0345910700090601 Takeuchi‐Sato, T., Arima, T., Mew, M., & Svensson, P. (2019). Relationships between craniofacial morphology and masticatory muscle activity during isometric contraction at different interocclusal distances. Archives of Oral Biology, 98, 52–60. https://doi.org/10. 1016/j.archoralbio.2018.10.030 Kiliaridis, S., Mills, C. M., & Antonarakis, G. S. (2010). Masseter muscle thickness as a predictive variable in treatment outcome of the twin‐ block appliance and masseteric thickness changes during treatment. Orthodontics & Craniofacial Research, 13(4), 203–213. https://doi. org/10.1111/j.1601-6343.2010.01496.x Tulloch, J. F., Medland, W., & Tuncay, O. C. (1990). Methods used to evaluate growth modification in Class II malocclusion. American Journal of Orthodontics and Dentofacial Orthopedics, 98(4), 340–347. https://doi.org/10.1016/s0889-5406(05)81491-x Kubota, M., Nakano, H., Sanjo, I., Satoh, K., Sanjo, T., Kamegai, T., & Ishikawa, F. (1998). Maxillofacial morphology and masseter muscle thickness in adults. European Journal of Orthodontics, 20(5), 535–542. https://doi.org/10.1093/ejo/20.5.535 Vargervik, K., & Harvold, E. P. (1985). Response to activator treatment in Class II malocclusions. American Journal of Orthodontics, 88(3), 242–251. https://doi.org/10.1016/s0002-9416(85)90219-2 Kumar, S. A., Shetty, K. S., & Prakash, A. T. (2013). Growth modulation using functional appliances—cephalometric predictors of successful response. Orthodontics, 14(1), e50–e59. https://doi.org/10.11607/ ortho.917 Weijs, W. A., & Hillen, B. (1984). Relationships between masticatory muscle cross‐section and skull shape. Journal of Dental Research, 63(9), 1154–1157. REFERENCES Antonarakis, G. S., & Kiliaridis, S. (2015). Predictive value of masseter muscle thickness and bite force on Class II functional appliance treatment: A prospective controlled study. European Journal of Orthodontics, 37(6), 570–577. https://doi.org/10.1093/ejo/cju089 Dibbets, J. M. (1992). One century of Wolff's law, in bone biodynamics in orthodontic and orthopaedic treatment. Craniofacial Growth Series, 27, 1–13. Farella, M., Bakke, M., Michelotti, A., Rapuano, A., & Martina, R. (2003). Masseter thickness, endurance and exercise‐induced pain in subjects with different vertical craniofacial morphology. European Journal of Oral Sciences, 111(3), 183–188. https://doi.org/10.1034/j.1600- 0722.2003.00035.x Antonarakis, G. S., Kjellberg, H., & Kiliaridis, S. (2012). Predictive value of molar bite force on Class II functional appliance treatment outcomes. European Journal of Orthodontics, 34(2), 244–249. https://doi.org/10.1093/ejo/cjq184 Azaroual, M. F., Fikri, M., Abouqal, R., Benyahya, H., & Zaoui, F. (2014). Relationship between dimensions of muscles of mastication (masseter and lateral pterygoid) and skeletal dimensions: Study of 40 cases. International Orthodontics, 12(1), 111–124. https://doi. org/10.1016/j.ortho.2013.09.001 Ferrario, V. F., Serrao, G., Dellavia, C., Caruso, E., & Sforza, C. (2002). Relationship between the number of occlusal contacts and masticatory muscle activity in healthy young adults. Cranio, 20(2), 91–98. https://doi.org/10.1080/08869634.2002.11746196 Franchi, L., & Baccetti, T. (2006). Prediction of individual mandibular changes induced by functional jaw orthopedics followed by fixed appliances in Class II patients. Angle Orthodontist, 76(6), 950–954. https://doi.org/10.2319/110205-385 Bakke, M., Tuxen, A., Vilmann, P., Jensen, B. R., Vilmann, A., & Toft, M. (1992). Ultrasound image of human masseter muscle related to bite force, electromyography, facial morphology, and occlusal factors. Scandinavian Journal of Dental Research, 100(3), 164–171. https:// doi.org/10.1111/j.1600-0722.1992.tb01734.x Graber, T. M., & Neumann, B. (1984). Removable Orthodontic Appliances. Saunders. Barton, S., & Cook, P. A. (1997). Predicting functional appliance treatment outcome in Class II malocclusions—a review. American Journal of Orthodontics and Dentofacial Orthopedics, 112(3), 282–286. https:// doi.org/10.1016/s0889-5406(97)70257-9 Hannam, A. G., & Wood, W. W. (1989). Relationships between the size and spatial morphology of human masseter and medial pterygoid muscles, the craniofacial skeleton, and jaw biomechanics. American 8 TENTOLOURI ET AL. Journal of Physical Anthropology, 80(4), 429–445. https://doi.org/10. 1002/ajpa.1330800404 8 \| 8 8 \| https://doi.org/10.1177/002203458406300 91201 Lione, R., Franchi, L., Noviello, A., Bollero, P., Fanucci, E., & Cozza, P. (2013). Three‐dimensional evaluation of masseter muscle in different vertical facial patterns: A cross‐sectional study in growing children. Ultrasonic Imaging, 35(4), 307–317. https://doi.org/10. 1177/0161734613502468 Weijs, W. A., & Hillen, B. (1986). Correlations between the cross‐sectional area of the jaw muscles and craniofacial size and shape. American Journal of Physical Anthropology, 70(4), 423–431. https://doi.org/10. 1002/ajpa.1330700403 Macey‐Dare, L. V., & Nixon, F. (1999). Functional appliances: Mode of action and clinical use. Dental Update, 26(6), 240–244. https://doi. org/10.12968/denu.1999.26.6.240 Wolff, J. (1870). Ueber die innere Architectur der Knochen und ihre Bedeutung für die Frage vom Knochenwachsthum. Archiv für pathologische Anatomie und Physiologie und für klinische Medicin, 50(3), 389–450. Patel, H. P., Moseley, H. C., & Noar, J. H. (2002). Cephalometric determinants of successful functional appliance therapy. Angle Orthodontist, 72(5), 410–417. https://doi.org/10.1043/0003-3219 (2002)072%3C0410:Cdosfa%3E2.0.Co;2 Wong, A., Woods, M. G., & Stella, D. (2016). Three‐dimensional computed tomographic assessment of mandibular muscles in growing subjects with different vertical facial patterns. Australian Orthodontic Journal, 32(1), 2–17. Proffit, W. R., & Fields, H. W. (1983). Occlusal forces in normal‐ and long‐ face children. Journal of Dental Research, 62(5), 571–574. https://doi. org/10.1177/00220345830620051301 Woodside, D. G. (1998). Do functional appliances have an orthopedic effect? American Journal of Orthodontics and Dentofacial Orthopedics, 113(1), 11–14. https://doi.org/10.1016/s0889-5406(98)70270-7 Raadsheer, M. C., Van Eijden, T. M., Van Spronsen, P. H., Van Ginkel, F. C., Kiliaridis, S., & Prahl‐Andersen, B. (1994). A comparison of human masseter muscle thickness measured by ultrasonography and magnetic resonance imaging. Archives of Oral Biology, 39(12), 1079–1084. https://doi.org/10.1016/0003-9969(94)90061-2 How to cite this article: Tentolouri, E., Antonarakis, G. S., Georgiakaki, I., & Kiliaridis, S. (2022). Masseter muscle thickness and vertical cephalometric characteristics in children with Class II malocclusion. Clinical and Experimental Dental Research, 1–8. https://doi.org/10.1002/cre2.528 Raadsheer, M. C., Kiliaridis, S., Van Eijden, T. M., Van Ginkel, F. C., & Prahl‐ Andersen, B. (1996). Masseter muscle thickness in growing individuals and its relation to facial morphology. Archives of Oral Biology, 41(4), 323–332. https://doi.org/10.1016/0003-9969(95) 00136-0 Rani, S., & Ravi, M. S. (2010). Masseter muscle thickness in different skeletal morphology: An ultrasonographic study. Indian Journal of
https://openalex.org/W3198390285	https://researchonline.lshtm.ac.uk/id/eprint/4663299/1/1-s2.0-S0140673621016998-main.pdf	English	null	Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: a substudy of two randomised controlled trials (COV001 and COV002)	Lancet	2,021	cc-by	10,372	Lancet 2021; 398: 981–90 Lancet 2021; 398: 981–90 Published Online September 1, 2021 https://doi.org/10.1016/ S0140-6736(21)01699-8 See Comment page 933 Contributed equally Jenner Institute, Nuffield Department of Medicine (A Flaxman DPhil, D Jenkin MRCP, J Aboagye BSc, S Belij-Rammerstorfer PhD, M Bittaye PhD, F Cappuccini PhD, S Davies BSc, K J Ewer PhD, P M Folegatti MD, J Fowler BA, Prof A V S Hill FRS, A M Minassian DPhil, H Sanders BSc, E Sheehan PhD, H Smith MSc, D Woods BSc, Prof S C Gilbert FMedSci, T Lambe PhD), Oxford Vaccine Methods In this substudy, volunteers aged 18–55 years who were enrolled in the phase 1/2 (COV001) controlled trial in the UK and had received either a single dose or two doses of 5 × 10¹⁰ viral particles were invited back for vaccination. Here we report the reactogenicity and immunogenicity of a delayed second dose (44–45 weeks after first dose) or a third dose of the vaccine (28–38 weeks after second dose). Data from volunteers aged 18–55 years who were enrolled in either the phase 1/2 (COV001) or phase 2/3 (COV002), single-blinded, randomised controlled trials of ChAdOx1 nCoV-19 and who had previously received a single dose or two doses of 5 × 10¹⁰ viral particles are used for comparison purposes. COV001 is registered with ClinicalTrials.gov, NCT04324606, and ISRCTN, 15281137, and COV002 is registered with ClinicalTrials.gov, NCT04400838, and ISRCTN, 15281137, and both are continuing but not recruiting. Findings Between March 11 and 21, 2021, 90 participants were enrolled in the third-dose boost substudy, of whom 80 (89%) were assessable for reactogenicity, 75 (83%) were assessable for evaluation of antibodies, and 15 (17%) were assessable for T-cells responses. The two-dose cohort comprised 321 participants who had reactogenicity data (with prime-boost interval of 8–12 weeks: 267 [83%] of 321; 15–25 weeks: 24 [7%]; or 44–45 weeks: 30 [9%]) and 261 who had immunogenicity data (interval of 8–12 weeks: 115 [44%] of 261; 15–25 weeks: 116 [44%]; and 44–45 weeks: 30 [11%]). 480 participants from the single-dose cohort were assessable for immunogenicity up to 44–45 weeks after vaccination. Antibody titres after a single dose measured approximately 320 days after vaccination remained higher than the titres measured at baseline (geometric mean titre of 66·00 ELISA units [EUs; 95% CI 47·83–91·08] vs 1·75 EUs [1·60–1·93]). 32 participants received a late second dose of vaccine 44–45 weeks after the first dose, of whom 30 were included in immunogenicity and reactogenicity analyses. Articles Articles Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: a substudy of two randomised controlled trials (COV001 and COV002) Amy Flaxman, Natalie G Marchevsky, Daniel Jenkin, Jeremy Aboagye, Parvinder K Aley, Brian Angus, Sandra Belij-Rammerstorfer, Sagida Bibi, Mustapha Bittaye, Federica Cappuccini, Paola Cicconi, Elizabeth A Clutterbuck, Sophie Davies, Wanwisa Dejnirattisai, Christina Dold, Katie J Ewer, Pedro M Folegatti, Jamie Fowler, Adrian V S Hill, Simon Kerridge, Angela M Minassian, Juthathip Mongkolsapaya, Yama F Mujadidi, Emma Plested, Maheshi N Ramasamy, Hannah Robinson, Helen Sanders, Emma Sheehan, Holly Smith, Matthew D Snape, Rinn Song, Danielle Woods, Gavin Screaton, Sarah C Gilbert, Merryn Voysey, Andrew J Pollard, Teresa Lambe, and the Oxford COVID Vaccine Trial group Summary y Background COVID-19 vaccine supply shortages are causing concerns about compromised immunity in some countries as the interval between the first and second dose becomes longer. Conversely, countries with no supply constraints are considering administering a third dose. We assessed the persistence of immunogenicity after a single dose of ChAdOx1 nCoV-19 (AZD1222), immunity after an extended interval (44–45 weeks) between the first and second dose, and response to a third dose as a booster given 28–38 weeks after the second dose. Paediatrics (N G Marchevsky MSc, P K Aley PhD, B Angus MD, S Bibi PhD, P Cicconi PhD, E A Clutterbuck PhD, C Dold PhD, S Kerridge MSc, Y F Mujadidi MSc, E Plested, M N Ramasamy DPhil, H Robinson DipHE, M D Snape MD, R Song MD, M Voysey DPhil, Prof Sir A J Pollard FMedSci), and Chinese Academy of Medical Science (CAMS) Oxford Institute (T Lambe), University of Oxford, Oxford, UK; NIHR Oxford Biomedical Research Centre, Oxford, UK (N G Marchevsky, S Bibi, E A Clutterbuck, C Dold, M Voysey, Prof Sir A J Pollard); Wellcome Centre for Human Genetics, University of Oxford, Oxford, UK (W Dejnirattisai PhD, J Mongkolsapaya PhD, Prof G Screaton DPhil) Lancet 2021; 398: 981–90 Published Online September 1, 2021 https://doi.org/10.1016/ S0140-6736(21)01699-8 See Comment page 933 *Contributed equally Jenner Institute, Nuffield Department of Medicine (A Flaxman DPhil, D Jenkin MRCP, J Aboagye BSc, S Belij-Rammerstorfer PhD, M Bittaye PhD, F Cappuccini PhD, S Davies BSc, K J Ewer PhD, P M Folegatti MD, J Fowler BA, Prof A V S Hill FRS, A M Minassian DPhil, H Sanders BSc, E Sheehan PhD, H Smith MSc, D Woods BSc, Prof S C Gilbert FMedSci, T Lambe PhD), Oxford Vaccine Group, Department of Paediatrics (N G Marchevsky MSc, P K Aley PhD, B Angus MD, S Bibi PhD, P Cicconi PhD, E A Clutterbuck PhD, C Dold PhD, S Kerridge MSc, Y F Mujadidi MSc, E Plested, M N Ramasamy DPhil, H Robinson DipHE, M D Snape MD, R Song MD, M Voysey DPhil, Prof Sir A J Pollard FMedSci), and Chinese Academy of Medical Science (CAMS) Oxford Institute (T Lambe), University of Oxford, Oxford, UK; NIHR Oxford Biomedical Research Centre, Oxford, UK (N G Marchevsky, S Bibi, E A Clutterbuck, C Dold, M Voysey, Prof Sir A J Pollard); Wellcome Centre for Human Genetics, University of Oxford, Oxford, UK (W Dejnirattisai PhD, J Mongkolsapaya PhD, Prof G Screaton DPhil) Correspondence to: Dr Amy Flaxman, Jenner Institute, Nuffield Department of Medicine, University of Oxford, Headington, Oxford, OX3 7DQ, UK amy.flaxman@ndm.ox.ac.uk For more on COVAX see https://www.who.int/initiatives/ act-accelerator/covax Evidence before this study homologous third dose of BNT162b2, 6 or 12 months after the second dose. Multiple vaccines against SARS-CoV-2 have now been authorised for use in various countries. Most vaccines are given in a two-dose primary schedule, and further doses might be required to maintain protective immunity or control emerging variants. We searched PubMed for research articles published between database inception and June 23, 2021, using the search terms “SARS-CoV-2”, “vaccine”, “clinical trial”, AND (“third dose” OR “late boost”) with no language restrictions. We identified animal studies using combinations of three-dose vaccine delivery in prime-boost schedules. Additionally, we identified three clinical trials of three-dose delivery, including two in solid organ transplant recipients. In the first study in transplant recipients, antibody titres increased after the third dose of either BNT162b2 (Pfizer–BioNTech) or mRNA-1273 (Moderna) vaccines in a third of patients who had negative antibody titres and in all patients who had low-positive antibody titres. In the second study in transplant recipients, prevalence of antibody titres increased from 44% after a second dose to 68% after a third dose. In a phase 1 and 2 trial of a protein subunit vaccine ZF2001, the safety and immunogenicity data support the use of a 25 μg dose in a three-dose schedule. A number of clinical studies are measuring the effect of a third dose of vaccine, including a phase 1 study of 144 participants who received a Articles Biomedical Research Centre, Chinese Academy of Medical Sciences Innovation Fund for Medical Science, Thames Valley and South Midlands NIHR Clinical Research Network, AstraZeneca, and Wellcome. Biomedical Research Centre, Chinese Academy of Medical Sciences Innovation Fund for Medical Science, Thames Valley and South Midlands NIHR Clinical Research Network, AstraZeneca, and Wellcome. Biomedical Research Centre, Chinese Academy of Medical Sciences Innovation Fund for Medical Science, Thames Valley and South Midlands NIHR Clinical Research Network, AstraZeneca, and Wellcome. Copyright © 2021 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4 Introduction the possible risk of breakthrough infection as new variants emerge. The COVID-19 pandemic continues to put a substantial burden on health-care systems and a massive global effort is underway to protect populations through vaccination. COVID-19 vaccine supply shortages in many countries are causing concern about com promised immunity as the interval between the first and second dose extends beyond 12 weeks.1 WHO recommends that the second dose of the ChAdOx1 nCoV-19 vaccine is given 8–12 weeks after the first dose because the clinical trial data provide support for good levels of protection with this interval;2,3 however, many countries cannot obtain sufficient supplies to allow second doses to be administered by 12 weeks. These supply shortages are leading to longer intervals and uncertainty among policy makers about whether protection against COVID-192,3 will be maintained because no data exist on the efficacy of the immunisation schedules with intervals between the first and second dose that extend beyond this limit. ChAdOx1 nCoV-19 (AZD1222), a replication deficient adenoviral vectored vaccine that encodes the SARS-CoV-2 spike protein, is one of the most widely used vaccines globally. More than half a billion doses have been distributed to more than 168 countries across six continents, including provision through the COVAX Facility. Here, we describe tolerability and immune response to a late second dose (44–45 weeks after the first dose) of ChAdOx1 nCoV-19, and after a third dose (28–38 weeks after the second dose). We also report the persistence of antibody and cellular responses at 182 days and for antibodies up to 320 days after first dose of ChAdOx1 nCoV-19. Study design and participants y g In this substudy, we extended the data already collected as part of the UK COV001 and COV002 trials. In these trials, participants were randomly assigned to receive ChAdOx1 nCoV-19 or a meningococcal conjugate vaccine (MenACWY) as a control. Procedures, safety, immune responses, and efficacy before late vaccination and after Conversely, some high-income countries with highly vaccinated populations are considering administration of a third dose of a COVID-19 vaccine because of uncertainty about duration of immunity after the first two doses and Added value of this study We report immune responses to ChAdOx1 nCoV-19 following a second dose after an extended interval between the first and second dose, and after a third dose with an extended interval between the second and third dose. The extended interval between the first two doses (44–45 weeks) resulted in higher antibody titres after the second dose than with a shortened interval. A third dose given 28–38 weeks after the primary series increased the antibody titres to above those after a second dose with a shortened interval. Reactogenicity was lower after the second or third dose than after the first dose. Lancet 2021; 398: 981–90 Antibody titres were higher 28 days after vaccination in those with a longer interval between first and second dose than for those with a short interval (median total IgG titre: 923 EUs [IQR 525–1764] with an 8–12 week interval; 1860 EUs [917–4934] with a 15–25 week interval; and 3738 EUs [1824–6625] with a 44–45 week interval). Among participants who received a third dose of vaccine, antibody titres (measured in 73 [81%] participants for whom samples were available) were significantly higher 28 days after a third dose (median total IgG titre: 3746 EUs [IQR 2047–6420]) than 28 days after a second dose (median 1792 EUs [IQR 899–4634]; Wilcoxon signed rank test p=0·0043). T-cell responses were also boosted after a third dose (median response increased from 200 spot forming units [SFUs] per million peripheral blood mononuclear cells [PBMCs; IQR 127–389] immediately before the third dose to 399 SFUs per milion PBMCs [314–662] by day 28 after the third dose; Wilcoxon signed rank test p=0·012). Reactogenicity after a late second dose or a third dose was lower than reactogenicity after a first dose. Interpretation An extended interval before the second dose of ChAdOx1 nCoV-19 leads to increased antibody titres. A third dose of ChAdOx1 nCoV-19 induces antibodies to a level that correlates with high efficacy after second dose and boosts T-cell responses. Funding UK Research and Innovation, Engineering and Physical Sciences Research Council, National Institute for Health Research, Coalition for Epidemic Preparedness Innovations, National Institute for Health Research Oxford 981 www.thelancet.com Vol 398 September 11, 2021 Articles Articles Procedures Participants who were included as part of this substudy were vaccinated with a standard dose of ChAdOx1 nCoV-19 (5 × 10¹⁰ viral particles). For control participants who had previously received either one or two doses of MenACWY, this vaccination was their first dose of ChAdOx1 nCoV-19. For participants who had previously received ChAdOx1 nCoV-19, this vaccination was either their second dose (44–45 weeks after the first) or their third dose. These late vaccinations occurred 10 months (plus or minus 56 days) from enrolment. 7 days after vaccination participants were unmasked to treatment allocation, so that those who had received only one dose of ChAdOx1 nCoV-19 could subsequently receive a second dose, in line with national vaccination roll-out in the UK. For analysis of immunogenicity after a single dose of ChAdOx1 nCoV-19, we included all participants in COV001 and COV002 who had yet to receive a second dose of vaccine and for whom immunogenicity data were available. A substudy was added to the COV001 trial as a protocol amendment on March 1, 2021, to investigate the immunogenicity and tolerability of a third dose of the vaccine. Participants who had previously received two doses of ChAdOx1 nCoV-19 were recruited for this substudy, along with control participants, who had received two doses of MenACWY previously, to maintain blinding of reactogenicity data. All recruits received ChAdOx1 nCoV-19; for the control participants this was their first dose. For participants from this three-dose cohort to be eligible for inclusion in these analyses, they had to have an 8–16 week interval between first and second doses. Participants enrolled in the substudy had blood samples taken on the day of vaccination, and then at 14 days and 28 days after vaccination to allow immunogenicity assess ments to be made. Binding antibody titres were measured using standardised single dilution total IgG ELISAs as previously described.5 This assay was used to measure antibody responses before and after vaccination to Victoria/01/2020 SARS-CoV-2 spike protein and adapted to measure responses to beta (B.1.351) SARS-CoV-2 protein. Implications of all the available evidence Vaccine shortages have resulted in some people receiving a first dose of ChAdOx1 nCoV-19 without receiving the second dose within the recommended 4–12 week period. We report that increasing the interval up to 45 weeks results in increased antibody titres after the second dose, offering increased flexibility in vaccination schedules. A third dose at an extended interval after the second dose resulted in a further increase in antibody titres, mitigating concerns that antibodies raised against the ChAdOx1 vector would limit repeated use of the vaccine. 982 www.thelancet.com Vol 398 September 11, 2021 Articles April 19, 2020). The protocol for COV001 is provided in appendix 1 and the protocol for COV002 is provided in appendix 2. April 19, 2020). The protocol for COV001 is provided in appendix 1 and the protocol for COV002 is provided in appendix 2. second dose have been previously published.2,4,5 The trials were originally planned as single-dose vaccine studies, but the strong neutralising titres seen in COV001 induced by a second dose of vaccine4 prompted a protocol amendment to allow the addition of booster doses to most study participants across both trials. Most participants in both COV001 and COV002 were invited to receive a second dose from July, 2020, onwards. The timing of the second dose varied and allows for comparisons of immunogenicity between the recom mended vaccination schedule in the UK of 8–12 weeks and a longer interval of 15–25 weeks.3 The initial phase 1 immunogenicity group in COV001 was retained as a single-dose cohort to observe the persistence of immune responses after a single dose. See Online for appendix 1 See Online for appendix 2 www.thelancet.com Vol 398 September 11, 2021 See Online for appendix 1 See Online for appendix 2 Procedures ELISA assays to Victoria/01/2020 were performed on samples Figure 1: Trial profile for three-dose cohort 90 participants were enrolled and received their third dose of ChAdOx1 nCoV-19 80 participants assessable for reactogenicity 10 excluded because included in an open-label subgroup and interval between doses one and two was <28 days 75 assessable for antibody immunogenicity assessment 5 excluded because their prime-boost interval was outside of the deﬁned interval (8–16 weeks) 15 assessable for T-cell response immunogenicity assessments 60 excluded 56 no ELISPot data 4 insuﬃcient ELISpot data Some participants from the COV001 single-dose cohort were also offered a second dose at this time. The single dose cohort originally comprised a 1:1 ratio of ChAdOx1 nCoV-19 recipients to MenACWY controls. These participants were invited back in a 2:1 ratio, so participants who had previously received a single dose of ChAdOx1 nCoV-19 received their second dose, with an interval of 44–45 weeks, and those who had been controls received their first dose of ChAdOx1 nCoV-19. Participants were targeted for inclusion in the substudy if they had not previously been unmasked to treatment allocation or offered a vaccine as part of the UK Government COVID-19 vaccine programme. A subset of two-dose recipients (for whom reactogenicity or immunogenicity data, or both, were available) were selected for inclusion in analyses for comparison with those who received two doses 44–45 weeks apart. Participants who had a positive PCR test for SARS-CoV-2 were removed from the analysis if the infection occurred before the blood draw. Participants for this substudy were only enrolled at the Oxford site. In the UK, the COV001 and COV002 studies were approved by the South Central Berkshire Research Ethics Committee (COV001 reference 20/SC/0145, on March 23, 2020; COV002 reference 20/SC/0179; con ditional approval on April 8, 2020, and full approval on Figure 1: Trial profile for three-dose cohort www.thelancet.com Vol 398 September 11, 2021 983 Articles Articles Articles For the three-dose cohort, Figure 3: Antibody (A) and T-cell (B) persistence after one dose of ChAdOx1 nCoV-19 vaccine Figure 3: Antibody (A) and T-cell (B) persistence after one dose of ChAdOx1 nCoV-19 vaccine Datapoints represent individual participants and the solid line represents estimates from a linear regression model, with shaded areas showing the 95% CI. Antibody levels to SARS-CoV-2 Victoria/01/2020 spike measured by total IgG ELISA over 1 year after a single dose. Data are from 480 participants across COV001 and COV002 who received a standard dose of ChAdOx1 nCoV-19. Vaccine-induced T-cell responses against the SARS-CoV-2 spike insert were monitored up to day 182 in a cohort of 44 participants who received a single dose of ChAdOx1 nCoV-19. For participants who were excluded from these analyses due to positive PCR test result, second dose on trial, or external COVID-19 vaccination, no ELISA results or ELISpot results beyond the date of censoring were used. PBMCs=peripheral blood mononuclear cells. SFUs=spot-forming units Figure 3: Antibody (A) and T-cell (B) persistence after one dose of ChAdOx1 nCoV-19 vaccine Datapoints represent individual participants and the solid line represents estimates from a linear regression model, with shaded areas showing the 95% CI. Antibody levels to SARS-CoV-2 Victoria/01/2020 spike measured by total IgG ELISA over 1 year after a single dose. Data are from 480 participants across COV001 and COV002 who received a standard dose of ChAdOx1 nCoV-19. Vaccine-induced T-cell responses against the SARS-CoV-2 spike insert were monitored up to day 182 in a cohort of 44 participants who received a single dose of ChAdOx1 nCoV-19. For participants who were excluded from these analyses due to positive PCR test result, second dose on trial, or external COVID-19 vaccination, no ELISA results or ELISpot results beyond the date of censoring were used. PBMCs=peripheral blood mononuclear cells. SFUs=spot-forming units data were included only for those who had an interval of 8–16 weeks between first and second doses. See Online for appendix 3 Participants were asked to complete a diary card for 7 days after each vaccination to record solicited local (induration, itch, pain, redness, swelling, tenderness, and warmth at the injection site) and systemic (chills, fatigue, fever of ≥38°C, feverish [self-reported feeling of feverish ness, whereas fever is an objective fever measurement], headache, joint pain, malaise, muscle ache, and nausea) adverse reactions. Articles Mild Moderate Severe Potentially life-threatening or requiring admission to hospital Severity 0 25 50 75 100 Proportion (%) 0 25 50 75 100 Proportion (%) 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Interval between ﬁrst and second doses of vaccine (weeks) Interval between ﬁrst and second doses of vaccine (weeks) A Induration Itch Pain Redness Swelling Tenderness Warmth Local adverse reactions 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Systemic adverse reactions 0 25 50 75 100 Proportion (%) B Induration Itch Pain Redness Swelling Tenderness Warmth Local adverse reactions 0 25 50 75 100 Proportion (%) First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Systemic adverse reactions Mild Moderate Severe Potentially life-threatening or requiring admission to hospital Severity 0 25 50 75 100 Proportion (%) 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Interval between ﬁrst and second doses of vaccine (weeks) A Induration Itch Pain Redness Swelling Tenderness Warmth Local adverse reactions Proportion (%) Proportion (%) 984 www.thelancet.com Vol 398 September 11, 2021 0 0 25 50 75 100 Proportion (%) 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Interval between ﬁrst and second doses of vaccine (weeks) Interval between ﬁrst and second doses of vaccine (weeks) 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Systemic adverse reactions 0 25 50 75 100 Proportion (%) B Induration Itch Pain Redness Swelling Tenderness Warmth Local adverse reactions 0 25 50 75 100 Proportion (%) First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Systemic adverse reactions 0 25 50 75 100 Proportion (%) Interval between ﬁrst and second doses of vaccine (weeks) 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 8–12 15–25 44–45 Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Systemic adverse reactions B Proportion (%) Proportion (%) B 0 25 50 75 100 Proportion (%) B Induration Itch Pain Redness Swelling Tenderness Warmth Local adverse reactions First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose Systemic adverse reactions Proportion (%) Proportion (%) Systemic adverse reactions 0 25 50 75 100 Proportion (%) First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose First dose Second dose Third dose Chills Fever (≥38°C) Feverish Headache Joint pain Malaise Muscle ache Nausea Fatigue Proportion (%) Proportion (%) www.thelancet.com Vol 398 September 11, 2021 984 Articles Figure 3: Antibody (A) and T-cell (B) persistence after one dose of ChAdOx1 nCoV-19 vaccine Datapoints represent individual participants and the solid line represents estimates from a linear regression model, with shaded areas showing the 95% CI. Articles Neutralisation assays were done in a randomly selected subset of participants (due to laboratory capacity) who received a third dose of vaccine. Timepoints assessed were 28 days after second vaccination and 28 days after third vaccination. from single-dose recipients up to 1 year after vaccination, in two-dose recipients up to 6 months after the second vaccination, and in three-dose recipients up to 28 days after the third vaccination. ELISA assays on the beta SARS-CoV-2 variant were only done on samples from participants recruited to the substudy who received either a late second vaccination or a third vaccination, up to 28 days after the late vaccination. Meso Scale Discovery multiplex immunoassay was used to assess antibody titres against spike proteins from different variants (Victoria/01/2020, D614G, alpha [B.1.17], beta, and gamma [P.1]). V-PLEX SARS-CoV-2 Panel 6 (IgG) kits were used following manufacturer’s instructions (Meso Scale Discovery, K15433U; full details are in appendix 3 [p 1]). Meso Scale Discovery assays were done on samples from participants recruited to the substudy who received a late second dose of vaccine (44–45 weeks after the first dose. Ex-vivo IFN-γ ELISpot assays were done as previously described5 to assess T-cell responses to Victoria/01/2020 SARS-CoV-2 spike overlapping peptide pools before and after vaccination. Isolated peripheral blood mononuclear cells (PBMCs) were stimulated overnight with peptides spanning the SARS-CoV-2 spike insert. ELISpot assays were done on samples from participants in the single dose cohort up to 182 days after vaccination. ELISpot assays were also done in some participants (due to laboratory capacity) recruited to the substudy who received a third dose, up to 28 days after third dose. Focus reduction neutralisation assays were done as described previously6 to measure neutralising antibody titres against alpha, beta, and delta (B.1.617.2) SARS-CoV-2 viral variants. Neutralisation assays were done in a randomly selected subset of participants (due to laboratory capacity) who received a third dose of vaccine. Timepoints assessed were 28 days after second vaccination and 28 days after third vaccination. A 1 10 100 1000 10000 0 50 100 150 200 250 300 350 ELISA units B Time since ﬁrst dose of vaccine (days) 10 100 1000 10000 0 50 100 150 200 SFU per million PBMCs For all immunogenicity assessments, data were excluded upon earliest occurrence of a positive PCR test result or external COVID-19 vaccination. For single dose immunogenicity assessments, data were excluded from after receipt of second dose. Figure 2: Solicited adverse reactions up to 7 days after ChAdOx1 nCoV-19 vaccination by interval between first and second doses (A) and after the first, second, and third dose for participants who received a third dose of vaccine (B) Figure shows maximum severity of respective solicited adverse event recorded for each participant during days 0–7 after vaccination. In panel A, reactogenicity data after the second dose are shown for 263 participants for fever (≥38°C) and 267 participants for all other symptoms for the 8–12 week interval, for 23 participants for fever (≥38°C) and 24 participants for all other symptoms for the 15–25 week interval, and 28 participants for fever (≥38°C) and 30 participants for all other symptoms for the 44–45 week interval. In panel B, reactogenicity data are from after each dose recorded by participants who received a third dose of vaccine, with data available for 80 participants for all symptoms after dose 1; 15 participants for all symptoms after dose 2; and 77 participants for fever and 80 participants for all symptoms after dose 3. Participants included in panel B received their third dose 20–38 weeks after the second dose (median of 30 weeks [IQR 30–30]). www.thelancet.com Vol 398 September 11, 2021 Articles Antibody levels to SARS-CoV-2 Victoria/01/2020 spike measured by total IgG ELISA over 1 year after a single dose. Data are from 480 participants across COV001 and COV002 who received a standard dose of ChAdOx1 nCoV-19. Vaccine-induced T-cell responses against the SARS-CoV-2 spike insert were monitored up to day 182 in a cohort of 44 participants who received a single dose of ChAdOx1 nCoV-19. For participants who were excluded from these analyses due to positive PCR test result, second dose on trial, or external COVID-19 vaccination, no ELISA results or ELISpot results beyond the date of censoring were used. PBMCs=peripheral blood mononuclear cells. SFUs=spot-forming units A B 1 10 100 1000 10000 0 50 100 150 200 250 300 350 ELISA units Time since ﬁrst dose of vaccine (days) 10 100 1000 10000 0 50 100 150 200 SFU per million PBMCs from single-dose recipients up to 1 year after vaccination, in two-dose recipients up to 6 months after the second vaccination, and in three-dose recipients up to 28 days after the third vaccination. ELISA assays on the beta SARS-CoV-2 variant were only done on samples from participants recruited to the substudy who received either a late second vaccination or a third vaccination, up to 28 days after the late vaccination. Meso Scale Discovery multiplex immunoassay was used to assess antibody titres against spike proteins from different variants (Victoria/01/2020, D614G, alpha [B.1.17], beta, and gamma [P.1]). V-PLEX SARS-CoV-2 Panel 6 (IgG) kits were used following manufacturer’s instructions (Meso Scale Discovery, K15433U; full details are in appendix 3 [p 1]). Meso Scale Discovery assays were done on samples from participants recruited to the substudy who received a late second dose of vaccine (44–45 weeks after the first dose. Ex-vivo IFN-γ ELISpot assays were done as previously described5 to assess T-cell responses to Victoria/01/2020 SARS-CoV-2 spike overlapping peptide pools before and after vaccination. Isolated peripheral blood mononuclear cells (PBMCs) were stimulated overnight with peptides spanning the SARS-CoV-2 spike insert. ELISpot assays were done on samples from participants in the single dose cohort up to 182 days after vaccination. ELISpot assays were also done in some participants (due to laboratory capacity) recruited to the substudy who received a third dose, up to 28 days after third dose. Focus reduction neutralisation assays were done as described previously6 to measure neutralising antibody titres against alpha, beta, and delta (B.1.617.2) SARS-CoV-2 viral variants. See Online for appendix 3 Articles between cohorts, participants were excluded from the two-dose reactogenicity cohort if the interval between first and second dose was outside of the intervals defined in the two-dose immunogenicity cohort. All analyses of COV002 participants were restricted to those aged 18–55 years, to align with the inclusion criteria of the COV001 study, which only enrolled participants in this age range. Figure 4: Antibody response by interval between first and second vaccination Datapoints are medians for each group, with error bars showing IQRs. Antibody levels to SARS-CoV-2 Victoria/01/2020 spike measured by total IgG ELISA. Data are shown for 115 participants for the 8–12 week interval; 116 participants for the 15–25 week interval, and 30 participants for the 44–45 week interval. Unadjusted and age-adjusted geometric mean ratios are shown in appendix 3 (p 16). 1 10 100 1000 10 000 −300 −250 −200 −150 −100 −50 0 50 100 150 200 Total IgG ELISA units Time since second dose of vaccine (days) 8–12 weeks 15–25 weeks 44–45 weeks Vaccination interval 1 10 100 1000 10 000 −300 −250 −200 −150 −100 −50 0 50 100 150 200 Total IgG ELISA units Time since second dose of vaccine (days) 8–12 weeks 15–25 weeks 44–45 weeks Vaccination interval Total IgG ELISA units Our analysis of the decay of antibodies and T-cell responses over time after a single dose included all available data from timepoints up until the booster dose was administered (ie, day 28 and the day of the second dose, which varied across all participants). We modelled data using an unadjusted restricted-maximum likelihood-based mixed-effects regression approach (SAS proc mixed) with participant-level random inter cepts fitted to log-transformed antibody values. We used the variance components covariance structure. We estimated GMRs and GMTs from the linear combination of model parameters. We chose the linear models after comparison with quadratic models and generalised additive (smoothed) models (GAM). The quadratic term was non-significant in the linear models and the GAM results were similar and did not substantially improve the model fits (compared using Akaike information criterion [known as AIC] statistics) from the linear models; therefore, the linear models were retained. Figure 4: Antibody response by interval between first and second vaccination Datapoints are medians for each group, with error bars showing IQRs. Antibody levels to SARS-CoV-2 Victoria/01/2020 spike measured by total IgG ELISA. Role of the funding source AstraZeneca reviewed the manuscript before submission, but the academic authors retained editorial control. All other funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. We used the Wilcoxon rank sum and Kruskal-Wallis tests for comparisons between independent groups and we used the Wilcoxon sign rank test to compare paired data. Geometric mean ratios (GMRs) with 95% CIs were produced when comparing groups. When appropriate, adjusted GMTs and GMRs were also presented to adjust for the effect of age. We did not do a sample size calculation for the immunogenicity subgroups in this analysis because of logistical considerations, including laboratory capacity. Sample size calculations for COV001 and COV002 were based on the primary efficacy outcome, which have been previously reported. Statistical analysis We present summary statistics for individuals vaccinated with one, two, or three doses of ChAdOx1 nCoV-19 as median (IQR) or geometric mean titre (GMT) with 95% CIs. We do not include data from control participants (who had previously received one or two doses of MenACWY); they received ChAdOx1 nCoV-19 on recruitment to the substudy. For the purposes of ensuring trial personnel were masked to treatment assignment, data for both vaccinnees and controls were collected. Unmasking information was only available to those performing the final data analyses. Upon unmasking of participants and study personnel, control participants were excluded. We did all statistical analyses using R (version 4.0.2 or later) and SAS (version 9.2). p values of less than 0·05 were considered to be significant and we made no adjustments for multiple comparisons. COV001 is registered with ClinicalTrials.gov, NCT04324606, and ISRCTN, 15281137, and COV002 is registered with ClinicalTrials.gov, NCT04400838, and ISRCTN, 15281137, and both are no longer recruiting. Articles Data are shown for 115 participants for the 8–12 week interval; 116 participants for the 15–25 week interval, and 30 participants for the 44–45 week interval. Unadjusted and age-adjusted geometric mean ratios are shown in appendix 3 (p 16). e 4: Antibody response by interval between first and second vaccination www.thelancet.com Vol 398 September 11, 2021 Articles Participants reported the severity of their adverse reactions as mild, moderate, severe, or life threaten ing as per definitions provided (appendix 1 pp 92–94). 985 www.thelancet.com Vol 398 September 11, 2021 Articles Articles Results pp p ) Baseline characteristics for the one-dose, two-dose, and three-dose cohorts are shown in appendix 3 (pp 4–5). More than 90% of participants were White. There were small differences in the median age of reactogenicity cohorts and immunogenicity cohorts. The median age of participants in the two-dose cohort antibody immunogenicity cohort with an 8–12 week interval between the first and second dose was 39 years (IQR 30–49), in the 15–25 weeks interval group was 36 years (30–43), and in the 44–45 weeks interval group was 32 years (25–44). In the three-dose cohort, the median age of participants in the reactogenicity cohort was 37 years (IQR 31–42), in the immunogenicity antibody cohort was 37 years (31–42), and in the immunogenicity T-cell response cohort 40 years (32–44). h f l l d l d d Baseline characteristics for the one-dose, two-dose, and three-dose cohorts are shown in appendix 3 (pp 4–5). More than 90% of participants were White. There were small differences in the median age of reactogenicity cohorts and immunogenicity cohorts. The median age of participants in the two-dose cohort antibody immunogenicity cohort with an 8–12 week interval between the first and second dose was 39 years (IQR 30–49), in the 15–25 weeks interval group was 36 years (30–43), and in the 44–45 weeks interval group was 32 years (25–44). In the three-dose cohort, the median age of participants in the reactogenicity cohort was 37 years (IQR 31–42), in the immunogenicity antibody cohort was 37 years (31–42), and in the immunogenicity T-cell response cohort 40 years (32–44). The severity of local and systemic solicited adverse reactions 7 days after a second dose were mostly mild to moderate irrespective of the interval between doses. Local symptoms occurred after a second dose in 201 (75%) of 267 participants in the 8–12 week interval group, 15 (63%) of 24 participants in the 15–25 week interval group, and 23 (77%) of 30 participants in the 44–45 wee interval Figure 5: Antibody responses in participants who received a third dose of ChAdOx1 nCoV-19 (A) Antibody levels to SARS-CoV-2 Victoria/01/2020 spike protein measured by total IgG ELISA (n=75). Datapoints in lighter colours represent individual participants and darker datapoints show median values with error bars showing the IQRs and with solid lines connecting these median values. Results Between March 11 and 21, 2021, 90 participants were enrolled into the COV001 third-dose boost substudy and vaccinated with ChAdOx1 nCoV-19. Ten participants from this substudy were excluded from analyses because they were enrolled from an open-label subgroup and the interval between their first and second doses was shorter than 28 days. The remaining 80 participants who were assessable for reactogenicity received their third dose of vaccine 20–38 weeks after their second dose (median 30 weeks [IQR 30–30]). A further five participants were excluded from all immunogenicity assessments because their prime-boost interval was outside the defined range of 8–16 weeks. For The reactogenicity cohorts included masked participants who received at least two standard doses of ChAdOx1 nCoV-19 in the two-dose cohort or three standard doses of ChAdOx1 nCoV-19 in the three-dose cohort, and had completed at least one entry in their adverse event diary after each dose. For consistency 986 www.thelancet.com Vol 398 September 11, 2021 Articles Figure 5: Antibody responses in participants who received a third dose of ChAdOx1 nCoV-19 (A) Antibody levels to SARS-CoV-2 Victoria/01/2020 spike protein measured by total IgG ELISA (n=75). Datapoints in lighter colours represent individual participants and darker datapoints show median values with error bars showing the IQRs and with solid lines connecting these median values. (B) Neutralisation titres from a randomly selected subset of participants (45 of 75 participants who received a third dose of vaccine and who had an interval of 8–16 weeks between their first and second dose). Datapoints represent individual participants for the three variants of concern investigated. FRNT50=focus reduction neutralisation titres with 50% neutralisation cutoff. 1 10 100 1000 10000 First dose 28 days after ﬁrst dose Second dose 28 days after second dose Third dose 14 days after third dose 28 days after third dose Timepoint Total IgG ELISA units 28 days after second dose 28 days after third dose 100 101 102 103 28 days after second dose 28 days after third dose 28 days after second dose 28 days after third dose 104 FRNT50 (reciprocal serum dilution) B A Timepoint Timepoint Timepoint Alpha (B.1.1.7) variant Beta (B.1.351) variant Delta (B.1.617.2) variant immunogenicity analysis of T-cell responses, another 60 participants were excluded because they did not have or had insufficient ELISpot data, leaving an analysable cohort of 15 participants (figure 1). www.thelancet.com Vol 398 September 11, 2021 Results To maintain blinding at the time of vaccination, 40 control participants (who had previously received two doses of MenACWY) were also recruited and vaccinated with ChAdOx1 nCoV-19. Data for these participants have not been included in these analyses. 1 10 100 1000 10000 First dose 28 days after ﬁrst dose Second dose 28 days after second dose Third dose 14 days after third dose 28 days after third dose Timepoint Total IgG ELISA units A Of 1110 participants from the COV001 and COV002 studies who had received a single dose of vaccine, 480 were included in the single-dose immunogenicity assessment of antibodies and 44 were included in the immunogenicity assessment of T-cell responses (appendix 3 p 2). 66 participants from the single-dose cohort of COV001 were also offered a second dose, of whom 44 were ChAdOx1 nCoV-19 recipients and 22 were control group participants who had received MenACWY. 32 (73%) of 44 participants who had previously received a single dose of ChAdOx1 nCoV-19 received their second dose, with an interval of 44–45 weeks. Two of these participants were subsequently excluded from reacto genicity and immunogenicity analyses because they had positive PCR tests for SARS-CoV-2 infection within the follow-up period, leaving 30 participants for inclusion in the two-dose cohort analyses. 28 days after second dose 28 days after third dose 100 101 102 103 28 days after second dose 28 days after third dose 28 days after second dose 28 days after third dose 104 FRNT50 (reciprocal serum dilution) B Timepoint Timepoint Timepoint Alpha (B.1.1.7) variant Beta (B.1.351) variant Delta (B.1.617.2) variant B 28 days after second dose 28 days after third dose 100 101 102 103 104 FRNT50 (reciprocal serum dilution) B Timepoint Alpha (B.1.1.7) variant 28 days after second dose 28 days after third dose Timepoint Beta (B.1.351) variant 28 days after second dose 28 days after third dose Timepoint Delta (B.1.617.2) variant Alpha (B.1.1.7) variant FRNT50 (reciprocal serum dilution) y The two-dose cohort comprised 321 participants from COV001 and COV002 with prime-boost intervals of 8–12 weeks (267 [83%] of 321), 15–25 weeks (24 [7%]), or 44–45 weeks (30 [9%]) who had reactogenicity data available and were included in our analyses, and 261 who had immunogenicity data available (115 [44%] of 261 had an 8–12 week interval, 116 [44%] had a 15–25 week interval, and 30 [11%] had a 44–45 week interval; appendix 3 p 3). Articles Figure 6: IFN-γ ELISpot responses in participants who received a third dose of ChAdOx1 nCoV-19 15 participants with an interval of 8 weeks between their first and second doses were assessed for ELISpot responses. These participants received their third dose 37–38 weeks after the second dose (median 38 weeks [IQR 38–38]). Datapoints in lighter colours represent individual participants and darker datapoints show median values with error bars showing the IQRs and with solid lines connecting these median values. The dotted horizontal line represents the lower limit of detection of the assay (48 SFU per million PBMCs). SFU=spot-forming unit. PBMC=peripheral blood mononuclear cells. 100 1000 10000 14 days after second dose 28 days after second dose Third dose 14 days after third dose 28 days after third dose Timepoint SFU per million PBMCs 6 months after the second dose of vaccine, antibody levels remained significantly higher in the group with a 15–25 week interval between doses compared with an 8–12 week interval (median 1240 EUs [IQR 432–2002] vs 278 EUs [166–499]; Wilcoxon rank sum test with con tinuity correction p<0·0001; figure 4; appendix 3 p 16). 100 1000 10000 14 days after second dose 28 days after second dose Third dose 14 days after third dose 28 days after third dose Timepoint SFU per million PBMCs y pi g pp p ) IgG binding titres to all four variants tested (D614G, alpha, beta, and gamma) were significantly greater after second dose than after the first dose (p<0·0001 for all comparisons [pairwise comparisons using Wilcoxon sign rank test]; appendix 3 pp 17, 19). Antibody responses after a third dose of vaccine were assessed in 75 participants who had received their first two doses with an interval of 8–16 weeks, and who subsequently received their third dose 28–38 weeks after the second (median 30 weeks [IQR 30–30]). Administering a third dose of vaccine boosted antibody response to Victoria/01/2020 SARS-CoV-2 spike protein (figure 5A; appendix 3 p 18). Antibody levels after the third dose were significantly higher than after the second dose (median total IgG titre was 1792 EUs [IQR 899–4634] at 28 days after the second dose vs 3746 EUs [2047–6420] 28 days after the third dose; pairwise comparison in 73 participants due to two samples not being available at these timepoints using Wilcoxon signed rank test p=0·0043). Articles Binding antibody titres to the beta variant increased after a third dose (appendix 3 p 19). Neutralising antibody titres after a third dose were higher than those after the second dose against alpha (p=0·0023), beta (p<0·0001), and delta (p<0·0001) variants (Wilcoxon signed rank test; figure 5B; appendix 3 p 20). Figure 6: IFN-γ ELISpot responses in participants who received a third dose of ChAdOx1 nCoV-19 15 participants with an interval of 8 weeks between their first and second doses were assessed for ELISpot responses. These participants received their third dose 37–38 weeks after the second dose (median 38 weeks [IQR 38–38]). Datapoints in lighter colours represent individual participants and darker datapoints show median values with error bars showing the IQRs and with solid lines connecting these median values. The dotted horizontal line represents the lower limit of detection of the assay (48 SFU per million PBMCs). SFU=spot-forming unit. PBMC=peripheral blood mononuclear cells. vaccination compared with 21 (7%) of 321 participants after second vaccinations (appendix 3 pp 6–9). Third dose vaccinations were also less reactogenic than first doses, with four (5%) of 80 participants in the three-dose cohort reporting more than two moderate-to-severe systemic symptoms after a third dose compared with 27 (34%) of 80 participants after the first dose (appendix 3 pp 10–14). Spike-specific cellular immune responses were measured after a third dose of ChAdOx1 nCoV-19 in 15 individuals. These individuals had received their first two doses with an interval of 8 weeks, and subsequently received their third dose 37–38 weeks after the second (median 38 weeks [IQR 38–38]). Median response increased from 200 spot-forming units (SFUs) per million PBMCs (IQR 127–389) immediately before the third dose to 264 SFUs per million PBMCs (131–452) 14 days after the third dose (p=0·57), and to 399 SFUs per million PBMCs (314–662) by 28 days after the third dose (p=0·012; figure 6; appendix 3 p 20). Peak responses at day 28 after the third dose were not significantly different to the responses after the second dose (p=0·060; appendix 3 p 20). Antibody responses after a single dose of vaccine and measured approximately 320 days after vaccination remained higher than responses measured at baseline (GMTs of 66·00 ELISA units [EUs; 95% CI 47·83–91·08 vs 1·75 EUs [1·60–1·93]). Articles At day 180, geometric mean antibody levels were half the levels observed at the day 28 peak (GMR 0·51 [95% CI 0·45–0·57]), and by day 320 were less than a third of the levels at the peak (GMR 0·27 [0·22–0·34; figure 3A; appendix 3 p 15). Vaccine-induced cellular immune responses after a single dose of ChAdOx1 nCoV-19 followed a similar pattern of decay as antibody responses. T-cell responses decreased over the course of 6 months but were maintained above baseline levels, and at day 180 geometric mean T-cell levels were half the levels observed at the day 28 peak (GMR 0·50 [95% CI 0·41–0·60]; figure 3B; appendix 3 p 15). www.thelancet.com Vol 398 September 11, 2021 Results (B) Neutralisation titres from a randomly selected subset of participants (45 of 75 participants who received a third dose of vaccine and who had an interval of 8–16 weeks between their first and second dose). Datapoints represent individual participants for the three variants of concern investigated. FRNT50=focus reduction neutralisation titres with 50% neutralisation cutoff. group (figure 2A; appendix 3 pp 6–9). Systemic reactions occurred in 190 (71%) of 267 participants in the 8–12 week interval group, 18 (75%) of 24 participants in the 15–25 week interval group, and 26 (87%) of 30 participants in the 44–45 week interval group (figure 2A; appendix 3 pp 6–9). 65 (81%) of 80 participants in the three-dose group reported at least one local symptom after a third dose (figure 2B; appendix 3 pp 10–14). Second dose vaccinations in the two-dose cohort were less reactogenic than first dose vaccinations; with 72 (22%) of 321 participants reporting more than two moderate-to-severe systemic symptoms after first 987 Articles Articles Discussion Antibody levels induced by a single dose of ChAdOx1 nCoV-19 decreased gradually but remained above baseline levels after 1 year. We have previously shown that administration of a second dose of vaccine induces higher antibody responses by 1 month after the second dose than before the second dose, with higher responses with a dose interval up to 3 months between the first two doses.3 Here, we found that a long extension of the dose interval (up to 45 weeks) between the first and second dose further Antibody levels 28 days after a second dose of vaccine were higher among those with longer intervals between doses than among those with shorter intervals between doses (median total IgG titre of 923 EUs [IQR 525–1764] for 8–12 week interval; 1860 EUs [917–4934] with 15–25 week interval; and 3738 EUs [1824–6625] with 44–45 week interval; Kruskal-Wallis test p<0·0001; figure 4; appendix 3 p 16). Age was not statistically significant in adjusted models (appendix 3 p 16). 988 Articles enhances the immune response to the second dose when compared with shorter dose intervals. Furthermore, for the first time, we showed that a third dose of ChAdOx1 nCoV-19 can induce a strong boost to immune responses to the transgene product, SARS-CoV-2 spike protein, and that these responses result in increased neutralising antibody titres and enhanced antibody activity against variants. shown to dampen vaccine-induced immune responses,12,13 here we found no evidence that repeated use of a replication-deficient simian adenoviral vector induces antivector immunity at a level sufficient to impair responses to further vaccination. The third dose was well tolerated by participants with lower reactogenicity than after the first dose. i Our study has several limitations, including a paucity of T-cell data after a late second dose, a paucity of tolerability data after the second dose for those who were recruited to receive a third dose, and the small number of participants who were available at 1 year after single dose who still had only received one dose (mainly due to being offered a second dose after unblinding, as per protocol). Thus far, data are only available 28 days after the third dose; however, follow-up at 6 months is planned. Participants were aged 18–55 years and caution should be taken when extrapolating our findings to beyond this age range. Discussion The sample size in this study is not sufficiently large to assess rare vaccine side-effects, but there were no tolerability concerns reported in those receiving a late second dose or a third dose booster. These results are from a mainly White population and cannot necessarily be generalised to other populations. The devastating impact of COVID-19 is most apparent in countries with low vaccine coverage and little health- care infrastructure, including low-income and middle- income countries. Global vaccine shortages and policy decisions implemented at national levels have curtailed vaccine supplies for some countries where substantial numbers of individuals have already received one dose of vaccine. We have previously shown that protection against symptomatic COVID-19 is maintained after a single dose of ChAdOx1 nCoV-19 for at least 3 months, despite some waning of antibody levels3 and we now report that the antibody levels remain above baseline for at least 1 year after single dose immunisation. These data are important for those countries where administration of a second dose is delayed because of a shortage of supply. We also showed here that an extended interval between the first and second dose of ChAdOx1 nCoV-19 results in a significantly higher antibody response 28 days after the second dose than with shorter intervals. This finding is consistent with previous data showing a longer interval between first and second dose of ChAdOx1 nCoV-19 resulted in an increase in antibody titres,3 thus providing further reassurance that a delay in administration of the second dose will not compromise the level of protection attained. Similar findings have been reported with other vaccines;1,7,8 a delayed two-dose regimen against HPV, given at least 6 months apart, results in as good or better antibody response than does three doses. A second dose of ChAdOx1 nCoV-19 is well tolerated in a delayed two-dose schedule and a third dose is also well tolerated. Reports have emerged of thrombosis and thrombocytopenia after the first dose of ChAdOx1 nCoV-19,9 and information from Public Health England indicates that this very rare event might not occur after a second dose.10 Here, we found that immunity induced by the viral vectored vaccine ChAdOx1 nCoV-19 is maintained for long periods after a first dose, with greater boosting of effects after the second dose after a longer interval between doses than shorter intervals. Contributors Contributors AJP, TL, MV, and DJ contributed to the writing and design of the protocol and design of the study. AJP is chief investigator. AF, JA, PKA, BA, SB-R, SB, MB, FC, PC, EAC, SD, WD, CD, KJE, PMF, JF, DJ, AMM, JM, YFM, EP, MNR, HR, HSa, ES, HSm, MDS, RS, and DW contributed to implementation of the study or laboratory experimentation, or both. AF, MV, and NGM accessed and verified the underlying study data and did the statistical analysis. TL, AF, AJP, SCG, FC, NGM, and MV contributed to the preparation of the report. TL, AF, AJP, SCG, and MV were responsible for the decision to submit the manuscript. All authors critically reviewed and approved the final version. All authors had full access to the data in the study and had final responsibility for the decision to submit for publication. g If booster vaccinations against SARS-CoV-2 will be required, perhaps to counter waning immunity or to augment protection against emerging variants, is not yet known. Here, we show that a third dose of ChAdOx1 nCoV-19 is well tolerated and significantly boosts antibody titres above those measured after the second dose to the level associated with 80% vaccine efficacy, or higher, after two vaccine doses (unpublished; preprint data available11). Higher titre neutralising antibodies against alpha, beta, and delta variants of SARS-CoV-2 were induced 28 days after a third dose vaccination than after the second dose. Spike-specific T-cell responses were boosted after a third dose of ChAdOx1 nCoV-19 and were similar in magnitude to the responses measured after two doses. Although pre- existing immunity to human adenoviral vectors has been Discussion Therefore, a single dose of ChAdOx1 nCoV-19 with a second dose given after an extended period might be an effective strategy3,14 in settings where vaccine supplies are scarce in the short term. A third dose resulted in a further increase in immune responses, including increased neutralisation of variant SARS-CoV-2 viruses, and could be used to increase vaccine efficacy against variants in susceptible populations. www.thelancet.com Vol 398 September 11, 2021 Contributors AJP, TL, MV, and DJ contributed to the writing and design of the protocol and design of the study. AJP is chief investigator. AF, JA, PKA, BA, SB-R, SB, MB, FC, PC, EAC, SD, WD, CD, KJE, PMF, JF, DJ, AMM, JM, YFM, EP, MNR, HR, HSa, ES, HSm, MDS, RS, and DW contributed to implementation of the study or laboratory experimentation, or both. AF, MV, and NGM accessed and verified the underlying study data and did the statistical analysis. TL, AF, AJP, SCG, FC, NGM, and MV contributed to the preparation of the report. TL, AF, AJP, SCG, and MV were responsible for the decision to submit the manuscript. All authors critically reviewed and approved the final version. All authors had full access to the data in the study and had final responsibility for the decision to submit for publication. References 1 WHO. Interim recommendations for use of the AZD1222 (ChAdOx1-S [recombinant]) vaccine against COVID19 developed by Oxford University and AstraZeneca. Geneva: World Health Organization, Feb 10, 2021. https://apps.who.int/iris/bitstream/ handle/10665/339477/WHO-2019-nCoV-vaccines-SAGE- recommendation-AZD1222-2021.1-eng.pdf?sequence=5&isAllowed=y (accessed June 23, 2021). Acknowledgments 5 Ramasamy MN, Minassian AM, Ewer KJ, et al. Safety and immunogenicity of ChAdOx1 nCoV-19 vaccine administered in a prime-boost regimen in young and old adults (COV002): a single- blind, randomised, controlled, phase 2/3 trial. Lancet 2021; 396: 1979–93. Acknowledgments Oxford University has entered into a partnership with AstraZeneca for further development of ChAdOx1 nCoV-19. We acknowledge funding for this study from the UK Research and Innovation (MC_PC_19055); Engineering and Physical Sciences Research Council (EP/R013756/1); National Institute for Health Research (COV19 OxfordVacc-01); Coalition for Epidemic Preparedness Innovations (Outbreak Response To Novel Coronavirus [COVID-19]); NIHR Oxford Biomedical Research Centre (BRC4 Vaccines Theme); The Chinese Academy of Medical Sciences (CAMS) Innovation Fund for Medical Science (CIFMS), China (2018-I2M-2-002; to AF and TL); Thames Valley and South Midlands NIHR Clinical Research Network; AstraZeneca; and the Wellcome Trust (220991/Z/20/Z). The views expressed in this publication are those of the authors and not necessarily those of the NIHR or the UK Department of Health and Social Care. Development of SARS-CoV-2 reagents was partially supported by the National Institute of Allergy and Infectious Diseases Centers of Excellence for Influenza Research and Surveillance (CEIRS) contract HHSN272201400008C. We thank all participants in the trial and the international Data and Safety Monitoring Board (DSMB) and the independent members of the Trial Steering Committee for their invaluable advice. We also thank the various teams within the University of Oxford including Medical Sciences Division, Nuffield Department of Medicine and Department of Paediatrics, the Oxford Immunology Network COVID Consortium, Clinical Trials Research Governance, Research Contracts, Public Affairs Directorate, and the Clinical Biomanufacturing Facility, as well as the Oxford University Hospitals NHS Foundation Trust, and Oxford Health NHS Foundation Trust and the trial sites. Named individuals we acknowledge are listed in appendix 3 (p 25). We are grateful for the input of the Protein Production Team at the Jenner Institute and the team at the Pirbright Institute. We acknowledge the philanthropic support of the donors to the University of Oxford’s COVID-19 Research Response Fund. 6 Zhou D, Dejnirattisai W, Supasa P, et al. Evidence of escape of SARS-CoV-2 variant B.1.351 from natural and vaccine-induced sera. Cell 2021; 184: 2348–61. 7 Baden LR, Karita E, Mutua G, et al. Assessment of the safety and immunogenicity of 2 novel vaccine platforms for HIV-1 prevention: a randomized trial. Ann Intern Med 2016; 164: 313–22. 8 Dobson SR, McNeil S, Dionne M, et al. Data sharing The current study protocol for COV001 is provided in appendix 1 and for COV002 in appendix 2. Anonymised participant data will be made available when the trial is complete, upon requests directed to the corresponding author. Proposals will be reviewed and approved by the sponsor, investigator, and collaborators on the basis of scientific merit. After approval of a proposal, data can be shared through a secure online platform after signing a data access agreement. All data will be made available for a minimum of 5 years from the end of the trial. 3 Voysey M, Costa Clemens SA, Madhi SA, et al. Single-dose administration and the influence of the timing of the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a pooled analysis of four randomised trials. Lancet 2021; 397: 881–91. 4 Folegatti PM, Ewer KJ, Aley PK, et al. Safety and immunogenicity of the ChAdOx1 nCoV-19 vaccine against SARS-CoV-2: a preliminary report of a phase 1/2, single-blind, randomised controlled trial. Lancet 2020; 396: 467–78. Articles 2 Advisory Group of Experts (SAGE). AJP is an NIHR Senior Investigator. All other authors declare no competing interests. Advisory Group of Experts (SAGE). AJP is an NIHR Senior Investigator. All other authors declare no competing interests. 2 Voysey M, Clemens SAC, Madhi SA, et al. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK. Lancet 2021; 397: 99–111. Acknowledgments Immunogenicity of 2 doses of HPV vaccine in younger adolescents vs 3 doses in young women: a randomized clinical trial. JAMA 2013; 309: 1793–802. 9 Schultz NH, Sørvoll IH, Michelsen AE, et al. Thrombosis and thrombocytopenia after ChAdOx1 nCoV-19 vaccination. N Engl J Med 2021; 384: 2124–30. 10 Public Health England. Vaccine update: issue 322, June 2021, COVID-19 phase 2 special edition. Rare syndrome related to AstraZeneca vaccine. UK Government, June 17, 2021. https://www. gov.uk/government/publications/vaccine-update-issue-322-june- 2021-covid-19-phase-2-special-edition/vaccine-update-issue-322- june-2021-covid-19-phase-2-special-edition (accessed June 24, 2021). 11 Feng S, Phillips DJ, White T, et al Correlates of protection against symptomatic and asymptomatic SARS-CoV-2 infection. medRxiv 2021; published online June 24. https://doi.org/10.1101/ 2021.06.21.21258528 (preprint). 12 Zhu FC, Guan XH, Li YH, et al. Immunogenicity and safety of a recombinant adenovirus type-5-vectored COVID-19 vaccine in healthy adults aged 18 years or older: a randomised, double-blind, placebo-controlled, phase 2 trial. Lancet 2020; 396: 479–88. 13 Roberts DM, Nanda A, Havenga MJ, et al. Hexon-chimaeric adenovirus serotype 5 vectors circumvent pre-existing anti-vector immunity. Nature 2006; 441: 239–43. 14 14 Public Health England. SARS-CoV_2 variants of concern and variants under investigation in England: technical briefing 16. UK Government, June 18, 2021. https://assets.publishing.service. gov.uk/government/uploads/system/uploads/attachment_data/ file/994839/Variants_of_Concern_VOC_Technical_Briefing_16.pdf (accessed June 23, 2021). www.thelancet.com Vol 398 September 11, 2021 Declaration of interests SCG and AVSH are cofounders of and shareholders in Vaccitech (collaborators in the early development of ChAdOx1 nCoV-19) and named as inventors on a patent covering use of ChAdOx1-vectored vaccines (PCT/GB2012/000467) and a patent application covering this SARS-CoV-2 vaccine (SCG only). TL is named as an inventor on a patent covering use of ChAdOx1-vectored vaccines (PCT/GB2012/000467) and was a consultant to Vaccitech. PMF is a consultant to Vaccitech. AJP is chair of the UK Department of Health and Social Care’s Joint Committee on Vaccination and Immunisation, but does not participate in policy advice on coronavirus vaccines, and is a member of the WHO Strategic 989 Articles Articles References 990 www.thelancet.com Vol 398 September 11, 2021
https://openalex.org/W3159003573	https://bmcgenomics.biomedcentral.com/counter/pdf/10.1186/s12864-021-07563-9	English	null	Strategies for detecting and identifying biological signals amidst the variation commonly found in RNA sequencing data	BMC genomics	2,021	cc-by	15,893	Wilfinger et al. BMC Genomics (2021) 22:322 https://doi.org/10.1186/s12864-021-07563-9 Wilfinger et al. BMC Genomics (2021) 22:322 https://doi.org/10.1186/s12864-021-07563-9 Open Access © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. * Correspondence: billw@mrcgene.com 1Molecular Research Center, Inc., Cincinnati, USA Full list of author information is available at the end of the article * Correspondence: billw@mrcgene.com 1Molecular Research Center, Inc., Cincinnati, USA Full list of author information is available at the end of the article Correspondence: billw@mrcgene.com 1Molecular Research Center, Inc., Cincinnati, USA Full list of author information is available at the en * Correspondence: billw@mrcgene.com 1Molecular Research Center, Inc., Cincinnati, USA Full list of author information is available at the end of the article Abstract Background: RNA sequencing analysis focus on the detection of differential gene expression changes that meet a two-fold minimum change between groups. The variability present in RNA sequencing data may obscure the detection of valuable information when specific genes within certain samples display large expression variability. This paper develops methods that apply variance and dispersion estimates to intra-group data to identify genes with expression values that diverge from the group envelope. STRING database analysis of the identified genes characterize gene affiliations involved in physiological regulatory networks that contribute to biological variability. Individuals with divergent gene groupings within network pathways can thereby be identified and judiciously evaluated prior to standard differential analysis. Results: A three-step process is presented for evaluating biological variability within a group in RNA sequencing data in which gene counts were: (1) scaled to minimize heteroscedasticity; (2) rank-ordered to detect potentially divergent “trendlines” for every gene in the data set; and (3) tested with the STRING database to identify statistically significant pathway associations among the genes displaying marked trendline variability and dispersion. This approach was used to identify the “trendline” profile of every gene in three test data sets. Control data from an in- house data set and two archived samples revealed that 65–70% of the sequenced genes displayed trendlines with minimal variation and dispersion across the sample group after rank-ordering the samples; this is referred to as a linear trendline. Smaller subsets of genes within the three data sets displayed markedly skewed trendlines, wide dispersion and variability. STRING database analysis of these genes identified interferon-mediated response networks in 11–20% of the individuals sampled at the time of blood collection. For example, in the three control data sets, 14 to 26 genes in the defense response to virus pathway were identified in 7 individuals at false discovery rates ≤1.92 E-15. Conclusions: This analysis provides a rationale for identifying and characterizing notable gene expression variability within a study group. The identification of highly variable genes and their network associations within specific individuals empowers more judicious inspection of the sample group prior to differential gene expression analysis. Keywords: Scaling, Rank-order, Trendline, Biological variability, Biological pathway analysis, RNA sequencing, STRI NG-db, Minimum value adjustment, White blood cells Background sequencing studies, we applied an approach similar to scaling normalization to evaluate RNA sequencing results. sequencing studies, we applied an approach similar to scaling normalization to evaluate RNA sequencing results. Blood from 35 healthy adults was extracted and proc- essed for RNA sequencing [10, 11]. The read counts were scaled to establish a uniform starting point across all genes and rank-ordered to characterize gene expression in the sample group as a “trendline” pattern for each gene. Excel-based tools were employed to analyze and catalogue the resulting gene trendlines [12]. Utilizing trendline ana- lysis, we determined that 65–70% of the genes in our con- trol data set follow a linear relationship with minimal variance when the genes were scaled and rank-ordered. However, other genes that did not follow this linear profile displayed markedly higher levels of dispersion and vari- ability that diverged significantly from the genes in a nor- mally distributed control sample. We identified standard statistical measures that characterize and catalogue these different trendlines and utilized this information to iden- tify factors that may contribute to this heightened bio- logical variability. When genes displaying the most variable and dispersed trendline expression patterns were evaluated with the STRING database [13–15], distinct bio- logical regulatory pathways were identified in some indi- viduals, thereby providing an explanation for some of the variability in the sample group. g A major goal of RNA-seq studies is to improve and extend our understanding of gene expression responses amidst the challenging variability commonly found in sequencing data. Although numerous factors are known to affect sequencing results such as the reference genome, the read processing pipeline, internal references, read fragment size, and the se- lected data analysis algorithms, among others [1], thus far it has been difficult to discern how these sequencing proce- dures combined with intrinsic biological variability might impact differential analysis. For example, many software packages commonly employ different normalization proce- dures that are designed to mitigate read count variability; however, these strategies are known to yield dissimilar dif- ferential expression analysis results [2–6]. Biological vari- ation is considered to be larger than technical variation [3, 6–8], but the biological implications associated with read count normalization are not well-understood. Previous studies have suggested that increasing the sequencing depth (read coverage) and/or the number of biological replicates generally improves estimates of biological variation [6–8]. Background g y p g [ ] Conclusions relating to biological variation are usually based on Analysis of Variance (ANOVA) Sums of Squares estimations. Although increasing the level of replication may increase the Between Sums of Square difference and provide a more definitive statistical conclusion about an identified biological response (e.g. larger F-value), an increase in the Sums of Squares does not identify the factor(s) contributing to the variability. More broadly untangling the impact of variability on each step of the RNA-seq pipeline is difficult. One must identify specific sources of biological variability in the data set and consider how the normalization process impacts the overall results. This problem becomes increasingly difficult to resolve in samples in which cell number and cell type fluctuate significantly. Identifying and quantifying significant variability within RNA se- quencing data sets would provide information that would be very useful for evaluating the robustness of computational steps, for example, devising and evaluat- ing methodologies for determining how normalization protocols impact technical and biological variation. We also demonstrate that the scaling normalization strategy employed in our study reduced gene expression heteroscedasticity within three different control data sets as previously demonstrated by van den Berg et al. [9]. Scaling adjustments in conjunction with rank-order ana- lysis clarify and extend the analysis of inter-individual variations relating to differential gene expression previ- ously described by Whitney et al. [16], Savelyeva et al. [17], Preininger et al. [18] and Jaffe et al. [19] to within- the-group analysis. STRING-db analysis of genes displaying the most variable and dispersed trendlines re- vealed that 11–20% of the individuals in our control sample and two archived control data sets, identified a prominent network of interferon-stimulated genes. The interferon-induced genes identified in this analysis play a pivotal regulatory role in three Gene Ontology pathways [20–22] that include response to virus, defense response to virus and the type I-interferon signaling/regulatory response pathways. The evaluation of gene trendline re- sponses within a group and across individuals identifies sources of previously unrecognized biological variability that now can be detected and appraised. This method of analysis can be applied to archived RNA sequencing data to detect previously unrecognized sources of biological variability that may have impacted differential analysis and physiological conclusions. © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 Page 2 of 19 Page 2 of 19 Results Rank-ordering RNA sequencing counts graphically portrays the impact of sample dispersion on gene trendline profiles Rank-ordering RNA sequencing counts graphically portrays the impact of sample dispersion on gene trendline profiles DeSeq-normalized TPM (Transcripts Per kilobase Million) gene counts for 35 individuals were processed through our q ( p ) gene counts for 35 individuals were processed through our g ( ) and plotting the counts for the 35 samples in ascending Fig. 1 Rank-ordering RNA sequencing counts identifies individuals displaying gene count divergence. a Box plots of sequencing counts for five genes INTS6, AKAP13, KCNJ2, IFIT3 and EIF1AY depicting increasing levels of sample dispersion with computed coefficient of variation values ranging from 17.9 to 171.2% of the unadjusted TPM gene counts (Mean ± 1SD). Box boundaries exclude individuals in the first and fourth quartile for each gene. b Rank-ordering the unadjusted counts of 35 individuals delineates different gene trendline patterns for the five genes. Gene rank- order position is established in relation to the gene expression level for an individual gene within the sample group, therefore the ranking order does not identify the same individual at each position along the various gene trendlines since the relative level of gene expression for an individual changes across genes. c Minimum Value Adjusted (MVA) gene counts significantly improve count heteroscedasticity (5-fold scale reduction) without altering the incremental trendline profiles within the sample group. Rank-order analysis extends the descriptive sample information available from a box plot by: defining the number of data points within the sample that deviate from the count level in the 2nd and 3rd quartiles; identifying their inflection point(s) and providing an estimate of the relative change in gene expression based on the computed slope ratio change. Black vertical lines identify quartiles 1, 2–3 and 4. See Additional file 1 for a more detailed discussion Fig. 1 Rank-ordering RNA sequencing counts identifies individuals displaying gene count divergence. a Box plots of sequencing counts for five genes INTS6, AKAP13, KCNJ2, IFIT3 and EIF1AY depicting increasing levels of sample dispersion with computed coefficient of variation values ranging from 17.9 to 171.2% of the unadjusted TPM gene counts (Mean ± 1SD). Box boundaries exclude individuals in the first and fourth quartile for each gene. b Rank-ordering the unadjusted counts of 35 individuals delineates different gene trendline patterns for the five genes. Background The methods outlined in this report will be useful in identifying within group variability commonly found in RNA sequencing data sets and when employed in conjunction with established Van den Berg et al. [9] have employed various scaling strategies to their metabolomics data and examined their usefulness in categorizing the relative importance of vari- ous metabolites identified in these studies. They deter- mined that scaling normalizations performed better than other strategies because they removed the dependence of the metabolites initial ranking based on the magnitude of a quantitative response. The scaled metabolites were eval- uated in relation to their sample-to-sample response range which also reduced the heteroscedasticity (mean and vari- ance dispersion) within the data set. Since these data sets were qualitatively similar to the data obtained in RNA Page 3 of 19 Page 3 of 19 Page 3 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 data processing pipelines, they are likely to improve the robustness of these studies. data processing pipelines, they are likely to improve the robustness of these studies. pipeline [23] and the count data were rank-ordered to con- struct a unique trendline for each gene. Figure 1a depicts a box plot of data for five example genes displaying increas- ing variance where the box boundaries identify gene counts in the 2ed and 3rd quartiles (25th–75th percentile). The breadth of the box illustrates the degree of count dispersion across the 35 data points for each gene. The mean for the INTS6 gene is 10.52 ± 1.88 (1 SD) counts and plotting the counts for the 35 samples in ascending Statistical characterization of trendline “tailing responses” identify gene pathway regulatory groupings that contribute to biological variability After rank-ordering unadjusted and MVA gene counts to create gene trendlines, standard Excel functions were used to perform a variety of statistical calculations [12]. Mean and median calculations measure aspects of dispersion and skewness, standard deviation, range, and slope measure dispersion, and skewness measures the unevenness of dispersion. Ranking these statistical pa- rameters characterizes the degree to which this disper- sion impacts gene expression levels for various genes. Calculations were computed for each of the 8746 genes and the results were ranked in descending order (Additional file 2, sheet 6). The 300 genes displaying the largest numerical values for each calculation were sub- jected to STRING-db analysis and the identified genes were surveyed for pathway affiliations (Additional file 2, sheet 7). The results were summarized and presented in Additional file 4A and B. The last two example genes, IFIT3 and EIF1AY, displayed much greater deviation from the linear trendline model (Fig. 1a; 21.96 ± 25.52 and 26.88 ± 46.03, respectively). The rank-ordered IFIT3 trendline depicted in Fig. 1b, identified individuals in quartile 4 with markedly different expression levels when compared to individuals in quartiles 1–3. The final example gene, EIF1AY, is located on the Y chromo- some and is expressed only in males. The gene trendline in Fig. 1b, shows an expected bimodal pattern with samples 24–35 comprising the eleven males in the sample group. The R2 values for these two genes were 0.429 and 0.5923, respectively, which denotes a significant deviation from lin- earity (CV 116.18 and 171.24%, respectively). y y These five example genes exhibit increasing degrees of gene expression variability among the individuals in quartiles 1 and 4. The observed trendline profiles illus- trate how rank-ordering of RNA sequencing counts can identify marked changes in gene expression variability among some of the 8746 protein coding genes identified in our study. Based on linear regression analysis, 65– 70% of the 8000 to 10,000 evaluated genes (3 data sets) displayed trendlines where the incremental difference in gene expression across the group followed a linear pat- tern resulting in R2 values that were ≥0.9 (e.g. INTS6, Fig. 1, panel b). Under ideal conditions with minimal within sample variation, one might expect all of the se- quenced genes in the control sample to follow this linear pattern but this is not the case. Results Therefore, we conclude that genes displaying a linear trendline profile across a defined range of expression values represent a “normally distributed control envelope” grouping of expres- sion values within the identified samplying window. The mean counts for genes AKAP13 and KCNJ2 were 18.26 ± 4.47 and 12.88 ± 3.82, respectively (Fig. 1a). While these genes showed slightly more dispersion across the 35 samples (Panels a and b, with CV values of 25.26 and 29.62% and R2 values of 0.8499 and 0.8418, respectively), rank-ordering the counts revealed more complex trendlines where the slope of the line for samples in quartiles 1 and/or 4 deviated from the slope of the line for samples in quartiles 2 plus 3 (Fig. 1, panel b). Results Gene rank- order position is established in relation to the gene expression level for an individual gene within the sample group, therefore the ranking order does not identify the same individual at each position along the various gene trendlines since the relative level of gene expression for an individual changes across genes. c Minimum Value Adjusted (MVA) gene counts significantly improve count heteroscedasticity (5-fold scale reduction) without altering the incremental trendline profiles within the sample group. Rank-order analysis extends the descriptive sample information available from a box plot by: defining the number of data points within the sample that deviate from the count level in the 2nd and 3rd quartiles; identifying their inflection point(s) and providing an estimate of the relative change in gene expression based on the computed slope ratio change. Black vertical lines identify quartiles 1, 2–3 and 4. See Additional file 1 for a more detailed discussion Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 Page 4 of 19 Page 4 of 19 however, the unique incremental sample-to-sample gene expression relationship of the 35 rank-ordered samples was maintained irrespective of the trendline profile (Fig. 1, panels b vs. c). When the quartile slopes for individuals in quartiles 1 and/or 4 deviates from those in quartiles 2 plus 3, a “tailing” profile was established as illustrated by the genes depicted in panels b and c of Fig. 1. Due to random chance, it would be difficult and unlikely to find several hundred genes displaying 4–8 “outliers” in a common subset of 35 individuals. Further- more, we will now demonstrate how these “tailing response” profiles, as illustrated for the IFIT3 gene, can be used to identify other genes sharing comparable trendline profiles, and thereby identify sources of biological variation among selected individuals in a sample group. rank-order created a linear INTS6 trendline as illustrated in Fig. 1b. A coefficient of variation (CV) of 17.9% and the co- efficient of determination (R2) of 0.9498 further supports the linear profile of the INTS6 trendline. This trendline profile was identical to the pattern obtained when numbers were randomly selected from a normally distributed popu- lation within a defined range of values and rank-ordered (see Additional file 1 for a detailed discussion). Statistical characterization of trendline “tailing responses” identify gene pathway regulatory groupings that contribute to biological variability Our subsequent analysis attempts to provide some explanation for the heightened variability noted for genes such as IFIT3 in Fig. 1. The unadjusted and MVA gene counts identified Bio- logical Gene Ontology (GO) pathways associated with cotranslational protein targeting to membrane (section 4A) or immune system process pathways (section 4B) when the largest means representing the various statis- tical calculations were evaluated for the two groups. The unadjusted mean counts identified gene pathway group- ings having the largest relative gene expression levels. When the gene counts are scaled by MVA to reflect the sample-to-sample incremental changes of each gene, the resulting trendline means identified immune pathway classifications rather than the highly expressed genes as- sociated with protein synthesis (Additional file 4, panel A vs. B). The identification of markedly different path- way affiliations following MVA is consistent with the findings reported by van den Berg et.al [9].. When the unadjusted gene counts were used for these calculations, parameters that measure the relative magnitude of the Figure 1c depicts the Minimum Value Adjusted (MVA) TPM counts which substantially reduce the range of gene expression (e.g. > 5-fold decrease in scale); Wilfinger et al. BMC Genomics (2021) 22:322 Page 5 of 19 Page 5 of 19 count, such as mean, standard deviation, maximum, me- dian, quartile 1, quartile 3, slope etc. all select highly expressed genes in Biological GO pathways associated with protein synthesis and targeting proteins to different areas of the cell (Panel 4A vs 4B). However, when statis- tical parameters such as range/median, skewness and kurtosis were used that characterize the “tailedness” and the unevenness of sample dispersion, identical pathway results were obtained with either unadjusted or MVA counts (Panel 4A vs 4B). Therefore, the type of measure- ment used for gene trendline characterization prior to STRING-db analysis impacts pathway selection if the heteroscedastic nature of the raw counts was not ad- dressed prior to pathway analysis. groupings that were selected using various combinations of sample size (e.g. 250–450 genes) and statistical par- ameter groupings (combine 1–3 measures for pathway selection). STRING-db analysis of 250–300 genes based on trendline kurtosis estimates selected identical path- ways (data not shown). Samples of 300 genes surveyed at various rank position locations, ranging from 1 to 6000, selected different GO pathways with lower FDR’s following STRING-db analysis. Sampling genes at lower gene rankings identified large pathways involved in cel- lular metabolism and function. Correlation analysis identifies genes displaying similar trendline profiles and regulatory pathway associations trendline profiles and regulatory pathway associations The previous analysis demonstrated that ranking certain statistical measures in a sample of 35 individuals identi- fied genes with “tailed” trendlines and affiliated pathway groupings. To further evaluate this result, we employed correlation analysis to identity genes that might display similar associations to the trendline profiles previously noted for the IFIT3 gene (Fig. 1b and c). We used Excel to perform Pearson correlation analysis on the MVA counts of 8746 genes in our study [12]. To limit the size of the correlation matrix (> 78 × 106 values) to a more discernable number of terms, estimated values for the highest correlation and anticorrelation range was used to provide a count of the number of genes displaying correlation values > or < input values and the number of genes assigned r values ≥or ≤the input terms were iden- tified [12]. After the initial analysis, the input correlation values are adjusted up or down to limit the number of genes assigned to a smaller correlation subset matrix. Using this rationale, we identified a subset of 500 genes with correlation values ≥0.95725 or ≤−0.524674. Within this group of genes, the IFIT3 gene was positively corre- lated with the largest cluster of genes including IFIT1 and 12 other genes. STRING-db analysis indicated that these 14 genes were associated with 24 GO pathways containing multiple regulatory protein associations as depicted in Fig. 2. The top 3 GO pathways with FDR ≤ E-15 were GO:0009615, response to virus, 5.33 E-21; The statistical parameters depicted in file 4 illustrate that some measures identified a larger number of gene as- sociations with lower False Discovery Rates (FDR) based on the observed “tailing” patterns. Range/Q3, range/me- dian and kurtosis measures detected 122, 113 and 105 im- mune system process (GO:0002376) pathway genes, respectively. Although all three parameters demonstrated proficiency in selecting genes with “tailing” profiles, only 8 of the top 10 pathways were identical among the three cal- culations and 7–14% fewer total genes were identified when either kurtosis or range/median measures were employed. Although a variety of calculations can be used for identifying gene pathway affiliations in addition to range/Q3, range/median and kurtosis, the other parame- ters selected fewer genes, different rank-orders, and al- ternative pathways when these parameters were employed to identify gene affiliations based on gene trendline tailing response profiles (Additional file 4). Statistical characterization of trendline “tailing responses” identify gene pathway regulatory groupings that contribute to biological variability These pathways involve thousands of genes and due to the size of the pathways much lower FDR’s were observed (e.g. FDR > E-15). The application of the MVA scaling reduced heterosce- dasticity as previously noted [9] while preserving important sample-to-sample incremental changes that contributed to the rank-ordered trendline profiles. In our sample of 35 individuals, MVA reduced Total Sums of square by 960- fold and Within Group Sums of Square by 303-fold (see Additional file 1). The various statistical parameters tested in our studies revealed that range/Q3, range/median and kurtosis were the most sensitive and robust parameters for identifying “tailedness” in unadjusted as well as MVA applications (Additional file 4B). Other statistical calculations that measure sample variability and trendline asymmetry such as coefficient of variation, maximum/minimum ratio, range/median, skew- ness, kurtosis, range/quartile 3, and R2 all identified immune-related GO pathways with FDR’s ranging from E-6 to E-32 (Panel 4B). The 300 genes displaying the largest range/Q3 (FDR = 6.22 E-32), range/median (FDR = 5.33 E-26) and kurtosis values (FDR = 6.85 E-27) detected the greatest trendline variability and had the smallest R2 values ranging from 0.2253 to 0.8754. These three statis- tical calculations selected trendline “tailing” patterns with the greatest fidelity that were similar to the profile previ- ously depicted by the IFIT3 gene in Fig. 1c. Correlation analysis identifies genes displaying similar trendline profiles and regulatory pathway associations Eight of the highlighted genes (red, blue and green) form statistically significant groupings with False Discovery Rates ranging from E−17 to E−21 that may collectively integrate the activity of all three pathways GO:0051607, defense response to virus, FDR 1.13 E-20 and GO:0060337, type 1 interferon signaling pathway, 2.64 E-17. The correlation results were identical when either the original counts or MVA counts were evaluated with an equivalent number of genes (i.e. 500). STRING- db analysis of the most highly correlated genes within the entire data set identified gene pathways that were activated in response to virus exposure. than 30 genes were identified with negative correla- tions (r ≤−.52465; TMEM38B, 43 genes; MMP9, 39 genes and CLEC4D,36 genes). The list of 43 genes associated with TMEM38B were evaluated with the STRING-db to determine if any of these genes shared pathway relationships and the results are depicted in Fig. 4. These 44 genes form associations with 145 different Biological GO pathways with PPI enrichment < 1.0 E-16 and they appear to be primarily involved in mediating immune responses (GO:0006955). Based on the STRING-bd results presented in Fig. 2, 7 genes displaying two or more pathway affiliations were se- lected and their expression profiles were plotted in the 35 unranked control samples. The gene expression profiles for our control group and two additional archived control data sets are presented in Fig. 3. The average baseline expression level for most of these genes is ~ 5 counts, so gene expres- sion levels of 30–110 counts represent markedly elevated levels of gene expression in certain individuals. Interferon induced IFI44L and ISG15 genes are markedly elevated in individuals 6, 9 and 12 in panel a, sample 7 in panel b and samples 3 and 4 in panel c, and the coordinated response is suggestive of individuals responding to the presence of a virus. It is important to emphasize that the elevated level of gene expression of these 7 genes is confined to specific indi- viduals in the sample group and the non-random nature of the response is unlikely due to methodological variability. Localization of highly correlated gene groupings in specific individuals is used to construct a scoring function The highly correlated cluster of genes identified in Fig. 2, and their coordinated expression responses within cer- tain individuals as depicted in Fig. 3, suggested a second avenue for analysis. Correlation analysis identifies genes displaying similar trendline profiles and regulatory pathway associations Changes in the order of the top 10 identified pathways were impacted by the number of known genes in a des- ignated pathway and the selected measure used to identify the pathway-related genes in the sample. For example, the identification of 50 genes in a pathway of 200 genes provides a lower FDR than the detection of 50 genes in a pathway containing 2000 genes. The identification of the top 300 computed trendline values, as outlined above, was also used to evaluate gene Page 6 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Fig. 2 Listing of highly correlated genes identified by correlation analysis and their known integrated network affiliations within the immune system. STRING database analysis of the 13 genes found to be highly correlated (r ≥0.95725) with the IFIT3 gene. This regulatory cluster is associated with 24 GO pathways that are primarily involved in response to virus (red, GO:0009615), defense response to virus (blue, GO:0051607) and type 1 interferon signaling (green, GO:0060337). Eight of the highlighted genes (red, blue and green) form statistically significant groupings with False Discovery Rates ranging from E−17 to E−21 that may collectively integrate the activity of all three pathways Fig. 2 Listing of highly correlated genes identified by correlation analysis and their known integrated network affiliations within the immune system. STRING database analysis of the 13 genes found to be highly correlated (r ≥0.95725) with the IFIT3 gene. This regulatory cluster is associated with 24 GO pathways that are primarily involved in response to virus (red, GO:0009615), defense response to virus (blue, GO:0051607) and type 1 interferon signaling (green, GO:0060337). Eight of the highlighted genes (red, blue and green) form statistically significant groupings with False Discovery Rates ranging from E−17 to E−21 that may collectively integrate the activity of all three pathways Fig. 2 Listing of highly correlated genes identified by correlation analysis and their known integrated network affiliations within the immune system. STRING database analysis of the 13 genes found to be highly correlated (r ≥0.95725) with the IFIT3 gene. This regulatory cluster is associated with 24 GO pathways that are primarily involved in response to virus (red, GO:0009615), defense response to virus (blue, GO:0051607) and type 1 interferon signaling (green, GO:0060337). Correlation analysis identifies genes displaying similar trendline profiles and regulatory pathway associations The rationale was based on the premise that the coordinated gene activity within a biological pathway would involve multiple genes and this should result in a higher rank-order position for the genes in the activated pathway as well as an increase in the relative number of positionally ranked genes repre- senting that pathway. To explore this possibility, a “Scor- ing Function” depicting the gene rank position listing was determined for every gene and this analysis is de- scribed in Additional file 2, sheet 7 and file 6. Table 1 provides an abbreviated summary of the results. Based In addition to the 14 positively correlated genes, there were also several gene clusters in which more Wilfinger et al. BMC Genomics (2021) 22:322 Page 7 of 19 Fig. 3 Highly correlated and functionally related gene networks are simultaneous elevated in specific individuals. Seven genes were selected from the highly correlated list of genes identified in Fig. 2 and their unranked expression profiles were plotted for the individuals in three different Control data sets (a, b, and c). In panel a (35 in house Controls), b (9 Controls, [24]) and c (12 Controls, [25]) the interferon induced IFI44L and ISG15 genes were specifically elevated in approximately 12% of the individuals (gene expression levels > 6-fold of baseline expression) Fig. 3 Highly correlated and functionally related gene networks are simultaneous elevated in specific individuals. Seven genes were selected from the highly correlated list of genes identified in Fig. 2 and their unranked expression profiles were plotted for the individuals in three different Control data sets (a, b, and c). In panel a (35 in house Controls), b (9 Controls, [24]) and c (12 Controls, [25]) the interferon induced IFI44L and ISG15 genes were specifically elevated in approximately 12% of the individuals (gene expression levels > 6-fold of baseline expression Fig. 3 Highly correlated and functionally related gene networks are simultaneous elevated in specific individuals. Seven genes were selected from the highly correlated list of genes identified in Fig. 2 and their unranked expression profiles were plotted for the individuals in three different Control data sets (a, b, and c). In panel a (35 in house Controls), b (9 Controls, [24]) and c (12 Controls, [25]) the interferon induced IFI44L and ISG15 genes were specifically elevated in approximately 12% of the individuals (gene expression levels > 6-fold of baseline expression IFIT3). Correlation analysis identifies genes displaying similar trendline profiles and regulatory pathway associations Eight of these 10 genes were previously identified in Fig. 2 with FDR’s ranging from E-15 to E-27 (see Additional files 6, 8 and 9). on STRING-db analysis, six individuals were identified with gene clusters representing multiple immune path- ways with False Discovery Rates (FDR) ≤E-15. Range /Q3 and kurtosis calculations identified individuals 4, 6, 9, 10, 12 and 33 with multiple immune pathways at FDR’s ≤E-15 to E-27 (Fig. 3, Table 1 and Additional file 6). The analysis of the 35 control samples identified 6 individ- uals or 17% of the sample group with genes displaying marked “tailedness”. Moreover, the genes identified in these individuals are involved in the regulation immune function pathways, such as defense response to virus (GO: 0051607) which was identified in 4 of the 6 individuals (11%). A Venn Plot of the genes identified in all three data sets (e.g. data set 1; samples 6, 9, 10, 12 data set 2; sample 7 and data set 3; samples 3 and 4) identified 10 genes common to all three data sets (e.g. HERC5, OAS3, RSAD2, OAS1, MX1, IFI6, IFI44L, IFIT1, OASL and Individuals responding to viruses and pronounced inflammatory responses resulting in elevated numbers of white blood cells contribute to biological variability The positionally nked genes were evaluated to determine if any genes with range/median, range/Q3, kurtosis and Q4/Q(2 + 3) slope values were within a group of the top 300 enes previously identified for each of the selection parameters. For example, in a list of 1000 positionally ranked genes, only genes with a range/median value ≥ the computed value of the 300th gene would be identified. The genes identified with these four parameters were subsequently evaluated with the STRING db Protein coding gene counts in three data sets were Minimum Value Adjusted (MVA) and rank ordered. The individual with the highest read count for any given gene was assigned a positional rank of 1, the second highest count was assigned a rank of 2, etc. until all of the samples received a positional gene assignment ranking (see Additional file 6). A scoring function was employed to identified a minimum of 1000 positionally ranked genes for each individual. The positionally ranked genes were evaluated to determine if any genes with range/median, range/Q3, kurtosis and Q4/Q(2 + 3) slope values were within a group of the top 300 genes previously identified for each of the selection parameters. For example, in a list of 1000 positionally ranked genes, only genes with a range/median value ≥ to the computed value of the 300th gene would be identified. The genes identified with these four parameters were subsequently evaluated with the STRING db to determine if they were associated with known biological pathways. The black font represents the number of times the designated pathway was ranked in the top 10 pathways in the 35 samples. The number highlighted in red font represents the number of individuals with a False Discover Rate (FDR) < E-15. Samples 4, 6, 9, 10, 12 and 33 all contained one or more pathways with FDR < E-15 to E-42 (Additional files 6, 8, 9). Range/Q3 and range/median calculations were the most robust parameters and identifying the largest number of pathway genes with the smallest FDR. Individuals responding to viruses and pronounced inflammatory responses resulting in elevated numbers of white blood cells contribute to biological variability Individuals responding to viruses and pronounced inflammatory responses resulting in elevated numbers of white blood cells contribute to biological variability Our analysis highlighted sample 33 with neutrophil and leukocyte activation pathways (Additional file 6) and we speculated whether WBC number might be influencing these responses [26, 27]. To address this question, we plotted the WBC differential cell counts for the 35 indi- viduals in our control sample and the results are pre- sented in Additional file 7. Sample 33 clearly contained the largest number of WBC’s and neutrophils. When the cell counts were rank-order, samples 33, 6 and 8 con- tained a proportionally larger number of WBC’s and Wilfinger et al. BMC Genomics (2021) 22:322 Page 8 of 19 Fig. 4 STRING database analysis of 44 genes found to be negatively correlated (r ≤−0.52465) with the TMEM38B gene. This regulatory cluste associated with 145 GO pathways that are primarily involved with immune responses such as leukocyte activation (blue, GO:0045321), neutro degranulation (red, GO:0043312) and immune system process (green, GO:0002376) Fig. 4 STRING database analysis of 44 genes found to be negatively correlated (r ≤−0.52465) with the TMEM38B gene. This regulatory cluster is associated with 145 GO pathways that are primarily involved with immune responses such as leukocyte activation (blue, GO:0045321), neutrophil degranulation (red, GO:0043312) and immune system process (green, GO:0002376) Fig. 4 STRING database analysis of 44 genes found to be negatively correlated (r ≤−0.52465) with the TMEM38B gene. This regulatory cluster is associated with 145 GO pathways that are primarily involved with immune responses such as leukocyte activation (blue, GO:0045321), neutrophil degranulation (red, GO:0043312) and immune system process (green, GO:0002376) Page 9 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Table 1 Summary of Sample Positional Rank Assignments otein coding gene counts in three data sets were Minimum Value Adjusted (MVA) and rank ordered. The individual with the highest read count for any given ene was assigned a positional rank of 1, the second highest count was assigned a rank of 2, etc. until all of the samples received a positional gene assignment nking (see Additional file 6). A scoring function was employed to identified a minimum of 1000 positionally ranked genes for each individual. Protein coding gene counts in three data sets were Minimum Value Adjusted (MVA) and rank ordered. The individual with the highest read count for any given gene was assigned a positional rank of 1, the second highest count was assigned a rank of 2, etc. until all of the samples received a positional gene assignment ranking (see Additional file 6). A scoring function was employed to identified a minimum of 1000 positionally ranked genes for each individual. The positionally ranked genes were evaluated to determine if any genes with range/median, range/Q3, kurtosis and Q4/Q(2 + 3) slope values were within a group of the top 300 genes previously identified for each of the selection parameters. For example, in a list of 1000 positionally ranked genes, only genes with a range/median value ≥ to the computed value of the 300th gene would be identified. The genes identified with these four parameters were subsequently evaluated with the STRING db to determine if they were associated with known biological pathways. The black font represents the number of times the designated pathway was ranked in the top 10 pathways in the 35 samples. The number highlighted in red font represents the number of individuals with a False Discover Rate (FDR) < E-15. Samples 4, 6, 9, 10, 12 and 33 all contained one or more pathways with FDR < E-15 to E-42 (Additional files 6, 8, 9). Range/Q3 and range/median calculations were the most robust parameters and identifying the largest number of pathway genes with the smallest FDR. Immune function pathways relating to defense response to virus, response to virus and type I interferon signaling pathways were highlighted in individuals 6, 9,10 and 12 with FDR’s from E-15 to E-26 when range/Q3, range/ median or kurtosis values were used for gene identification and STRING-db analysis Application of positional rank analysis to survey trendline patterns in control archived RNA-Seq data sets patterns in control archived RNA Seq data sets We retrieved archived Control peripheral blood RNA sequencing files from the GEO database [24, 25, 31] (GSE109313 and GSE112057), processed the raw counts through our pipeline and evaluated genes with mean TPM counts ≥0.5 in the two control data sets (Additional file 11). The resulting lists of genes were filtered to remove non- protein coding genes and the most robust trendline selec- tion measures including range/median, range/Q3, kurtosis and Q4/(Q2 + Q3) slope values were used to identify genes with the highest positional rank assignments as pre- viously outlined (Additional file 5). A detailed summary of the results is presented in (Additional files 6, 8 and 9). Since we previously identified individuals in our 35- member Control group that were likely responding to virus-mediated immune challenges (Table 1 and file 6), we evaluated two additional archived Control data sets to de- termine if the smaller sample groups also contained similar individuals. Employing the 4 parameters used in our previ- ous study (Table 1), a list of positionally ranked genes was assembled with each of the screening parameters. The as- sembled positionally ranked gene lists were evaluated with the STRING-db [13] to determine if any Biological GO pathways were specifically highlighted in these individuals. STRING-db analysis of the positionally ranked genes iden- tified one individual in Control group b (sample 7, Fig. 3), and 2 individuals in Control Group c (samples 3 and 4, Fig. 3) in which the defense response to virus pathway was significantly elevated (see Table 1 and Additional files 8 and 9 for a detailed summary of the pathway results). To further evaluate technical and biological variability, we normalized the CBX3, IFIT3, IFI44L and DEFA3 gene expression in relation to the stable ATG3 gene (see Additional file 10). This normalization was performed by calculating gene expression ratios for each of the 35 samples and evaluating the degree of count dispersion across samples when the counts are expressed in relation to a known stable gene. In panel a of Fig. 6, the CBX3/ ATG3 gene ratios for two stable genes are presented. The gene ratios for these two genes range from 1.2 in sample 8 to 2.4 in sample 14 (2X range). When the highly correlated and variable expression profiles of the IFIT3 and IFI44L genes, previously identified in Figs. Can the detected biological responses be explained by technical variability? Can the detected biological responses be explained by technical variability? It is always important to consider how the application of an analytical strategy may impact the results and conclu- sions of any study. In three separate RNA-seq studies, we noted that 65–70% of the genes follow linear trendline profiles with R2 values ≥0.9. We selected 4 genes with unadjusted mean counts ranging from 9 to 3836 to determine if consistent trendline linearity ex- tended across a broad range of gene expression values. The trendline expression profiles for these genes and 4 ERCC standards with similar expression levels are pre- sented in Fig. 5. Similar linear trendlines were observed for both the selected genes and the ERCC standards with the exception of sample 35, representing sample 7, (rank-ordered samples are not listed in numeric se- quence, 1–35), in which the ERCC values were consist- ently larger. The marked deviation from the computed trendline observed in ERCC sample 35 is not reflected in the 4 selected genes in the left panel of Fig. 5, which suggests a potential spike-in issue with little or no tech- nical impact on gene counts. Individuals responding to viruses and pronounced inflammatory responses resulting in elevated numbers of white blood cells contribute to biological variability Immune function pathways relating to defense response to virus, response to virus and type I interferon signaling pathways were highlighted in individuals 6, 9,10 and 12 with FDR’s from E-15 to E-26 when range/Q3, range/ median or kurtosis values were used for gene identification and STRING-db analysis neutrophils than the other 32 samples and the removal of those three individuals markedly improved the WBC vari- ation explained by the regression line (WBC R2 improves from 0.8492 to 0.9824). This analysis demonstrates that a disproportionate number of WBC’s can also impact gene pathway analysis when the cell numbers are elevated above 10 million WBC’s / ml of blood. The survey of gene tren- dlines such as IFIT3 based on range/Q3, range/median and kurtosis (Table 1) provided a strategy to identify pathways in which groups of genes appeared to display coordinated ex- pression patterns. The highly correlated group of genes identified in Fig. 2 form multiple protein interactions involv- ing an assortment of different biological pathways as previ- ously illustrated Additional files 4 and 6. The genes identified in Fig. 2 represent virus-induced interferon-stimulated genes [27–30]. When the individuals identified as having coordi- nated immune responses were removed from the analysis (e.g. individuals 6, 9, 10, 12 and 33, Additional file 6 and Table 1) and the correlations were re-evaluated in the remaining 30 samples, the genes previously correlated with IFIT1 (e.g. r ≥0.9579) were reduced from 14 genes to 0. Furthermore, the R2 values for the genes identified in Page 10 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Page 10 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Fig. 2 ranged from 0.295 to 0.426 in the sample of 35 indi- viduals were significantly improved to 0.931 to 0.965 after removing the 5 highlighted individuals from the analysis. Can the detected biological responses be explained by technical variability? Application of positional rank analysis to survey trendline patterns in control archived RNA-Seq data sets 2 and 3, were evaluated in relation to the ATG3 gene (Fig. 6, panels b and c) the sample gene ratios range from 0.59 in sample 21 to 10.62 in sample 9 (17.9 X range) for IFIT3 and from 0.08 in sample 2 to 4.1 in sample 9 (51 X range). It is also important to note that samples 9, 6 and 12 displayed the largest relative expres- sion levels for both genes as we previously reported in Fig. 3a. Although the highly variable DEFA3 gene dis- playing an R2 value of 0.3956 was not identified among the genes highlighted in Figs. 2 and 3, it also displayed DEFA3/ATG3 gene ratios that varied from 0.2 in sample 9 to 23.1 in sample 13 (116X range). The DEFA3 defen- sin gene belongs to a family of microbiocidal/cytotoxic peptides found in neutrophile granules that are thought to be involved in host defense responses. The gene ex- pression ratios for the IFIT3 and IFI44L genes identi- fied in samples 9, 6 and 12 (Fig. 6, panels b and c) were substantially greater than the range of values identified among the 35 CBX3/ATG3 gene ratios presented in The combined data summarized in Table 1, files 6, 8 and 9 and Fig. 3 demonstrate that positional rank ana- lysis identified from 11 to 17% of the individuals in the three control data sets with gene associations represent- ing virus activated immune pathways. We selected a relatively high benchmark with pathway FDR’s < E-15; but if PPI values < 1.0 E-16 were used for evaluating the gene listings, ~ 20% of the surveyed samples contained individuals in which viral induced immune pathways were identified. The individuals identified as undergo- ing viral-induced immune responses significantly im- pact gene expression levels in the pathways that were identified by our analysis, thereby increasing the bio- logical variability of the control sample groups. In addition, the noted increases in the number of WBC’s in some of these individuals were also identified as another source of biological variability in our data set (Additional file 7). Wilfinger et al. BMC Genomics (2021) 22:322 Page 11 of 19 Fig. 5 Trendline gene expression profiles remain consistent over a wide range of RNA-seq expression levels. Four genes displaying unadjusted gene counts ranging from 9 to 3836 counts were rank-ordered (left panel). Gene trendline linearity was independent of the level of gene expression. Discussion panel a of Fig. 6. Presumably, the heightened gene ra- tio responses reflect increased variability resulting from the coordinated biological responses impacting gene expression in samples 9, 6 and 12. In contrast, although the DEFA3/ATG3 gene expression ratios also displayed marked variability, the largest gene ra- tios were observed in different individuals (samples 13, 28, 34 and 35) suggesting that the biological vari- ability contributing to these changes were different to those impacting samples 9, 6 and 12. Representative sample gene ratio profiles of other stable genes as well as the various genes identified in Fig. 2 were identical to those depicted in Fig. 6, in panels a, b and c. This analysis demonstrates that the relative magnitude of the gene ratio responses identified in samples 9, 6 and 12 were much larger than the 2-fold range of sample-to-sample variation observed for 65– 70% of the sequenced genes as depicted in panel a of Fig. 6, and they are unlikely due to technical variability. panel a of Fig. 6. Presumably, the heightened gene ra- tio responses reflect increased variability resulting from the coordinated biological responses impacting gene expression in samples 9, 6 and 12. In contrast, although the DEFA3/ATG3 gene expression ratios also displayed marked variability, the largest gene ra- tios were observed in different individuals (samples 13, 28, 34 and 35) suggesting that the biological vari- ability contributing to these changes were different to those impacting samples 9, 6 and 12. Representative sample gene ratio profiles of other stable genes as well as the various genes identified in Fig. 2 were identical to those depicted in Fig. 6, in panels a, b and c. This analysis demonstrates that the relative magnitude of the gene ratio responses identified in samples 9, 6 and 12 were much larger than the 2-fold range of sample-to-sample variation observed for 65– 70% of the sequenced genes as depicted in panel a of Fig. 6, and they are unlikely due to technical variability. RNA-sequencing technology has the potential to con- tribute significantly to our ability to diagnose and treat many diseases. As clinical medicine relies more heavily on this technology for the application of personalized treatment strategies, it becomes increasingly important to identify and disentangle sources of technological error while preserving and identifying intrinsic sources of biological variability. Application of positional rank analysis to survey trendline patterns in control archived RNA-Seq data sets ERCC spike-in standards also showed linearity but the ERCC standard in index 35 (sample 7) was markedly elevated (right panel). However, sample 7 did not display similar deviations in any of the four genes depicted in the left panel of Fig. 5, which suggest a spike-in issue with little or no technical impact on gene counts. Red circles identify sample 7 Fig. 5 Trendline gene expression profiles remain consistent over a wide range of RNA-seq expression levels. Four genes displaying unadjusted gene counts ranging from 9 to 3836 counts were rank-ordered (left panel). Gene trendline linearity was independent of the level of gene expression. ERCC spike-in standards also showed linearity but the ERCC standard in index 35 (sample 7) was markedly elevated (right panel). However, sample 7 did not display similar deviations in any of the four genes depicted in the left panel of Fig. 5, which suggest a spike-in issue with little or no technical impact on gene counts. Red circles identify sample 7 Fig. 5 Trendline gene expression profiles remain consistent over a wide range of RNA-seq expression levels. Four genes displaying unadjusted gene counts ranging from 9 to 3836 counts were rank-ordered (left panel). Gene trendline linearity was independent of the level of gene expression. ERCC spike-in standards also showed linearity but the ERCC standard in index 35 (sample 7) was markedly elevated (right panel). However, sample 7 did not display similar deviations in any of the four genes depicted in the left panel of Fig. 5, which suggest a spike-in issue with little or no technical impact on gene counts. Red circles identify sample 7 Wilfinger et al. BMC Genomics (2021) 22:322 Page 12 of 19 Fig. 6 Gene count ratio estimates identify the specific individuals previously identified by String-db analysis. Stable genes identified in three data sets were used to normalize gene expression (see Additional file 10). In panel a, the CBX3/ATG3 gene ratios of two highly stable genes are plotted for each of the 35 samples in our study. A 2-fold range of variation is noted between samples 8 and 14. In contrast, when two of the interferon regulated genes were normalized in relation to the ATG3 gene (panels b and c) samples 9, 6 and 12 were highlighted with ratios 2 to 5-fold higher than noted in panel a. Application of positional rank analysis to survey trendline patterns in control archived RNA-Seq data sets The relative response profile of samples 9, 6 and 12 in panels b and c correspond to the genes correlated with IFIT3 gene previously identified in Fig. 3. In contrast to the IFIT3 and IFI44L response profiles, when the DEFA3 gene was normalized in relation to ATG3, samples 13, 34, 28 and 35 were highlighted. When evaluated in relation to highly stable genes within the data set, the nonlinear gene trendlines identified in our analysis highlight meaningful inter-individual changes in gene expression that cannot be explained on the basis of technical variation Fig. 6 Gene count ratio estimates identify the specific individuals previously identified by String-db analysis. Stable genes identified in three data sets were used to normalize gene expression (see Additional file 10). In panel a, the CBX3/ATG3 gene ratios of two highly stable genes are plotted for each of the 35 samples in our study. A 2-fold range of variation is noted between samples 8 and 14. In contrast, when two of the interferon regulated genes were normalized in relation to the ATG3 gene (panels b and c) samples 9, 6 and 12 were highlighted with ratios 2 to 5-fold higher than noted in panel a. The relative response profile of samples 9, 6 and 12 in panels b and c correspond to the genes correlated with IFIT3 gene previously identified in Fig. 3. In contrast to the IFIT3 and IFI44L response profiles, when the DEFA3 gene was normalized in relation to ATG3, samples 13, 34, 28 and 35 were highlighted. When evaluated in relation to highly stable genes within the data set, the nonlinear gene trendlines identified in our analysis highlight meaningful inter-individual changes in gene expression that cannot be explained on the basis of technical variation Discussion Nevertheless, the important incremental sample- to-sample gene expression changes were maintained after MVA scaling as illustrated by the gene profiles depicted in panels b and c of Fig. 1. After MVA, these incremental changes form the basis for identifying the trendline expres- sion profiles (Additional file 4, Panel A vs B). We note that while MVA-scaled data is suitable for trendline analysis, it is important to follow the correct scaling protocol for dif- ferential expression analysis by following the specific guide- lines of the software that is being employed. After performing MVA scaling on our data set, we determine that ~ 70% of the 8746 genes in our sample (Fig. 1c), contribute significantly to the variability in the control data set (see Additional file 1). The approach employed in our manuscript is designed to provide an explanatory (and visually inspectable) methodology that can augment existing tools and guide the decisions of the investigator. The statistics used are based on the quantile function [38] and higher moments of distribu- tions (skew) which are readily available in a number of standard statistical software packages. A dramatic ex- ample of the variability associated with these genes is depicted by the marked increase in their computed coef- ficient of variability (Additional file 1, Fig. 4). investigator. Moreover, any outlier detection by current software during DGE is performed at the gene level only. Our approach is focused on providing complementary information prior to DGE. We have evaluated the possi- bility of creating a methodology that uses existing pack- ages such as Deseq2 [37] or Cuffdiff [36] for intra-group analysis (for example, by using variable-leave-k-out comparisons, where k varies over sample subsets). In our analysis, such an approach was computationally very costly (exploring the set of subsets of 35 samples), and not as descriptively informative. Moreover, our approach extends beyond the gene-level view and reports on the potential impact of divergent values at the network level. y ( g ) By rank-ordering the incremental change in gene ex- pression across samples, we created a unique snapshot or “trendline”. Statistical evaluation of the trendlines identi- fied several robust measures that provided the greatest ability to characterize the biological variability or “tailed- ness” of the expression values. STRING-db analysis of the genes exhibiting the most pronounced “tailedness” expres- sion profile revealed that these genes were associated with important regulatory networks (Additional file 4). Discussion We also demonstrated that positional rank analysis could be used to further evaluate RNA-seq data and identify gene ex- pression variability within specific individuals in the group (see Additional files 6, 8 and 9). The ability to identify and characterize gene trendline properties provides a new and powerful strategy to detect pathway-affiliated genes, and quantify the significance of their associated biological re- sponses via the computed False Discovery Rate in any tis- sue or sample. To our knowledge, this is the first study in which variability that diverges from standard technical variability has been identified (e.g. The number of Ob- served Genes per Pathway are identified and statistically quantified by the calculation of False Discovery Rates). p p g RNA sequencing counts routinely display large differ- ences in their relative gene expression levels, which scales with the mean of the sequenced counts. In previous re- ports, we developed strategies for stabilizing RNA in whole blood samples and significantly improving RNA recovery during extraction [10, 11]. Here, we used ERCC spike-in standard concentration ratios to minimize amplification and dilutional errors across samples [23], noting that ERCC spike-in mixes were not used in the archived data sets. We also addressed heteroscedasticity by using MVA scaling. This method of scaling is suitable for variability analysis because individual sample-to-sample gene expres- sion levels are adjusted to a common starting point across samples while maintaining the incremental trendline fluc- tuations of individual genes [23]. In our Control sample, MVA of the 8746 genes reduced the mean and standard deviation by an average of 3.9-fold; however, in highly expressed genes such as B2M the mean and variance are scaled down by as much as 1316-fold (Additional file 2, sheet 3). Nevertheless, the important incremental sample- to-sample gene expression changes were maintained after MVA scaling as illustrated by the gene profiles depicted in panels b and c of Fig. 1. After MVA, these incremental changes form the basis for identifying the trendline expres- sion profiles (Additional file 4, Panel A vs B). We note that while MVA-scaled data is suitable for trendline analysis, it is important to follow the correct scaling protocol for dif- ferential expression analysis by following the specific guide- lines of the software that is being employed. Discussion As previously noted, differential ex- pression tools are impacted by a variety of factors and they are not universally robust to the presence of outlier results from divergent expression data [3, 4, 32–36]. While tools such as EBSeq [33], LFCseq [35], leave-one- out methods [34], median-based approaches, and other software [36, 37], may detect and manage outliers based on the software’s defined criteria, they operate primarily as a “black box” during the differential expression ana- lysis step. These tools may or may not report the outlier, and to our knowledge, they do not provide an intra- group outlier explanation that is readily available to the Page 13 of 19 Page 13 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 investigator. Moreover, any outlier detection by current software during DGE is performed at the gene level only. Our approach is focused on providing complementary information prior to DGE. We have evaluated the possi- bility of creating a methodology that uses existing pack- ages such as Deseq2 [37] or Cuffdiff [36] for intra-group analysis (for example, by using variable-leave-k-out comparisons, where k varies over sample subsets). In our analysis, such an approach was computationally very costly (exploring the set of subsets of 35 samples), and not as descriptively informative. Moreover, our approach extends beyond the gene-level view and reports on the potential impact of divergent values at the network level. RNA sequencing counts routinely display large differ- ences in their relative gene expression levels, which scales with the mean of the sequenced counts. In previous re- ports, we developed strategies for stabilizing RNA in whole blood samples and significantly improving RNA recovery during extraction [10, 11]. Here, we used ERCC spike-in standard concentration ratios to minimize amplification and dilutional errors across samples [23], noting that ERCC spike-in mixes were not used in the archived data sets. We also addressed heteroscedasticity by using MVA scaling. This method of scaling is suitable for variability analysis because individual sample-to-sample gene expres- sion levels are adjusted to a common starting point across samples while maintaining the incremental trendline fluc- tuations of individual genes [23]. In our Control sample, MVA of the 8746 genes reduced the mean and standard deviation by an average of 3.9-fold; however, in highly expressed genes such as B2M the mean and variance are scaled down by as much as 1316-fold (Additional file 2, sheet 3). Discussion Trendline slope analysis of our 35 control samples identified individuals with gene expression rates in quar- tiles 1 and 4 that were more than 6-fold greater than the computed slope in quartiles 2 + 3 (32 and 351 genes identified in Q1 and Q4, respectively). Although higher or lower rates of relative gene expression for any given gene in the sample set will contribute to increased vari- ability, rank-order quartile analysis was useful in further characterizing this variation (Fig. 1, panels b and c). One explanation for this response profile is that certain critical regulatory genes were governed by positive or negative regulatory signals [39]. After performing MVA scaling on our data set, we determine that ~ 70% of the 8746 genes in our sample group displayed trendline linearity as assessed by R2 values ≥0.9. The application MVA in conjunction with rank-order trendline analysis illustrated that gene ex- pression in this group of genes is consistent with the profile obtained from a normally distributed sample. Moreover, the remaining 30% of the genes that deviate from this linear profile were easily identified and evalu- ated due their increased variability and dispersion. Our analysis demonstrated that a subset on individuals with tailed gene trendlines in quartile 4, similar to IFIT3 The expression profile of genes such as IFIT3 (Fig. 1b) are difficult to explain on the basis of sample-to-sample variability. The “tailing” trendline observed for the IFIT3 gene indicates that gene expression in about 25% of the individuals was markedly different from the other members of the sample group. These gene trendlines displayed significant non-uniformity (high variability) as illustrated by range/median, range/Q3, skewness and Page 14 of 19 Page 14 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 kurtosis measures. Although, kurtosis and skewness cal- culations identified the degree of “tailedness” of the gene sample distribution, quartile slope analysis provided a more direct measure of these changes. Calculating the slope ratios of Q1/(Q2 + Q3) or Q4/(Q2 + Q3) identified individuals that deviated from the central core of the sample group. STRING-db analysis of the genes display- ing these non-linear trendline profiles identified highly integrated pathway associations, as depicted in Figs. 2, 3 and 4, that are involved in the detection and response to a virus. Discussion We note that the robustness of the slope ratio calculations is dependent on the size of the sample group (e.g. n ≥16, Additional file 3). analysis. In addition, other related pathways involving defense response (GO:0006952), response to virus (GO: 0009615) and the type 1 interferon signaling (GO: 0009615) pathways were also routinely found among the top 10 identified pathways in conjunction with the defense response to virus pathway. Identification of individuals that exhibit a defense-related response within an experimental group is consistent with the time- dependent activation and transition of the immune sys- tem from the detection of a foreign object to a defined immune response [27–30]. We have noted that changes in gene expression may represent 5 to 50-fold deviations from the median ex- pression level as illustrated for the IFIT3 gene in panel c of Fig. 1. Current analytical pipelines for DGE are de- signed to remove “outliers” which display more than 30% disagreement [1]. Based on our results, these proto- cols remove the gene counts of individuals displaying the greatest biological variability thereby diminishing the opportunity to detect cases of interest wherein a contrib- uting physiological basis for variability may be at work. The removal of specific individuals that appear to be responding to a viral challenge may remove a previously unrecognized source of biological variation. While our approach informs the investigator about these cases, it is beyond the scope of this report to determine how the mitigation of this source of biological variability will in- fluence DGE during principal component analysis. The characterized gene trendline patterns provided a strategy for evaluating gene associations displaying similar trendline profiles [12]. In our sample group, range/median, range/Q3 and kurtosis calculations were also used to identify gene trendlines displaying marked dispersion and “nonlinearity”. Gene clusters identified with these parameters were subsequently evaluated with the STRING-db and the results were summarized in Table 1. STRING-db analysis as well as correlation stud- ies demonstrated that the genes displaying pronounced non-linear trendline properties resembling the IFIT3 gene, play a prominent role in the mobilization and acti- vation of specific immune pathways (Additional files 4, 6, 8 and 9 and Figs. 2, 3 and 6). Discussion We identified specific individuals in the control data sets expressing a dispro- portionate number of genes in the defense response to virus pathway (GO:0051607) with Observed Gene Count pathway assignments ranging from 4 to 29 genes and FDR’s from E-2 to E-26 in a pathway containing only 181 genes. Our results are suggestive of intra-group differ- ences beyond technical variability. Under the assumption that “methodological variability” is mostly stochastic, the emergence of highly correlated regulatory pathways (Fig. 2) with significant FDR’s identified during STRING-db analysis (Additional file 6) would be highly unlikely. This observation was further supported by noting that when the interferon-mediated genes identified in Fig. 2 were expressed in relation to highly stable genes such as ATG3 (Fig. 6), samples 9, 6 and 12 were consistently identified with the largest gene ratios. Moreover, random selections of 300 genes from the entire list of 8746 genes never resulted in FDR’s ≤E-6 (n = 5, Additional file 3). While we cannot exclude the possibility that technological and methodological error may influence our findings, the results presented in Table 1 and in Figs. 5 and 6 support the conclusion that technical and methodological error had nominal impact on our findings. g p p p y Identifying and characterizing the genes assigned to the highest positional ranks enabled us to identify indi- viduals with an increased number of genes in certain GO pathways in our control group (Additional file 6) as well as in the archived data sets (files 8, and 9). Our ana- lysis identified asymptomatic individuals who may have been responding to an immune challenge; for example, a recent infection, an immunization, or a response to one of the many latent viruses we commonly harbor in our bodies [40]. A broad array of genes in alternative gene expression pathways were also identified in our analysis but they did not reach a statistical level of significance (see Additional files 6, 8 and 9). These genes also contribute to the biological variability in the various sample groups but they do not appear to introduce an inordinate degree of variability. Although the impact of differences in WBC number and the detection of interferon-regulated genes involved in inflammatory re- sponse have been previously described [26–30], we believe this is the first report in which these responses have been quantified and localized to specific individuals within control data sets. Discussion After MVA and establishing trendlines, several calcula- tions including estimates of R2 were prepared. R2, a measure of the variance explained by regression, pro- vided an estimate of the linearity of every gene in the Intra-group identification of 11–20% of the individuals in three separate data sets as responding to a viral- induced immune response is an important observation that should be considered prior to differential expression Page 15 of 19 Page 15 of 19 Page 15 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 data set. We have ranked the R2 values for the three data sets in descending order and identified 15 genes with R2 values > 0.96 that had raw mean counts > 5 and that were common to all three data sets (Additional file 10). Within this group of genes, MVA reduced the raw means by 12-fold and identified genes with very stable trendlines that may be useful as internal reference gene candidates for evaluating RNA sequencing results (e.g. Fig. 6) or as qRT-PCR gene markers as previously sug- gested by Stamova et al. [41]. Furthermore, the resulting range and magnitude of the incremental changes in gene expression following MVA were markedly similar even though the RNA was ex- tracted and processed differently in the three data sets evaluated in our study. While MVA may also reduce some of the variability that is commonly introduced when data sets are sequenced in different laboratories and at different times, its designated utility in our studies is in the analysis of intra-group comparisons. Sample collection, RNA extraction, sequencing and data analysis Sample collection, RNA extraction, sequencing and data analysis Blood samples were collected from 35 healthy adults ac- cording to a protocol approved by Chesapeake Research Review, LLC (#Pro:00009509). Blood samples were used for complete blood cell analysis and RNA extraction [10, 11]. RNA was DNase-treated and submitted to the University of Cincinnati Genomics, Epigenetics and Sequencing Core Fa- cility for RNA sequencing. The computational pipeline employed in this study has been previously described [23]. p y y p y The RNA extracted from the 35 control samples passed quality control assessment prior to the addition of Exter- nal RNA Controls Consortium ExFold RNA spike-in mixes (ERCC: Ambion, 4,456,739; Foster City CA) and re- moval of ribosomal and globin RNA from the samples (Illumina GZG1206; San Diego, CA). The cDNA libraries were processed according to standardized Illumina proto- cols and sequenced on the Illumina HiSeq 2000 platform. Fastq data files containing 53–77 million single-end reads were trimmed and processed for data analysis. Reads were aligned to reference genome GRCh37.p13[hg19] using the BowTie2 aligner supporting gapped alignments as previ- ously described [23, 36]. Conclusions Rank-order analysis of the MVA gene expression values in conjunction with R2 calculations revealed that 65– 70% of the sequenced genes display a linear “baseline” level of gene expression across the data set. Statistical measures, such as range/median, range/ Q3 and kurtosis, that characterize the “tailing” properties of some gene trendlines, were used in conjunction with databases such as STRING-db or PANTHER-db to identify and quantify gene pathways contributing to biological variability within and across three different sample groups. Pathways relating to viral-induced immunological responses were identified in 11–20% of the 54 individuals evaluated in our combined studies. This previously unrecognized source of variability may confound or obscure DGE results and mask important conclusions obtained from immuno- logical studies. Methods The application MVA in conjunction with rank-order trendline analysis provides a strategy for identifying previ- ously unrecognized sources of biological variation. How- ever, it should be noted that the expression levels of ERCC spike-in standards, as depicted by sample 7 in Fig. 5, may also be impacted by the initial composition of the RNA sample [3, 19]. Our approach is helpful towards tackling the difficult question of how variability may impact the conclusions of previous studies. This analysis is simple to perform and requires a minimal amount of computational time. Therefore, it may significantly con- tribute to the clarity of our understanding of the manner in which previously unrecognized sources of biological variability may have biased or confounded the experimen- tal analysis with minimal overhead. Although differential gene expression analysis remains an important application in RNA-seq studies, the detection of regulatory pathways containing genes that may not otherwise be identified as differentially expressed provides new insight into the complex mechanisms governing human diseases. Sample collection, RNA extraction, sequencing and data analysis Data processing, normalization strategies and statistical analysis In our control data set, an iterative correction of the length-adjusted ERCC spike-in concentration ratios was used to proportionally adjust for sample processing effects, pipetting errors, dilutional differences and other sources of methodological variability while the archived data sets that did not contain ERCC spike-ins were limited to size factor normalization using the median-to- ratio method as previously outlined [23]. Data were adjusted for sequencing depth and normalized using the Deseq2 median of ratios method [36] before the reads (in units of transcript per kilobase million or TPM) were used in trendline analysis. Our objective was to minimize the impact of methodological variability before applying trendline analysis. A primary objective of our study was to determine whether rank-order analysis was a useful strategy to identify sources of intra-group biological variability that otherwise remain difficult to detect in RNA-seq data. In order to omit nonprotein coding genes from the sequen- cing results, each gene was surveyed to determine whether the transcript was a known protein coding gene. We used the Gene Ontology website to provide a list of 19,623 known human protein coding genes [20–22]. Genes that were not recognized as protein coding genes are listed as “filtered units” and they can be further ana- lyzed if desired. Two markedly different gene trendline expression profiles are presented in supplement 1 to illustrate how linear regression analysis and other calculations can be used to analyze rank-ordered gene count data. Tren- dlines were constructed for each gene by ranking the ad- justed and unadjusted gene counts for the 35 samples in ascending order as described in Additional files 1, 2, 3 and 4. From our list of 35 samples, we identified 15 genes with the largest mean expression values before and after MVA (Additional file 2, sheets 2 & 3). MVA significantly reduced the mean and standard deviation among these genes by 67 to 1316-fold but the Coefficient of Variation was unchanged. Genes were identified in the unadjusted and MVA data sets with R2 > 0.9 (linear profile) and R2 < 0.9 (nonlinear) trendline patterns. These mathematical calculations were used to further characterize gene trendlines and identify gene groupings (clusters) that shared similar mathematical features. This analysis identified relational associations among genes that would otherwise be indiscernible using more classical analytical procedures. Supplement file 2 describes additional software programing features and examples of data output. Data processing, normalization strategies and statistical analysis An over- view of the data input and processing pipeline is summa- rized in sheet 1 of file 2. Read counts for various genes ranged over a 5 Log10 scale thereby creating large differences in the variance within the sequenced data set. To address the heteroscedastic nature of the raw data, we applied Mini- mum Value Adjustment (MVA) scaling normalization strategy to our counts. The smallest count for each gene was identified and used to proportionally adjust the remaining samples. MVA = Gene A (Sample Counts (1 to n) / Gene A Minimum Value Count). This adjustment as- signs a value of 1 to the smallest count value for every gene. After MVA, all of the incremental changes for the 8746 protein-coding genes across the 35-sample data set fall within a numeric range of 1–60 relative counts. This adjustment of every gene to its lowest common denomin- ator eliminated large comparative differences in the rela- tive magnitude of the observed counts between genes within and across individuals while maintaining the im- portant incremental changes when the adjusted counts are rank-ordered. In addition, the trendline starting point is identical for every gene in the analysis. The removal of in- ordinately low (counts < 0.5) and nonsignificant gene counts minimized the possibility of inflating the magni- tude of the sample-to-sample incremental changes that can be sensitive to very small outliers during scaling ad- justments [9]. Trendline identification and Analysis For statistical analysis, the Microsoft Windows Excel Platform and its statistical and mathematical functions were used. In order to organize and manage the data, software was developed to streamline and augment our analysis [12]. RNA sequencing counts were imported into Excel, duplicate gene entries were identified and re- moved, gene counts that were below a user defined value were checked for meaningful gene expression levels and nonprotein coding genes were removed (optional) before performing Minimum Value Adjustment scaling (see Additional file 2 for details and examples of data output). The final raw data processing step was the Minimum Value Adjustment scaling procedure described below (MVA, Additional files 1 and 3). More detained informa- tion relating to the data entry process is provided in Additional file 2. Processing RNA sequencing data files g q g Data files containing 10–13 thousand transcripts were imported into an Excel spreadsheet and matched to a list of human protein coding genes identified at the Gene Ontology website [20–22]. To optimize our analysis in databases such as STRING [13–15] and PATHER [42, 43], we removed noncoding genes (filter) from further consideration [12]. When noncoding genes were re- moved, False Discovery Rate (FDR) estimates were im- proved during STRING-db pathway analysis thereby increasing the sensitivity of gene pathway identification. Transcripts that were not identified as protein-coding genes were assigned to a separate “Discard” data sheet. After identifying the protein coding genes, samples with gene counts < 0.5 were replaced with 0 and genes with means < 0.5 counts were eliminated from further ana- lysis [12]. Minimum Value Adjustment (MVA) scaling signifi- cantly reduced the average mean and standard deviation in the data set by 3.9-fold among the 8746 protein- coding sequenced genes. However, MVA preserved the unique incremental sample-to-sample changes in gene expression across individuals following rank-ordering. Single genes that mapped to multiple genomic locations (Copy Number Variants, CNV [44]) were identified and Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 Page 16 of 19 Page 16 of 19 Trendline identification and Analysis read counts were combined under a single identification convention (e.g. The Control data set contained 66 CNV genes with duplicate/triplicate gene location listings). Analysis of archived RNA Seq data files An identical method of analysis was used to evaluate se- quencing data obtained from peripheral blood samples extracted and processed with different methodologies in order to determine whether the MVA scaling and rank- ordering methods could identify similar changes in gene pathway affiliations among Control samples in two archived data sets. We downloaded data files from the NCBI Gene Expression Omnibus site [31] containing sequencing data obtained from the peripheral blood of 9 Controls (GSE109313, [24]) in one data set and 12 Controls in a second data set (GSE112057, [25]). Blood samples in the first data set were extracted with the PAXgene RNA blood extraction kit and further proc- essed to prepare poly(A) selected and ribo-depleted RNA-seq libraries. In data set 2, blood was collected via Tempus Tubes and RNA was extracted with the Tempus Spin RNA isolation kit. Authors’ contributions KM and PC were involved in the design of the sample collection protocol and IRB approval process. WW, KM and PC participated in the sample extraction and processing of the RNA samples. HE assembled the data processing and analysis pipeline through which the sequenced data were processed. WW, HE and PC contributed to the development of the data analysis rationale outline in the manuscript. RM provided the programming for the RAnGER software. WW wrote the manuscript and all the authors assisted in the review, edits and final approval of the manuscript. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s12864-021-07563-9. The online version contains supplementary material available at https://doi. org/10.1186/s12864-021-07563-9. Additional file 1. Rationale for identifying sources of biological variation in RNA sequencing data. Additional file 1. Rationale for identifying sources of biological variation in RNA sequencing data. Additional file 2. Summary of Statistical Measures used to Evaluate Gene Expression before and after Minimum Value Adjustment (MVA). Additional file 3. The Identification of Gene Trendlines with “Tailing” Profiles Using Quartile Slope Analysis. Additional file 4. STRING db Pathway Identification Based on the Analysis of the Top 300 Genes Identified by the Designated Statistical Measures. Additional file 5. Examination of Intraindividual Gene Rankings to Identify Individuals Displaying Coordinated Gene Regulatory Activity. Additional file 6. STRING db Analysis of Intra-individual Positional Gene Rankings in 35 Control Samples Based On Range/Median, Range/Q3, Kur- tosis and Q4/Q(2 + 3) Slope Calculations. Additional file 7. Rank-Ordered Distribution of White Blood Cells Among 35 Control Samples. Additional file 8. STRING db Analysis of Intra-individual Positional Gene Rankings In 9 Archived Controls Based on Range/Median, Range/Q3, Kur- tosis and Q4/Q(2 + 3) Slope Calculations. Additional file 9. STRING db Analysis of Intra-individual Positional Gene Rankings In 12 Archived Control Samples Based On Range/Median, Range/Q3, Kurtosis and Q4/Q(2 + 3) Slope Calculations. Additional file 10. Reference Genes with Raw Counts Greater than 5 and R2 Values > 0.9 in three separate RNA sequencing Studies. Additional file 11. Control Reference Data Files. Additional file 11. Control Reference Data Files. Funding “Not applicable”. FDR: False Discovery Rate; GO: Gene Ontology; MVA: Minimum Value Adjustments; RNA-seq: RNA sequencing; SD: Standard Deviation; CV: Coefficient of Variation; db: Database; WBC: White blood cells; DGE: Differential gene expression Acknowledgements We acknowledge the We acknowledge the assistance of Dr. Xiang Zhang and his staff at the Genomics, Epigenomics and Sequencing Core, University of Cincinnati for their assistance in providing the RNA sequencing results used in this report. Gene pathway associations identified by comparing similarities in Trendline characteristics Statistical calculations were performed using the resi- dent statistical macros provided with Windows 10 Excel software. A variety of statistical calculations were performed on the trendline of each gene. Our results were based on a list of 8746 protein-coding genes with a mean baseline Page 17 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 count ≥0.5 counts that were rank-ordered to identify genes with the most prominent trendline property (e.g. range/Q3, kurtosis, R2, etc.). Prominent genes consid- ered for additional analysis fell among the top 300 genes characterized by the largest or smallest measurements for any of the computed parameters. For each computed statistical parameter, the identified genes were imported into the STRING database (db) to determine if that parameter linked genes to functionally related protein associations. For example, 300 genes that displayed the largest computed means were identified and imported into the STRING database version 11 [13–15] to deter- mine if the highlighted genes formed known associations within any biological pathway(s). STRING-db analysis was performed using the “high confidence setting of 0.7”. Using various statistical tests, the STRING-db iden- tified the Number of Observed Gene Counts within the original list of 300 submitted genes that were associated with known GO pathways and assigned a False Discov- ery Rate based on the Observed Gene Count and the total number of known genes in the pathway. Al- though algorithms employed to compute False Dis- covery Rates in the STRING-db and Panther-db [13– 15, 42, 43] databases are different, False Discovery Rates < E-15 are generally considered significant and the identified Observed Gene Count groupings cannot not be attributed to random gene associations (see file 2 for additional detail). Availability of data and materials C f h h d d y Copies of the three processed unadjusted data files that were used in this study are provided in Additional file 11. Sheet 1 contains our control data files and sheets 2 and 3 contain the Control data files from the Mangul et.al [24]. and Mo et.at. reports [25] that were processed through our pipeline. The original RNA sequencing data files for the 35 controls described in this study have been archived at the NCBI Gene Expression Omnibus site with accession number GSE169359 (https://www.ncbi.nlm.nih.gov/geo/query/acc. cgi?acc=GSE169359). The additional archived data files referenced in our study are available at NCBI Gene Expression Omnibus site (https://www.ncbi. nlm.nih.gov/geo/) containing sequencing data obtained from the peripheral blood of 9 Controls (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc= GSE109313) [24] in one data set and 12 Controls in a second data set (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE112057) [25]. References 22. The Gene Ontology Consortium. Expansion of the Gene Ontology knowledgebase and resources. Nucleic Acids Res. 2017;45(D1):D331–8. https://doi.org/10.1093/nar/gkw1108. 1. Conesa A, Madrigal P, Tarazona S, Gomez-Cabrero D, Cervera A, McPherson A, et al. A survey of best practices for RNA-seq data analysis. Genome Biol. 2016;17(1):13. https://doi.org/10.1186/s13059-016-0881-8. 1. Conesa A, Madrigal P, Tarazona S, Gomez-Cabrero D, Cervera A, McPherson A, et al. A survey of best practices for RNA-seq data analysis. Genome Biol. 2016;17(1):13. https://doi.org/10.1186/s13059-016-0881-8. 23. Eghbalnia HR, Wilfinger WW, Mackey K, Chomczynski P. Coordinated analysis of exon and intron data reveals novel differential gene expression changes. Sci Rep. 2020;10(1):15669. https://doi.org/10.1038/ s41598-020-72482-w. 2. Robinson MD, Oshlack A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol. 2010;11(3):R25. https:// doi.org/10.1186/gb-2010-11-3-r25. 2. Robinson MD, Oshlack A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol. 2010;11(3):R25. https:// doi.org/10.1186/gb-2010-11-3-r25. 3. Bullard JH, Purdom E, Hansen KD, Dudoit S. Evaluation of statistical methods for normalization and differential expression in mRNA-Seq experiments. BMC Bioinformatics. 2010;11(1). https://doi.org/10.1186/1471-2105-11-94. 3. Bullard JH, Purdom E, Hansen KD, Dudoit S. Evaluation of statistical methods for normalization and differential expression in mRNA-Seq experiments. BMC Bioinformatics. 2010;11(1). https://doi.org/10.1186/1471-2105-11-94. 24. Mangul S, Yang HT, Strauli N, Gruhl F, Porath HT, Hsieh K, et al. ROP: dumpster diving in RNA-sequencing to find the source of 1 trillion reads across diverse adult human tissues. Genome Biol. 2018;19(1):36. https://doi. org/10.1186/s13059-018-1403-7. 4. Dillies MA, Rau A, Aubert J, Hennequet-Antier C, Jeanmougin M, Servant N, et al. A comprehensive evaluation of normalization methods for Illumina high-throughput RNA sequencing data analysis. Brief Bioinform. 2013;14(6): 671–83. https://doi.org/10.1093/bib/bbs046. 25. Mo A, Marigorta UM, Arafat D, Chan LHK, Ponder L, Jang SR, et al. Disease- specific regulation of gene expression in a comparative analysis of juvenile idiopathic arthritis and inflammatory bowel disease. Genome Med. 2018; 10(1):48. https://doi.org/10.1186/s13073-018-0558-x. 5. Zyprych-Walczak J, Szabelska A, Handschuh L, Górczak K, Klamecka K, Figlerowicz M, et al. The impact of normalization methods on RNA-Seq data Analysis. Biomed Res Int. 2015;2015:1–10. https://doi.org/10.1155/2 015/621690. 26. Zhang X, Kluger Y, Nakayama Y, Poddar R, Whitney C, DeTora A, et al. Gene expression in mature neutrophils: early responses to inflammatory stimuli. J Leuk Biol. 2004;75(2):358–72. https://doi.org/10.1189/jlb.0903412. 6. Rapaport F, Khanin R, Liang Y, Pirun M, Zumbo P, Mason CE, et al. Comprehensive evaluation of differential gene expression analysis methods for RNA-seq data. Genome Biol. 2013. https://doi.org/10.1186/ gb-2013-14-9-r95. 27. References Jiang K, Sun X, Chen Y, Shen Y, Jarvis JN. RNA sequencing from human neutrophils reveals distinct transcriptional differences associated with chronic inflammatory states. BMC Med Genet. 2015;8(1):55. https://doi.org/1 0.1186/s12920-015-0128-7. 7. McIntyre LM, Lopiano KK, Morse AM, Amin V, Oberg AL, Young LJ, et al. RNA-seq: technical variability and sampling. BMC Genomics. 2011;12(1). https://doi.org/10.1186/1471-2164-12-293. 28. Schoggins JW, Rice CM. Interferon-stimulated genes and their antiviral effector functions. Curr Opin Virol. 2011;1(6):519–25. https://doi.org/10.1016/ j.coviro.2011.10.008. 8. Hansen K, Wu Z, Irizarry R, Leek J. Sequencing technology does not eliminate biological variability. Nat Biotechnol. 2011;29(7):572–3. https://doi. org/10.1038/nbt.1910. 29. Schreiber G. The molecular basis for differential type 1 interferon signaling. JBC. 2017;292(18):7285–94. https://doi.org/10.1074/jbc.R116.774562. 30. Schoggins JW. Interferon-stimulated genes: what do they all do? Ann Rev Virol. 2019;6(1):567–84. https://doi.org/10.1146/annurev-virology-092 818-015756. 9. van den Berg RA, Hoefsloot HC, Westerhuis JA, Smilde AK, van der Werf MJ. Centering, scaling, and transformations: improving the biological information content of metabolomics data. BMC Genomics. 2006;7(1). https://doi.org/10.1186/1471-2164-7-142. 31. Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, Marshall KA, Phillippy KH, Sherman PM, Holko M, Yefanov A, Lee H, Zhang N, Robertson CL, Serova N, Davis S, Soboleva A. NCBI GEO: archive for functional genomics data sets--update. Nucleic Acids Res. 2013 Jan; 41(Database issue):D991-5. Gene Expression Omnibus database, a public functional data repository. (https://www.ncbi.nlm.nih.gov/sites/GDSbrowser/). 10. Chomczynski P, Wilfinger WW, Kennedy A, Rymaszewski M, Mackey K. RNAzol©BD: a reagent for the effective isolation of RNA from whole blood. Nat Methods. 2013;10(5):ii. https://doi.org/10.1038/nmeth.f.364. 11. Chomczynski P, Wilfinger WW, Eghbalnia HR, Kennedy A, Rymaszewski M, Mackey K. Inter-individual differences in RNA levels in human peripheral blood. PLoS One. 2016;11(2):e0148260. https://doi.org/10.1371/journal.pone.0148260. 32. Charlotte S, Mauro D. A comparison of methods for differential expression analysis on RNA-seq data. BMC Bioinformatics. 2013. https://doi.org/10.1093/ bioinformatics/btt087. 12. Miller R. RAnGER© (Rank, Analyze, Graph and Evaluate Results), 2021 (http:// www.rmiller-llc.com). 33. Ning L, Dawson John A, Thomson James A, Victor R, Rissman Anna I, Smits Bart MG, et al. EBSeq: an empirical bayes hierarchical model for inference in RNA-seq experiments. Bioinformatics. 2013. https://doi.org/10.1093/ bioinformatics/btt087. 13. Protein-Protein Interaction Networks Functional Enrichment Analysis, STRI NG-db (https://string-db.org/cgi/input?sessionId=bpH5iTmmOFi1&input_pa ge_active_form=multiple_identifiers), Version 11.0. 14. Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, et al. STRI NG v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–13. https://doi.org/10.1093/nar/gky1131. 34. George NI, Bowyer JF, Crabtree NM, Chang CW. Declarations 15. Szklarczyk D, Morris JH, Cook H, Kuhn M, Wyder S, Simonovic M, et al. The STRING database in 2017: quality-controlled protein-protein association networks, made broadly accessible. Nucleic Acids Res. 2017. https://doi. org/10.1093/nar/gkw937. Competing interests The authors declare that they have no competing interests. 19. Jaffe AE, Irizarry RA. Accounting for cellular heterogeneity is critical in epigenome-wide association studies. Genome Biol. 2014;15(2):R31. https:// doi.org/10.1186/gb-2014-15-2-r31. Abbreviations FDR F l Di Page 18 of 19 Page 18 of 19 Page 18 of 19 Page 18 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics Received: 5 November 2020 Accepted: 29 March 2021 Received: 5 November 2020 Accepted: 29 March 2021 21. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000. https://doi.org/10.1038/75556. Consent for publication Consent for publication “Not applicable”. 18. Preininger M, Arafat D, Kim J, Nath AP, Idaghdour Y, Brigham KL, et al. Blood-informative transcripts define nine common axes of peripheral blood gene expression. PLoS Genet. 2013;9(3):e1003362. https://doi.org/10.1371/ journal.pgen.1003362. “Not applicable”. Author details 1 1Molecular Research Center, Inc., Cincinnati, USA. 2Robert Miller Enterprises, LLC, Cincinnati, USA. 3University of Wisconsin-Madison, Madison, USA. 4University of Cincinnati, Cincinnati, USA. 20. Carbon S, Ireland A, Mungall CJ, Shu S, Marshall B, Lewis S, AmiGO Hub, Web Presence Working Group. AmiGO: online access to ontology and annotation data. Bioinformatics. 2009;25(2):288-9 The Gene Ontology Resource (http://geneontology.org). Ethics approval and consent to participate The blood samples reported in this study were collected in accordance with the approved protocol provided by the Chesapeake Research Review, LLC. CIRBI Protocol # Pro 00009509. Participants provided witnessed signed informed consent documents specifically approved by Chesapeake IRB guidelines. The approval for the other archived data cited in this report is included in the original publications. 16. Whitney AR, Diehn M, Popper SJ, Alizadeh AA, Boldrick JC, Relman DA, et al. Individuality and variation in gene expression patterns in human blood. PNAS. 2003;100(4):1896–901. https://doi.org/10.1073/pnas.252784499. 17. Savelyeva AV, Kuligina EV, Bariakin DN, Kozlov VV, Ryabchikova EI, Richter VA, et al. Variety of RNAs in peripheral blood cells, plasma, and plasma fractions. Biomed Res Int. 2017;2017:1–10. https://doi.org/10.11 55/2017/7404912. References 2010. https://doi.org/10.1126/science.1197005. 44. Sudmant PH, Kitzman JO, Antonacci F, Alkan C, Malig M, Tsalenko A, et al. Diversity of human copy number variation and multicopy genes. Science. 2010. https://doi.org/10.1126/science.1197005. References An iterative leave-one-out approach to outlier detection in RNA-Seq data. PLoS One. 2015;10(6): e0125224. https://doi.org/10.1371/journal.pone.0125224. Page 19 of 19 Page 19 of 19 Wilfinger et al. BMC Genomics (2021) 22:322 Wilfinger et al. BMC Genomics (2021) 22:322 35. Lin B, Li-Feng Z, Xin C. LFCseq: a nonparametric approach for differential expression analysis of RNA-seq data. BMC Genomics. 2014;15(Suppl 10):S7. https://doi.org/10.1186/1471-2164-15-S10-S7. 36. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11(10):R106. https://doi.org/10.1186/gb-2010-11-10-r106. 37. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. https://doi.org/10.1186/s13059-014-0550-8. 38. Gilchrist WG. Statistical Modelling with Quantile Functions. London: Chapman & Hall/CRC. 2000. p. 1–52. 39. Anisur R, Abhinav T, Jatin N. Hickling timothy. importance of feedback and feedforward loops to adaptive immune response modeling. CPT Pharmacometr Syst Pharmacol. 2018. https://doi.org/10.1002/psp4.12352. 39. Anisur R, Abhinav T, Jatin N. Hickling timothy. importance of feedback and feedforward loops to adaptive immune response modeling. CPT Pharmacometr Syst Pharmacol. 2018. https://doi.org/10.1002/psp4.12352. 40. Wylie KM, Mihindukulasuriya KA, Zhou Y, Sodergren E, Storch GA, Weinstock GM. Metagenomic analysis of double-stranded DNA viruses in healthy adults. BMC Biol. 2014;12(1):71. https://doi.org/10.1186/s12915-014-0071-7. 40. Wylie KM, Mihindukulasuriya KA, Zhou Y, Sodergren E, Storch GA, Weinstock GM. Metagenomic analysis of double-stranded DNA viruses in healthy adults. BMC Biol. 2014;12(1):71. https://doi.org/10.1186/s12915-014-0071-7. 41. Stamova BS, Apperson M, Walker WL, Tian Y, Xu H, Adamczy P, et al. Identification and validation of suitable endogenous reference genes for gene expression studies in human peripheral blood. BMC Med Genet. 2009; 2(1). https://doi.org/10.1186/1755-8794-2-49. 42. Attrill H, Gaudet P, Huntley RP, Lovering RC, Engel SR, Poux S, Van Auken KM, et al. The Gene Ontology Consortium, Annotation of gene product function from high-throughput studies using the Gene Ontology, Database, Volume 2019, 2019, baz007, https://doi.org/10.1093/database/baz007 Panther Classification System Tool for grafting sequences (http://www.pantherdb.org/). 43. Mi H, Muruganujan A, Huang X, Ebert D, Mills C, Guo X, et al. Protocol Update for large-scale genome and gene function analysis with the PANT HER classification system (v.14.0). Nat Protoc. 2019. https://doi.org/10.1038/ s41596-019-0128-8. 43. Mi H, Muruganujan A, Huang X, Ebert D, Mills C, Guo X, et al. Protocol Update for large-scale genome and gene function analysis with the PANT HER classification system (v.14.0). Nat Protoc. 2019. https://doi.org/10.1038/ s41596-019-0128-8. 44. Sudmant PH, Kitzman JO, Antonacci F, Alkan C, Malig M, Tsalenko A, et al. Diversity of human copy number variation and multicopy genes. Science. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
https://openalex.org/W4360985341	https://www.frontiersin.org/articles/10.3389/fnins.2023.1130831/pdf	English	null	A machine learning model for visualization and dynamic clinical prediction of stroke recurrence in acute ischemic stroke patients: A real-world retrospective study	Frontiers in neuroscience	2,023	cc-by	7,896	A machine learning model for visualization and dynamic clinical prediction of stroke recurrence in acute ischemic stroke patients: A real-world retrospective study OPEN ACCESS EDITED BY Ming Li, Hong Kong Polytechnic University, Hong Kong SAR, China REVIEWED BY Xiaofei Hu, Army Medical University, China Xinyu Yu, Huazhong University of Science and Technology, China Dingkang Xu, Chinese Academy of Medical Sciences and Peking Union Medical College, China *CORRESPONDENCE Wenle Li drlee0910@163.com Liangqun Rong rongliangqun@163.com Xiu’e Wei wxeqq@163.com †These authors have contributed equally to this work and share ﬁrst authorship SPECIALTY SECTION This article was submitted to Neural Technology, a section of the journal Frontiers in Neuroscience RECEIVED 23 December 2022 ACCEPTED 27 February 2023 PUBLISHED 27 March 2023 CITATION Wang K, Shi Q, Sun C, Liu W, Yau V, Xu C, Liu H, Sun C, Yin C, Wei X, Li W and Rong L (2023) A machine learning model for visualization and dynamic clinical prediction of stroke recurrence in acute ischemic stroke patients: A real-world retrospective study. Front. Neurosci. 17:1130831. doi: 10.3389/fnins.2023.1130831 1Department of Neurology, The Second Afﬁliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu, China, 2Key Laboratory of Neurological Diseases, The Second Afﬁliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu, China, 3State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics & Center for Molecular Imaging and Translational Medicine, School of Public Health, Xiamen University, Xiamen, China, 4Department of Neurosurgery, The Second Afﬁliated Hospital of Soochow University, Suzhou, China, 5Department of Orthopaedic Surgery, The First Afﬁliated Hospital of Nanchang University, Nanchang, China, 6Division of Oral and Maxillofacial Surgery, Columbia University Irving Medical Center, New York, NY, United States, 7Department of Dermatology, Xianyang Central Hospital, Xianyang, China, 8Faculty of Medicine, Macau University of Science and Technology, Macau, China Background and purpose: Recurrent stroke accounts for 25–30% of all preventable strokes, and this study was conducted to establish a machine learning-based clinical predictive rice idol for predicting stroke recurrence within 1 year in patients with acute ischemic stroke (AIS). Wang K, Shi Q, Sun C, Liu W, Yau V, Xu C, Liu H, Sun C, Yin C, Wei X, Li W and Rong L (2023) A machine learning model for visualization and dynamic clinical prediction of stroke recurrence in acute ischemic stroke patients: A real-world retrospective study. Front. Neurosci. 17:1130831. doi: 10.3389/fnins.2023.1130831 Methods: A total of 645 AIS patients at The Second Afﬁliated Hospital of Xuzhou Medical University were screened, included and followed up for 1 year for comprehensive clinical data. TYPE Original Research PUBLISHED 27 March 2023 DOI 10.3389/fnins.2023.1130831 TYPE Original Research PUBLISHED 27 March 2023 DOI 10.3389/fnins.2023.1130831 1. Introduction application of models to predict recurrent stroke using regression or other statistical methods is often limited by the narrow range of variables (Chaudhary et al., 2019) as studies have shown that the area under the receiver operating characteristic (ROC) curve for multivariate logistic models developed using clinical and retinal characteristics for recurrent stroke within 1 year is 0.71–0.74 [higher area under curve (AUC) values indicate better model predictive power] (Yuanyuan et al., 2020). Similarly, when machine learning (ML) is used with single- or multi-omics medical data, more details can be mined from the data and better diagnostic and prognostic tools can be developed compared to traditional statistical regression models (Bersanelli et al., 2016; Erickson et al., 2017; Dias-Audibert et al., 2020; Fleuren et al., 2020). Studies have demonstrated that ML can successfully predict favorable outcomes for up to 3 months after acute stroke event and that the area under the curve of deep neural network models is signiﬁcantly higher than the Astral score (0.888 vs. 0.839; P < 0.001) (Heo et al., 2019). In addition, ML can be used to eﬃciently label stroke patients in the emergency setting to facilitate triage (Abedi et al., 2020), as well as to monitor predictive models for long-term recurrent stroke (5 years) by using six of its algorithmic models (Abedi et al., 2021). Stroke is characterized by acute focal injury of the central nervous system (i.e., brain, retina, or spinal cord) resulting in neurological dysfunction due to sudden rupture of blood vessels or obstruction of blood ﬂow. It is categorized into ischemic and hemorrhagic stroke, while the incidence of the former is higher than that of the latter, accounting for 60–70% of all strokes (Sacco et al., 2013). The major clinical manifestation of stroke is the sudden onset of focal neurological deﬁcits, whose clinical diagnosis is further complemented with imaging of the brain and its vascular trees (Campbell and Khatri, 2020). Epidemiological data suggest that stroke is the second leading cause of death and disability worldwide, causing tremendous burden shared by low- and middle- income countries (Saini et al., 2021). As suggested by 2016 global burden of disease data that one in four people will have a stroke in their lifetime (GBD 2016 Neurology Collaborators, 2019), its prevalence is approximately equal in men and women. A machine learning model for visualization and dynamic clinical prediction of stroke recurrence in acute ischemic stroke patients: A real-world retrospective study Univariate and multivariate logistic regression (LR) were used to screen the risk factors of stroke recurrence. The data set was randomly divided into training set and test set according to the ratio of 7:3, and the following six prediction models were established by machine algorithm: random forest (RF), Naive Bayes model (NBC), decision tree (DT), extreme gradient boosting (XGB), gradient boosting machine (GBM) and LR. The model with the strongest prediction performance was selected by 10-fold cross-validation and receiver operating characteristic (ROC) curves, and the models were investigated for interpretability by SHAP. Finally, the models were constructed to be visualized using a web calculator. COPYRIGHT © 2023 Wang, Shi, Sun, Liu, Yau, Xu, Liu, Sun, Yin, Wei, Li and Rong. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: Logistic regression analysis showed that right hemisphere, homocysteine (HCY), C-reactive protein (CRP), and stroke severity (SS) were independent risk factors for the development of stroke recurrence in AIS patients. In 10-fold cross- validation, area under curve (AUC) ranked from 0.777 to 0.959. In ROC curve analysis, AUC ranged from 0.887 to 0.946. RF model has the best ability to predict stroke recurrence, and HCY has the largest contribution to the model. A web- 01 Frontiers in Neuroscience frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 based calculator https://mlmedicine-re-stroke2-re-stroke2-baylee.streamlitapp. com/ has been developed accordingly. Conclusion: This study identiﬁed four independent risk factors affecting recurrence within 1 year in stroke patients, and the constructed RF-based prediction model had good performance. 1. Introduction However, incidence of stoke is higher in older women (more than 50% higher comparing to men aged 75 years or older), and among some races (e.g., 1.91 per 1,000 in Black or African American and 0.88 per 1,000 in Caucasians) (Virani et al., 2020). Data from 2010 to 2017 showed a 5.3% increase in stroke morbidity and mortality and a 19.3% increase in prevalence, respectively (Goldstein, 2020). In the current study, we constructed six diﬀerent prediction models by adding observational indicators and explored factors inﬂuencing recurrence in all stroke patients based on 1-year follow-up data, evaluated their performances based on sensitivity, speciﬁcity, accuracy, and subject operating characteristic curve (ROC), and analyzed the relative importance and interpretability of diﬀerent factors on the models. We aimed to provide a reference for identifying stroke patients at high risk of recurrence, which is conducive to early diagnosis, and treatment of stroke recurrence, leading to improved survival and recovery of patients. Despite eﬀective treatment approaches, stroke patients are still at measurable risk of recurrent episodes after initial recovery. Recurrent strokes account for 25–30% of all preventable strokes, a majority of which are ischemic strokes, and their onset lead to a higher mortality and disability rate than the initial episode (Luengo-Fernandez et al., 2012). The risk rates for early recurrence of ischemic stroke are approximately 5% at 7 days, and 10% at 14 days, respectively; the long-term recurrence risk rates are approximately 11.1% [95% conﬁdence interval (CI) 9–0 to 13.3] at 1 year (Hankey, 2014). Therefore, the identiﬁcation of risk factors for stroke recurrence is beneﬁcial to identify populations of high-risk recurrence, ensuring early intervention to reduce the morbidity and mortality. Previous studies have shown that pathophysiological factors and lifestyle factors are all inﬂuential factors leading to stroke recurrence. In addition, history of previous cerebrovascular events and stroke subtypes are also important risk factors for recurrence (Chin et al., 2018). Therefore, it is crucial to develop predictive models for eﬀective secondary prevention and management. Frontiers in Neuroscience KEYWORDS stroke, recurrence, machine learning, SHAP, web calculator stroke, recurrence, machine learning, SHAP, web calculator 2.1. Data sources, inclusion criteria, exclusion criteria The data of this study were obtained from patients who were diagnosed with acute ischemic stroke (AIS) at The Second Aﬃliated Hospital of Xuzhou Medical University from August 2017 to July 2019. The inclusion criteria for AIS patients included: ischemic stroke diagnoses following the World Health Organization criteria and from onset of symptom to hospitalization less than 24 h. Exclusion criteria were as follows: (1) incomplete With rapid development of precision medicine in the recent years, data science and predictive analytics take on signiﬁcant roles for physicians to deliver individualized care. However, clinical 02 frontiersin.org frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 algorithm (RF), gradient augmentation algorithm (GBM), and xgboost (XGB) algorithm. Based on the training set data, average AUC values were calculated and the accuracy of the ML-based model algorithms was veriﬁed using the 10-fold crossover method. In addition, the ROC curves of various algorithmic models under the test set were plotted for external validation, while radar plots characterizing sensitivity, accuracy, and speciﬁcity of the models were provided to comprehensively evaluate performance of the models. The algorithm presenting the highest average AUC value was selected as the optimal algorithm. Then, contribution of each variable in the optimal model were calculated by the interpretable model SHAP to determine importance of the variables and the positive or negative contribution to the model. Finally, a web calculator was built on this basis to enable input of patient data and disease prediction to help physicians assess the risk of stroke recurrence within 1 year. clinical information. (2) Patients with severe organ dysfunction. (3) Inadequate ancillary tests. (4) Follow-up time for less than 1 year. (5) Patients with disturbance of consciousness and severe aphasia. The study was approved by the Ethics Committee of The Second Aﬃliated Hospital of Xuzhou Medical University [ethics number: (2020) 081603], and all patients signed a written consent form. 2.2. Statistical methods The collected clinicopathological and biochemical data were subjected to statistical analysis and model construction using R (version4.0.5)1 and Python (version3.8). Firstly, based on various types of data, continuous variables were expressed as mean ± standard deviation and compared using an independent samples t-test; categorical variables were expressed as frequency (percentage, %) and analyzed with χ2 test. Logistic regression (LR) analysis was used to identify risk factors independently associated with stroke recurrence. Variables with P-values less than 0.05 in the results of univariate LR analysis were included in multivariate LR analysis. Finally, factors with a P-value < 0.05 in the results of multivariate LR analysis were identiﬁed as independent risk factors for stroke recurrence, and the odds ratio (OR) and 95% CI were calculated for each variable. 2.3. Model building and validation In our study, prediction models based on six diﬀerent ML algorithms were used to analyze our data: linear regression algorithm (LR), plain Bayesian classiﬁcation algorithm Naive Bayes model (NBC), decision tree algorithm (DT), random forest 3.1. Baseline population characteristics A total of 832 patients with AIS were included in this study, and 48 patients with incomplete clinical data and inadequate ancillary tests, 64 patients with disturbance of consciousness and severe aphasia, 18 patients with severe abnormal organ function, and 57 patients with less than 1 year of follow-up for various reasons were excluded. The ﬁnal 645 patients with AIS with or without stroke recurrence within 1 year were included (Figure 1). A total of 84 patients experienced recurrent stroke. The rate of stroke recurrence was 13%. Table 1 shows that diﬀerences in the side of hemisphere (SOH), HCY, CRP, NSE, S100β, anticoagulation, PPI, dysphagia, and SS were statistically signiﬁcant in the presence or absence of stroke recurrence. It is suggested that these variables may be independent risk factors for stroke recurrence. Frontiers in Neuroscience 3.2. Univariate and multivariate logistic regression Univariate logistics regression analysis showed that right hemisphere, HCY, CRP, NSE, S100β, Anticoagulation, PPI, dysphagia, and SS all had statistically signiﬁcant correlation with the occurrence of stroke recurrence. The nine variables mentioned above had P < 0.05 after univariate analysis were included in a multivariate logistics regression analysis, suggesting that right hemisphere, HCY, CRP, and SS, might be independent risk factors for stroke recurrence. The diﬀerences were all statistically signiﬁcant [right (OR: 0.25, 95% CI 0.09–0.73, P < 0.05), HCY (1.6, 1.44–1.79, P < 0.05), CRP (1.06, 1.01–1.11, P < 0.05) and SS (3.98, 1.98–7.97, P < 0.001)]. 1 https://www.r-project.org/ 2.1. Data sources, inclusion criteria, exclusion criteria Data of enrolled patients were collected, including demographic data (age, sex); vascular risk factors (hypertension, diabetes, ischemic heart disease); baseline blood pressure [SBP and diastolic blood pressure (DBP)]; trial of org 10 172 in acute stroke treatment (TOAST) (large-artery atherosclerosis, cardioembolism, small-vessel occlusion, acute stroke of other determined etiology, stroke of undetermined etiology); stroke severity (SS) [based on the National Institutes of Health NIHSS, NIHSS score ≤8 for mild stroke, NIHSS score ≥9 for moderate to severe stroke (Muchada et al., 2014), all assessments completed on admission]; MRI readings [stroke distribution (anterior circulation, posterior circulation, anterior/posterior circulation), hemispheric laterality (left, right, bilateral), number of stroke lesions (single, multiple); stroke lesion site (cortical, cortico-subcortical, subcortical, subcortical, subcortical) (subcortical, subcortical, brainstem and cerebellum)], laboratory tests [total cholesterol, triglycerides, low-density lipoprotein (LDL), fasting blood glucose (FBG), homocysteine (HCY), uric acid (UA), ﬁbrinogen (FIB), myoglobin (MB), C-reactive protein (CRP), d-dimer, brain natriuretic peptide (BNP), HBALC, neuronal speciﬁc enolase (NSE), S- 100β], and clinical treatment [thrombolysis, thrombectomy, antiplatelet, anticoagulation, statin, pump inhibitor therapy (PPI)]. complications of stroke [dysphagia (Banda et al., 2022), stroke-associated pneumonia (SAP) (Qiu et al., 2022)]. 3.3. Model building and performance validation Four signiﬁcantly diﬀerent factors were deﬁned as variables of the model by single-factor and multi-factor screening, and the 03 frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 FIGURE 1 Flowchart of the whole experiment. 04 Frontiers in Neuroscience frontiersin.org frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 TABLE 1 Baseline demographic and clinicopathological characteristics of all patients (stroke recurrence group and non-stroke recurrence group). TABLE 1 Baseline demographic and clinicopathological characteristics of all patients (stroke recurrence group and non-stroke recurrence group). NA, not available; SD, stroke distribution; SOH, side of hemisphere; NOS, number of stroke lesions; SOS, site of stroke lesions; LDL, low-density lipoprotein; FBG, fasting blood glucose; HCY, homocysteine; UA, uric acid; FIB, ﬁbrinogen; MB, myoglobin; CRP, C-reactive protein; BNP, brain natriuretic peptide; NSE, neuron-speciﬁc enolase; PPI, proton pump inhibitor therapy; SS, stroke severity, SAP, stroke-associated pneumonia; DBP, diastolic blood pressure. Frontiers in Neuroscience Frontiers in Neuroscience 05 frontiersin.org 3.3. Model building and performance validation 10.3389/fnins.2023.1130831 model was initially constructed by randomly slicing original data set into a training set and a test set with the ratio of 7:3, with 452 patients in training set and 193 patients in test set. Six ML algorithms including linear regression algorithm (LR), plain Bayesian classiﬁcation algorithm (NBC), decision tree algorithm (DT), random forest algorithm (RF), GBM, and xgboost (XGB) algorithm were executed in the training set to build a prediction model. To avoid training chance errors, the performance of the models in ﬁtting the original data was measured, visually compared. To improve the model prediction eﬀectiveness, a 10-fold cross- validation method was used for internal validation (Figure 2). The results showed that the RF model was the best predictor of stroke recurrence (mean AUC 0.959, standard deviation 0.017) and the LR model had the lowest AUC value of 0.777. FIGURE 3 Receiver operating characteristic (ROC) curve of machine algorithm model under the test set. In addition, the ROC curves of various algorithmic models under the test set (Figure 3) were plotted for external validation in this study, where the RF algorithm prediction model had the largest area under the AUC curve (AUC = 0.946), the LR algorithm prediction model had the lowest (AUC = 0.887), and the other models were in between, indicating that the RF model had a better performance in terms of data ﬁtting eﬀect. The radar plot of prediction model sensitivity and speciﬁcity showed (Figure 4) that the LR and RF models had better accuracy and sensitivity in their prediction ability, while the NBC model had higher speciﬁcity. However, the RF model approach was more eﬀective (F1 = 0.585) when considering the precision and recall rates (F1) together. Finally, the performance of the six algorithmic models was compiled into a table (Table 3). Therefore, we chose the RF model as the ﬁnal prediction model. Receiver operating characteristic (ROC) curve of machine algorithm model under the test set. FIGURE 4 Radar plots of sensitivity and speciﬁcity of 6 machine algorithm models. 3.3. Model building and performance validation Characteristics Level Overall (N = 645) No (N = 561) Yes (N = 84) P Age, n (%) < 60 362 (56.1) 311 (55.4) 51 (60.7) 0.429 ≥60 283 (43.9) 250 (44.6) 33 (39.3) Gender, n (%) Female 263 (40.8) 233 (41.5) 30 (35.7) 0.372 Male 382 (59.2) 328 (58.5) 54 (64.3) SBP, median (Q1, Q3) NA 143.0 (132.0, 156.0) 143.0 (132.0, 156.0) 145.0 (132.0, 157.0) 0.705 DBP, median (Q1, Q3) NA 87.0 (74.0, 97.0) 86.0 (74.0, 97.0) 88.0 (75.8, 96.0) 0.466 SD, n (%) Anterior circulation 258 (40.0) 220 (39.2) 38 (45.2) 0.482 Posterior circulation 235 (36.4) 209 (37.3) 26 (31.0) Anterior/Posterior circulation 152 (23.6) 132 (23.5) 20 (23.8) SOH, n (%) Left 271 (42.0) 228 (40.6) 43 (51.2) 0.014 Right 256 (39.7) 221 (39.4) 35 (41.7) Bilateral 118 (18.3) 112 (20.0) 6 (7.1) SOS, n (%) Cortex 149 (23.1) 131 (23.4) 18 (21.4) 0.716 Cortex-subcortex 147 (22.8) 123 (21.9) 24 (28.6) Subcortex 176 (27.3) 155 (27.6) 21 (25.0) Brainstem 100 (15.5) 89 (15.9) 11 (13.1) Cerebellum 73 (11.3) 63 (11.2) 10 (11.9) NOS, n (%) Single stroke lesion 453 (70.2) 390 (69.5) 63 (75.0) 0.37 Multiple stroke lesions 192 (29.8) 171 (30.5) 21 (25.0) Cholesterol, median (Q1, Q3) NA 5.3 (4.4, 6.2) 5.2 (4.4, 6.2) 5.6 (4.8, 6.0) 0.267 Triglyceride, median (Q1, Q3) NA 2.2 (1.9, 2.4) 2.2 (1.9, 2.4) 2.2 (2.1, 2.4) 0.059 LDL, median (Q1, Q3) NA 4.8 (4.3, 4.9) 4.7 (4.3, 5.0) 4.8 (4.4, 4.9) 0.772 FBG, median (Q1, Q3) NA 5.2 (4.6, 5.8) 5.2 (4.6, 5.8) 5.3 (4.6, 6.0) 0.388 HBALC, median (Q1, Q3) NA 5.6 (5.3, 5.9) 5.6 (5.3, 5.9) 5.7 (5.3, 6.0) 0.364 HCY, median (Q1, Q3) NA 15.5 (12.6, 19.1) 14.7 (12.3, 18.0) 20.7 (18.7, 22.0) < 0.001 UA, median (Q1, Q3) NA 350.1 (310.8, 407.6) 350.1 (311.3, 406.8) 349.7 (307.5, 411.9) 0.809 MB, median (Q1, Q3) NA 97.0 (74.9, 144.7) 97.7 (75.3, 146.7) 94.9 (71.9, 143.4) 0.58 CRP, median (Q1, Q3) NA 11.9 (7.5, 17.1) 11.4 (7.3, 16.6) 14.9 (9.3, 20.1) < 0.001 FIB, median (Q1, Q3) NA 4.4 (4.0, 4.8) 4.3 (4.0, 4.7) 4.4 (4.0, 4.8) 0.449 D-dimer, median (Q1, Q3) NA 174.0 (133.0, 221.0) 175.0 (134.0, 222.0) 158.5 (126.8, 201.5) 0.31 BNP, median (Q1, Q3) NA 93.0 (73.0, 162.0) 93.0 (74.0, 161.0) 86.0 (69.8, 168.5) 0.792 NSE, median (Q1, Q3) NA 16.2 (12.7, 18.6) 15.9 (12.6, 18.4) 17.7 (15.0, 19.0) 0.002 S100β, median (Q1, Q3) NA 275.0 (223.0, 289.0) 274.0 (223.0, 288.0) 283.5 (241.5, 306.0) 0.012 Thrombolysis, n (%) No 448 (69.5) 393 (70.1) 55 (65.5) 0.47 Yes 197 (30.5) 168 (29.9) 29 (34.5) Thrombectomy, n (%) No 614 (95.2) 535 (95.4) 79 (94.0) 0.584 Yes 31 (4.8) 26 (4.6) 5 (6.0) Antiplatelet, n (%) No 117 (18.1) 100 (17.8) 17 (20.2) 0.701 Yes 528 (81.9) 461 (82.2) 67 (79.8) Anticoagulation, n (%) No 553 (85.7) 495 (88.2) 58 (69.0) < 0.001 Yes 92 (14.3) 66 (11.8) 26 (31.0) Statin, n (%) No 98 (15.2) 88 (15.7) 10 (11.9) 0.461 Yes 547 (84.8) 473 (84.3) 74 (88.1) PPI, n (%) No 519 (80.5) 461 (82.2) 58 (69.0) 0.007 Yes 126 (19.5) 100 (17.8) 26 (31.0) Dysphagia, n (%) No 525 (81.4) 464 (82.7) 61 (72.6) 0.039 Yes 120 (18.6) 97 (17.3) 23 (27.4) SS, n (%) No 380 (58.9) 356 (63.5) 24 (28.6) < 0.001 Yes 265 (41.1) 205 (36.5) 60 (71.4) SAP, n (%) No 494 (76.6) 435 (77.5) 59 (70.2) 0.182 Yes 151 (23.4) 126 (22.5) 25 (29.8) NA, not available; SD, stroke distribution; SOH, side of hemisphere; NOS, number of stroke lesions; SOS, site of stroke lesions; LDL, low-density lipoprotein; FBG, fasting blood glucose; HCY, 05 Frontiers in Neuroscience frontiersin.org Wang et al. 3.4. Interpretability study of variables in the RF model Considering clinical utility, we focused on the contribution of each variable to the ﬁnal prediction outcome in the ML-based RF prediction model for AIS patients using the interpretable model SHAP. With each variable as a participant and the model output as a collaborative outcome, the contribution, or SHAP value, was FIGURE 2 Ten-fold cross-validation within the training set of the machine algorithm. FIGURE 4 Radar plots of sensitivity and speciﬁcity of 6 machine algorithm models. calculated. As can be seen in Figure 5 (left), HCY, CRP, SS, and SOH are in descending order. Meanwhile, Figure 5 (right) shows the magnitude of the four variables taking values in diﬀerent colors, with higher values corresponding to red and vice versa. The ﬁgure takes the SHAP value of zero as the origin, and the negative left and positive right indicate that the variable contributes negatively or positively to the output prediction results. Therefore, we conclude that HCY has the greatest impact on the model, HCY, SS, and CRP all contribute positively to the RF model output prediction results, and SOH contributes negatively to the RF model output prediction results. Ten-fold cross-validation within the training set of the machine algorithm. Frontiers in Neuroscience 06 frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 TABLE 2 Univariate and multivariate logistic regression analysis of patients with recurrent stroke. TABLE 2 Univariate and multivariate logistic regression analysis of patients with recurrent stroke. Frontiers in Neuroscience 3.5. Web-based calculator RF model learning utilizes large-scale, diverse datasets to build useful patterns by running complex algorithms, and has an important role in the biomedical ﬁeld for disease detection, diagnosis, prevention, and treatment. Its development leads to more accurate early diagnosis, individualized treatment, and continuous monitoring, as well as eﬀective screening for disease-related risk factors and prediction of disease recurrence (Goecks et al., 2020). Machine learning is particularly useful when datasets are too large or complex for human analysis, and/or when you want to automate the data analysis process to build reproducible and time-saving pipelines. The RF model used in the study, for example, has the advantage of knowing how important each element is to the prediction, and the individual DTs are human-readable, making them easier to train and adjust. But there are also shortcomings that are not suitable for regression and many DTs are diﬃcult to explain (Greener et al., 2022). The RF-based model performed best among the six models. Therefore, we built a web-based calculator1 to facilitate the clinical application of this predictive model (Figure 6). 3.4. Interpretability study of variables in the RF model Characteristics Univariate logistic analysis Multivariate logistic analysis OR 95% CI P-value OR 95% CI P-value Age < 60 Reference Reference ≥60 0.81 (0.50–1.28) 0.367 Gender Female Reference Reference Male 1.28 (0.80–2.08) 0.315 SBP 1 (0.99–1.01) 0.99 DBP 1.01 (0.99–1.02) 0.522 SD Anterior circulation Reference Reference Posterior circulation 0.72 (0.42–1.23) 0.23 Anterior/posterior circulation 0.88 (0.48–1.56) 0.668 SOH Left Reference Reference Reference Reference Right 0.84 (0.52–1.36) 0.482 0.62 (0.33–1.15) 0.129 Bilateral 0.29 (0.11–0.66) 0.002 0.25 (0.09–0.73) 0.011 SOS Cortex Reference Reference Cortex-subcortex 1.42 (0.73–2.78) 0.302 Subcortex 0.98 (0.50–1.95) 0.965 Brainstem 0.9 (0.39–1.99) 0.806 Cerebellum 1.16 (0.49–2.64) 0.728 NOS Single stroke lesion Reference Reference Multiple stroke lesions 0.76 (0.44–1.27) 0.309 Cholesterol 1.08 (0.91–1.29) 0.381 Triglyceride 1.69 (0.93–3.08) 0.086 LDL 1.1 (0.80–1.52) 0.544 FBG 1.06 (0.84–1.35) 0.605 HBALC 1.23 (0.70–2.16) 0.476 HCY 1.57 (1.42–1.73) < 0.001 1.6 (1.44–1.79) < 0.001 UA 1 (1.00–1.00) 0.687 MB 1 (1.00–1.00) 0.655 CRP 1.07 (1.04–1.11) < 0.001 1.06 (1.01–1.11) 0.012 FIB 1.1 (0.76–1.58) 0.622 D-dimer 1 (1.00–1.00) 0.767 BNP 1 (1.00–1.00) 0.476 NSE 1.1 (1.03–1.17) 0.005 0.98 (0.88–1.09) 0.704 S100β 1.01 (1.00–1.01) 0.016 1 (1–1.01) 0.379 Thrombolysis No Reference Reference Yes 1.24 (0.75–2.00) 0.398 (Continued) 07 frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 TABLE 2 (Continued) Characteristics Univariate logistic analysis Multivariate logistic analysis OR 95% CI P-value OR 95% CI P-value Thrombectomy No Reference Reference Yes 1.33 (0.43–3.32) 0.583 Antiplatelet No Reference Reference Yes 0.85 (0.49–1.56) 0.585 Anticoagulation No Reference Reference Reference Reference Yes 3.36 (1.96–5.68) < 0.001 1.9 (0.9–4.02) 0.094 Statin No Reference Reference Yes 1.36 (0.70–2.91) 0.377 PPI No Reference Reference Reference Reference Yes 2.07 (1.22–3.42) 0.007 0.81 (0.37–1.78) 0.598 Dysphagia No Reference Reference Reference Reference Yes 1.81 (1.05–3.03) 0.034 0.9 (0.41–1.97) 0.789 SS No Reference Reference Reference Reference Yes 4.32 (2.64–7.27) < 0.001 3.98 (1.98–7.97) < 0.001 SAP No Reference Reference Yes 1.47 (0.87–2.42) 0.149 1.47 Frontiers in Neuroscience 4. Discussion Stroke prevalence is universal, but it is more likely to occur in the middle-aged and elderly population with higher rates of disability and mortality. Most strokes are predominantly ischemic strokes caused by arterial occlusion, and stroke recurrence leads to a higher risk of death and disability than their ﬁrst occurrence (Tu et al., 2021, 2022). Therefore, there is an urgent need to identify risk factors for stroke recurrence in such patients to improve prevention, reduce recurrence and disability rates, and prolong patient survival. Traditional data mining and statistical methods usually require feature engineering to obtain eﬀective and more robust features, and then construct prediction or clustering models. With complex data and a lack of suﬃcient domain knowledge, both steps present many challenges (Miotto et al., 2018). Machine Throughout the study, we followed 645 patients with AIS, 84 of whom experienced recurrence. The rate of stroke recurrence was consistent with previous result (13% vs. 11.1%) (Hankey, 2014). After evaluating baseline characteristics of multiple clinical variables collected with stroke recurrence and identifying four independent risk factors for stroke recurrence by univariate and multivariate LR analysis (i.e., right hemisphere, HCY, CRP, and SS), 08 frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 10.3389/fnins.2023.1130831 10.3389/fnins.2023.1130831 TABLE 3 Summary of speciﬁc performance of six machine algorithm models. TABLE 3 Summary of speciﬁc performance of six machine algorithm models. As a pro-inﬂammatory marker, the inﬂammatory process has an important role in pathophysiology of ischemic stroke, and elevation of HCY is considered a risk factor for ischemic stroke (Chen et al., 2017). The data in Table 1 show that HCY was lower in patients with recurrent stroke than in patients with non- recurrent stroke (13.2 vs. 16.2, P = 0.002). Previously, it was shown that hypertension with high homocysteine (HHcy) (H-type hypertension) and CRP can increase the incidence of ischemic stroke. Later data demonstrated that recurrent ischemic stroke (RIS) is associated with advanced age, male sex, diabetes, H-type hypertension, and C-reactive protein. In contrast, controlling H-type hypertension and C-reactive protein levels reduce the risk of RIS (Zhang et al., 2016). Under normal conditions, the levels of NSE and S100β in body ﬂuids are extremely low. When neuronal injury or necrosis occurs, NSE and S100β rapidly spill from the cells into the cerebrospinal ﬂuid and enter the blood through the damaged blood-brain barrier, resulting in elevated serum NSE and S100β concentrations. 4. Discussion The levels of which reﬂect the extent of neuronal damage, so elevated NSE and S100β suggest possible relapse. In addition, other serum biomarkers, such as serum Copeptin levels, are associated with recurrent stroke events and are predictors of severity at admission and 1-year stroke recurrence in stroke patients (Tang et al., 2017). In addition, other biomarkers such as serum fatty acid binding protein 4 (FABP4) (Li et al., 2019), serum CXCL12 levels (Gu et al., 2016), interleukin-37 (Zhang et al., 2021), and cystatin C (Liu et al., 2021) have been reported to be associated with stroke recurrence. Scoring LR NBC DT RF GBM XGB AUC 0.887 0.923 0.923 0.946 0.924 0.930 accuracy 0.845 0.758 0.737 0.825 0.778 0.778 Sensitivity (recall) 0.880 0.731 0.719 0.814 0.754 0.766 speciﬁcity 0.630 0.926 0.852 0.889 0.926 0.852 F1 0.531 0.515 0.474 0.585 0.538 0.517 AUC, area under curve. AUC, area under curve. six ML methods were used: LR, NBC, DT, XGB, GBM, and RF. They were used to develop prediction models for individualized prediction of stroke recurrence in AIS patients (Figure 1). The results showed that the RF model had the best predictive ability, and the mean AUC of all 10-fold cross-validation results were > 0.8. After a comprehensive evaluation of the performance of the six models (Table 3), it was concluded that the RF model performed better, and the relative importance of the four risk variables in the RF model was compared from highest to lowest, HCY, CRP, SS, and SOH. Finally, an online web calculator was created to facilitate clinical application. Table 1 shows the baseline characteristics of patients with AIS. There were no signiﬁcant diﬀerences between non-stroke recurrent patients and stroke recurrent patients in terms of gender and age; however, there were signiﬁcant diﬀerences in the side of the hemisphere (SOH; P = 0.014 < 0.05), blood homocysteine (HCY; P < 0.0 01), C-reactive protein (CRP; P < 0.001), neuron-speciﬁc enolase (NSE; P = 0.002 < 0.05), central neurospeciﬁc protein (S100β; P = 0.012 < 0.05), anticoagulation therapy (P < 0.001), proton pump inhibitor therapy (PPI; P = 0.007 < 0.05), dysphagia (P = 0.039 < 0.05), and stroke severity (SS; P < 0.001). Serum C-reactive protein, an acute chronotropic reactive protein elevated in the presence of infection, is second only to HCY in RF models in terms of relative importance. 4. Discussion risk factor for stroke recurrence, has a statistically signiﬁcant diﬀerence between the baseline characteristics of patients with and without stroke recurrence.Importance of primary prevention in patients with the ﬁrst stroke and secondary prevention in recurrent stroke is stressed in the current study. Primary prevention treatment is anticoagulation for atrial ﬁbrillation, antihypertensive treatment for hypertension and controlling glucose for diabete, etc (Diener and Hankey, 2020). The primary aim of secondary prevention is to prevent or reduce the risk of recurrent stroke and to reduce the degree of disability. Eﬀective treatments include antithrombotic and anticoagulant therapy, revascularization, and implementation of structured evaluation and intervention (Hankey, 2014). Although eﬀective for secondary prevention of ischemic stroke with aspirin, increases the risk of hemorrhagic stroke, upper gastrointestinal bleeding (UGIB), and dyspepsia. Prophylactic administration of proton pump inhibitors (PPIs) may reduce the risk of these digestive symptoms (Takabayashi et al., 2015). There is evidence that some proton pump inhibitors can attenuate the antiplatelet eﬀects of clopidogrel, but after multivariate adjustment, the data show that the use of proton pump inhibitors is not associated with a signiﬁcantly increased risk of recurrent stroke or death (Juurlink et al., 2011). Again, this is consistent with the conclusions reached in this study. FIGURE 6 Online calculator for predicting stroke recurrence. with the risk of post-stroke depression (PSD) 1 year after stroke onset, and the combination of these two factors adds prognostic information to early assessment of PSD (Cheng et al., 2018). The study in Table 2 found that the diﬀerences in NSE, S100β, Anticoagulation, PPI, and dysphagia were statistically signiﬁcant in the univariate LR analysis, but the diﬀerences were not found to be statistically signiﬁcant when the above variables were included in the multivariate LR analysis, which may be due to the sample size of the study population, selection bias. SS was a good independent predictor of stroke recurrence (Table 2), with a risk ratio of 3.98 for recurrence in stroke patients, as well as being a relatively important factor in the RF model. In a 2016 study analyzing the regression after ischemic stroke and its associated factors in elderly patients, it was shown that at 12 months of stroke, moderate stroke was associated with dependency and severe stroke was associated with dependency and recurrence (Wu et al., 2016). In addition, a meta-analysis of stroke recurrence rates was recently performed in a retrospective study of patients with ﬁrst ischemic stroke. Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. 4. Discussion It is also a non-speciﬁc marker of systemic inﬂammation reﬂecting various infectious and non-infectious inﬂammatory conditions in the organism. A retrospective review showed that 26 studies reported an association of CRP with recurrent stroke, of which 12 (46%) described a positive association (McCabe et al., 2021), a result that is consistent with what we obtained. In addition to the above- mentioned control of CRP with H hypertension that reduces the risk of RIS, elevated serum Hs-CRP and HCY levels are associated HCY, CRP, NSE, and S100β are all serum proteins. HCY is a sulfur-containing amino acid in the body and an important intermediate in the metabolism of methionine and cysteine, which itself is not involved in protein synthesis. Elevated HCY has become an independent risk factor for the development of atherosclerosis. FIGURE 5 Relative importance of variables based on SHAP for RF prediction model. FIGURE 5 Relative importance of variables based on SHAP for RF prediction model. Relative importance of variables based on SHAP for RF prediction model. 09 09 Frontiers in Neuroscience frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 10.3389/fnins.2023.1130831 risk factor for stroke recurrence, has a statistically signiﬁcant diﬀerence between the baseline characteristics of patients with and without stroke recurrence.Importance of primary prevention in patients with the ﬁrst stroke and secondary prevention in recurrent stroke is stressed in the current study. Primary prevention treatment is anticoagulation for atrial ﬁbrillation, antihypertensive treatment for hypertension and controlling glucose for diabete, etc (Diener and Hankey, 2020). The primary aim of secondary prevention is to prevent or reduce the risk of recurrent stroke and to reduce the degree of disability. Eﬀective treatments include antithrombotic and anticoagulant therapy, revascularization, and implementation of structured evaluation and intervention (Hankey, 2014). Although eﬀective for secondary prevention of ischemic stroke with aspirin, increases the risk of hemorrhagic stroke, upper gastrointestinal bleeding (UGIB), and dyspepsia. Prophylactic administration of proton pump inhibitors (PPIs) may reduce the risk of these digestive symptoms (Takabayashi et al., 2015). There is evidence that some proton pump inhibitors can attenuate the antiplatelet eﬀects of clopidogrel, but after multivariate adjustment, the data show that the use of proton pump inhibitors is not associated with a signiﬁcantly increased risk of recurrent stroke or death (Juurlink et al., 2011). Again, this is consistent with the conclusions reached in this study. 5. Conclusion In conclusion, we constructed six risk prediction models for stroke recurrence in patients with AIS by machine learning algorithm (ML), introducing four independent risk factors associated with stroke recurrence (i.e., right hemisphere, HCY, CRP, and SS). Among them, we found that RF model made promising prediction, as it performed the best in both internal validation and external validation combined, with comparable accuracy, sensitivity, and speciﬁcity. It is hoped that this web-based calculator can serve as an eﬀective predictive tool to help stroke patients prevent recurrence and assist physicians in adjudication. The right hemisphere is the last relative importance in the RF model. Stroke patients experience impairments such as contralateral motor deﬁcits and interhemispheric imbalances including hyperexcitability of the contralateral hemisphere after stroke. Since the recovery of cerebral hemispheres through motor dysfunction can be achieved by increasing excitability of the aﬀected hemisphere or decreasing the excitability of the unaﬀected hemisphere, current brain treatments for stroke patients include a brain-computer interface (BCI) and transcranial magnetic stimulation (TMS) therapies to reduce mortality and alleviate the degree of disability in patients. Studies have shown that bilateral hemisphere treatment by TMS facilitates motor recovery of paralyzed hands in stroke patients (Takeuchi et al., 2009). In contrast, when patients present with bilateral focal hemispheres, there may be an interruption of the axis between the central nervous system and the gastrointestinal system, leading to secondary symptoms such as dysphagia and gastrointestinal bleeding (Schaller et al., 2006). Similarly, our study data show that dysphagia, although not an independent 4. Discussion The results of this study showed that hypertension, diabetes mellitus, atrial ﬁbrillation, transient ischemic attack, and high SS were independent risk factors for recurrence (Kolmos et al., 2021). However, there are some limitations that need to be addressed in the future. First, the ML algorithm model we developed is limited to one hospital, which may limit its widespread use in other regions. Second, the sample size of this study has some limitations and there is room for extending the follow-up period. Finally, this study is retrospective and suﬀers from the inherent data bias of retrospective studies, which may lead to bias in the data. We will conduct further multicenter and prospective studies in the future. Frontiers in Neuroscience References Abedi, V., Avula, V., Chaudhary, D., Shahjouei, S., Khan, A., Griessenauer, C. J., et al. (2021). Prediction of long-term stroke recurrence using machine learning models. J. Clin. Med. 10:1286. doi: 10.3390/jcm10061286 systematic review and meta-analysis of diagnostic test accuracy. Intensive Care Med. 46, 383–400. doi: 10.1007/s00134-019-05872-y Abedi, V., Avula, V., Chaudhary, D., Shahjouei, S., Khan, A., Griessenauer, C. J., et al. (2021). Prediction of long-term stroke recurrence using machine learning models. J. Clin. Med. 10:1286. doi: 10.3390/jcm10061286 GBD 2016 Neurology Collaborators (2019). Global, regional, and national burden of neurological disorders, 1990-2016: A systematic analysis for the global burden of disease study 2016. Lancet Neurol. 18, 459–480. Abedi, V., Khan, A., Chaudhary, D., Misra, D., Avula, V., Mathrawala, D., et al. (2020). Using artiﬁcial intelligence for improving stroke diagnosis in emergency departments: A practical framework. Ther. Adv. Neurol. Disord. 13:1756286420938962. doi: 10.1177/1756286420938962 Goecks, J., Jalili, V., Heiser, L. M., and Gray, J. W. (2020). How machine learning will transform biomedicine. Cell 181, 92–101. doi: 10.1016/j.cell.2020.03.022 Banda, K. J., Chu, H., Kang, X. L., Liu, D., Pien, L. C., Jen, H. J., et al. (2022). Prevalence of dysphagia and risk of pneumonia and mortality in acute stroke patients: A meta-analysis. BMC Geriatr. 22:420. doi: 10.1186/s12877-022-02 960-5 Goldstein, L. B. (2020). Introduction for focused updates in cerebrovascular disease. Stroke 51, 708–710. doi: 10.1161/STROKEAHA.119.024159 Greener, J. G., Kandathil, S. M., Moﬀat, L., and Jones, D. T. (2022). A guide to machine learning for biologists. Nat. Rev. Mol. Cell Biol. 23, 40–55. doi: 10.1038/ s41580-021-00407-0 Bersanelli, M., Mosca, E., Remondini, D., Giampieri, E., Sala, C., Castellani, G., et al. (2016). Methods for the integration of multi-omics data: Mathematical aspects. BMC Bioinformatics 17(Suppl. 2):15. doi: 10.1186/s12859-015-0857-9 Gu, X. L., Liu, L., Lu, X. D., and Liu, Z. R. (2016). Serum CXCL12 levels as a novel predictor of future stroke recurrence in patients with acute ischemic stroke. Mol. Neurobiol. 53, 2807–2814. doi: 10.1007/s12035-015-9151-0 Campbell, B. C. V., and Khatri, P. (2020). Stroke. Lancet 396, 129–142. doi: 10.1016/ S0140-6736(20)31179-X Hankey, G. J. (2014). Secondary stroke prevention. Lancet Neurol. 13, 178–194. doi: 10.1016/S1474-4422(13)70255-2 Chaudhary, D., Abedi, V., Li, J., Schirmer, C. M., Griessenauer, C. J., and Zand, R. (2019). Clinical risk score for predicting recurrence following a cerebral ischemic event. Front. Neurol. 10:1106. doi: 10.3389/fneur.2019.01106 Heo, J., Yoon, J. G., Park, H., Kim, Y. D., Nam, H. S., and Heo, J. H. (2019). References Machine learning-based model for prediction of outcomes in acute stroke. Stroke 50, 1263–1265. doi: 10.1161/STROKEAHA.118.024293 Chen, S., Dong, Z., Cheng, M., Zhao, Y., Wang, M., Sai, N., et al. (2017). Homocysteine exaggerates microglia activation and neuroinﬂammation through microglia localized STAT3 overactivation following ischemic stroke. J. Neuroinﬂammation 14:187. doi: 10.1186/s12974-017-0963-x Juurlink, D. N., Gomes, T., Mamdani, M. M., Gladstone, D. J., and Kapral, M. K. (2011). The safety of proton pump inhibitors and clopidogrel in patients after stroke. Stroke 42, 128–132. doi: 10.1161/STROKEAHA.110.596643 Cheng, L. S., Tu, W. J., Shen, Y., Zhang, L. J., and Ji, K. (2018). Combination of high- sensitivity c-reactive protein and homocysteine predicts the post-stroke depression in patients with ischemic stroke. Mol. Neurobiol. 55, 2952–2958. doi: 10.1007/s12035- 017-0549-8 Kolmos, M., Christoﬀersen, L., and Kruuse, C. (2021). Recurrent ischemic stroke— a systematic review and meta-analysis. J. Stroke Cerebrovasc. Dis. 30:105935. doi: 10.1016/j.jstrokecerebrovasdis.2021.105935 Chin, Y. Y., Sakinah, H., Aryati, A., and Hassan, B. M. (2018). Prevalence, risk factors and secondary prevention of stroke recurrence in eight countries from south, east and Southeast Asia: A scoping review. Med. J. Malaysia 73, 90–99. Li, B., Wu, J., Jiang, P., Li, M., Liu, Q., Cao, Y., et al. (2019). Serum fatty acid binding protein 4 is positively associated with early stroke recurrence in nondiabetic ischemic stroke. Aging 11, 1977–1989. doi: 10.18632/aging.101886 Dias-Audibert, F. L., Navarro, L. C., de Oliveira, D. N., Delaﬁori, J., Melo, C. F. O. R., Guerreiro, T. M., et al. (2020). Combining machine learning and metabolomics to identify weight gain biomarkers. Front. Bioeng. Biotechnol. 8:6. doi: 10.3389/fbioe. 2020.00006 Liu, H., Qian, S., Zhong, C., Wang, A., Peng, Y., Peng, H., et al. (2021). Predictive Value of cystatin c for stroke recurrence in patients with acute ischemic stroke. Circ. J. 85, 213–219. doi: 10.1253/circj.CJ-20-0771 Luengo-Fernandez, R., Gray, A. M., Rothwell, P. M., and Oxford Vascular Study (2012). A population-based study of hospital care costs during 5 years after transient ischemic attack and stroke. Stroke 43, 3343–3351. doi: 10.1161/STROKEAHA.112. 667204 Diener, H. C., and Hankey, G. J. (2020). Primary and secondary prevention of ischemic stroke and cerebral hemorrhage: JACC focus seminar. J. Am. Coll. Cardiol. 75, 1804–1818. doi: 10.1016/j.jacc.2019.12.072 Erickson, B. J., Korﬁatis, P., Akkus, Z., and Kline, T. L. (2017). Machine learning for medical imaging. Radiographics 37, 505–515. doi: 10.1148/rg.2017160130 McCabe, J. J., O’Reilly, E., Coveney, S., Collins, R., Healy, L., McManus, J., et al. (2021). Funding This study was supported by the Scientiﬁc Research Project of Jiangsu Health Committee (No. H2019054), the Xuzhou Science and Technology Planning Project (No. KC21220), the Science and Ethics statement The studies involving human participants were reviewed and approved by the Ethics Committee of The Second Aﬃliated 10 frontiersin.org Wang et al. 10.3389/fnins.2023.1130831 Hospital of Xuzhou Medical University. The patients/participants provided their written informed consent to participate in this study. Technology Development Fund of Aﬃliated Hospital of Xuzhou Medical University (No. XYFY202250), and the Shaanxi Provincial Health and Health Research Fund Project (No. 2022E006). Technology Development Fund of Aﬃliated Hospital of Xuzhou Medical University (No. XYFY202250), and the Shaanxi Provincial Health and Health Research Fund Project (No. 2022E006). Conﬂict of interest WlL, LQR, and XW completed the study design. KW, WcL, and WlL performed the study and collected and analyzed the data. QS, CaS, and WlL drafted the manuscript. LQR, XW, KW, and HL provided the expert consultations and suggestions. CX and CY conceived of the study, participated in its design and coordination, and helped to embellish language. All authors reviewed the ﬁnal version of the manuscript and approved the submitted version. The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their aﬃliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. References Interleukin-6, C-reactive protein, ﬁbrinogen, and risk of recurrence after ischaemic stroke: Systematic review and meta-analysis. Eur. Stroke J. 6, 62–71. doi: 10.1177/2396987320984003 Fleuren, L. M., Klausch, T. L. T., Zwager, C. L., Schoonmade, L. J., Guo, T., Roggeveen, L. F., et al. (2020). Machine learning for the prediction of sepsis: A 11 Frontiers in Neuroscience frontiersin.org Wang et al. Wang et al. 10.3389/fnins.2023.1130831 10.3389/fnins.2023.1130831 Miotto, R., Wang, F., Wang, S., Jiang, X., and Dudley, J. T. (2018). Deep learning for healthcare: Review, opportunities and challenges. Brief Bioinform. 19, 1236–1246. Tang, W. Z., Wang, X. B., Li, H. T., Dong, M., and Ji, X. (2017). Serum copeptin predicts severity and recurrent stroke in ischemic stroke patients. Neurotox. Res. 32, 420–425. doi: 10.1007/s12640-017-9754-5 Muchada, M., Rubiera, M., Rodriguez-Luna, D., Pagola, J., Flores, A., Kallas, J., et al. (2014). Baseline national institutes of health stroke scale-adjusted time window for intravenous tissue-type plasminogen activator in acute ischemic stroke. Stroke 45, 1059–1063. doi: 10.1161/STROKEAHA.113.004307 Tu, W. J., Chao, B. H., Ma, L., Yan, F., Cao, L., Qiu, H., et al. (2021). Case-fatality, disability and recurrence rates after ﬁrst-ever stroke: A study from bigdata observatory platform for stroke of China. Brain Res. Bull. 175, 130–135. doi: 10.1016/j.brainresbull. 2021.07.020 Qiu, H., Song, J., Hu, J., Wang, L., Qiu, L., Liu, H., et al. (2022). Low serum transthyretin levels predict stroke-associated pneumonia. Nutr. Metab. Cardiovasc. Dis. 32, 632–640. doi: 10.1016/j.numecd.2021.12.008 Tu, W. J., Hua, Y., Yan, F., Bian, H., Yang, Y., Lou, M., et al. (2022). Prevalence of stroke in China, 2013-2019: A population-based study. Lancet Reg. Health West. Pac. 28:100550. doi: 10.1016/j.lanwpc.2022.100550 Sacco, R. L., Kasner, S. E., Broderick, J. P., Caplan, L. R., Connors, J. J., Culebras, A., et al. (2013). An updated deﬁnition of stroke for the 21st century: A statement for healthcare professionals from the American heart association/American stroke association. Stroke 44, 2064–2089. Virani, S. S., Alonso, A., Benjamin, E. J., Bittencourt, M. S., Callaway, C. W., Carson, A. P., et al. (2020). Heart disease and stroke statistics-2020 update: A report from the American heart association. Circulation 141, e139–e596. doi: 10.1161/CIR. 0000000000000757 Saini, V., Guada, L., and Yavagal, D. R. (2021). Global epidemiology of stroke and access to acute ischemic stroke interventions. Neurology 97, S6–S16. doi: 10.1212/ WNL.0000000000012781 Wu, Q., Zou, C., Wu, C., Zhang, S., and Huang, Z. (2016). Risk factors of outcomes in elderly patients with acute ischemic stroke in China. References Aging Clin. Exp. Res. 28, 705–711. doi: 10.1007/s40520-015-0478-1 Schaller, B. J., Graf, R., and Jacobs, A. H. (2006). Pathophysiological changes of the gastrointestinal tract in ischemic stroke. Am. J. Gastroenterol. 101, 1655–1655. doi: 10.1111/j.1572-0241.2006.00540.x Yuanyuan, Z., Jiaman, W., Yimin, Q., Haibo, Y., Weiqu, Y., and Zhuoxin, Y. (2020). Comparison of prediction models based on risk factors and retinal characteristics associated with recurrence one year after ischemic stroke. J. Stroke Cerebrovasc. Dis. 29:104581. doi: 10.1016/j.jstrokecerebrovasdis.2019.104581 Takabayashi, N., Murata, K., Tanaka, S., and Kawakami, K. (2015). Cost-eﬀectiveness of proton pump inhibitor co-therapy in patients taking aspirin for secondary prevention of ischemic stroke. Pharmacoeconomics 33, 1091–1100. doi: 10.1007/ s40273-015-0289-4 Zhang, Q., Qiu, D. X., Fu, R. L., Xu, T. F., Jing, M. J., Zhang, H. S., et al. (2016). H-type hypertension and c reactive protein in recurrence of ischemic stroke. Int. J. Environ. Res. Public Health 13:477. doi: 10.3390/ijerph13050477 Takeuchi, N., Tada, T., Toshima, M., Matsuo, Y., and Ikoma, K. (2009). Repetitive transcranial magnetic stimulation over bilateral hemispheres enhances motor function and training eﬀect of paretic hand in patients after stroke. J. Rehabil. Med. 41, 1049–1054. doi: 10.2340/16501977-0454 Zhang, Y., Xu, C., Wang, H., and Nan, S. (2021). Serum interleukin-37 increases in patients after ischemic stroke and is associated with stroke recurrence. Oxid. Med. Cell. Longev. 2021:5546991. doi: 10.1155/2021/5546991 12 12 frontiersin.org Frontiers in Neuroscience
https://openalex.org/W2807057423	https://publicatio.bibl.u-szeged.hu/13570/1/7493418%20%283%29.pdf	English	null	The Role of Uron and Chlorobenzene Derivatives, as Potential Endocrine Disrupting Compounds, in the Secretion of ACTH and PRL	International journal of endocrinology	2,018	cc-by	7,554	1First Department of Medicine, Faculty of Medicine, University of Szeged, Szeged, Hungary 2Department of Biometrics and Agricultural Informatics, Faculty of Horticultural Science, Szent István University, Budapest Hungary First Department of Medicine, Faculty of Medicine, University of Szeged, Szeged, Hungary Department of Biometrics and Agricultural Informatics, Faculty of Horticultural Science, Szent István University, Budapest Hungary Department of Environmental Biology and Education, Institute of Environmental and Technological Sciences Juhász Gyula Faculty of Education, University of Szeged, Szeged, Hungary sz Gyula Faculty of Education, University of Szeged, Szeged, Hungary Juhász Gyula Faculty of Education, University of Szeged, Szeged, Hungary Department of Inorganic and Analytical Chemistry, Faculty of Science and Informatics, University of Szeged, Szeg orrespondence should be addressed to Krisztian Sepp; sepp.krisztian@med.u-szeged.hu Correspondence should be addressed to Krisztian Sepp; sepp.krisztian@med.u-szeged.hu Received 7 March 2018; Revised 22 April 2018; Accepted 3 May 2018; Published 29 May 2018 Received 7 March 2018; Revised 22 April 2018; Accepted 3 May 2018; Published 29 May 2018 Academic Editor: Rosaria Meccariello Academic Editor: Rosaria Meccariello Copyright © 2018 Krisztian Sepp et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Uron herbicides polluting the environment represent a serious concern for environmental health and may be regarded as endocrine-disrupting compounds (EDCs), which inﬂuence the regulation of human homeostasis. We aimed to investigate the eﬀect of EDC urons (phenuron: PU, monuron: MU, and diuron: DU) and chlorobenzenes on the basal release of the adrenocorticotropic hormone (ACTH), which is a part of the adenohypophysis-adrenocortical axis. Hormone secretion in the presence of EDC was studied in two cell types: normal adenohypophysis cells (AdH) and cells of prolactinomas (PRLOMA). PRLOMA was induced in female Wistar rats by subcutaneously injecting them with estrone acetate for 6 months. AdH and PRLOMA were separated from treated and untreated experimental animals, dissociated enzymatically and mechanically in order to create monolayer cell cultures, which served as an experimental in vitro model. We investigated the eﬀects of ED agents separately and in combination on ACTH and prolactin (PRL) release through the hypophyseal-adrenal axis. Hormone determination was carried out by the luminescent immunoassay and the radioimmunoassay methods. Our results showed that (1) uron agents separately did not change ACTH and PRL release in AdH culture; (2) ACTH secretion in arginine vasopressin- (AVP-) activated AdH cells was signiﬁcantly increased by EDC treatment; (3) ED agents increased the basal hormone release (ACTH, PRL) in PRLOMA cells; and (4) EDC exposure increased ACTH release in AVP-activated PRLOMA cells. Hindawi International Journal of Endocrinology Volume 2018, Article ID 7493418, 7 pages https://doi.org/10.1155/2018/7493418 Hindawi International Journal of Endocrinology Volume 2018, Article ID 7493418, 7 pages https://doi.org/10.1155/2018/7493418 Hindawi 1First Department of Medicine, Faculty of Medicine, University of Szeged, Szeged, Hungary 2Department of Biometrics and Agricultural Informatics, Faculty of Horticultural Science, Szent István University, Budapest Hungary We conclude that the herbicides PU, MU, and DU carry EDC eﬀects and show human toxicity potential. Krisztian Sepp ,1 Anna M. Laszlo,2 Zsolt Molnar,3 Andrea Serester,3 Tunde Alapi,4 Marta Galfi,3 Zsuzsanna Valkusz,1 and Marianna Radacs3 1First Department of Medicine, Faculty of Medicine, University of Szeged, Szeged, Hungary 2Department of Biometrics and Agricultural Informatics, Faculty of Horticultural Science, Szent István University, Budapest, Hungary 3 Krisztian Sepp ,1 Anna M. Laszlo,2 Zsolt Molnar,3 Andrea Serester,3 Tunde Alapi,4 Marta Galfi,3 Zsuzsanna Valkusz,1 and Marianna Radacs3 1. Introduction process [11, 12] which result in serious homeostatic alteration even in healthy individuals [13, 14]. Human adaptation patterns induced by environmental burdens are obviously modiﬁed when the exposition reaches an individual who suﬀers from functional and/or structural disorders (illnesses) [15, 16]. An open dynamic system which forms a unit with its environment, for example, a living human organism, can stabilize its equilibrium pro- cesses as determined by the direct environmental conditions (attraction range) [17] deﬁned by its genetic and functional Chemical agents (e.g., halogenated hydrocarbons and uron herbicides [1–3]) which pollute the environment represent a serious concern for environmental health [4, 5] and may be regarded as endocrine-disrupting compounds (EDCs), which inﬂuence the regulation of human homeostasis [6]. They may change the potential [7, 8] and capacity [9, 10] of the psycho-neuro-endocrine-immune regulation net- work and may bring about disturbances in the regulatory International Journal of Endocrinology 2 adaptation potential (algorithmic networks characterizing the local properties of the living system [18]). In human homeostasis, hormones are the creative elements of the neuroendocrine regulation [19, 20]. Human neuroendocrine regulation can be interpreted as a network of open, dynamic biological systems [21] in the outlined context. Biological cycles that can be described with the “AND” function are those essential for life (e.g., human hypothalamus- adenohypophysis-adrenal cortex axis functional disorder) [22]. The “OR” function-related control systems are not essential at the organizational level of the given individual (e.g., PRL); life functions can be maintained in their absence. The disturbance of the healthy (control) processes of the “OR” cycle will aﬀect the “AND” cycles. Chronic changing of the “OR” cycle may lead to structural disturbance, for example, cellular proliferation, which is sustained by contin- uous feedback information [23]. and androgenic functions have been discovered [41, 42]. The human toxicity potential (HTP) is an internationally standardized impact category set in LCA (life cycle analysis) standards, used to express various environmental eﬀects [43], in which the eﬀects of environmental 1,4-dichlorobenzene on human health are the benchmark. The determination and relation of ecotoxicity potential (ETP) to dClB eﬀects are performed in a similar fashion. Our aims were to investigate the eﬀect of ED compounds ((PU, MU, and DU) and dClB and 1,2,4-trichlorobenzene + hexachlorobenzene (chlorobenzene mixture, mClB)) on the basal release of the ACTH involved in the functioning of the adenohypophyseal-adrenocortical axis. 2. Materials and Methods 2.1. Experimental Animals. Certiﬁed healthy female rats were used in our experiments (Wistar strains weighing 120–250 g, 4–6 weeks old at the onset of the study) (Charles River, Isaszeg, Hungary). During the experimental period, animals were kept in a controlled (55–65% relative humidity, 22 ± 2°C air temperature), automated diurnal environment (12 h daytime, 12h night illumination cycles) in 32 × 40× 18cm cages (5 animals/cage). The diet required for experimental animals (CRLT/N, Charles River, Hungary) and drinking water were available ad libitum. The animals involved in the study were treated in accordance with Gov. Ordinance Number 40/2013 (II. 14.) on animal experiments. 2.2. Induced Prolactinoma and In Vitro Experimental Models. The PRLOMA models were made from Wistar rats (♀, n = 20) which were subcutaneously injected with estrone acetate for 6 months (CAS registration number 901-93-9, Sigma, Germany, 150 μg/kg/week) [44]. After the pentobarbital (4.5 mg/kg, Nembutal, Abbott, USA) anesthesia, the animals were decapitated and AdH was separated; the tissue was enzymatically (trypsin: Sigma, Germany, 0.2% for 30 min; collagenase: Sigma, Germany, 30 μg/ml for 40min; dispase: Sigma, Germany, 50μg/ml for 40 min; phosphate-buﬀered saline (PBS-A) was used for the solutions, temperature: 37°C) and mechanically (83 μm and 48 μm pore size nylon blutex ﬁlter) dissoci- ated. The cell viability was ≥95% (trypan blue staining). The cells were suspended in the following medium: Dulbecco’s Modiﬁed Essential Medium (DMEM, Sigma, Germany) +20% fetal calf serum (FCS, Sigma, Germany) + 1.0 IU/ml penicillin + streptomycin (Sigma, Germany). Then, they were placed into surface-treated (5% collagen) 24-well plastic culture vessels (Nunc, Germany) and put in a thermostat (temperature = 37°C, pCO2: 5%). Cell cultures were washed every 3 days after adherence until they became conﬂuent. When investigating the neuroendocrine aspects of ED agents (e.g., uron herbicides and ClB), the experimental layout is based on preliminary experiments (dose and time kinetics) [30] and the already proven physical, chemical, and biological eﬀects of the compounds. Urons are substituted phenylureas of high chemical stability, which are used in agriculture as photosynthesis-inhibiting herbi- cides [31]. Due to their long chemical half-life (a few months to one year in soil, 2–6 weeks in water), the food chain can be severely aﬀected [32, 33]. 1. Introduction This work focused on hormone secretion in the presence of EDC in two cell types: normal adenohypophysis (AdH) cells and pro- lactinoma (PRLOMA) cells. In this regard, the question was, Can EDC modify the feedback mechanism of ACTH release governed by arginine vasopressin (AVP) and corticosterone (B) (“AND” cycles)? Observing EDC eﬀects on PRL hormone secretion was also a goal in this research (“OR” cycle). Herbicides [24] and halogenated hydrocarbons [25] are widespread substances that have an EDC eﬀect. Phenuron (PU), monuron (MU), and diuron (DU) compounds may be viewed as a halogenated homologous series of herbicides [26]. Chlorobenzenes (ClB) represent a halogenated aro- matic hydrocarbon group, of which 1,4-dichlorobenzene (dClB) is known as an international reference compound due to its ecotoxicological and human toxicity potential [27]. It is a primary question from the point of view of medical practice, whether the cellular follow-up of ACTH regulation is a suitable test system for studying the eﬀect of the environmental pollutants. Any exposure can be inter- preted as a stressor in the human neuroendocrinium [28]. The outlined circuit (in mathematical terms: attractor) maintains processes indispensable in human homeostasis, since there is no life without adrenocortical hormones. The regulatory disturbances (e.g., feedback disturbances) may become causative factors, for example, they may lead to benign cell proliferation. The most commonly occurring human adenohypophysis cell proliferation is prolactinoma, which generates prolactin overproduction. In the develop- ment of the tumor, a signiﬁcant pathophysiological role is attributed to the estrogenic eﬀect [29]. 2. Materials and Methods + B + B + AVP +AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) + B + B + AVP + AVP +AVP + B Basal Figure 2: Functional assay for ACTH release in rat prolactinoma cultures (PRLOMA), in vitro (n = 11 −12 in each group); mean level ± SEM (pg ACTH/mg protein): compared with normal AdH cells (Figure 1). ACTH release always shows a similarly signiﬁcant (p < 0 001) increase in PRLOMA cells (basal: control; AVP: 8- arginine vasopressin; B: corticosterone). 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) + B + B + AVP + AVP +AVP + B Basal Figure 2: Functional assay for ACTH release in rat prolactinoma cultures (PRLOMA), in vitro (n = 11 −12 in each group); mean level ± SEM (pg ACTH/mg protein): compared with normal AdH cells (Figure 1). ACTH release always shows a similarly signiﬁcant (p < 0 001) increase in PRLOMA cells (basal: control; AVP: 8- arginine vasopressin; B: corticosterone). 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) + B + B + AVP + AVP +AVP + B Basal Figure 1: Functional assay for ACTH release in normal rat adenohypophysis cultures, in vitro (n = 8 – 12 in each group); mean level ± SEM (pg ACTH/mg protein): AVP signiﬁcantly increases (p < 0 001), whereas corticosterone alone does not alter ACTH release when compared to control; preincubation with +B inhibited AVP activation (basal: control; AVP: 8-arginine vasopressin; B: corticosterone). Figure 2: Functional assay for ACTH release in rat prolactinoma cultures (PRLOMA), in vitro (n = 11 −12 in each group); mean level ± SEM (pg ACTH/mg protein): compared with normal AdH cells (Figure 1). ACTH release always shows a similarly signiﬁcant (p < 0 001) increase in PRLOMA cells (basal: control; AVP: 8- arginine vasopressin; B: corticosterone). of Szeged (Immulite 2000, Siemens Healthcare Diagnostic, Deerﬁeld, IL, USA and DPC kit/L2KAC-02, Euro DPC Ltd., Glyn Rhonwy, United Kingdom). The protein content of the samples was determined using a modiﬁed Lowry method [45] and Pierce BCA Protein Assay Kit (Thermo Fisher Scientiﬁc Inc., Rockford, USA). Speciﬁc functional standardization for ACTH release was regulated by 1 μg/ml corticosterone (B) and 10−6 M AVP treatments. 2. Materials and Methods In humans, low to moderate toxic- ity is associated with spleen and liver involvement, whereas carcinogenicity [34] has also been demonstrated, and in the case of diurons, antiandrogenic properties have been described as well [35, 36]. Chlorobenzenes are also highly persistent chlorine-substituted aromatic hydrocarbons form- ing a homologous series [2, 37, 38]. Toxicity studies report liver and kidney impairment, but their roles as carcinogenic (breast, liver, and kidney) agents are also known [39, 40]. As EDC eﬀect, the alterations of thyroid hormone synthesis International Journal of Endocrinology 3 + B + B + AVP +AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) Figure 1: Functional assay for ACTH release in normal rat adenohypophysis cultures, in vitro (n = 8 – 12 in each group); mean level ± SEM (pg ACTH/mg protein): AVP signiﬁcantly increases (p < 0 001), whereas corticosterone alone does not alter ACTH release when compared to control; preincubation with +B inhibited AVP activation (basal: control; AVP: 8-arginine vasopressin; B: corticosterone). 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) + B + B + AVP + AVP +AVP + B Basal Figure 2: Functional assay for ACTH release in rat prolactinoma cultures (PRLOMA), in vitro (n = 11 −12 in each group); mean level ± SEM (pg ACTH/mg protein): compared with normal AdH cells (Figure 1). ACTH release always shows a similarly signiﬁcant (p < 0 001) increase in PRLOMA cells (basal: control; AVP: 8- arginine vasopressin; B: corticosterone). + B + B + AVP +AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) Figure 1: Functional assay for ACTH release in normal rat adenohypophysis cultures, in vitro (n = 8 – 12 in each group); mean level ± SEM (pg ACTH/mg protein): AVP signiﬁcantly increases (p < 0 001), whereas corticosterone alone does not alter ACTH release when compared to control; preincubation with +B inhibited AVP activation (basal: control; AVP: 8-arginine vasopressin; B: corticosterone). 2. Materials and Methods The ACTH release cycle (“AND” cycles) was activated by AVP; this mechanism was inhibited by a 20-minute preincubation with B (Figures 1 and 2). 2.5. Statistical Analysis. Measurements (n = 8 – 12 per group on 24 lots: pooled samples on AdH cell cultures) of ACTH and PRL hormone release by disease (PRLOMA versus normal AdH) in various EDC groups (control, dClB, mClB, PU, MU, and DU) by regulation (basal, +AVP, +B, +B +AVP, and +AVP+B) were compared using mixed models on rats [46, 47]. The regulation cycle was veriﬁed in a mixed model for the comparison of the control groups of EDC for ACTH in the 5 regulation phases, using disease and regulation as ﬁxed eﬀects and random intercept for the lots. For ACTH data, a mixed model was applied with disease, EDC, and regulation (only basal, +AVP, and +B+AVP) as ﬁxed factors and random intercept for the lots. For PRL measurements, a mixed model was applied with disease and EDC as ﬁxed fac- tors and random intercept for the lots for basal regulation data. In the analysis models, the reference group was the normal (healthy AdH), control (no EDC treatment), and basal (no regulation) group. Restricted maximum likelihood estimation and Kenward-Roger method for adjusting the degrees of free- dom were applied in all models with unstructured covariance matrix. Pairwise comparisons were estimated by least squares means using Sidak p value adjustment. Model residuals were displayed on quantile-quantile plots to check normality assumptions. Statistical analyses were performed in SAS (ver- sion 9.3 SAS Institute Inc., Cary, NC, USA), where p values of <0.05 were considered to indicate statistical signiﬁcance [48]. 2.3. Experimental Protocol. Time and dose kinetic assays, determining the appropriate arrangements, were performed on standardized AdH and PRLOMA in vitro cultures. In the present study, AdH and PRLOMA cell cultures were ﬁrst treated for 60 minutes with chlorobenzene (dClB = 0.1 ng/ml; chlorobenzene mix (mClB) =0.1 ng/ml; hexachlorobenzene and 1,2,4-trichlorobenzene 1 : 1) and with urons (PU: 10−6 M, MU: 10−6 M, DU: 10−6 M). At the end of the EDC treatment, samples were obtained from the supernatant media of the cell cultures for the determination of ACTH and PRL hormones. 2. Materials and Methods When studying the ACTH-mediated role of EDC agents in the hypophysis/adrenal cortex regula- tion, the EDC agents were coadministered with (10−6 M) AVP and after the 60-minute treatment period, samples were taken from the supernatant media of both AdH and PRLOMA, in the regulation cycle of AVP/B feedback studies in Figures 1 and 2. EDCs were added together with B. AVP was adminis- tered after a 20-minute pretreatment with EDC + B, and at the end of the treatment period, the supernatant media were sampled. In order to follow PRL hormone release, AdH and PRLOMA cell cultures were individually treated with ED agents for 60 minutes. Next, the supernatant media were used to measure PRL. 2.4. Hormone Assays. PRL assay was performed by radioimmunoassay from samples obtained according to the experimental protocol [44]. Determination of ACTH from samples was carried out by the luminescent immunoassay method, using the apparatus of the Endocrinology Unit, First Dept. of Internal Medicine, Faculty of Medicine, University 3. Results Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP administration by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for +AVP. Figure 4: EDC eﬀects in the regulation of ACTH release in rat PRLOMA cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal, +AVP, and +B+AVP. 0 5 10 15 20 25 ng PRL/mg protein (AdH) Control dCIB mCIB PU MU DU Figure 5: Eﬀects of EDC on PRL release in normal rat AdH cultures, in vitro. Mean (PRL/prolactin/level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: dClB = 1,4- dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. 0 5 10 15 20 25 ng PRL/mg protein (AdH) Control dCIB mCIB PU MU DU in the following cases: basal, AVP activated (+AVP), and the corticosterone-inhibited AVP activation (+B+AVP) in AdH cultures. It can be seen that ACTH release was not altered by ED agents (mean level± SEM (pg ACTH/mg protein): dClB 1567.91± 3.09; mClB 1585.33 ±2.72; PU 1533.67 ± 2.52; MU 1553.17 ± 3.40; and DU 1566.33 ±2.30) compared to the control group (1528.25 ± 6.14pg ACTH/mg protein). In the AVP-activated samples, ACTH release showed a signiﬁcant increase compared to the control group for each EDC (mean level± SEM (pg ACTH/mg protein): control 10,220.88 ±20.36; dClB 14,430.08 ± 3.01; mClB 14,488.90± 3.57; PU 11,845.67± 7.02; MU 13,008.25 ± 10.18; and DU 13,658.75 ±15.83). 3. Results Control dCIB mCIB PU MU DU + B + AVP + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) Figure 4: EDC eﬀects in the regulation of ACTH release in rat PRLOMA cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal, +AVP, and +B+AVP. + AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) Control dCIB mCIB PU MU DU Figure 3: EDC eﬀects in the regulation of ACTH release on normal rat AdH cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP administration by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for +AVP. Control dCIB mCIB PU MU DU + B + AVP + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) + AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) Control dCIB mCIB PU MU DU Figure 4: EDC eﬀects in the regulation of ACTH release in rat PRLOMA cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal, +AVP, and +B+AVP. Figure 3: EDC eﬀects in the regulation of ACTH release on normal rat AdH cultures, in vitro. 3. Results Figure 3 shows the eﬀect of various ED compounds (dClB, mClB, PU, MU, and DU) on ACTH release in AdH cultures International Journal of Endocrinology 4 Control dCIB mCIB PU MU DU + B + AVP + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) Figure 4: EDC eﬀects in the regulation of ACTH release in rat PRLOMA cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal, +AVP, and +B+AVP. + AVP + B + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (AdH) Control dCIB mCIB PU MU DU Figure 3: EDC eﬀects in the regulation of ACTH release on normal rat AdH cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP administration by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for +AVP. Control dCIB mCIB PU MU DU + B + AVP + AVP Basal 0 2500 5000 7500 10,000 12,500 15,000 pg ACTH/mg protein (PRLOMA) Figure 4: EDC eﬀects in the regulation of ACTH release in rat PRLOMA cultures, in vitro. Mean (ACTH level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: B = corticosterone: 1 μg/ml; AVP = 8-arginine vasopressin: 10−6 M, +B+AVP: in combination therapy B precedes AVP by 20 minutes; dClB = 1,4-dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal, +AVP, and +B+AVP. 3. Results In the case of regulatory eﬀect (+B +AVP-feedback), large deviations could not be detected in the presence of EDC (mean level ±SEM (pg ACTH/mg protein): control 1524.67 ±3.46; dClB 1542.00 ± 1.22; mClB 1566.92 ±2.40; PU 1540.08 ± 2.86; MU 1560.08 ±1.88; and DU 1578.08 ±1.26). Figure 5: Eﬀects of EDC on PRL release in normal rat AdH cultures, in vitro. Mean (PRL/prolactin/level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: dClB = 1,4- dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. Figure 4 shows the eﬀects of dClB, mClB, PU, MU, and DU on ACTH release in PRLOMA cultures in the following cases: basal, AVP activation (+AVP), and corticosterone- inhibited AVP activation (+B+AVP). It can be seen that ED agents modulate ACTH release compared to the control of the basal group (2193.64 ± 1.92 pg ACTH/mg protein): dClB: 2624.30±7.60pg ACTH/mg protein; mClB: 2956.08± 4.71pg ACTH/mg protein; PU: 2427.33±6.08pg ACTH/mg protein; MU: 2535.17±5.14pg ACTH/mg protein; and DU: 2705.33±4.63pg ACTH/mg protein. AVP-activated ACTH release of PRLOMA cultures shows a signiﬁcant increase for each ED compound used (mean level±SEM (pg ACTH/mg protein): control 12,674.50±7.23; dClB 14,620.58±5.61; mClB 14,830.50±7.42; PU 13,129.67±6.06; MU 14,954.17±11.22; and DU 15,197.58±4.99). EDC eﬀects were detected in the regulation model (+B+AVP-feedback) (mean level±SEM (pg ACTH/mg protein): control 2195.50±4.69; dClB 2579.42± 3.98; mClB 2513.00±2.94; PU 2421.00±2.28; MU 2553.83± 7.87; and DU 2690.42±4.14). Figure 5 shows the PRL release in AdH cultures in the presence of the ED compounds tested. It can be seen that the ED compounds did not trigger relevant diﬀerences in PRL release (mean level± SEM (ng PRL/mg protein): con- trol: 7.13 ±0.04; dClB 7.28 ±0.03; mClB 7.26 ±0.01; PU 7.12 ± 0.02; MU 7.02 ±0.02; and DU 7.14 ± 0.01). International Journal of Endocrinology International Journal of Endocrinology 5 0 5 10 15 20 25 ng PRL/mg protein (PRLOMA) Control dCIB mCIB PU MU DU Figure 6: Eﬀects of EDC on PRL release in rat PRLOMA cultures, in vitro. Mean (PRL/prolactin/level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: dClB = 1,4- dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal release. 3. Results agents caused increased ACTH release in the AVP- activated PRLOMA cells as well (Figure 4, +AVP group). It is well known that in biological complexities (biological networks), regulations appear as the resultants of activating and inhibiting functions [51]. Both in the normal AdH and PRLOMA cell types, the negative feedback eﬀect of cortico- sterone was modeled by the inhibition of AVP-activated ACTH release [52]. Figure 3 demonstrates that AdH cells were able to maintain their inhibitory functions despite a treatment with EDC (Figure 3, +B+AVP group). Thus, the adaptive potential of healthy AdH cells in cases of chemical environmental load modeled by EDC remained reversibly regulated. The regulation of PRLOMA cells was incomplete in the presence of EDC (Figure 4, +B+AVP group); although the inhibitory eﬀect of corticosterone on ACTH secretion is also present in PRLOMA, the maintenance of this eﬀect is strongly disrupted by ED compounds (Figure 4). In this context, the results can be interpreted as an environmental disruption leading to irreversible processes. The presence of ED compounds enhanced the already overexpressed ACTH secretion in PRLOMA cells (Figure 4), which was further increased by AVP activation. Therefore, it can be assumed that both the structure and the endocrine cell function of PRLOMA were damaged [53]. Due to the role ACTH plays in essential life functions, this result may have signiﬁcance in the environmental exposition of prolactinoma patients and in the development of adaptational potential disorders in healthy individuals [54], as ACTH regulation is present in human adaptation as an open dynamic requirement, which is modeled as an “AND” logical function algorithm. Our results show that PRL secretion in normal AdH cells was not inﬂuenced by EDC in the applied experimental system (Figure 5). However, the already elevated PRL secre- tion of PRLOMA cells was further enhanced by EDC (Figure 6). 0 5 10 15 20 25 ng PRL/mg protein (PRLOMA) Control dCIB mCIB PU MU DU Figure 6: Eﬀects of EDC on PRL release in rat PRLOMA cultures, in vitro. Mean (PRL/prolactin/level) ± SEM. The mean and SEM are calculated from n = 12. Abbreviations: dClB = 1,4- dichlorobenzene: 0.1 ng/ml; mClB = chlorobenzene mix: 0.1 ng/ml; PU = phenuron: 10−6 M; MU = monuron: 10−6 M; DU = diuron: 10−6 M. All EDC groups diﬀer signiﬁcantly from the control (p < 0 001) for basal release. 3. Results Figure 6 shows the PRL release of rat PRLOMA cultures in the presence of ED chemical agents as described in the experimental protocol. All the examined compounds showed a signiﬁcant stimulating eﬀect (dClB: 22.47 ± 0.03; mClB: 23.17 ± 0.02; PU: 19.82 ± 0.01; MU: 21.50 ± 0.02; and DU: 22.41 ± 0.02 ng PRL/mg protein) on the release of PRL com- pared to that of the control (17.14 ±0.02 ng PRL/mg protein) in PRLOMA cultures. 4. Discussion and Conclusions We have studied the eﬀects of PU, MU, DU, dClB, and mClB as potential environmental factors, on the basal release of ACTH (Figure 3) (creative element: mobile network junction [49]), which plays a role in the essential functioning of the AdH/AC axis [30] (interpreted in the human homeostasis network as an “AND” function). The ED eﬀects of the applied compounds could be detected, since each chemical agent generated signiﬁcant changes in the ACTH secretion of PRLOMA cells (Figure 4) in the basal group. PRL release can be increased by enhancing peripheral estrogene levels, which can be the “OR” cycle in connection to ACTH release. In our work, we modeled a homeostatic regulatory neuroen- docrine network both under normal regulation (AdH model) and under a neuroendocrine disturbance (+ER). (First, cell cycle change was induced via autoregulatory cell dysfunction [29].) In the PRLOMA cells that were already altered by the control cycle, the level of not only PRL (Figure 6), but also that of ACTH increased (Figure 4). The events of the neuro- endocrine regulatory cycles in our investigation suggest that the outlined mechanisms can be regarded as homeostatic biological network elements [50]. HTP is deﬁned in relation to dClB with a standard approach [43] and uniform risk assessment. The authors wished to provide an opportunity for comparison by follow- ing dClB eﬀects too. Accordingly, uron/dClB relations were determined in AdH and PRLOMA cells. Our results show that the EDC classiﬁcation of the examined uron compounds strongly approximates the eﬀects of dClB in an AdH model (Figures 3 and 5). However, basal ACTH release of PRLOMA is more eﬀective in the presence of DU (uron/dClB →0.97– 0.99) than dClB (+DU/dClB = 1.05). In baseline PRL secre- tion, the uron eﬀects on AdH cells approximated those of dClB (urons/dClB: 0.96–0.98). In the case of PRLOMA, the EDC eﬀects of the investigated uron compounds on PRL secretion approximated those of dClB (urons/dClB: 0.96– 0.98). However, in the case of AVP-activated ACTH release in the PRLOMA model, dClB eﬀect was exceeded by the agents MU and DU (PRLOMA: +AVP + MU/+AVP + dClB = 1.02; PRLOMA: +AVP + DU/+AVP +dClB =1.03). The HTP values of PRLOMA can be interpreted with double risk classiﬁcation according to our present study, because when the doses of ED compounds exceed those necessary for irreversible eﬀects (such as suspension of corticosterone inhibition), regulatory dysfunctions may present diﬃcult-to-treat disease processes. References [15] L. E. Brace, S. C. Vose, K. Stanya et al., “Increased oxidative phosphorylation in response to acute and chronic DNA damage,” npj Aging and Mechanisms of Disease, vol. 2, no. 1, p. 16022, 2016. [1] M. Gavrilescu, K. Demnerová, J. Aamand, S. Agathos, and F. Fava, “Emerging pollutants in the environment: present and future challenges in biomonitoring, ecological risks and bioremediation,” New Biotechnology, vol. 32, no. 1, pp. 147– 156, 2015. [16] Z. Kasanova, D. Hernaus, T. Vaessen et al., “Early-life stress aﬀects stress-related prefrontal dopamine activity in healthy adults, but not in individuals with psychotic disorder,” PLoS One, vol. 11, no. 3, article e0150746, 2016. [2] J. L. Barber, A. J. Sweetman, D. van Wijk, and K. C. Jones, “Hexachlorobenzene in the global environment: emissions, levels, distribution, trends and processes,” Science of the Total Environment, vol. 349, no. 1–3, pp. 1–44, 2005. [17] T. Veitshans, D. Klimov, and D. Thirumalai, “Protein folding kinetics: timescales, pathways and energy landscapes in terms of sequence-dependent properties,” Folding and Design, vol. 2, no. 1, pp. 1–22, 1997. [3] G. Holmes, “Australia’s pesticide environmental risk assess- ment failure: the case of diuron and sugarcane,” Marine Pollution Bulletin, vol. 88, no. 1-2, pp. 7–13, 2014. [18] S. Maslov and K. Sneppen, “Speciﬁcity and stability in topology of protein networks,” Science, vol. 296, no. 5569, pp. 910–913, 2002. [4] A. Bergman, J. J. Heindel, S. Jobling, K. A. Kidd, and R. T. Zoeller, State of the Science of Endocrine Disrupting Chemicals - 2012: An Assessment of the State of the Science of Endocrine Disruptors Prepared by a Group of Experts for the United Nations Environment Programme (UNEP) and WHO, United Nations Environment Programme and the World Health Organization, 2013. [19] G. Kotronoulas, A. Stamatakis, and F. Stylianopoulou, “Hormones, hormonal agents, and neuropeptides involved in the neuroendocrine regulation of sleep in humans,” Hormones, vol. 8, no. 4, pp. 232–248, 2009. [20] M. P. Cornejo, S. T. Hentges, M. Maliqueo, H. Coirini, D. Becu-Villalobos, and C. F. Elias, “Neuroendocrine regula- tion of metabolism,” Journal of Neuroendocrinology, vol. 28, no. 7, 2016. [5] W. H. Goodson III, L. Lowe, D. O. Carpenter et al., “Assessing the carcinogenic potential of low-dose exposures to chemical mixtures in the environment: the challenge ahead,” Carcino- genesis, vol. 36, Supplement 1, pp. S254–S296, 2015. [21] C. B. Nemeroﬀ, “Psychoneuroimmunoendocrinology: the bio- logical basis of mind-body physiology and pathophysiology,” Depression and Anxiety, vol. International Journal of Endocrinology 6 It can be seen from the above how important researching human homeostatic network disorders is when one wishes to evaluate the health consequences of environmental factors. [9] J. A. McLachlan, “Environmental signaling: what embryos and evolution teach us about endocrine disrupting chemicals,” Endocrine Reviews, vol. 22, no. 3, pp. 319–341, 2001. [10] A. Mokarizadeh, M. R. Faryabi, M. A. Rezvanfar, and M. Abdollahi, “A comprehensive review of pesticides and the immune dysregulation: mechanisms, evidence and conse- quences,” Toxicology Mechanisms and Methods, vol. 25, no. 4, pp. 258–278, 2015. Data Availability Requests for data will be considered by the corresponding author. [11] A. Nadal, I. Quesada, E. Tuduri, R. Nogueiras, and P. Alonso- Magdalena, “Endocrine-disrupting chemicals and the regula- tion of energy balance,” Nature Reviews Endocrinology, vol. 13, no. 9, pp. 536–546, 2017. The authors declare that they have no conﬂict of interest. The authors declare that they have no conﬂict of interest. [12] M. Preciados, C. Yoo, and D. Roy, “Estrogenic endocrine disrupting chemicals inﬂuencing NRF1 regulated gene net- works in the development of complex human brain diseases,” International Journal of Molecular Sciences, vol. 17, no. 12, 2016. Acknowledgments This research was supported by the European Union and the State of Hungary, coﬁnanced by the European Social Fund in the framework of TÁMOP-4.2.4.A/2-11/1-2012- 0001 “National Excellence Program,” TÁMOP-4.2.6-15/1- 2015-0002, EFOP-3.6.1-16-2016-00008, and EFOP-3.4.3- 16-2016-00014. [13] C. López-Otín, M. A. Blasco, L. Partridge, M. Serrano, and G. Kroemer, “The hallmarks of aging,” Cell, vol. 153, no. 6, pp. 1194–1217, 2013. [14] J. M. Vicencio, L. Galluzzi, N. Tajeddine et al., “Senescence, apoptosis or autophagy? When a damaged cell must decide its path—a mini-review,” Gerontology, vol. 54, no. 2, pp. 92– 99, 2008. 4. Discussion and Conclusions The AVP-activated ACTH hormone secretion of AdH cells was signiﬁcantly increased by all ED compounds used when compared to the control (Figure 3, +AVP group). ED International Journal of Endocrinology Randall, “Protein measurement with the Folin phenol reagent,” The Journal of Biological Chemistry, vol. 193, no. 1, pp. 265–275, 1951. [31] J. Liu, “Phenylurea herbicides,” in Hayes’ Handbook of Pesticide Toxicology, R. Krieger, Ed., pp. 1725–1731, Academic Press, San Diego, CA, USA, 3rd edition, 2010. [46] H. Brown and R. Prescott, Applied Mixed Models in Medicine, Edinburgh: John Wiley & Sons, 2nd edition, 2006. [32] C. Federico, M. Pappalardo, G. C. Leotta, Z. Minniti, V. Librando, and S. Saccone, “Phenylurea herbicides: chemical properties and genotoxic eﬀects,” in Handbook on Herbi- cides, D. Kobayashi and E. Watanabe, Eds., pp. 111–132, Hauppauge: Nova Science Publishers, 2014. [47] J. Singer and J. Willett, Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence, Oxford University Press, New York, NY, USA, 2003. [48] SAS, SAS/STAT 9.3 User’s Guide, Cary: SAS Institute Inc., 2011. [33] J. Lintelmann, A. Katayama, N. Kurihara, L. Shore, and A. Wenzel, “Endocrine disruptors in the environment (IUPAC technical report),” Pure and Applied Chemistry, vol. 75, no. 5, pp. 631–681, 2003. [49] F. L. Strand, K. J. Rose, J. A. King, A. C. Segarra, and L. A. Zuccarelli, “ACTH modulation of nerve development and regeneration,” Progress in Neurobiology, vol. 33, no. 1, pp. 45–85, 1989. [34] K. A. Lewis, J. Tzilivakis, D. J. Warner, and A. Green, “An international database for pesticide risk assessments and management,” Human and Ecological Risk Assessment: An International Journal, vol. 22, no. 4, pp. 1050–1064, 2016. [50] A. Gądek-Michalska, J. Spyrka, P. Rachwalska, J. Tadeusz, and J. Bugajski, “Inﬂuence of chronic stress on brain corticosteroid receptors and HPA axis activity,” Pharmacological Reports, vol. 65, no. 5, pp. 1163–1175, 2013. [35] W. Mnif, A. I. H. Hassine, A. Bouaziz, A. Bartegi, O. Thomas, and B. Roig, “Eﬀect of endocrine disruptor pesticides: a review,” International Journal of Environmental Research and Public Health, vol. 8, no. 6, pp. 2265–2303, 2011. [51] J. A. McLachlan, “Environmental signaling: from environmen- tal estrogens to endocrine-disrupting chemicals and beyond,” Andrology, vol. 4, no. 4, pp. 684–694, 2016. [52] N. Jasnic, J. Djordjevic, P. Vujovic, I. Lakic, S. Djurasevic, and G. Cvijic, “The eﬀect of vasopressin 1b receptor (V1bR) block- ade on HPA axis activity in rats exposed to acute heat stress,” Journal of Experimental Biology, vol. 216, no. 12, pp. 2302– 2307, 2013. [36] H. Kojima, E. Katsura, S. Takeuchi, K. Niiyama, and K. International Journal of Endocrinology 7 Toxicology Program Technical Report Series, vol. 319, pp. 1–198, 1987. [25] Z. Valkusz, G. Nagyéri, M. Radács et al., “Further analysis of behavioral and endocrine consequences of chronic exposure of male Wistar rats to subtoxic doses of endocrine disruptor chlorobenzenes,” Physiology & Behavior, vol. 103, no. 5, pp. 421–430, 2011. [40] A. S. Randi, C. Cocca, V. Carbone et al., “Hexachlorobenzene is a tumor co-carcinogen and induces alterations in insulin- growth factors signaling pathway in the rat mammary gland,” Toxicological Sciences, vol. 89, no. 1, pp. 83–92, 2006. [26] M. A. Oturan, M. C. Edelahi, N. Oturan, K. El kacemi, and J. J. Aaron, “Kinetics of oxidative degradation/mineralization pathways of the phenylurea herbicides diuron, monuron and fenuron in water during application of the electro-Fenton process,” Applied Catalysis B: Environmental, vol. 97, no. 1-2, pp. 82–89, 2010. [41] J. L. Ralph, M. C. Orgebin-Crist, J. J. Lareyre, and C. C. Nelson, “Disruption of androgen regulation in the prostate by the envi- ronmental contaminant hexachlorobenzene,” Environmental Health Perspectives, vol. 111, no. 4, pp. 461–466, 2003. [42] C. den Besten, J. J. R. M. Vet, H. T. Besselink et al., “The liver, kidney, and thyroid toxicity of chlorinated benzenes,” Toxicol- ogy and Applied Pharmacology, vol. 111, no. 1, pp. 69–81, 1991. [27] H. Motejlová and V. Koci, “Application of 1,4-dichloroben- zene as a reference substance in the LCA methodology (life cycle assessment),” Acta Environmentalica Universitatis Comenianae (Bratislava), vol. 19, pp. 243–246, 2011. [43] International Standard ISO 14044, “Environmental manage- ment. Life cycle assessment. Requirements and guidelines,” International Organization for Standardization, 2006. [28] E. R. De Kloet, M. Joels, and F. Holsboer, “Stress and the brain: from adaptation to disease,” Nature Reviews Neuroscience, vol. 6, no. 6, pp. 463–475, 2005. [44] Z. Molnár, R. Pálföldi, A. László et al., “The eﬀects of hypoka- laemia on the hormone exocytosis in adenohypophysis and prolactinoma cell culture model systems,” Experimental and Clinical Endocrinology & Diabetes, vol. 122, no. 10, pp. 575– 581, 2014. [29] D. K. Sarkar, “Genesis of prolactinomas: studies using estrogen-treated animals,” Frontiers of Hormone Research, vol. 35, pp. 32–49, 2006. [30] Z. Molnár, R. Pálföldi, A. László et al., “Eﬀects of chronic and subtoxic chlorobenzenes on adrenocorticotrophic hor- mone release,” Journal of Environmental Sciences, vol. 34, pp. 165–170, 2015. [45] O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. References 30, no. 4, pp. 285–287, 2013. [6] U. Rajamani, A. R. Gross, C. Ocampo, A. M. Andres, R. A. Gottlieb, and D. Sareen, “Endocrine disruptors induce pertur- bations in endoplasmic reticulum and mitochondria of human pluripotent stem cell derivatives,” Nature Communications, vol. 8, no. 1, p. 219, 2017. [22] A. Papadimitriou and K. N. Priftis, “Regulation of the hypothalamic-pituitary-adrenal axis,” Neuroimmunomodula- tion, vol. 16, no. 5, pp. 265–271, 2009. [23] R. S. Wang, A. Saadatpour, and R. Albert, “Boolean model- ing in systems biology: an overview of methodology and applications,” Physical Biology, vol. 9, no. 5, article 055001, 2012. [7] A. P. Santos-Silva, M. N. Andrade, P. Pereira-Rodrigues et al., “Frontiers in endocrine disruption: impacts of organotin on the hypothalamus-pituitary-thyroid axis,” Molecular and Cellular Endocrinology, vol. 460, pp. 246–257, 2018. [24] F. Orton, I. Lutz, W. Kloas, and E. J. Routledge, “Endocrine disrupting eﬀects of herbicides and pentachlorophenol: in vitro and in vivo evidence,” Environmental Science & Technology, vol. 43, no. 6, pp. 2144–2150, 2009. [8] T. T. Schug, A. Janesick, B. Blumberg, and J. J. Heindel, “Endocrine disrupting chemicals and disease susceptibility,” The Journal of Steroid Biochemistry and Molecular Biology, vol. 127, no. 3–5, pp. 204–215, 2011. International Journal of Endocrinology Kobayashi, “Screening for estrogen and androgen receptor activities in 200 pesticides by in vitro reporter gene assays using Chinese hamster ovary cells,” Environmental Health Perspectives, vol. 112, no. 5, pp. 524–531, 2004. [53] M. C. Oh and M. K. Aghi, “Dopamine agonist-resistant prolactinomas,” Journal of Neurosurgery, vol. 114, no. 5, pp. 1369–1379, 2011. [37] M. Morita, “Chlorinated benzenes in the environment,” Ecotoxicology and Environmental Safety, vol. 1, no. 1, pp. 1–6, 1977. [54] R. L. Spencer and T. Deak, “A users guide to HPA axis research,” Physiology and Behavior, vol. 178, pp. 43–65, 2017. [38] D. Mackay, W. Y. Shiu, K. C. Ma, and S. C. Lee, Handbook of Physical-Chemical Properties and Environmental Fate for Organic Chemicals, CRC Press, Taylor & Francis, Boca Raton, FL, USA, 2nd edition, 2006. [39] National Toxicology Program, “Toxicology and carcinogen- esis studies of 1,4-dichlorobenzene (CAS no. International Journal of Endocrinology 106-46-7) in F344/N rats and B6C3F1 mice (gavage studies),” National Stem Cells International Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 MEDIATORS INFLAMMATION of Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Disease Markers Hindawi www.hindawi.com Volume 2018 BioMed Research International Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Hindawi www.hindawi.com Volume 2018 PPAR Research Hindawi Publishing Corporation http://www.hindawi.com Volume 2013 Hindawi www.hindawi.com The Scientific World Journal Volume 2018 Immunology Research Hindawi www.hindawi.com Volume 2018 Journal of Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Diabetes Research Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Gastroenterology Research and Practice Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Submit your manuscripts at www.hindawi.com Hindawi www.hindawi.com Volume 2018 MEDIATORS INFLAMMATION of Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Disease Markers Hindawi Publishing Corporation http://www.hindawi.com Volume 2013 Hindawi www.hindawi.com The Scientific World Journal Volume 2018 Immunology Research Hindawi www.hindawi.com Volume 2018 Journal of Diabetes Research Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Gastroenterology Research and Practice Submit your manuscripts at www.hindawi.com The Scientific World Journal Hindawi Publishing Corporation Hindawi The Scientific World Journal Diabetes Research Journal of Hindawi Endocrinology International Journal of Hindawi www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Hindawi www.hindawi.com Volume 2018 BioMed Research International Submit your manuscripts at www.hindawi.com Stem Cells International Hindawi www.hindawi.com Volume 2018 Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Stem Cells International Hindawi www.hindawi.com Volume 2018 Oncology Journal of Hindawi www.hindawi.com Volume 2013 Hindawi www.hindawi.com Volume 2018 Oxidative Medicine and Cellular Longevity Obesity Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Computational and Mathematical Methods in Medicine Hindawi www.hindawi.com Volume 2018 Behavioural Neurology Ophthalmology Journal of Hindawi www.hindawi.com Volume 2018 Hindawi www.hindawi.com Volume 2018 Research and Treatment AIDS Hindawi www.hindawi.com Volume 2018 Parkinson’s Disease Evidence-Based Complementary and Alternative Medicine Volume 2018 Hindawi www.hindawi.com Obesity Journal of Hindawi www.hindawi.com Volume 2018
https://openalex.org/W2777500966	https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0188810&type=printable	English	null	Efficacy and safety of selective glucocorticoid receptor modulators in comparison to glucocorticoids in arthritis, a systematic review	PloS one	2,017	cc-by	6,476	RESEARCH ARTICLE Editor: Irina Budunova, Northwestern University, UNITED STATES Editor: Irina Budunova, Northwestern University, UNITED STATES Received: August 25, 2017 Accepted: October 23, 2017 Published: December 21, 2017 Received: August 25, 2017 Accepted: October 23, 2017 Published: December 21, 2017 Efficacy and safety of selective glucocorticoid receptor modulators in comparison to glucocorticoids in arthritis, a systematic review M. Safy1, M. J. H. de Hair1, J. W. G. Jacobs1, F. Buttgereit2, M. C. Kraan3, J. M. van Laar1 1 Department of Rheumatology & Clinical Immunology, University Medical Center Utrecht, Utrecht, the Netherlands, 2 Department of Rheumatology and Clinical Immunology, Charite´—Universita¨tsmedizin Berlin, Berlin, Germany, 3 Department of Rheumatology and Inflammation Research at Institute of Medicine, University of Gothenburg, Gothenburg, Sweden a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 M.Safy@umcutrecht.nl Objective To systematically review the current literature on efficacy and safety of oral SGRMs in com- parison to GCs in arthritis. Copyright: © 2017 Safy et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Background OPEN ACCESS Citation: Safy M, de Hair MJH, Jacobs JWG, Buttgereit F, Kraan MC, van Laar JM (2017) Efficacy and safety of selective glucocorticoid receptor modulators in comparison to glucocorticoids in arthritis, a systematic review. PLoS ONE 12(12): e0188810. https://doi.org/ 10.1371/journal.pone.0188810 Long-term treatment with glucocorticoids (GCs) plays an important role in the management of arthritis patients, although the efficacy/safety balance is unfavorable. Alternatives with less (severe) adverse effects but with good efficacy are needed. Selective GC receptor modulators (SGRMs) are designed to engage the GC receptor with dissociative characteris- tics: transactivation of genes, which is mainly responsible for unwanted effects, is less strong while trans-repression of genes, reducing inflammation, is maintained. It is expected that SGRMs thus have a better efficacy/safety balance than GCs. A systematic review pro- viding an overview of the evidence in arthritis is lacking. Methods A search was performed in Medline, Embase and the Cochrane Library, from inception dates of databases until May 2017. Experimental studies involving animal arthritis models or human material of arthritis patients, as well as clinical studies in arthritis patients were included, provided they reported original data. All types of arthritis were included. Data was extracted on the SGRM studied and on the GC used as reference standard; the design or setting of the study was extracted as well as the efficacy and safety results. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: MS was supported with a research grant from AstraZeneca. AstraZeneca did not have any additional role in the study design, data collection and analysis, decision to publish or preparation of the manuscript. The specific roles of this author are articulated in the ‘author contributions’ section. Conclusion Studies assessing both efficacy and safety of SGRMs are scarce. There is limited evidence for dissociation of anti-inflammatory and metabolic effects of the SGRMs studied. Develop- ment of many SGRMs is haltered in a preclinical phase. One SGRM showed a better clinical efficacy/safety balance. Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis using human material, and 10 studies involved pre-clinical animal models (acute and/or chronic arthritis induced in mice or rats). PF-04171327, the only compound investigated in a clinical trial setting, had a better efficacy/safety balance compared to GCs: better clinical anti-inflammatory efficacy and similar safety. Competing interests: We have read the journal’s policy and the authors of this manuscript have the following competing interests: MS received a student grant from AstraZeneca. MK is holder of AstraZeneca stock and received a honorary from AstraZeneca as well. AstraZeneca was not involved in interpretation and analysis of the data. FB reported receiving consultancy fees, honoraria and travel expenses from Pfizer. He served as principal investigator in a Pfizer sponsored trial in rheumatoid arthritis investigating the effects of fosdagrocorate. JMvL received honoraria from MSD, Roche, Pfizer, BMS, Eli Lilly. This does not alter our adherence to PLOS ONE policies on sharing data and materials. Results A total of 207 articles was retrieved of which 17 articles were eligible for our analysis. Two studies concerned randomized controlled trials (RCT), five studies were pre-clinical studies 1 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 Introduction Glucocorticoids (GCs) are the most commonly used anti-inflammatory drugs worldwide, applied in arthritic diseases, inflammatory bowel disease, and chronic pulmonary disease, for example [1–3]. In rheumatoid arthritis (RA) between 56% and 68% of the patients are treated with GCs [4–6]. GCs not only exhibit anti-inflammatory effects, but also have proven disease modifying effects as they halter radiological damage and improve physical disability in RA patients in addition to reducing disease activity [7–9]. Despite their proven beneficial effects, GCs potentially cause adverse effects. The most common adverse effects associated with GC use are cardiovascular events, endocrine/metabolic effects (weight gain, dysregulation of glu- cose metabolism and development of diabetes), infections, gastro-intestinal events and osteo- porosis [10–11]. These unwanted effects especially occur when used long-term (>6 months) and in high-dose (>10 mg/daily),and limit the dosing and duration of GC treatment [12]. Hence, the quest for alternatives with a better efficacy/safety balance continues, such as selec- tive GC receptor modulators (SGRMs). SGRMs are specifically designed to engage the GC receptor (GR) with dissociative characteristics: after binding to the GR, GCs may either bind to and activate transcription from gene promoters (transactivation) or interact with other tran- scription factors to change their function (transrepression). It is assumed that SGRMs promote transrepression over transactivation [13]. Transrepression is most critical for the anti-inflammatory effects of GCs, as it leads to decreased production of pro-inflammatory transcription factors such as nuclear factor-kappa B (NF-κB) and activator protein 1 (AP-1). On the contrary, transactivation is thought to cause detrimental effects of GCs [14]. Upon binding of a GC to GC response elements (GRE) trans- activation in various gene promotors occurs, such as glucose-6-phosphatase (G6Pase), phos- phoenolpyruvate carboxykinase (PEPCK), fatty acid synthase (FAS) and tyrosine aminotransferase (TAT). The protein products of these genes are involved in carbohydrate, lipid and protein metabolism [15]. As such, activation of these genes could lead to aforemen- tioned (metabolic) side effects. Transrepression is most critical for the anti-inflammatory effects of GCs, as it leads to decreased production of pro-inflammatory transcription factors such as nuclear factor-kappa B (NF-κB) and activator protein 1 (AP-1). On the contrary, transactivation is thought to cause detrimental effects of GCs [14]. Upon binding of a GC to GC response elements (GRE) trans- activation in various gene promotors occurs, such as glucose-6-phosphatase (G6Pase), phos- phoenolpyruvate carboxykinase (PEPCK), fatty acid synthase (FAS) and tyrosine aminotransferase (TAT). Introduction The protein products of these genes are involved in carbohydrate, lipid and protein metabolism [15]. As such, activation of these genes could lead to aforemen- tioned (metabolic) side effects. SGRMs may have an improved efficacy/safety balance compared to conventional GCs, by their potentially disparate effects on transrepression and transactivation. However, till date, no SGRM has entered the market yet, suggesting that development of SGRMs meets challenges. A systematic review providing an in-depth overview of both the efficacy and safety of SGRMs is lacking. Our aim was therefore to systematically investigate whether oral SGRMs have a supe- rior efficacy/safety balance compared to conventional GCs in arthritis in (pre)clinical settings. PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 2 / 12 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Search and selection A systematic literature search was performed, to assess efficacy and safety of oral SGRMs in arthritis, compared to GCs. MEDLINE (PubMed), Embase and the Cochrane Library were searched until May 2017. The search (S1 Box) was established after consultation of a librarian at the University Medical Center (UMC) Utrecht with expertise in systematic literature searches (P.H.W.). Duplicates were excluded. Two authors (M.S. and M.J.H.H.) independently screened titles and abstracts for eligibility. Studies were included if fulfilling the following criteria: investi- gating efficacy and safety of an oral SGRM; studying GC as reference compound; performed in arthritis. Both in vivo and in vitro, were included. Subsequently, the same authors indepen- dently screened full texts of eligible articles. Selection was based on mutual agreement. Studies were excluded if not investigating a GC as reference compound; if performed in non-arthritic disease(s); if investigating non-selective GRMs, or if investigating SGRMs with administration route other than oral. Review articles without presentation of original data were also excluded. Of the selected articles, references and citing publications were additionally screened. Data extraction Data was extracted using the SYstematic Review Centre for Laboratory animal Experimenta- tion (SYRCLE’s) guideline ([16]. This guideline is adapted from the Cochrane risk of bias tool [17] and focuses on laboratory animal studies. We extracted data on the SGRM investigated (experimental compound), the GC that was investigated as reference compound, the animal model or setting of the study, and efficacy and safety results. Initial data extraction was per- formed by one author (M.S.) and extracted data was re-assed by the second author (M.J.H.H.). The efficacy results concerned pro- and anti-inflammatory effects and the safety results con- cerned any adverse effect reported, including effects on glucose, fat and bone metabolism, as well as mineralocorticoid effects. For clinical studies, also results on adverse effects of GC were extracted, such as cardiovascular events and infections. Results were reported following the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) checklist [18]. Search and selection The search and selection is presented in Fig 1. A total of 207 reports was retrieved by the initial search. Excluding duplicates and reports other than articles resulted in 81 articles, of which title and abstract were screened. This resulted in 40 articles of which full text was screened. Finally, 17 articles were eligible for inclusion and analysis. Risk of bias was not assessed, because of very high heterogeneity in study types and in study design and information presented. Data extraction Study characteristics are shown in Table 1. Results are reported using the PRISMA checklist (S1 Table). The following SGRMs were investigated: Compound A, PF-04171327, LGD-5552, Compounds 4, 5, and 14, Compounds (R)-16, (R)-18, (R)-21, (R)-35, and (R)-37, Ginsenoside Rg1 and Org 214007–0 [19–35]. Of the 17 studies, 2 studies concerned randomized controlled trials (RCT), 3 were pre-clini- cal studies using only human material, 10 studies were performed using only a pre-clinical ani- mal model (acute and/or chronic arthritis induced mice or rat model) and 2 pre-clinical 3 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 studies used both animal as well as human material Regarding the animal models an acute Fig 1. Flow chart of search and selection of studies on efficacy and safety of SGRMs. SGRMs: selective GRMs; GRMs: glucocorticoid receptor modulators; GCs: glucocorticoids. https://doi.org/10.1371/journal.pone.0188810.g001 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Fig 1. Flow chart of search and selection of studies on efficacy and safety of SGRMs. SGRMs: selective GRMs; GRMs: glucocorticoid receptor modulators; GCs: glucocorticoids. https://doi.org/10.1371/journal.pone.0188810.g001 https://doi.org/10.1371/journal.pone.0188810.g001 studies used both animal as well as human material. Regarding the animal models, an acute arthritis induced model was used to measure pro-inflammatory cytokines, and a chronic induced arthritis model to measure a clinical outcome, such as paw swelling. Dexamethasone was used as reference GC in 7 studies, and prednisone in 11 studies. studies used both animal as well as human material. Regarding the animal models, an acute arthritis induced model was used to measure pro-inflammatory cytokines, and a chronic induced arthritis model to measure a clinical outcome, such as paw swelling. Dexamethasone was used as reference GC in 7 studies, and prednisone in 11 studies. Seven studies showed better safety of the studied SGRM compared to dexamethasone or prednisone, with similar efficacy. Four compounds showed similar efficacy of the studied 4 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Table 1. Overview of studies with efficacy or safety results of a selective glucocorticoid receptor modulator in comparison to a glucocorticoid. Author, year SGRM tested Reference GC Setting Safety Conclusion on efficacy and safety in comparison to a glucocorticoid Glucose homeostasis/ fat metabolism Bone markers Dewint et al., 2008 Compound A Dex B, C X Similar efficacy. Better safety of SGRM. Gossye et al., 2009 Compound A Dex B Similar efficacy. No safety data shown. Gossye et al., 2010 Compound A Dex B, C Lower efficacy of SGRM. No safety data shown. Rauch et al.,2011 Compound A Dex B X Similar efficacy. Better safety of SGRM. Rauner et al., 2013 Compound A Dex C X Lower efficacy of SGRM. Better safety of SGRM. Malaise et al., 2015 Compound A Pred B X Similar efficacy. Better safety of SGRM. Yang et al.,2015 Compound 4 and 5 Pred C Similar efficacy of compound 4, better efficacy of compound 5. No safety data shown. Razavi et al., 2014 Compound 14 Pred C X Similar efficacy. Better safety of SGRM. Riether et al., 2010 Compounds (R)-16 and (R)-37 Pred C X Similar efficacy. Better safety of SGRM. Harcken et al., 2014 Compounds (R)-18 and (R)-21 Pred C X X Similar efficacy. Better safety of SGRMs. Weinstein et al., 2011 Compound35 and 37 Pred, Dex C Better efficacy of SGRMs. No safety data shown. Miner et al., 2007 LGD-5552 Pred C Similar efficacy. No safety data shown. Lopez et al.,2008 LGD-5552 Pred C Similar efficacy. No safety data shown. Du et al., 2011 Ginsenoside Rg1 Dex C X X Similar efficacy. Better safety of SGRM. Van Lierop et al., 2012 Org 214007–0 Pred C Similar efficacy. No safety data shown. Conrado et al., 2015 PF-04171327 Pred D Similar efficacy. No safety data shown. Stock et al., 2017 PF-04171327 Pred A X X Better efficacy of SGRM. Similar safety. Studies are sorted on type of SGRM. Efficacy was measured in all 17 studies, safety was measured in nine studies. SGRM: selective glucocorticoid receptor modulator; GC: glucocorticoid; Pred: prednisone; Dex: dexamethasone; A: randomized controlled trial (RCT); B: pre-clinical study with human material; C: pre-clinical study with animal material/model; D: stochastic simulations based on non-published RCT Studies are sorted on type of SGRM. Efficacy was measured in all 17 studies, safety was measured in nine studies. SGRM: selective glucocorticoid receptor modulator; GC: glucocorticoid; Pred: prednisone; Dex: dexamethasone; A: randomized controlled trial (RCT); B: pre-clinical study with human material; C: pre-clinical study with animal material/model; D: stochastic simulations based on non-published RCT Studies are sorted on type of SGRM. Efficacy was measured in all 17 studies, safety was measured in nine studies. SGRM: selective glucocorticoid receptor modulator; GC: glucocorticoid; Pred: prednisone; Dex: dexamethasone; A: randomized controlled trial (RCT); B: pre-clinical study with human material; C: pre-clinical study with animal material/model; D: stochastic simulations based on non-published RCT SGRM compared to prednisone, but no safety data was provided. Three compounds showed better efficacy of the studies SGRM than prednisone or dexamethasone, but no safety data was provided for two compounds. In depth results of 6 studies that assessed both efficacy and safety of SGRMs in comparison to GCs are depicted in Table 2 and only studies reporting efficacy and safety results of both SGRMs and GCs were included in this table. Fosdagrocorat (PF- 04171327), was the only compound investigated in a clinical setting [35], and of this SGRM both safety and efficacy data was available. In this phase 2 study, 86 RA patients were random- ized to receiving either 10 mg or 25 mg fosdagrocorat, or 5 mg prednisone or placebo. A signif- icantly better improvement in DAS28-CRP was observed after two weeks of treatment with 25 mg fosdagrocorat compared to 5 mg prednisone and placebo. Treatment with the 10 mg dose of fosdagrocorat was only compared to placebo, not prednisone. Plasma cortisol levels 5 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 SGRM: selective glucocorticoid receptor modulator; GC: glucocorticoid; pred: prednisone, dex: dexamethasone; FLS: fibroblast-like synoviocytes; CIA: collagen induced arthritis; cDNA: copy DNA; TNF: tumor necrosis factor; MMP: matrix metalloproteinase; P1NP: N-terminal propeptide of type 1 collagen; CTX1: collagen type 1 cross-linked C-telopeptide; PBMCs: peripheral blood mononuclear cells; IFN-α: interferon alpha; NF-κβ: Nuclear Factor kappa-light-chain-enhancer of activated B cells; mRNA: messenger ribonucleic acid; RA: rheumatoid arthritis; DAS28-4 (CRP): disease activity score using 28 joints and c-reactive protein with 4 variables; IL-6: interleukin 6; uNTX-1: N-terminal telopeptide 1 in urine; NA: not statistically analyzed; NS: not significant; S: significant; RCT: randomized controlled trial; ": results higher for SGRM compared to GC; #: results lower for SGRM compared to GC; =: results similar for SGRM compared to GC. https://doi.org/10.1371/journal.pone.0188810.t002 https://doi.org/10.1371/journal.pone.0188810.t002 decreased significantly more in the group of patients treated with 10 mg and 25 mg fosdagro- corat compared to 5 mg prednisone. The number of adverse events was similar between the group of patients receiving 25 mg fosdagrocorat compared to 5 mg prednisone. Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Table 2. Details of studies with data on both efficacy and safety of selective glucocorticoid receptor modulator in comparison to a glucocorticoid. Author, year SGRM tested GC Setting/model Efficacy (SGRM compared to GC) Safety (SGRM compared to GC) Dewint et al., 2008 Compound A Dex FLS cells derived from RA patients Amount of cDNA of TNF = NA Amount of cDNA of MMP1 = NA Amount of cDNA of MMP3 = NA CIA mice Arthritis score at day 8 of arthritis " NA Serum levels of insulin # NA Paw swelling at day 8 of arthritis = NA Normal histology of knee joints # NA mRNA G6P # NA mRNA PEPCK # NA Rauner et al., 2013 Compound A Dex CIA mice Arthritis score " NA Bone loss # NA Paw swelling " NA Serum P1NP # NA Paw temperature # NA Serum CTX-1 # NA Cellular infiltration in paws # NA Cartilage destruction # NA Inhibition of number of osteoclasts # NA Supernatant from PBMCs from mice, ex vivo stimulated with collagen type II TNF = NA IFN-a = NA NF-κβ = NA mRNA expression in joint tissue from CIA mice TNF " NA IL-6 # NA Razavi et al., 2014 Compound 14 Pred Mice inflammation model, LPS stimulated IL-6 # NA TNF = NA CIA mice Arthritis score " NA Insulin " NA Body fat " NA Triglycerides # NA Free fatty acids " NA Harcken et al., 2014 Compound R18 and 21 Pred CIA mice Arthritis score # NA Osteocalcin " NA Body fat " NA Triglycerides " NA Free fatty acids " NA Insulin " NA Femur cortical thickness = S Du et al., 2011 Ginsenoside Rg1 Dex Inflamed paw model Paw swelling " NS CIA mice Arthritis score = NA Body weight # NA Blood glucose levels # S Bone cortical thickness " NA Bone content " S Trabecular tibial number " S (Continued) PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 6 / 12 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis Table 2. (Continued) Author, year SGRM tested GC Setting/model Efficacy (SGRM compared to GC) Safety (SGRM compared to GC) Trabecular tibial thickness " NS Trabecular tibial separation " S Stock et al., 2017 PF-04171327 Pred Phase 2 RCT, 86 RA patients, 2 weeks treatment DAS28-4 (CRP) improvement " S Fasting glucose = NS Plasma cortisol " S Adverse events = NA Mean osteocalcin levels = NA Mean uNTX-1 levels = NA Six studies reporting results on both efficacy and safety of SGRMs compared to GCs are depicted, only studies that reported efficacy and safety of both SGRM and GC are shown in this table. Results have been summarized for each SGRM if multiple dosing schemes were used. SGRM: selective glucocorticoid receptor modulator; GC: glucocorticoid; pred: prednisone, dex: dexamethasone; FLS: fibroblast-like synoviocytes; CIA: collagen induced arthritis; cDNA: copy DNA; TNF: tumor necrosis factor; MMP: matrix metalloproteinase; P1NP: N-terminal propeptide of type 1 collagen; CTX1: collagen type 1 cross-linked C-telopeptide; PBMCs: peripheral blood mononuclear cells; IFN-α: interferon alpha; NF-κβ: Nuclear Factor kappa-light-chain-enhancer of activated B cells; mRNA: messenger ribonucleic acid; RA: rheumatoid arthritis; DAS28-4 (CRP): disease activity score using 28 joints and c-reactive protein with 4 variables; IL-6: interleukin 6; uNTX-1: N-terminal telopeptide 1 in urine; NA: not statistically analyzed; NS: not significant; S: significant; RCT: randomized controlled trial; ": results higher for SGRM compared to GC; #: results lower for SGRM compared to GC; =: results similar for SGRM compared to GC. Six studies reporting results on both efficacy and safety of SGRMs compared to GCs are depicted, only studies that reported efficacy and safety of both SGRM and GC are shown in this table. Results have been summarized for each SGRM if multiple dosing schemes were used. PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis transrepression for the desirable anti-inflammatory effects (full dissociation). Besides the described classical genomic mechanisms of action, which require several hours to take place, SGRMs also act by very rapid non-genomic mechanisms, especially at higher doses [40]. These non-genomic mechanisms of action are thought to be mediated by affecting the physicochemi- cal property of cell membranes, or through binding to intracellular or membrane-bound GC receptor, causing inflammatory signal transduction cascades (mitogen-activated protein kinases (MAPK), neutrophil degranulation and phagocytosis by macrophages) [41–44]. Com- bined with epigenetic effects of SGRMs, these two mechanisms also contribute to the lack of dissociative effects of SGRMs. Furthermore, in vitro studies use simplified GRE reporter sys- tems compared to the more complex GRE systems present in in vivo gene promoters [45]. Another important predicament in development of SGRMs is to establish equipotent doses of GCs and SGRMs. It has been shown that with increasing SGRM dosage, effects but also the SGRM-induced adverse effects increase [26,30] and vice versa. A case in point is deflazacort, an oxazoline derivative of prednisolone that was believed to have similar efficacy as prednisone but with fewer adverse effects, but in fact this actually proved to be at a lower than equipotent dosage; deflazacort even showed increased adverse effects compared to prednisolone in really equipotent dosages [46–47]. Furthermore, adverse effects measured in most of the experimen- tal studies, such as increased glucose levels and changes in cortical bone thickness, are in fact surrogate markers for clinical adverse effects in patients, respectively development of diabetes mellitus and osteoporosis. Thus, these parameters in preclinical studies do not fully reflect the clinical GC-related adverse effects. The only SGRM that did manage to enter a clinical phase in RA patients is PF-04171327 (NCT01393639), of which the first results of 12-week follow-up, were presented at the Annual European League Against Rheumatism (EULAR) Congress in 2015 [48]. In 323 RA patients, 15 mg of PF-04171327 daily showed similar efficacy as prednisone 10 mg daily, assessed by American College of Rheumatology (ACR) 20 response and Disease Activity Score 28 (DAS28), while (unwanted) effects on bone formation and plasma glucose level were similar as 5 mg of prednisone daily. These preliminary results suggest that development of SGRMs with a better efficacy/safety balance compared to GCs (better clinical anti-inflammatory efficacy and similar safety) is feasible. The strengths of our study include the thorough search across several databases and inclu- sion of pre-clinical and clinical studies. The present systematic review is the first to investigate the benefit and risks of oral SGRMs compared to GCs in arthritis. A limitation of our review is the heterogeneity of the reported studies which made it difficult to compare these studies. We investigated efficacy (transrepression) by measuring effects of the SGRMs on inflammatory markers, and safety (transactivation) by measuring effects on glucose and bone metabolism. However, measurable effects on bone metabolism are more difficult to detect compared to effects on glucose levels in studies with short duration. This could be an explanation why only three of the 17 studies examined effect on bone markers. In conclusion, studies assessing both efficacy and safety parameters of SGRMs in arthritis are scarce. There is limited evidence for dissociation of anti-inflammatory and metabolic effects of the SGRMs studied. Development of many SGRMs is haltered in a preclinical phase. One SGRM showed a better clinical efficacy/safety balance, compared to prednisone. Discussion This paper aimed to systematically review the efficacy and safety of SGRMs compared to con- ventional GCs. We found 17 studies which investigated a SGRM compared to a GC in arthritis, of which seven showed similar efficacy and better safety compared to GCs. However only one SGRM, fosdagrocorat/PF-04171327, was investigated in a clinical setting. There are several possible explanations as to why most of these SGRMs did not enter the clinical phase of drug development. One of them being the fact that some adverse effects asso- ciated with GC treatment are presumed to be caused by transrepression rather than by transac- tivation. For example, the immunosuppressive effects of GCs, leading to an increased risk of infections, are predominantly caused by transrepression rather than transactivation and there- fore this clinically important adverse effect will not be reduced by a dissociative compound [36]. Other side effects, such as osteoporosis, are mediated by both transrepression (osteocalcin transcription) and transactivation (osteoblast apoptosis) [14]. Furthermore, transactivation is not only associated with negative effects, as it has been demonstrated that some genes that are upregulated by transactivation, such as mitogen-activated protein kinase phosphatase-1 (MKP-1, a crucial anti-inflammatory gene), GC-induced leucine zipper (GILZ, a protein which inhibits NFκB and AP-1) and the anti-inflammatory interleukine IL-10, have anti- inflammatory functions [37–39]. Thus, the actual effects of transrepression and transactivation are much more complex than suggested by the hypothesized working mechanism of SGRMs, in which it is claimed that transactivation is solely responsible for the adverse effects and 7 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 S1 Box. PubMed search for studies on efficacy and safety of selective glucocorticoid recep- tor modulators in comparison to glucocorticoids, in arthritis. (DOCX) Acknowledgments We thank Paulien Wiersma, librarian at University Medical Center Utrecht, the Netherlands, as well as Kicki Johansson, Lena Ripa, and Per A¨berg, employed at AstraZeneca, Mo¨lndal, Swe- den, for their contribution to the design of the literature search. Project administration: M. Safy. Resources: M. Safy. Software: M. Safy. Supervision: M. Safy, M. J. H. de Hair, J. M. van Laar. Validation: M. Safy. Validation: M. Safy. Visualization: M. Safy. Visualization: M. Safy. Writing – original draft: M. Safy. Writing – original draft: M. Safy. Writing – review & editing: M. Safy, M. J. H. de Hair, J. W. G. Jacobs, F. Buttgereit, M. C. Kraan, J. M. van Laar. Supporting information S1 Box. PubMed search for studies on efficacy and safety of selective glucocorticoid recep- tor modulators in comparison to glucocorticoids, in arthritis. (DOCX) 8 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis S1 Table. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Checklist. (DOC) Author Contributions Conceptualization: M. Safy, J. W. G. Jacobs, J. M. van Laar. Conceptualization: M. Safy, J. W. G. Jacobs, J. M. van Laar. Formal analysis: M. Safy, M. J. H. de Hair. Investigation: M. Safy, M. J. H. de Hair. Methodology: M. Safy, M. J. H. de Hair, J. W. G. Jacobs. References 1. Barnes PJ. How corticosteroids control inflammation: Quintiles Prize Lecture 2005. Br J Pharmacol. 2006; 148:245–54. https://doi.org/10.1038/sj.bjp.0706736 PMID: 16604091 2. Normansell R, Kew KM, Mansour G. Different oral corticosteroid regimens for acute asthma. Cochrane database Syst Rev. 2016;CD011801. https://doi.org/10.1002/14651858.CD011801.pub2 PMID: 27176676 3. Buchner AM, Blonski W, Lichtenstein GR. Update on the management of Crohn’s disease. Curr Gastro- enterol Rep. 2011; 13:465–74. https://doi.org/10.1007/s11894-011-0220-x PMID: 21792543 4. Weinblatt ME, Kremer JM, Coblyn JS, Maier AL, Helfgott SM, Morrell M, et al. Pharmacokinetics, safety, and efficacy of combination treatment with methotrexate and leflunomide in patients with active rheuma- toid arthritis. Arthritis Rheum. 1999; 42:1322–8. https://doi.org/10.1002/1529-0131(199907) 42:7<1322::AID-ANR4>3.0.CO;2-P PMID: 10403258 5. Zink A, Listing J, Niewerth M, Zeidler H. The national database of the German Collaborative Arthritis Centres:II, Treatment of patients with rheumatoid arthritis. Ann Rheum Dis. 2001; 60:207–13. https:// doi.org/10.1136/ard.60.3.207 PMID: 11171679 6. Buttgereit F, Straub RH, Wehling M, Burmester GR. Glucocorticoids in the treatment of rheumatic dis- eases an update on the mechanisms of action. Arthritis Rheum. 2004; 50:3408–17. https://doi.org/10. 1002/art.20583 PMID: 15529366 7. Bakker MF, Jacobs JW, Welsing PM, Verstappen SM, Tekstra J, Ton E, et al. Low-dose prednisone inclusion in a methotrexate-based, tight control strategy for early rheumatoid arthritis: a randomized 9 / 12 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis trial. Ann Intern Med. 2012; 156:329–39. https://doi.org/10.7326/0003-4819-156-5-201203060-00004 PMID: 22393128 trial. Ann Intern Med. 2012; 156:329–39. https://doi.org/10.7326/0003-4819-156-5-201203060-00004 PMID: 22393128 8. 8. den Uyl D, ter Wee M, Boers M, Kerstens P, Voskuyl A, Nurmohamed M, et al. A non-inferiority trial of an attenuated combination strategy (“COBRA-light”) compared to the original COBRA strategy: clinical results after 26 weeks. Ann Rheum Dis. 2014; 73:1071–8. https://doi.org/10.1136/annrheumdis-2012- 202818 PMID: 23606682 9. Goekoop-Ruiterman YP, de Vries-Bouwstra JK, Allaart CF, van Zeben D, Kerstens PJ, Hazes JM, et al. Clinical and radiographic outcomes of four different treatment strategies in patients with early rheuma- toid arthritis (the BeSt study): a randomized, controlled trial. Arthritis Rheum. 2005; 52:3381–90. https:// doi.org/10.1002/art.21405 PMID: 16258899 10. Hoes JN, Jacobs JWG, Buttgereit F, Bijlsma JW. Current view of glucocorticoid co-therapy with DMARDs in rheumatoid arthritis. Nat Rheumatol Rev. 2010; 6:693–702. 11. Hoes JN, Jacobs JWG, Boers M, Boumpas D, Buttgereit F, Caeyers N, et al. EULAR evidence-based recommendations on the management of systemic glucocorticoid therapy in rheumatic diseases. Ann Rheum Dis. 2007; 66:1560–7. https://doi.org/10.1136/ard.2007.072157 PMID: 17660219 12. References Smolen JS, Landewe´ R, Bijlsma JW, Burmester G, Chatzidionysiou K, Dougados M, et al. EULAR rec- ommendations for the management of rheumatoid arthritis with synthetic and biological disease-modify- ing antirheumatic drugs: 2016 update. Ann Rheum Dis. 2017; 76:960–77. https://doi.org/10.1136/ annrheumdis-2016-210715 PMID: 28264816 13. Sedwick C. Wanted: a new model for glucocorticoid receptor transactivation and transrepression. PLoS Biol. 2014; 12:e1001814. https://doi.org/10.1371/journal.pbio.1001814 PMID: 24642535 14. Schacke H, Docke WD, Asadullah K. Mechanisms involved in the side effects of glucocorticoids. Phar- macol Ther. 2002; 96:23–43. PMID: 12441176 15. Jantzen HM, Stra¨hle U, Gloss B, Stewart F, Schmid W, Boshart M, et al. Cooperativity of glucocorticoid response located far upstream of the tyrosine aminotransferase. Cell. 1987; 49:29–38. PMID: 2881624 16. Hooijmans CR, Wever KE, de Vries RBM. SYRCLE’s starting guide for systematic reviews of preclinical animal intervention studies. 2016;1. Available from:https://issuu.com/radboudumc/docs/syrcle_ starting_guide_tool?e=28355229/45970684 17. Higgins JP, Altman DG, Gøtzsche PC, Ju¨ni P, Moher D, Oxman AD, et al. The Cochrane Collabora- tion’s tool for assessing risk of bias in randomised trials. BMJ. 2011; 343:d5928. https://doi.org/10.1136/ bmj.d5928 PMID: 22008217 18. Moher D, Liberati A, Tetzlaff J, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009; 6:e1000097. https://doi.org/10.1371/journal. pmed.1000097 PMID: 19621072 19. Dewint P, Gossye V, De Bosscher K, Vanden Berghe W, Van Beneden K, Deforce D, et al. A plant- derived ligand favoring monomeric glucocorticoid receptor conformation with impaired transactivation potential attenuates collagen-induced arthritis. J Immunol. 2008; 180:2608–15. PMID: 18250472 20. Gossye V, Elewaut D, Bougarne N, Bracke D, Van Calenbergh S, Haegeman G, et al. Differential mech- anism of NF-kappaB inhibition by two glucocorticoid receptor modulators in rheumatoid arthritis synovial fibroblasts. Arthritis Rheum. 2009; 60:3241–50. https://doi.org/10.1002/art.24963 PMID: 19877072 21. Gossye V, Elewaut D, Van Beneden K, Dewint P, Haegeman G, De Bosscher K, et al. A plant-derived glucocorticoid receptor modulator attenuates inflammation without provoking ligand-induced resistance. Ann Rheum Dis. 2010; 69:291–6. https://doi.org/10.1136/ard.2008.102871 PMID: 19204014 22. Rauch A, Gossye V, Bracke D, Gevaert E, Jacques P, Van Beneden K, et al. An anti-inflammatory selective glucocorticoid receptor modulator preserves osteoblast differentiation. FASEB J. 2011; 25:1323–32. https://doi.org/10.1096/fj.10-173393 PMID: 21233489 23. Rauner M, Thiele S, Sinningen K, Winzer M, Salbach-Hirsch J, Gloe I, et al. Effects of the selective glu- cocorticoid receptor modulator compound a on bone metabolism and inflammation in male mice with collagen-induced arthritis. Endocrinology. 2013; 154:3719–28. https://doi.org/10.1210/en.2012-2221 PMID: 23885015 24. PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 References Malaise O, Relic B, Quesada-Calvo F, Charlier E, Zeddou M, Neuville S et al. Selective glucocorticoid receptor modulator compound A, in contrast to prednisolone, does not induce leptin or the leptin recep- tor in human osteoarthritis synovial fibroblasts. Rheumatol. 2015; 54:1087–92. 25. Yang MG, Dhar TGM, Xiao Z, Xiao HY, Duan JJ, Jiang B, et al. Improving the pharmacokinetic and CYP inhibition profiles of azaxanthene-based glucocorticoid receptor modulators-identification of (S)-5- (2-(9-fluoro-2-(4-(2-hydroxypropan-2-yl)phenyl)-5H-chromeno[2,3-b]pyridin-5-yl)-2-methylpropana- mido)-N-(tetrahydro-2H-pyran-4-yl)-1,3,4-thiadiazole-2-carboxamide (BMS-341). J Med Chem. 2015; 58:4278–90. https://doi.org/10.1021/acs.jmedchem.5b00257 PMID: 25905990 PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 10 / 12 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis 26. Razavi H, Riether D, Harcken C, Bentzien J, Dinallo RM, Souza D, et al. Discovery of a potent and dis- sociated non-steroidal glucocorticoid receptor agonist containing an alkyl carbinol pharmacophore. Bioorganic Med Chem Lett. 2014; 24:1934–40. 27. Riether D, Harcken C, Razavi H, Kuzmich D, Gilmore T, Bentzien J, et al. Nonsteroidal dissociated glu- cocorticoid agonists containing azaindoles as steroid A-ring mimetics. J Med Chem. 2010; 53:6681–98. https://doi.org/10.1021/jm100751q PMID: 20735001 28. Harcken C, Riether D, Kuzmich D, Liu P, Betageri R, Ralph M, et al. Identification of highly efficacious glucocorticoid receptor agonists with a potential for reduced clinical bone side effects. J Med Chem. 2014; 57:1583–98. https://doi.org/10.1021/jm4019178 PMID: 24506830 29. Weinstein DS, Gong H, Doweyko AM, Cunningham M, Habte S, Wang JH, et al. Azaxanthene Based Selective Glucocorticoid Receptor Modulators: Design, Synthesis, and Pharmacological Evaluation of (S)-4-(5-(1-((1,3,4-Thiadiazol-2-yl)amino)-2-methyl-1-oxopropan-2-yl)-5H-chromeno[2,3- b]pyridin-2- yl)-2-fluoro-N, N-dimethylbenzamide (BMS-776532) and its methylene homologue (BMS-791826). J Med Chem. 2011; 54:7318–33. https://doi.org/10.1021/jm200879j PMID: 21899328 30. Miner JN, Ardecky B, Benbatoul K, Griffiths K, Larson CJ, Mais DE, et al. Antiinflammatory glucocorti- coid receptor ligand with reduced side effects exhibits an altered protein-protein interaction profile. Proc Natl Acad Sci. 2007; 104:19244–9. https://doi.org/10.1073/pnas.0705517104 PMID: 18032610 31. Lopez FJ, Ardecky RJ, Bebo B, Benbatoul K, De Grandpre L, Liu S, et al. LGD-5552, an antiinflamma- tory glucocorticoid receptor ligand with reduced side effects, in vivo. Endocrinology. 2008; 149:2080–9. https://doi.org/10.1210/en.2007-1353 PMID: 18218700 32. Du J, Cheng B, Zhu X, Ling C. Ginsenoside Rg1, a novel glucocorticoid receptor agonist of plant origin, maintains glucocorticoid efficacy with reduced side effects. J Immunol. 2011; 187:942–50. https://doi. org/10.4049/jimmunol.1002579 PMID: 21666059 33. Van Lierop MJC, Alkema W, Laskewitz AJ, Dijkema R, van der Maaden HM, Smit MJ, et al. References Org 214007–0: A novel non-steroidal selective glucocorticoid receptor modulator with full anti-inflammatory properties and improved therapeutic index. Org 214007–0: a novel non-steroidal selective glucocorti- coid receptor modulator with full anti-inflammatory properties and improved therapeutic index. PLoS One.2012; 7:e48385. https://doi.org/10.1371/journal.pone.0048385 PMID: 23152771 34. Conrado DJ, Krishnaswami S, Shoji S, Kolluri S, Hey-Hadavi J, McCabe D, et al. Predicting the proba- bility of successful efficacy of a dissociated agonist of the glucocorticoid receptor from dose–response analysis. J Pharmacokinet Pharmacodyn. 2016; 43:325–41. https://doi.org/10.1007/s10928-016-9475- z PMID: 27178257 35. Stock T, Fleishaker D, Wang X, Mukherjee A, Mebus C. Improved disease activity with fosdagrocorat (PF-04171327), a partial agonist of the glucocorticoid receptor, in patients with rheumatoid arthritis: a Phase 2 randomized study. Int J Rheum Dis. 2017 Mar 22. https://doi.org/10.1111/1756-185X.13053 PMID: 28328159 36. Strehl C, van der Goes MC, Bijlsma JW, Jacobs JW, Buttgereit F. Glucocorticoid-targeted therapies for the treatment of rheumatoid arthritis. Expert Opin Investig Drugs. 2017; 26:187–195. https://doi.org/10. 1080/13543784.2017.1276562 PMID: 28043173 37. Ratman D, Vanden Berghe W, Dejager L, Libert C, Tavernier J, Beck IM, et al. How glucocorticoid receptors modulate the activity of other transcription factors: A scope beyond tethering. Mol Cell Endo- crinol. 2013; 380:41–54. https://doi.org/10.1016/j.mce.2012.12.014 PMID: 23267834 38. Vandevyver S, Dejager L, Tuckermann J, Libert C. New insights into the anti-inflammatory mechanisms of glucocorticoids: an emerging role for glucocorticoid-receptor-mediated transactivation. Endocrinol- ogy. 2013; 154:993–1007. https://doi.org/10.1210/en.2012-2045 PMID: 23384835 39. Buttgereit F, Bijlsma JWJ, Strehl C. Will we ever have better glucocorticoids? Clin Immunol. 2017 Jul 27. https://doi.org/10.1016/j.clim.2017.07.023 PMID: 28757452 40. Lo¨wenberg M, Stahn C, Hommes DW, Buttgereit F. Novel insights into mechanisms of glucocorticoid action and the development of new glucocorticoid receptor ligands. Steroids. 2008; 73:1025–9. https:// doi.org/10.1016/j.steroids.2007.12.002 PMID: 18221974 41. Stellato C. Post-transcriptional and nongenomic effects of glucocorticoids. Proc Am Thorac Soc. 2004; 1:255–63. https://doi.org/10.1513/pats.200402-015MS PMID: 16113443 42. Long F, Wang YX, Liu L, Zhou J, Cui RY, Jiang CL. Rapid nongenomic inhibitory effects of glucocorti- coids on phagocytosis and superoxide anion production by macrophages. Steroids. 2005; 70:55–61. https://doi.org/10.1016/j.steroids.2004.10.004 PMID: 15610897 43. Liu L, Wang YX, Zhou J, Long F, Sun HW, Liu Y. Rapid non-genomic inhibitory effects of glucocorticoids on human neutrophil degranulation. Inflamm Res. 2005; 54:37–41. https://doi.org/10.1007/s00011-004- 1320-y PMID: 15723203 44. Solito E, Mulla A, Morris JF, Christian HC, Flower RJ, Buckingham JC. PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 References Dexamethasone induces rapid serine-phosphorylation and membrane translocation of annexin 1 in a human folliculostellate cell line PLOS ONE \| https://doi.org/10.1371/journal.pone.0188810 December 21, 2017 11 / 12 Efficacy and safety of selective glucocorticoid receptor modulators compared to glucocorticoids in arthritis via a novel nongenomic mechanism involving the glucocorticoid receptor, protein kinase C, phosphati- dylinositol 3-kinase, and mitogen-activated protein kinase. Endocrinology. 2003; 144:1164–74. https:// doi.org/10.1210/en.2002-220592 PMID: 12639897 via a novel nongenomic mechanism involving the glucocorticoid receptor, protein kinase C, phosphati- dylinositol 3-kinase, and mitogen-activated protein kinase. Endocrinology. 2003; 144:1164–74. https:// doi.org/10.1210/en.2002-220592 PMID: 12639897 via a novel nongenomic mechanism involving the glucocorticoid receptor, protein kinase C, phosphati- dylinositol 3-kinase, and mitogen-activated protein kinase. Endocrinology. 2003; 144:1164–74. https:// doi.org/10.1210/en.2002-220592 PMID: 12639897 via a novel nongenomic mechanism involving the glucocorticoid receptor, protein kinase C, phosphati- dylinositol 3-kinase, and mitogen-activated protein kinase. Endocrinology. 2003; 144:1164–74. https:// doi.org/10.1210/en.2002-220592 PMID: 12639897 45. De Bosscher K, Beck IM, Haegeman G. Classic glucocorticoids versus non-steroidal glucocorticoid receptor modulators: Survival of the fittest regulator of the immune system? Brain Behav Immun. 2010; 24:1035–42. https://doi.org/10.1016/j.bbi.2010.06.010 PMID: 20600811 46. Krogsgaard MR, Thamsborg G, Lund B. Changes in bone mass during low dose corticosteroid treat- ment in patients with polymyalgia rheumatica: a double blind, prospective comparison between prednis- olone and deflazacort. Ann Rheum Dis. 1996; 55:143–46. PMID: 8712867 47. Jacobs JW, Bijlsma JW. Innovative combination strategy to enhance effect and diminish adverse effects of glucocorticoids: another promise? Arthritis Res Ther. 2009 Feb 27. https://doi.org/10.1186/ ar2615 PMID: 19291268 47. Jacobs JW, Bijlsma JW. Innovative combination strategy to enhance effect and diminish adverse effects of glucocorticoids: another promise? Arthritis Res Ther. 2009 Feb 27. https://doi.org/10.1186/ ar2615 PMID: 19291268 48. Buttgereit F, Strand V, Lee EB, McCabe D, Kolluri S, Tammara B. Efficacy and Safety of PF-04171327, a Novel Dissociated Agonist of the Glucocorticoid Receptor (DAGR): Results of a Phase 2, Random- ized, Double-Blind Study. Ann Rheum Dis. 2015; 74:737–38. 48. Buttgereit F, Strand V, Lee EB, McCabe D, Kolluri S, Tammara B. Efficacy and Safety of PF-04171327, a Novel Dissociated Agonist of the Glucocorticoid Receptor (DAGR): Results of a Phase 2, Random- ized, Double-Blind Study. Ann Rheum Dis. 2015; 74:737–38. 12 / 12
https://openalex.org/W3127385822	https://inria.hal.science/hal-03132416/document	English	null	Jacobian Free Methods for Coupling Transport with Chemistry in Heterogenous Porous Media	Water	2,021	cc-by	13,328	To cite this version: Laila Amir, Michel Kern. Jacobian Free Methods for Coupling Transport with Chemistry in Het- erogenous Porous Media. Water, 2021, 13 (3), pp.370. 10.3390/w13030370. hal-03132416 Distributed under a Creative Commons Attribution 4.0 International License Keywords: reactive transport; ﬁnite volume methods; Newton–Krylov methods; porous media ﬂow and transport Citation: Amir, L.; Kern, M. Jacobian Free Methods for Coupling Transport with Chemistry in Heterogenous Porous Media. Water 2021, 13, 370. https://doi.org/10.3390/w13030370 Laila Amir 1 and Michel Kern 2,3,* Laila Amir 1 and Michel Kern 2,3,* 1 Laboratoire d’Ingénierie Informatique et Systèmes (L2IS), Faculté des Sciences et Techniques, UCAM, 4000 Marrakech, Morrocco; l.amir@uca.ma 2 Inria, 2 Rue Simone Iff, 75589 Paris, France 1 Laboratoire d’Ingénierie Informatique et Systèmes (L2IS), Faculté des Sciences et Techniques, UCAM, 4000 Marrakech, Morrocco; l.amir@uca.ma * Correspondence: michel.kern@inria.fr Abstract: Reactive transport plays an important role in various subsurface applications, includ- ing carbon dioxide sequestration, nuclear waste storage, biogeochemistry and the simulation of hydro–thermal reservoirs. The model couples a set of partial differential equations, describing the transport of chemical species, to nonlinear algebraic or differential equations, describing the chemical reactions. Solution methods for the resulting large nonlinear system can be either fully coupled or can iterate between transport and chemistry. This paper extends previous work by the authors where an approach based on the Newton–Krylov method applied to a reduced system has been developed. The main feature of the approach is to solve the nonlinear system in a fully coupled manner while keeping transport and chemistry modules separate. Here we extend the method in two directions. First, we take into account mineral precipitation and dissolution reactions by using an interior point Newton method, so as to avoid the usual combinatorial approach. Second, we study two-dimensional heterogeneous geometries. We show how the method can make use of an existing transport solver, used as a black box. We detail the methods and algorithms for the individual modules, and for the coupling step. We show the performance of the method on synthetic examples. water water HAL Id: hal-03132416 https://inria.hal.science/hal-03132416v1 Submitted on 5 Feb 2021 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License water 1. Introduction Reactive transport plays an important role in various subsurface applications [1–3] including carbon dioxide sequestration, nuclear waste storage bio-geochemistry and the simulation of hydro-thermal reservoirs. Academic Editor: Anis Younes Received: 24 December 2020 Accepted: 25 January 2021 Published: 31 January 2021 The numerical simulation of reactive transport has been the topic of numerous work. The survey by Yeh and Tripathi [4] has been very inﬂuential in establishing a mathematical formalism for setting up models, and also for establishing the “operator splitting” approach (see below) as a standard. More recent surveys, detailing several widely used computer codes and their applications, can be found in the book [1] and the survey article [2]. Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institutional afﬁliations. Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institutional afﬁliations. pp y Methods for coupling transport and chemistry fall into two broad categories: • Sequential Iterative Approach (SIA), where transport and chemistry are solved alter- natively and one iterates to convergence [5–10]. • Sequential Iterative Approach (SIA), where transport and chemistry are solved alter- natively and one iterates to convergence [5–10]. • Globally Implicit Approach (GIA), where one solves the fully coupled system in one shot [11–14]. • Globally Implicit Approach (GIA), where one solves the fully coupled system in one shot [11–14]. Copyright: © 2021 by the authors. Li- censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/). Recently, methods that solve coupled problem without direct substitution, thus main- taining a separation between transport and chemistry, have been introduced [15–21]. In previous work [22,23], the authors introduced a method for the simulation of reactive transport that falls in the latter category. The method is a globally coupled approach, where transport and chemistry are solved together, but where the transport and https://www.mdpi.com/journal/water Water 2021, 13, 370. https://doi.org/10.3390/w13030370 Water 2021, 13, 370 2 of 26 chemistry modules are kept separate at the software level. At each time step, the large nonlinear system of algebraic equations representing the coupling of all chemical species at all mesh points is solved by a Newton–Krylov method [24]. The Newton–Krylov method has recently been used for closely related applications to reactive transport, see [25–27]. 1. Introduction In these works, the method is applied to a system where the mass action laws are subsituted into the mass balance transport equations, while in our work we solve a ﬁxed point system using solution operators for transport and chemistry. In the Newton–Krylov approach the linearized system at each Newton iteration is solved by an iterative method (usually GMRES), so that the full Jacobian matrix does not have to be formed. All that is required is a procedure to multiply the Jacobian matrix by a given vector, and this is where the speciﬁc structure of the coupled problem is exploited. It was also shown in [23] that a suitable non-linear preconditioning made the method both more robust and more efﬁcient, as the number of both linear and nonlinear iterations became independent of the mesh size. p The purpose of the present paper is to extend the approach to a larger set of models. Speciﬁcally, we target two different extensions: Mineral species: We extend the capabilities of the equilibrium chemical solver to handle mineral precipitation and dissolution; 2D geometries: All the examples in [22,23] were one-dimensional. Here we show 2D examples. 2D geometries: All the examples in [22,23] were one-dimensional. Here we show 2D examples. 2D geometries: All the examples in [22,23] were one-dimensional. Here we show 2D examples. Mineral precipitation and dissolution reactions lead to non-smooth problems, where it is not known a priori which species are dissolved. The corresponding equilibrium problem can be formulated in an elegant, and mathematically consistent, formulation as a complementarity problem, as for example in [28–30]. However, instead of using a Newton–min method as in those references, we solve the non-linear system by an interior point method [31] or [32], adapted to nonlinear equations as in [33,34]. The approach is validated by computing the potential-pH precipitiation diagram for iron. The chemical module is coupled to a transport module based on the recently devel- oped ComPASS code [35]. ComPASS targets multiphase, multicomponent geothermal systems. The discretization uses a fully implicit time integration combined with the Vertex Approximate Gradient (VAG) ﬁnite volume scheme [36], which is adapted to polyhedral meshes and anisotropic heterogeneous media. The extension of the coupling method to 2D is non-trivial as the cost of the transport step now becomes signiﬁcant. 1. Introduction The code is validated on two examples: The code is validated on two examples: • The potential pH diagram for iron is used to validate the interior point Newton method; • A 2D variant of the ion exchange system used in Example 11 of the PhreeqC docu- mentation [37]. This test case involves 8 chemical species undergoing 3 ion-exchange reactions, in a simple 2D geometry. p g y • The so-called SHPCO2 benchmark that involves 11 chemical species, distributed among 3 phases, and undergoing 4 reactions. One species is a mineral, and one is a gas. The results show that the proposed approach does handle the large class of models, bu that further work is still required to ascertain its efﬁciency on large scale examples. This topic is currently being pursued. An outline of the paper is as follows: Section 2 presents the model and its hypotheses. Section 3 shows how to obtain a reduced coupled problem, and discusses solution methods at the semi-discrete level. In Section 4, we present the numerical methods used for solving the chemical and transport sub-problems. Section 5 discusses the numerical examples. 2.1. The Model We consider a set of species subject to transport by advection and diffusion and to chemical reactions in a porous medium occupying a domain Ω⊂Rd (d = 1, 2, 3). The Water 2021, 13, 370 3 of 26 chemical phenomena involve both homogeneous and heterogeneous reactions. Homoge- neous reactions, in the aqueous phase, include water dissociation, acid–base reactions and redox reactions, whereas heterogeneous reactions occur between the aqueous and solid phases, and include surface complexation, ion exchange and precipitation and dissolution of minerals (see [38] for details on the modeling of speciﬁc chemical phenomena). Ac- cordingly, we assume there are Na aqueous (mobile) species (Xa j )j=1,...,Na in the aqueous phase undergoing Nr homogeneous reactions, Ns solid (immobile) species (Xs j )j=1,...,Ns in the solid phase undergoing Nr surface sorption reactions or ion exchange reactions, and Npprecipitation – dissolution reactions with the additional assumption that each min- eral reaction involves only aqueous species and only one mineral species, denoted by (Xp j )j=1,...,Np j j p In this work, we only consider equilibrium reactions, which means that the chemical phenomena occur on a much faster scale than transport phenomena, see [39]. This assump- tion is justiﬁed for aqueous phase and ion-exchange reactions, but may be questionable for reactions involving minerals that could also be modeled as kinetic reactions, with speciﬁc rate laws [13,40]. [ , ] Under the above assumptions, we can write the chemical system as Na ∑ j=1 (Saa)ijXa j ⇆0 i = 1, . . . , Nr homogeneous reactions, Na ∑ j=1 (Ssa)ijXa j + Ns ∑ j=1 (Sss)ijXs j ⇆0 i = 1, . . . Nr heterogeneous sorption reactions, Na ∑ j=1 (Spa)ijXa j −Xp i ⇆0 i = 1, . . . Np precipitation – dissolution reactions, Na ∑ j=1 (Saa)ijXa j ⇆0 i = 1, . . . , Nr homogeneous reactions, Na ∑ j=1 (Ssa)ijXa j + Ns ∑ j=1 (Sss)ijXs j ⇆0 i = 1, . . . Nr heterogeneous sorption reactions, −Xp i ⇆0 i = 1, . . . Np precipitation – dissolution reactions, or in condensed form S   Xa Xs Xp  ⇆   0 0 0  , where S =   Saa 0 0 Ssa Sss 0 Spa 0 −I   (1) (1) denotes the stoichiometric matrix, with the sub-matrices Saa ∈RNr×Na, Ssa ∈RNr×Na, Sss ∈RNr×Ns and Spa ∈RNp×Na. 2.1. The Model p y g g j For the aqueous and sorption reactions we write the corresponding mass action law Saa 0 Ssa Sss log c log s = log Ka log Ks , (2) (2) (2) where Ka ∈RNr, Ks ∈RNr are the equilibrium constants for their respective reactions. where Ka ∈RNr, Ks ∈RNr are the equilibrium constants for their respective reactions. For reactions involving minerals, we deﬁne the (logarithmic form of) the solubility product, with Kp ∈RNp denoting again the equilibrium constant where Ka ∈RNr, Ks ∈RNr are the equilibrium constants for their respective reactions. For reactions involving minerals, we deﬁne the (logarithmic form of) the solubility product, with Kp ∈RNp denoting again the equilibrium constant For reactions involving minerals, we deﬁne the (logarithmic form of) the solubility product, with Kp ∈RNp denoting again the equilibrium constant Π = log Kp −Spa log c. (3) (3) Mineral reactions can only be written if the mineral is present, and the mass action law must then be written as the following alternative: Mineral reactions can only be written if the mineral is present, and the mass action law must then be written as the following alternative: (Π = 0 and p ≥0) or (Π ≥0 and p = 0). (4) (4) Next, we write the mass conservation for each species, considering both transport by advection, diffusion and chemical reaction terms: φ∂tc + Lc = ST aara + ST sars + ST parp, ρs(1 −φ)∂ts = ST ssrs, ρs(1 −φ)∂tp = −rp. in Ω× [0, Tf ] (5) (5) where L denotes the advection–diffusion operator: where L denotes the advection–diffusion operator: L(c) = ∇· (qc −D∇c), L(c) = ∇· (qc −D∇c), φ is the porosity (fraction of void in a Representative Elementary Volume available for the ﬂow), q is the Darcy velocity (we assume here permanent ﬂow, so that q is consid- ered as known), D is a diffusion–dispersion coefﬁcient and ρs is the density of the rock matrix. The vectors ra ∈RNr, rs ∈RNr and rp ∈RNp contain the (unknown) reaction rates. We assume that the diffusion coefﬁcient is independent of the species. This is a strong restriction on the model, but one that is commonly assumed to hold [4,16,22,41]. To complete the model, we add boundary conditions, with a partition of the boundary ∂Ω= ΓD ∪ΓN, ΓD ∩ΓN = ∅. 2.1. The Model We assume that both the global stoichiometric matrix S and the “aqueous” stoichiometric matrix Saa are of full rank. Since the heterogeneous reactions in the examples used in Section 5 have either all surface reactions or all mineral reactions, we illustrate the above concepts with an example from [41] that has both kinds of reactions. The chemical reactions are as follows, the ﬁrst two reactions are homogeneous, the third one is a surface reaction and the fourth one is a mineral reaction: R1 H+ + OH– −−→ ←−−H2O R2 CO 2– 3 + H+ −−→ ←−−HCO – 3 R3 X2Ca + 2Na+ −−→ ←−−2XNa + Ca + 2 R4 Ca + 2 + CO 2– 3 −−→ ←−−CaCO3(s). R1 H+ + OH– −−→ ←−−H2O R2 CO 2– 3 + H+ −−→ ←−−HCO – 3 R3 X2Ca + 2Na+ −−→ ←−−2XNa + Ca + 2 R4 Ca + 2 + CO 2– 3 −−→ ←−−CaCO3(s). The aqueous species are Xa = H+, OH−, CO 2− 3 , Na+, Ca + 2 , H2O, HCO − 3 T , the sorbed species are Xs = (X2Ca, XNa)T and the mineral species is Xp = (CaCO3)T. The cor- responding matrices are: Saa = 1 1 0 0 0 −1 0 1 0 1 0 0 0 −1 , Ssa = 0 0 0 2 −1 0 0 , Sss = 1 −2 , Spa = 0 0 1 0 1 0 0 . Water 2021, 13, 370 4 of 26 Each reaction gives rise to a mass action law, linking the activities of the species. For simplicity, we ignore the activity correction, so that the activities of aqueous and sorption species are equal to their concentrations (if water is included as a species, as in example Section 5.3, its activity is taken equal to 1), while the activity of mineral species is by convention equal to 1. We denote by cj (resp. sj, pj) the concentration of species Xa j (resp. Xs j , Xp j ). It will be convenient to write the mass action law in logarithmic form: given a vector c with positive entries we denote by log c the vector with entries log cj. Dirichlet boundary: c = cD on ΓD × [0, Tf ], Neumann boundary: ∇c · n = 0 on ΓN × [0, Tf ], where n is the outgoing normal to Ω along ΓN. Neumann boundary: ∇c · n = 0 on ΓN × [0, Tf ], where n is the outgoing normal to Ω along ΓN. Finally, an initial condition c(x, 0) = c0(x) in Ω 2.1. The Model For simplicity we only consider either given concentration, or zero diffusive ﬂux boundaries: Dirichlet boundary: c = cD on ΓD × [0, Tf ], 2.2. Elimination of the Reaction and Coupled Problem 2.2. Elimination of the Reaction and Coupled Problem . Elimination of the Reaction and Coupled Problem With this kernel matrix, we can eliminate the reaction terms in Equation (5). We start by deﬁning the total analytic concentration for the mobile and immobile species respectively (these are the same as various total quantities deﬁned in the classical survey by Yeh and Tripathi [4]) T T = Uaa Uas Uap 0 Uss 0   c s p  . (7) (7) We denote by Nc = Na −Nr and Nc = Ns −Nr the respective dimensions of T and T. We also deﬁne the total mobile and immobile concentrations for the species in the aqueous phase We denote by Nc = Na −Nr and Nc = Ns −Nr the respective dimensions of T and T. We also deﬁne the total mobile and immobile concentrations for the species in the aqueous phase C = Uaa c, C = Uas s + Uapp, (8) (8) so that the total concentrations are given by so that the total concentrations are given by T = C + C = Uaa c + Uas s + Uapp. (9) (9) We now multiply system (5) on the left by U. Because of our assumption that D is the same for all chemical species, multiplication by U commutes with the differential operator, and the system can be rewritten as φ∂tC + ρs(1 −φ)∂tC + LC = 0, ρs(1 −φ)∂tT = 0. (10) (10) is given. g Since we assume that all reactions are at equilibrium the reaction rates are unknown, but they can be eliminated. The process is by now well known, and will not be repeated here. We adapt the construction given in [23], to take into account the extra structure due Water 2021, 13, 370 5 of 26 to the presence of mineral species. See [28] for a related derivation, and [16,17] for a more general framework. This approach, originating from [41], makes use of a kernel matrix U such that UST = 0, i.e. such that the columns of UT form a basis for the null-space of S. Among the many possible choices there exists one such that U = Uaa Uas Uap 0 Uss 0 Uaa ∈R(Na−Nr)×Na, Uas ∈R(Na−Nr)×Ns, Upa ∈R(Na−Nr)×Np Uss ∈R(Ns−Nr)×Ns (6) (6) For the example system given above, one ﬁnds that the kernel matrices are given by Uaa =       1 0 0 0 0 1 1 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0       , Usa =       0 0 0 0 0 0 −2 0 0 −1/2       , Uss = 2 1 , Upa =       0 0 −1 0 1       . 3. The Coupled Problem The coupled system consists of the (Na −Nr) + (Ns −Nr) conservation Partial Diffren- tial Equations (PDEs) and Ordinary Differential Equations (ODEs) (10), together with the Nr + Nr + Np mass action laws (2), the relations connecting concentration and totals (8) and the second line of (7), for the Na + Ns + Np concentrations and 2(Na −Nr) + Ns −Nr totals. Note that the ODEs for T are decoupled from the rest of the system. Water 2021, 13, 370 6 of 26 3.1. Reduction of the Coupled Problem 3.1. Reduction of the Coupled Problem It is possible to solve directly the coupled system for the concentrations of the species. This is what the Direct Substitution Approach (DSA) does. Actually this is usually done after a further reduction of the chemical system, by introducing primary and secondary species such that each mass action law expresses the formation of one secondary species from the primary species. One substitutes the mass action laws (2) and (4) into the mass conservation PDEs (10), obtaining a large system of nonlinear equations for the primary species (see [25] or [42] for representative examples of this approach). p p p pp We prefer, however, to separate the local chemical equations from the PDEs. In order to do this, it will be useful to eliminate the concentrations of the individual species, and keep only the totals as main unknowns. Towards this goal, we notice that given the totals T ∈Nc, T ∈NC, we can solve the “local chemical equilibrium problem” made up of the mass action laws (2) and (4) together with the deﬁnition of the totals (7). Knowing the species concentrations c, s and p, we can compute C ∈Nc, C ∈NC through Equation (8). Because T is constant, it is convenient to remove it from the system, and we deﬁne the chemical solution operator by ΨC : RNc →RNc T →C (11) (11) The end result is that the coupled problem given by Equations (2), (4), (7), (8) and (10) is equivalent to the following system involving only the totals C and C: φ∂tC + ρs(1 −φ)∂tC + LC = 0, C −ΨC(C + C) = 0. (12) (12) Remark 1. Note that one can equivalently add T as an unkwon with the added equation T = C + C. As we recall below, this is how the SIA method is deﬁned, but it turns out to be more efﬁcient to work directly with the smaller system (12) (see [23] for a comparison of the various formulations of the coupled problem). Before we can fully specify how to solve the reduced system (12), we need to make precise the numerical methods used for solving the transport equations and the local chemical system. We can then rewrite the coupled problem as We can then rewrite the coupled problem as rite the coupled problem as Cn+1 = ΨT Cn, −ρs(1 −φ)Cn+1 −Cn ∆tn ! Cn+1 = ΨC(Cn+1 + Cn+1). (15) (15) Among the various mathematically equivalent formulations, it was shown in [23] that eliminating Cn+1 leads to an efﬁcient formulation. The resulting problem involves only the unknown Cn+1 and is written as Cn+1 = def F(Cn+1) = ΨC Cn+1 + ΨT " Cn, −ρs(1 −φ)Cn+1 −Cn ∆tn #! . (16) (16) This is a ﬁxed point problem for Cn+1. We recall two ways by which this problem can be solved: This is a ﬁxed point problem for Cn+1. We recall two ways by which this problem can be solved: Fixed point A ﬁxed point iteration for solving (16) is known in the reactive transport litera- ture as the Sequential Iterative Approach (SIA, see [4,43–45] among many references). At each time step n, one iterates between transport and chemistry. More precisely, for each iteration k, one computes C(n+1,k+1) = F(C(n+1,k)) = ΨC C(n+1,k) + ΨT " Cn, −ρs(1 −φ) C(n+1,k) −Cn ∆tn #! . (17) (17) The iterations are stopped when the relative progress indicator ηk = ∥Cn+1,k+1 −Cn+1,k∥ ∥Cn+1,k∥ ηk = ∥Cn+1,k+1 −Cn+1,k∥ ∥Cn+1,k∥ becomes small enough. One can easily show that this is equivalent with the more standard description of SIA involving three unknowns (Cn+1, Tn+1 and Cn+1). Newton–Krylov As was shown in previous work by the authors [22,23], the coupled system (15) can also be solved by a Newton–Krylov method, leading to a more efﬁcient algorithm. Newton–Krylov As was shown in previous work by the authors [22,23], the coupled system (15) can also be solved by a Newton–Krylov method, leading to a more efﬁcient algorithm. Recall that at each step of the “pure” form of Newton’s method for solving F(Cn+1) = 0, one should compute the Jacobian matrix J = F′(C(n+1,k)), solve the linear system Recall that at each step of the “pure” form of Newton’s method for solving F(Cn+1) = 0, one should compute the Jacobian matrix J = F′(C(n+1,k)), solve the linear system J δC = −F(C(n+1,k)), (18) (18) usually by Gaussian elimination, and then set C(n+1.k+1) = C(n+1,k) + δC. In practice, one should use some form of globalization procedure in order to ensure convergence from an arbitrary starting point. 3.2. Semi-Discrete System in Time We ﬁrst keep space continuous, and describe our approach on a semi-discrete sys- tem. We introduce a discretization of the time interval [0, Tf ] into Nt intervals [tn, tn+1], n = 0, . . . , Nt −1, with t0 = 0 and tNt = Tf , and we denote ∆tn = tn+1 −tn. We take the simplest backward Euler discretization and approximate the coupled system (12) by φCn+1 −Cn ∆tn + ρs(1 −φ)Cn+1 −Cn ∆tn + LCn+1 = 0, Cn+1 −ΨC(Cn+1 + Cn+1) = 0. (13) (13) In analogy with chemical solution operator ΨC deﬁned in (11) we now introduce the transport solution operator as follows: given an initial concentration cn at time tn and a source term f n+1, we let ΨT(cn, f n+1) = cn+1 where cn+1 solves the transport problem over one time-step: 1 φcn+1 −cn ∆tn + Lcn+1 = f n+1. (14) (14) Water 2021, 13, 370 7 of 26 We can then rewrite the coupled problem as We can then rewrite the coupled problem as 4. Numerical Methods for Transport Furthermore, Chemistry 4.1. Solution of the Chemical Equilibrium Problem by an Interior Point Method In this section, we show how to solve the local equilibrium problem. We recall that, given T ∈RNC and T ∈RNC, we need to compute c ∈RNa, s ∈RNs and p ∈RNp, such that Equations (2), (4) and (7) hold. Once the concentrations have been computed a simple post-processing as in Equations (8) gives the chemical operator ΨC. g g In this section, in order to simplify the notation, we will not distinguish between aqueous and sorbed species. Alternatively, this is the same as assuming that all immobile species are mineral species. Additionally, we assume that the aqueous part of stoichiometric matrix Saa has been reduced to the “Morel” form Saa = I −˜Saa . Note that under the assumption that Saa is of full rank, this is always possible (see [41], or [23,48] for more general reductions). This means that the aqueous concentrations have been split between primary species cp and secondary species cs, so that each reaction produces one secondary species, the mass action laws take the form log cs = ˜Saa log cp + log Ka, and the mass conservation Equation (7) become T = cp + ˜ST aacs + ST pap. We restate the problem to be solved with this new notation. As mentioned previously, we take as unknowns not the aqueous concentrations themselves but their logarithms. This has the obvious advantage that the concentrations will always be positive, and also reduces the ill-conditioning of the non-linear system (see [22,49,50]). We denote by ξp (resp. ξs) the vector such that ξp i = log cp i (resp. ξs i = log cs i), and we will write ξp = log cp or cp = exp(ξp). The problem to be solved for local chemical equilibrium is: N N N exp ξp + ˜ST aa exp(ξs) + ST pap −T = 0, with ξs = ˜Saaξp −log Ka, (19) (log Kp −Spaξ = 0 and p ≥0) or (log Kp −Spaξ ≥0 and p = 0). (20) (19) (20) (20) Equation (19) combines the mass action law and the conservation of mass. Equation (20) takes the form of a complementarity condition [51,52] and expresses the fact that either the mineral is under-saturated and its concentration is zero, or the mineral is present and the solubility product equals 1. We can then rewrite the coupled problem as If a line search is used, the last step should be replaced by C(n+1,k+1) = C(n+1,k) + λδC, where λ is determined by the line search procedure [46]. In our context, however, solving the linearized system (18) exactly, or even computing the J matrix, is not possible as both ΨC and ΨT are implicitly deﬁned functions. It was shown in our previous work (see [22,23]) that this difﬁculty could be circumvented by resorting to the Newton–Krylov method (see [24,25,46], to which our work is closely related). The Newton–Krylov method is a variant of Newton’s method where the linear system (18) is solved by an iterative method (of Krylov type), for instance GMRES [47]. The advantage of this choice is that the Jacobian matrix J is no longer required. Instead one needs a procedure to compute the product Jv for any given vector v. Water 2021, 13, 370 8 of 26 8 of 26 In the next section, we complete the description of the method by giving details on the space discretization, and we give additional implementation details on the solution of the coupled problem. 4. Numerical Methods for Transport Furthermore, Chemistry Computing the Newton step Each Newton iteration requires the solution of a linear system Computing the Newton step Each Newton iteration requires the solution of a linear system ( ) Jcs = −Fc(X), Jcs = −Fc(X), where the Jacobian of system (22) is given by: Jc =   I + ˜ST aaY ˜Saa ST pa 0 Spa 0 I 0 Z P  , where Y = diag exp(ξS), and I is an identity matrix of the appropriate size. These systems can become severely ill-conditioned (see [49]) so that Gaussian elimination may give inaccurate solutions. We monitor the condition number of the Jacobian and if it is larger than 1015, so that the solution may have no correct digit, we perform a singular value decomposition and compute a least squares solution. Damping After the Newton step has been computed, we apply two steps of damping: the ﬁrst one reduces the step to ensure that the iterates for p and z remain positive, and the second one is a line search that makes sure the size of the residual is reduced along the Newton direction. 4. Numerical Methods for Transport Furthermore, Chemistry The classical method to take minerals into account involves a systematic search across possible states (the procedure is detailed in [53], and is used by most reactive transport codes [37,54] for example). More recently, methods based on a more consistent mathematical approach have been used with success: the semi-smooth Newton’s method in [28,29]) and an interior point method that originates from inequality constrained optimization in [33,55]. See also [32] for an overview of solution methods for both equilibrium and kinetic reaction modeling. More recently, methods based on a more consistent mathematical approach have been used with success: the semi-smooth Newton’s method in [28,29]) and an interior point method that originates from inequality constrained optimization in [33,55]. See also [32] for an overview of solution methods for both equilibrium and kinetic reaction modeling. In this work we follow [33], who adapted to the case of nonlinear equations with inequality constraints an interior point method originally proposed in [31]. q g In this work we follow [33], who adapted to the case of nonlinear equations with inequality constraints an interior point method originally proposed in [31]. To formulate the algorithm, we rewrite (19)–(20) by introducing a slack variable z ∈RNp (and we denote the system in the ﬁrst line of (19) by Fe(ξp, p) = 0): Water 2021, 13, 370 9 of 26 Fe(ξp, p) = 0 Spaξp + log Kp + z = 0 p · z = 0 , with p ≥0, z ≥0, (21) (21) where the inequalities are to be understood component-wise. where the inequalities are to be understood component-wise. The idea of the interior point method in [31,33] is to relax the last equation of (21) by introducing a parameter µ that will be driven to 0 by the algorithm and to apply Newton’s method to the relaxed system Fc(X) = 0 with unknowns X = (ξp, p, z) ∈RNa+2Np: Fe(ξp, p) = 0 Spaξp + log Kp + z = 0 PZe −µe = 0 , with p ≥0, z ≥0. (22) (22) We have used a notational device that is standard in the optimization literature: we let P = diag p, Z = diag z, and we also let e = (1, . . . , 1)T ∈RNp. A detailed description of the algorithm, together with a convergence proof, can be found in the given references. We just note a few important implementation details. 4.2. The VAG Scheme The Vertex Approximate Gradient (VAG) scheme has been introduced in [36,56,57]. Its purpose is to provide a consistent dicretization method for anisotropic diffusion schemes on general (polyhedral) meshes, at a reasonable cost. We only give a brief presentation here, referring to the above references for full details. In particular, we only describe the scheme in 2D where important simpliﬁcation arise. p p Let T be a mesh of Ωmade up of (open, convex) disjoint polygonal cells K such that ∪K∈T K = Ω. For all K ∈T , let xK ∈K be a point such that K is star-shaped wih respect to xK. We denote the set of vertices of the mesh by V, for each K ∈T by VK the set of its vertices, and for each v ∈V by TK the set of elements containing v as a vertex. y g The discrete unknowns of the VAG scheme are values at the vertices of the mesh and at the center of each element. Following [56], we deﬁne a discrete space as WT = {(cs)s∈V, (cK)K∈T , cs = 0 for s ∈ΓD}, and we denote by NT its dimension (the sum of the number of vertices and the number of cells, up to the vertices located on the Dirichlet part of the boundary). Water 2021, 13, 370 10 of 26 10 of 26 Given a cell K ∈T , we consider the triangular submesh where, for each edge σ ∈EK Given a cell K ∈T , we consider the triangular submesh where, for each edge σ ∈EK a triangle is formed by joining the center of the cell to each of the two vertices at the ends of σ (see Figure 1). g g a triangle is formed by joining the center of the cell to each of the two vertices at the ends of σ (see Figure 1). xK s s′ σ Figure 1. Triangular submesh for a cell K. xK s s′ σ Figure 1. Triangular submesh for a cell K. Figure 1. Triangular submesh for a cell K. We can now deﬁne a discrete gradient for each discrete unknown cT ∈W, as the gradient of the P1 ﬁnite element function on the submesh deﬁned above. This gradient is a constant vector on each triangular sub-cell. For future reference, we denote by ηs, s ∈V (resp. 4.3. The Discrete Coupled Problem We now put the two previous sections and Section 3.2 together. The framework for doing this is provided by previous work by the authors [23]. The main issue is to realize that we are solving a problem that couples all concentrations at all degrees of freedom (DOFs), both vertices and cell centers, but that transport only couples spatial DOFs while chemistry is local to each DOF. The method proposed in [22,23] attempts to decouple transport and chemistry, without sacriﬁcing the numerical efﬁciency afforded by Newton’s method on the coupled problem. The unknowns of the discrete problems are vectors in X = R(Nc+Np)×NT , and we denote them by the same letter as their continuous counterpart, written in bold face. For a vector C ∈X , we follow [22] and denote by • C:,j the column vector of concentrations of all chemical species at the DOF j. • C:,j the column vector of concentrations of all chemical species at the j • Ci,: the row vector of concentrations of species i at all DOFs. We now extend the transport and chemical operators to operate on vectors in X . j • Ci,: the row vector of concentrations of species i at all DOFs. We now extend the transport and chemical operators to operate on vectors in X . ΨC : X 7→X T →ΨC(T), and ΨT : X 7→X C →ΨC(C), (28) (28) with ΨC(T):,j = ψC(T:,j), for j = 1, . . . , NT , ΨT(C, f)i,: = ψT(Ci,:, fi,:), for i = 1, . . . , Nc + Np. (29) (29) We also note that ΨT can be made explicit, in matrix form, by making use of the Kronecker product [59]. As ﬁrst explained in [21] the action of the transport operator Cn+1 = ΨT(Cn, fn+1) Cn+1 = ΨT(Cn, fn+1) is equivalent to the solution of the linear system is equivalent to the solution of the linear system (Φ ⊗I)Cn+1 −Cn ∆t + (AV ⊗I)Cn+1 = fn+1. Of course the Kronecker product is just a notational device, the full matrix AV ⊗I is never actually formed. Rather, the implementation of the transport operator is done with a loop over all species, in which one time step of a transport problem is solved (see [23] for details). Similarly, the chemical solution operator ΨC can be evaluated independently for each degree of freedom of the mesh. 4.2. The VAG Scheme ηK, K ∈T ) the basis functions of the P1 space associated to the vertices (resp. the cell centers). The VAG scheme is deﬁned by a variational formulation on the discrete space WT using the gradient deﬁned above. The gradient leads to the deﬁnition of ﬂuxes from a cell K ∈T to any of its vertices v ∈VK. The diffusion ﬂuxes are deﬁned as Fd K,v(cn+1 T ) = ∑ v′∈VK Tvv′ K (cn+1 K −cn+1 v′ ), (23) (23) where the transmissivities Tvv′ K can be computed locally on element K. Their values are given in [56,57]. g One also needs advection ﬂuxes, deﬁned as One also needs advection ﬂuxes, deﬁned as Fa K,v(cn+1 T ) = Fd K,v(pT )cn+1 K,v , (24) (24) where Fd K,v(pT ) is the ﬂux computed as in (23) with the pressure from the ﬂow step, and cK,v is the upwind concentration deﬁned as (cf [56]) cK,v = ( cK if Fd K,v(pT ) ≥0, cv if Fd K,v(pT ) < 0. (25) (25) The discrete problem is now deﬁned by the following system:            φK\|K\|cn+1 K cn K ∆t + ∑ v,v′∈VK Tv,v′ K (cn+1 K −cn+1 v′ ) + ∑ v∈VK FK,v(pT )cn+1 K,v = fK, ∀K ∈T φK\|K\|cn+1 v cn v ∆t − ∑ K∈Vv,v′∈VK Tv,v′ K (cn+1 K −cn+1 v′ ) −∑ K∈Tv FK,v(pT )cn+1 K,v = fv, ∀v ∈V (26 (26) where φK is the average porosity on cell K, fK = R K f (x)ηK(x) dx and fv = R Ωf (x)ηs(x) dx. In a more condensed notation, we have to solve a linear system Φcn+1 T −cn T ∆t + AVcn+1 T = f n+1 T . (27) (27) One notes that the ﬁrst equation in (26) is local to each element. The cell unkowns can thus be eliminated locally, so that the resulting system involves only the values of the concentration at the vertices. However, in a time dependent problem, the cell unknowns Water 2021, 13, 370 11 of 26 11 of 26 must be recomputed at each time step. This will be important in the coupling with the local chemical problem in Section 4.3. must be recomputed at each time step. This will be important in the coupling with the local chemical problem in Section 4.3. 4.2. The VAG Scheme With a slight abuse of notation, we denote again by ΨT the discrete version of the transport operator deﬁned in (14), so that (27) is equivalent to cn+1 T = ΨT(cn T , fT ). In the examples reported in Section 5, we have used ComPASS [35,58] as a black box for the transport solver. ComPASS is an open source platform for the simulation of multiphase compositional ﬂow on generic unstructured meshes, targeting hydrothermal systems. 4.3. The Discrete Coupled Problem We noted in Section 3.2 that the implementation of the Newton–Krylov method required the solution of a linear system with the jacobian of the function F deﬁned in (16). We emphasize that this function is deﬁned implicitly through the evaluation of both solution operators ΨC and ΨT. Alternatively, evaluating F requires the solution of 1. One step of a transport problem for each chemical species; 2. One chemical equilibrium problem for each vertex and each cell. As explained above, computing the Jacobian matrix is impractical, as it requires the inverse of the two solution operators. However, GMRES does not need the Jacobian matrix itslef, but only its action on a vector. The product Jv is interpreted as a directional Water 2021, 13, 370 12 of 26 12 of 26 deritvative. We have shown in [23] how to compute this matrix vector product from the Jacobians of the transport operator (in our case, this is a linear operator) and and the chemical equilibrium operator (where the local Jacobian is needed as part of an inner Newton loop). In this work, we have resorted to a Jacobian free version of the Newton–Krylov method (see [24] or [25–27]), where the required matrix by vector product is approximated bya ﬁnite difference quotient. For a small value of δ (see [46] for guidelines on how to choose δ), there holds: Jv = F′(C(n+1,k))v ≈F(C(n+1,k) + δv) −F(C(n+1,k)) δ . F(C(n+1,k) + δv) −F(C(n+1,k)). The consequence is that the overall method can be implemented as soon as one can evaluate the transport and chemistry operators. Not only is this a mininal requirement, it is also what is needed to implement the more common Standard Iterative Approach recalled in (17). The tradeoff between both methods is an expected reduction in the number of iterations versus an increase in function evaluations for solving the linear systems at each Newton step (but notice that the Newton–Krylov methods are a special case of Inexact Newton methods, so that the linear system can be solved with an accuracy that depends on the progress of the outer Newton loop). 5. Numerical Results 5.1. Validation Example for the Interior Point Method: The Iron Precipitation Diagram As a validation for the Newton–interior point method for computing chemical equi- librium in the presence of precipitation, we chose to recompute the pE–pH diagram for iron, also known as a Eh–pH or Pourbaix diagram ([60], p. 73). The precise set of species used and the value of the equilibrium constants can be found in the PhD. thesis by Carrayrou [61]. It comprises nine aqueous species and four solid species. p For each pair of values of the pH (between 0 and 14) and the pE (between −14 and 20), a chemical equilibrium is solved by the method detailed in Section 4.1. We ﬁrst report on the number of iterations needed for the method to converge. On the left, Figure 2 shows a map of the number of iterations for each pE–pH pair, and on the right a histogram of the total number of iterations. Both images show that the method usually converges in less than 30 iterations (the average is 31.6 and the median is 19), but that it failed to converge in a signiﬁcant number of cases, and we see from the left image that these cases correspond to the upper-right corner of the diagram, that is too large values of both pE and pH. Figure 2. Number of iterations for the iron precipitation example. (Left) map as a function of pH and pE, (Right) histogram of iteration number. Figure 2. Number of iterations for the iron precipitation example. (Left) map as a function of pH and pE, (Right) histogram of iteration number. Water 2021, 13, 370 13 of 26 13 of 26 The diagram itself is shown on Figure 3. We see again in the upper right corner, and also on the rightmost band, that the Newton–IP method had difﬁculties with this range of values. However, the general shape of the computed diagram is in qualitative agreement with the expected behavior. Figure 3. pH–pE diagram (Pourbaix diagram) for iron. In each region, the dominant species is shown. 5 2 2D T t C E l 11 f Ph QC Figure 3. pH–pE diagram (Pourbaix diagram) for iron. In each region, the dominant species is shown. 5.2. 2D Test Case: Example 11 of PhreeQC 5.2.1. Description of the Test Case 5.2.1. Description of the Test Case Ca2+ Na+ Cl− K+ X log K CaX2 1 0 0 0 2 3.4576 NaX 0 1 0 0 1 0 KX 0 0 0 1 1 0.7 Total TCa TNa TCl TK W where TCa, TNa, TCl, TK are the total aqueous concentrations and W is the total ﬁxed concentrations (in this example W is a constant equal to 1.1 × 10−3). The domain is a rectangle that initially contains a solution of potassium and sodium, with CaCl2 injected on all or part of the left boundary (the other boundary conditions are zero diffusive ﬂux). We assume that the ﬂow undergoes a uniform horizontal velocity. ) g y We consider two different conﬁgurations depending on where CaCl2 is injected on the left boundary of the domain: 1D conﬁguration: The injection is made on the entire left boundary of the domain, which leads to a solution that depends only on x. This allows for a quick check of the method; 2D conﬁguration: The injection is made only on a part of the left boundary of the domain, which gives rise to a genuinely two-dimensional solution. 5.2.1. Description of the Test Case 5.2.1. Description of the Test Case The following example simulates an advective transport in the presence of a cation exchanger. This example is adopted as a test case to validate our method in the two- dimensional case. Its 1D version represents the well known example 11 of PHREEQC documentation [37]. The example describes an experiment in a column whose chemical composition contains a cation exchanger. p g Initially, the column contains a sodium-potassium-nitrate solution in equilibrium with the exchanger. The column is ﬂushed along its left side with three pore volumes of a calcium chloride solution. Calcium, potassium, and sodium react to equilibrium with the exchanger at all times. g The ﬂow and transport parameters used in this example are presented in Table 1, with some additional parameters to represent a 2D geometry. The initial and injected concentrations are listed in Table 2. The cation exchanger capacity is 1.1 mmol/L. g p y The chemical reactions for this example can be written as The chemical reactions for this example can be written as Na+ + X−−−→ ←−−NaX K+ + X−−−→ ←−−KX 1 2Ca2+ + X−−−→ ←−− 1 2CaX2 Water 2021, 13, 370 14 of 26 Table 1. Physical parameters. Parameter Value Darcy Velocity (m/s) 2.78 × 10−6 Diffusion coeff (m2/s) 5.56 × 10−9 Domain dimensions (m) 0.1 × 0.015 Mesh size (m) 0.002 × 0.0015 Time duration (day) 1 Table 2. Initial and injected aqueous concentrations (mmol). Comp. Cinit Cinﬂow Ca2+ 0 0.6 × 10−3 Cl– 0 1.2 × 10−3 K+ 2 × 10−4 0 Na+ 1 × 10−3 0 Table 2. Initial and injected aqueous concentrations (mmol). Comp. Cinit Cinﬂow Ca2+ 0 0.6 × 10−3 Cl– 0 1.2 × 10−3 K+ 2 × 10−4 0 Na+ 1 × 10−3 0 Table 2. Initial and injected aqueous concentrations (mmol). KX, NaX, and CaX2 are the complexes formed with the rock, while X indicates the ion exchange site with a charge −1. The chemical description (mass action and conservation laws) can be represented in the Morel tableau in Table 3: Table 3. Moreal tableau for Phreeqc example. Table 3. Moreal tableau for Phreeqc example. 1D Conﬁguration In this case, CaCl is injected along the full left boundary, so that the concentration depend only on x. This allows for a comparison with the reference solution from the PhreeqC documentation [37]. q Figure 4 shows the concentrations along a horizontal line at t = 6 h. Water 2021, 13, 370 15 of 26 Figure 4. Total aqueous concentrations for y = 0.0075, at t = 6 h. Then, we plot on Figure 5 the elution curves (concentrations at the end of the column as a function of pore volume), so as to compare with the reference solution (cf. page 241 of the PhreeqC documentation [37]). We recall that chloride is a tracer that does not react chemically, and is only subject to transport. Its concentration is given by the well known analytical solution of Ogata and Banks [62]. Figure 4. Total aqueous concentrations for y = 0.0075, at t = 6 h. Then, we plot on Figure 5 the elution curves (concentrations at the end of the column as a function of pore volume), so as to compare with the reference solution (cf. page 241 of the PhreeqC documentation [37]). We recall that chloride is a tracer that does not react chemically, and is only subject to transport. Its concentration is given by the well known analytical solution of Ogata and Banks [62]. Then, we plot on Figure 5 the elution curves (concentrations at the end of the column as a function of pore volume), so as to compare with the reference solution (cf. page 241 of the PhreeqC documentation [37]). We recall that chloride is a tracer that does not react chemically, and is only subject to transport. Its concentration is given by the well known analytical solution of Ogata and Banks [62]. Figure 5. Evolution of concentrations at the end of the column. The results show a good agreement with those in [37]. Figure 5. Evolution of concentrations at the end of the column. Figure 5. Evolution of concentrations at the end of the column. The results show a good agreement with those in [37]. The results show a good agreement with those in [37]. The results show a good agreement with those in [37]. Water 2021, 13, 370 16 of 26 16 of 26 2D Conﬁguration 2D Conﬁguration In this second case, to show real 2D effects, we inject the solution of CaCl2 on only a part of the left boundary. We consider the same values of the concentrations at the limits, and the same initial conditions as those of the 1D conﬁguration. 2 part of the left boundary. We consider the same values of the concentrations at the limits, and the same initial conditions as those of the 1D conﬁguration. Figure 6 shows the aqueous concentration of Ca2+, Na+, Cl– , and K+, obtained at time t = 6 h, while Figure 7 shows the aqueous concentration as a function of x along the middle of the domain. Figure 6. Total aqueous concentrations at t = 6 h. Figure 7. Total aqueous concentrations for y = 0.0075 m, at t = 6 h. g Figure 6 shows the aqueous concentration of Ca2+, Na+, Cl– , and K+, obtained at time t = 6 h, while Figure 7 shows the aqueous concentration as a function of x along the middle of the domain. Figure 6. Total aqueous concentrations at t = 6 h. Figure 6. Total aqueous concentrations at t = 6 h. Figure 6. Total aqueous concentrations at t = 6 h. Figure 6. Total aqueous concentrations at t = 6 h. Figure 7. Total aqueous concentrations for y = 0.0075 m, at t = 6 h. Figure 7. Total aqueous concentrations for y = 0.0075 m, at t = 6 h. Water 2021, 13, 370 17 of 26 The results are consistent with physical understanding. Here also chloride remains a perfect tracer whose solution curve can be compared with an analytical solution [63]. The numerical chloride results show a good agreement with this analytical solution. The be- havior of the other species is similar to the 1D case, with the addition of diffusion along the vertical direction. 5.2.3. Performance of the Elimination Method In this section we compare the methods discussed in Section 4 for solving the coupled system: the Sequential Iterative Approach and the global approach using Newton–Krylov method (NKM in what follows), that is the nonlinear elimination strategy described in Sections 3.1 and 4.3. Our main criterion will be the number of iterations required by each method. We apply the global and the splitting approaches to the 2D test case described in the previous section using several spatial resolutions (ranging from 80 to 2000 grid cells), see Figure 8. Figure 8. Number of iterations per time-step for the first 20 time-steps for the various solution methods. Figure 8. Number of iterations per time-step for the first 20 time-steps for the various solution methods. Figure 8 shows the number of iterations required by each method per time-step for the ﬁrst 20 time steps, as the mesh resolution is reﬁned. The shape of the curves shows that the both methods have difﬁculty in converging for the ﬁrst two time steps, but from the third time step (from time = 0.3 h), and for a given mesh resolution, the convergence becomes stable for almost the same number of iterations during the simulation. Now if we compare the two methods, SIA requires at each time step more iterations than the NKM method (the nonlinear elimination strategy). This number of iterations increases with the mesh resolution for the SIA method while it remains stable and small for the NKM method. The NKM method requires only two nonlinear iterations on average at each time step for different mesh resolution. Figure 9 shows the results for a typical time step (at t = 0.5 h). As predicted, the elimination method gives a convergence independent of the mesh size also in this 2D case (we had shown in [23] that this was the case for the 1D conﬁguration). Water 2021, 13, 370 18 of 26 Figure 9. Number of Iterations as a function of the mesh size for the various solution methods at t = 0.5 h. Figure 9. Number of Iterations as a function of the mesh size for the various solution methods at t = 0.5 h. Remark 2. Here, we do not compare the CPU time, nor the number of evaluation of the objective function. 5.2.3. Performance of the Elimination Method Certainly the NKM method whose Jacobian* vector product is approximated by ﬁnite differences at each linear iteration requires a larger number of function evaluations. The number of evaluation of the function will be proportional to the number of the non-linear iterations for the NKM method times the number of GMRES iterations. We are currently investigating the use of an exact Jacobian times vector product, as this should drastically reduce the CPU time for a simulation. 5.3.1. Description of the Test Case The aim of this study is to test the capabilities of the equilibrium chemical solver to handle mineral precipitation and dissolution as well as taking into account chemical reactions in a spatially heterogeneous medium. The test case has been described in [64], and results can be found in [65]. The model is a simpliﬁed scenario for CO2 injection assuming the gas remain immobile, so that a one-phase ﬂow model can be used and gaseous carbon dioxide CO2(g) is considered as a ﬁxed species. The reservoir model is heterogeneous, with low permeability barriers designed to create a complex ﬂow pattern At the same time, two injectors at different locations induce a complex mixing: one is upstream directly in contact with the bubble of CO2 and the second one is downstream of the CO2 bubble. At time t = 0, supercritical CO2(g) is injected in a speciﬁc zone and is in equilibrium with the concentrations of other chemical species. p p p 3 • One gas component: gaseous carbon dioxide (CO2(g)). Geometry and Physical Data In this test we consider a two dimensional rectangular domain. A detailed geometry of the domain is shown in Figure 10. The aquifer is divided into ﬁve zones, the three green rectangle zones are low permeability barriers, the orange disc zone is the injection zone of gaseous CO2 and has the same properties as the remaining bulk zone (in white). The ﬂow in the porous medium is computed by solving a single phase ﬂow for the Darcy velocity where the boundary conditions are 110 bars in injector 1, 105 bars in injector 2 and 110 bars in producer as labeled in Figure 10. The other boundary conditions are no ﬂow, an impermeable Neumann boundary conditions. Water 2021, 13, 370 19 of 26 Figure 10. Geometry of domain for the SHPCO2 Benchmark. Figure 10. Geometry of domain for the SHPCO2 Benchmark. Figure 10. Geometry of domain for the SHPCO2 Benchmark. Figure 11. Pressure and velocity for the SHPCO2 model. Figure 11. Pressure and velocity for the SHPCO2 model. The values of the ﬂow parameters are given in Table 4. Figure 11 shows the pressure and the velocity ﬁelds, and one can see clearly the effect of the low permeability barriers. Table 4. Flow parameters. Constants Bulk Barriers Permeability K [m2/Pas] 10−10 10−12 Diffusion coefﬁcient D [m/s2] 10−9 10−9 Porosity φ 0.2 0.2 Ch i l S Table 4. Flow parameters. Initial State For the transport problem, initially in the domain we have two zones with different initial values of total concentrations. The ﬁrst zone is the circular area containing gaseous CO2 (in orange on Figure 10). In the rest of the domain, concentration of CO2(g) is initially equal to zero. The initial concentrations of liquid primary components are given in Table 6 Table 6. Initial total concentrations for the SHPCO2 Benchmark. Table 6. Initial total concentrations for the SHPCO2 Benchmark. Components Tinit, out of bubble [mol/L] Tinit, in bubble [mol/L] H2O 77.19 77.21 H+ −43.36 −43.38 CO2(l) 21.68 24.37 Ca2+ 21.72 21.73 Cl– 0 1 Table 5. Morel table for SHPCO2 system Table 5. Morel table for SHPCO2 system H2O H+ CO2(l) Ca2+ Cl− log K OH– 1 −1 0 0 0 −13.2354 HCO – 3 1 −1 1 0 0 −6.22 CO2(g) 0 0 1 0 0 −0.2226 CaCO3 1 −2 1 1 0 −7.7454 Total TH2O TH+ TCO2(l) TCa2+ TCl− Table 5. Morel table for SHPCO2 system Table 5. Morel table for SHPCO2 system Initial State Chemical System Chemical System y The chemical system in this test consists of eight components: The chemical system in this test consists of eight components: • Five liquid primary components: water (H2O), hydrogen ion (H+), carbon dioxide dissolved in water (CO2(l)), calcium ion (Ca2+), chlorine (Cl– ); 2 • Two secondary liquid components: hydroxide ion (OH– ), hydrogen carbonate (HCO – 3 ); • One ﬁxed precipitated component: calcite (CaCO3); 2 • Two secondary liquid components: hydroxide ion (OH– ), hydrogen carbonate (HCO – 3 ); • Two secondary liquid components: hydroxide ion (OH– ), hydrogen carbonate (HCO – 3 ); • One ﬁxed precipitated component: calcite (CaCO3); • One gas component: gaseous carbon dioxide (CO2(g)). Water 2021, 13, 370 20 of 26 In this test, we make the assumption that the gas phase is immobile and therefore gaseous carbon dioxide CO2(g) can be considered like a ﬁxed precipitated component, and because pressure variations are small throughout the domain, the partial pressure can be incorporated in the equilibrium constant for CO2 dissolution. p q Among the components of the system there are four chemical reactions: • Two aqueous equilibrium reactions: • Two aqueous equilibrium reactions: OH−−−→ ←−−H2O −H+ HCO − 3 −−→ ←−−H2O −H+ + CO2(l) • Two precipitation–dissolution reactions (taken here to be at equilibrium): • Two precipitation–dissolution reactions (taken here to be at equilibrium): CO2(g) −−→ ←−−CO2(l) CaCO3 −−→ ←−−H2O −2 H+ + CO2(l) + Ca2+ Chlorine Cl– does not participate in chemical reactions and is used as tracer. Chemical description of the test can be represented in the Morel tableau in Table 5. The values of equilibrium constants log K are given under the condition that we use moles as units for concentration of components. 5.3.2. Numerical Results 5.3.2. Numerical Results 5.3.2. Numerical Results For this preliminary study, we have ran the simulations over the relatively short period of 1900 years. For the results reported below, we could not obtain convergence of the Newton–Krylov method. Because they appear to be qualitatively reasonable, we have nevertheless decided to include them. More investigations are underway to improve the behavior of the NKM method in the presence of minerals, and to ascertain those results. Water 2021, 13, 370 21 of 26 21 of 26 One particular direction we are currently following is to improve the computation of the Jacobian matrix by vector product, by exploiting the known block structure of the matrix. Evolution of Concentrations Figure 12 shows the initial condition, while Figures 13–15 show snapshots of the con- centrations of several species (gaseous CO2, calcite, H+ and CL– ) at 302 years, 1352 years and 1852 years respectively (for a mesh with 95 × 60 grid cells). The gas bubble disappears as CO2(g) dissolves into water. Here it appears that the dissolution time is somewhat less than 2000 years. At the same time, calcite precipitates, though the effect is quite small, and the medium becomes more acidic. Indeed, one can see that the concentration of H+ behaves somewhat differently than that of the tracer. y Variable time steps are used, as described in Table 7. The time steps are chosen so as to have a small time step at the initialization when strong variations happen and to let the time step increase when the solution varies more slowly. The results can be compared with those reported by [65]. It appears that the dissolution of the gas is faster on our simulations. On the other hand, the evolution of H+ happens on a similar scale as that in [65]. This is also the case for Cl– , as it is a tracer, and thus only subject to transport. Table 7. Values for variable time step simulation. Start Time (Year) End Time (Year) Timestep (Year) 0 2 0.1 2 20 1 20 400 10 400 1210 50 1210 1900 100 Figure 12. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 0 year. Table 7. Values for variable time step simulation. Start Time (Year) End Time (Year) Timestep (Year) 0 2 0.1 2 20 1 20 400 10 400 1210 50 1210 1900 100 Table 7. Values for variable time step simulation. Table 7. Values for variable time step simulation. 2 20 1 20 400 10 400 1210 50 1210 1900 100 Figure 12. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 0 year. Figure 12. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 0 year. 22 of 26 Water 2021, 13, 370 Figure 13. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 302 year. Figure 13. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 302 year. Evolution of Concentrations Figure 14. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 1352 year. Figure 14. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 1352 year. 23 of 26 Water 2021, 13, 370 Figure 15. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 1852 year. Figure 15. Concentrations of the gas CO2(g), the calcite CaCO3, H+ and the tracer Cl– at t = 1852 year. Comparison of the Dissolution CO2(g) on Two Different Meshes We investigate the effect of the mesh reﬁnement on the results. We present on Figures 16 to 17 the CO2 concentrations obtained with two different meshes (a coarse mesh with 38 × 24 cells and a ﬁne mesh with 95 × 60 cells) at two different times. Figure 16. CO2(g), at t = 302 y (left image: coarse mesh, right image: ﬁne mesh). Figure 17. CO2(g), at t = 800 y (left image: coarse mesh, right image: ﬁne mesh). One can see the expected better resolution of the bubble (less numerical diffusion), but also that the speed of dissolution of the bubble changes as the mesh is reﬁned. This is Figure 16. CO2(g), at t = 302 y (left image: coarse mesh, right image: ﬁne mesh). Figure 16. CO2(g), at t = 302 y (left image: coarse mesh, right image: ﬁne mesh). Figure 16. CO2(g), at t = 302 y (left image: coarse mesh, right image: ﬁne mesh). Figure 17. CO2(g), at t = 800 y (left image: coarse mesh, right image: ﬁne mesh). One can see the expected better resolution of the bubble (less numerical diffusion), but also that the speed of dissolution of the bubble changes as the mesh is reﬁned. This is Figure 17. CO2(g), at t = 800 y (left image: coarse mesh, right image: ﬁne mesh). One can see the expected better resolution of the bubble (less numerical diffusion), but also that the speed of dissolution of the bubble changes as the mesh is reﬁned. This is Figure 17. CO2(g), at t = 800 y (left image: coarse mesh, right image: ﬁne mesh). One can see the expected better resolution of the bubble (less numerical diffusion), but also that the speed of dissolution of the bubble changes as the mesh is reﬁned. one possible explanation for the apparent shorter dissolution times that we observed as compared to [65]. Author Contributions: Both authors contributed equally to all parts of the work. Both authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Data Availability Statement: The code and data used in the experiments can be found on https: //gitlab.inria.fr/charms/pynkrt. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Zhang, F.; Yeh, G.; Parker, J., Eds. Groundwater Reactive Transport Models; Bentham Publishers: Sharjah, United Arab Emirates, 2012. 2. Steefel, C.I.; Appelo, C.A.J.; Arora, B.; Jacques, D.; Kalbacher, T.; Kolditz, O.; Lagneau, V.; Lichtner, P.C.; Mayer, K.U.; Meeussen, J.C.L.; et al. Reactive transport codes for subsurface environmental simulation. Comput. Geosci. 2015, 19, 445– 478. [CrossRef] 3. Steefel, C.; DePaolo, D.; Lichtner, P. Reactive transport modeling: An essential tool and a new research approach for the Earth sciences. Earth Planet. Sci. Lett. 2005, 240, 539–558. [CrossRef] 4. Yeh, G.T.; Tripathi, V.S. A critical evaluation of recent developments in hydrogeochemical transp chemical components. Water Resour. Res. 1989, 25, 93–108. [CrossRef] p 5. Carrayrou, J.; Mosé, R.; Behra, P. Operator-splitting procedures for reactive transport and comparison of mass balance errors. J. Contam. Hydrol. 2004, 68, 239–268. [CrossRef] 6. Lagneau, V.; van der Lee, J. HYTEC results of the MoMas reactive transport benchmark. Comput. Geosci. 2010, 14, 435–449. [CrossRef] [ ] 7. Mayer, K.U.; Frind, E.O.; Blowes, D.W. Multicomponent reactive transport modeling in variably saturated porous media using a generalized formulation for kinetically controlled reactions. Water Resour. Res. 2002, 38, 1174, 7. Mayer, K.U.; Frind, E.O.; Blowes, D.W. Multicomponent reactive transport modeling in variably generalized formulation for kinetically controlled reactions. Water Resour. Res. 2002, 38, 1174, 8. Samper, J.; Xu, T.; Yang, C. A sequential partly iterative approach for multicomponent reactive transport with CORE2D. Comput. Geosci. 2008, 13, 301–316. [CrossRef] 9. Yeh, G.T.; Tripathi, V.S. A Model for Simulating Transport of Reactive Multispecies Components: Model Development and Demonstration. Water Resour. Res. 1991, 27, 3075–3094. [CrossRef] 10. Su, D.; Mayer, K.U.; MacQuarrie, K.T.B. MIN3P-HPC: A High-Performance Unstructured Grid Code Reactive Transport Simulation. Math. Geosci. 2020. [CrossRef] MacQuarrie, K.T.B. MIN3P-HPC: A High-Performance Unstructured Grid Code for Subsurface Flow and Simulation. Math. Geosci. 2020. [CrossRef] 11. Xu, T.; Sonnenthal, E.; Spycher, N.; Pruess, K. TOUGHREACT—A simulation program for non-isothermal multiphase reac- tive geochemical transport in variably saturated geologic media: Applications to geothermal injectivity and CO2 geological sequestration. Comput. Geosci. 2006, 32, 145—165. [CrossRef] q p 12. Yapparova, A.; Gabellone, T.; Whitaker, F.; Kulik, D.A.; Matthäi, S.K. Reactive Transport Modelling of Dolomitisation Using the New CSMP++GEM Coupled Code: Governing Equations, Solution Method and Benchmarking Results. Transp. Porous Med. 2017, 117, 385–413. [CrossRef] 13. Fan, Y.; Durlofsky, L.J.; Tchelepi, H.A. A fully-coupled ﬂow-reactive-transport formulation based on element conservation, with application to CO2 storage simulations. Adv. Water Resour. Evolution of Concentrations This is One can see the expected better resolution of the bubble (less numerical diffusion), but also that the speed of dissolution of the bubble changes as the mesh is reﬁned. This is Water 2021, 13, 370 24 of 26 24 of 26 one possible explanation for the apparent shorter dissolution times that we observed as compared to [65]. 21. Erhel, J.; Sabit, S.; de Dieuleveult, C. Solving Partial Differential Algebraic Equations and Reactive Transport Models. In Develop- ments in Parallel, Distributed, Grid and Cloud Computing for Engineering; Topping, B.H.V., Ivànyi, P., Eds.; Computational Science, Engineering and Technology Series; Saxe Coburg Publications: London, UK, 2013; pp. 151–169. References A global method for coupling transport with chemistry in heterogeneous porous media. Comput. Geosci. 2010, 14, 465–481. [CrossRef] 23. Amir, L.; Kern, M. Preconditioning a Coupled Model for Reactive Transport in Porous Media. Int. J. Numer. Anal. Model. 2019, 16, 18–48. 24. Knoll, D.A.; Keyes, D.E. Jacobian-free Newton-Krylov methods: A survey of approaches and applications. J. Comput. Phys. 2004, 193, 357–397. [CrossRef] 25. Hammond, G.E.; Valocchi, A.; Lichtner, P. Application of Jacobian-free Newton–Krylov with physics-based preconditioning to biogeochemical transport. Adv. Water Resour. 2005, 28, 359–376. [CrossRef] 26. Carr, E.; Turner, I.; Perré, P. A variable-stepsize Jacobian-free exponential integrator for simulating transport in heterogeneous porous media: Application to wood drying. J. Comput. Phys. 2013, 233, 66–82. [CrossRef] 27. Guo, L.; Huang, H.; Gaston, D.R.; Permann, C.J.; Andrs, D.; Redden, G.D.; Lu, C.; Fox, D.T.; Fujita, Y. A parallel, fully coupled, fully implicit solution to reactive transport in porous media using the preconditioned Jacobian-Free Newton-Krylov Method. Adv. Water Resour. 2013, 53, 101–108. [CrossRef] 28. Kräutle, S. The semismooth Newton method for multicomponent reactive transport with minerals. Adv. Water Resour. 2011, 34, 137–151. [CrossRef] 29. Buchholzer, H.; Kanzow, C.; Knabner, P.; Kräutle, S. The semismooth Newton method for the solut problems including mineral precipitation-dissolution reactions. Comput. Optim. Appl. 2011, 50, 193–22 problems including mineral precipitation-dissolution reactions. Comput. Optim. Appl. 2011, 50, 193–221. [CrossRef] 30. Kräutle, S. Existence of global solutions of multicomponent reactive transport problems with mass action kinetics in porous media. J. Appl. Anal. Comput. 2011, 1, 497–515. p g p p p p pp 30. Kräutle, S. Existence of global solutions of multicomponent reactive transport problems with mass action kinetics in porous media. J. Appl. Anal. Comput. 2011, 1, 497–515. 30. Kräutle, S. Existence of global solutions of multicomponent reactive transport problems with m media. J. Appl. Anal. Comput. 2011, 1, 497–515. J pp p 31. El-Bakry, A.S.; Tapia, R.A.; Tsuchiya, T.; Zhang, Y. On the formulation and theory of the Newton interior-point method for nonlinear programming. J. Optim. Theory Appl. 1996, 89, 507–541. [CrossRef] 32. Leal, A.M.M.; Kulik, D.A.; Smith, W.R.; Saar, M.O. An overview of computational methods for chemical equilibrium and kinetic calculations for geochemical and reactive transport modeling. Pure Appl. Chem. 2017, 89, 597–643. [CrossRef] 33 Saaf F E A Study of Reactive Transport Phenomena in Porous Media Ph D Thesis Rice University Houston TX USA 1997 y p y 34. References Arbogast, T.; Bryant, S.; Dawson, C.; Saaf, F.; Wang, C.; Wheeler, M. Computational methods for multiphase ﬂow and reactive transport problems arising in subsurface contaminant remediation. J. Comput. Appl. Math. 1996, 74, 19–32. [CrossRef] 35. Lopez, S.; Masson, R.; Beaude, L.; Birgle, N.; Brenner, K.; Kern, M.; Smaï, F.; Xing, F. Geothermal Modeling in Complex Geological Systems with the ComPASS Code. In Stanford Geothermal Workshop 2018—43rd Workshop on Geothermal Reservoir Engineering; Stanford University: Stanford, CA, USA, 2018. y 36. Eymard, R.; Guichard, C.; Herbin, R. Small-stencil 3D schemes for diffusive ﬂows in porous media. ESAIM Math. Model. Numer. Anal. 2012, 46, 265–290. [CrossRef] 37. Parkhurst, D.L.; Appelo, C. User’s Guide to PHREEQC (Version 2)—A Computer Program for Speciation, Batch-Reaction, One- Dimensional Transport, and Inverse Geochemical Calculations; Technical Report 99-4259; USGS: Reston, VA, USA, 1999. p p J.; Postma, D. Geochemistry, Groundwater and Pollution, 2nd ed.; CRC Press: Boca Raton, FL, USA, 2005. p p 38. Appelo, C.A.J.; Postma, D. Geochemistry, Groundwater and Pollution, 2nd ed.; CRC Press: Boca Raton 39. Rubin, J. Transport of reacting Solutes in Porous Media: Relation Between Mathematical Nature of Problem Formulation and Chemical Nature of reactions. Water Resour. Res. 1983, 19, 1231–1252. [CrossRef] 40. Bear, J.; Cheng, A.H.D. Modeling Groundwater Flow and Contaminant Transport; Theory and Applications of Transport in Porous Media; Springer: Berlin/Heidelberg, Germany, 2010; Volume 23. 41. Saaltink, M.; Ayora, C.; Carrera, J. A Mathematical Formulation for Reactive Transport that Eliminates Mineral Concentrations. Water Resour. Res. 1998, 34, 1649–1656. [CrossRef] 42. Fahs, M.; Carrayrou, J.; Younes, A.; Ackerer, P. On the efﬁciency of the direct substitution approach for reactive transport problems in porous media. Water Air Soil Pollut. 2008, 193, 299–308. [CrossRef] 43. Saaltink, M.; Carrera, J.; Ayora, C. On the behavior of approaches to simulate reactive transpo 48, 213–235. [CrossRef] 43. Saaltink, M.; Carrera, J.; Ayora, C. On the behavior of approaches to simulate reactive transport. J. Contam. Hydrol. 2001, 48, 213–235. [CrossRef] 44. Kanney, J.F.; Miller, C.T.; Kelley, C.T. Convergence of Iterative Split Operator Approaches for Approximating Nonlinear Reactive , [ ] 44. Kanney, J.F.; Miller, C.T.; Kelley, C.T. Convergence of Iterative Split Operator Approaches for Approximating Nonlinear Reactive Transport Problems. Adv. Water Resour. 2003, 26, 247–261. [CrossRef] 44. Kanney, J.F.; Miller, C.T.; Kelley, C.T. Convergence of Iterative Split Operator Approaches for Approxim Transport Problems. Adv. Water Resour. 2003, 26, 247–261. [CrossRef] 45. Saaltink, M.; Carrera, J.; Ayora, C. References 2012, 42, 47–61. [CrossRef] 14. Ahusborde, E.; El Ossmani, M.; Moulay, M.I. A fully implicit ﬁnite volume scheme for single phase ﬂo in porous media. Math. Comput. Simul. 2019, 164, 3–23. in porous media. Math. Comput. Simul. 2019, 164, 3 23. 15. Hoffmann, J.; Kräutle, S.; Knabner, P. A parallel global-implicit 2-D solver for reactive transport problems in porous media based on a reduction scheme and its application to the MoMaS benchmark problem Comput. Geosci. 2009, 14, 421–433 [CrossRef] 15. Hoffmann, J.; Kräutle, S.; Knabner, P. A parallel global-implicit 2-D solver for reactive transport problem on a reduction scheme and its application to the MoMaS benchmark problem. Comput. Geosci. 2009, 14 15. Hoffmann, J.; Kräutle, S.; Knabner, P. A parallel global-implicit 2-D solver for reactive transport problems in porous media based on a reduction scheme and its application to the MoMaS benchmark problem. Comput. Geosci. 2009, 14, 421–433. [CrossRef] 16. Kräutle, S.; Knabner, P. A new numerical reduction scheme for fully coupled multicomponent transport-reaction problems in porous media. Water Resour. Res. 2005, 41. [CrossRef] on a reduction scheme and its application to the MoMaS benchmark problem. Comput. Geosci. 2009, 14, 421–433. [CrossRef] 16. Kräutle, S.; Knabner, P. A new numerical reduction scheme for fully coupled multicomponent transport-reaction problems in porous media. Water Resour. Res. 2005, 41. [CrossRef] p 17. Kräutle, S.; Knabner, P. A reduction scheme for coupled multicomponent transport-reaction problems in porous media: Generalization to problems with heterogeneous equilibrium reactions. Water Resour. Res. 2007, 43. [CrossRef] 18. de Dieuleveult, C.; Erhel, J.; Kern, M. A global strategy for solving reactive transport equations. J. Comput. Phys. 2009, 228, 6395–6410. [CrossRef] 19. De Dieuleveult, C.; Erhel, J. A global approach to reactive transport: Application to the MoMas benchmark. Comput. Geosci. 2010, 14, 451–464. [CrossRef] . Analysis of a Global Reactive Transport Model and Results for the MoMaS Benchmark. Math. Comput. Si 98. [CrossRef] 20. Erhel, J.; Sabit, S. Analysis of a Global Reactive Transport Model and Results for the MoMaS Bench 2017, 137, 286–298. [CrossRef] 21. Erhel, J.; Sabit, S.; de Dieuleveult, C. Solving Partial Differential Algebraic Equations and Reactive Transport Models. In Develop- ments in Parallel, Distributed, Grid and Cloud Computing for Engineering; Topping, B.H.V., Ivànyi, P., Eds.; Computational Science, Engineering and Technology Series; Saxe Coburg Publications: London, UK, 2013; pp. 151–169. 25 of 26 Water 2021, 13, 370 22. Amir, L.; Kern, M. References A Comparison of Two Approaches for Reactive Transport Modelling. J. Geochem. Explor. 2000, 69–70, 97–101. [CrossRef] 46. Kelley, C.T. Iterative Methods for Linear and Nonlinear Equations; With Separately Available Software, Frontiers in Applied Mathematics; SIAM: Philadelphia, PA, USA, 1995; Volume 16, pp. 154–165. 47. Saad, Y.; Schultz, M.H. GMRES: A generalized minimal residual algorithm for solving nonsymmetric linear systems. SIAM J. Sci. Stat. Comput. 1986, 7, 856–869. [CrossRef] 48. Hoffmann, J.; Kräutle, S.; Knabner, P. A general reduction scheme for reactive transport in porous media. Comput. Geosci. 2012, 16, 1081–1099. [CrossRef] yrou, J.; Lucas, Y.; Ackerer, P. Thermodynamic Equilibrium Solutions Through a Modiﬁed Newton Raphson 016, 63, 1246–1262. [CrossRef] 49. Marinoni, M.; Carrayrou, J.; Lucas, Y.; Ackerer, P. Thermodynamic Equilibrium Solutions Through Method. AIChE J. 2016, 63, 1246–1262. [CrossRef] J 50. Wolery, T.; Jarek, R.L. Software User’s Manual; EQ3/6, Version 8.0; Sandia National Laboratories: Albuquerque, Bernalillo, 2003. 50. Wolery, T.; Jarek, R.L. Software User’s Manual; EQ3/6, Version 8.0; Sandia National Laboratories: Albuquerque, Bernalillo, 2003. 51. Facchinei, F.; Pang, J.S. Finite-Dimensional Variational Inequalities and Complementarity Problems; Springer Series in Operations Research; Springer: New York, NY, USA, 2003; Volume I, pp. 34, 624, I69. 51. Facchinei, F.; Pang, J.S. Finite-Dimensional Variational Inequalities and Complementarity Problems; Springer Series in Operations Research; Springer: New York, NY, USA, 2003; Volume I, pp. 34, 624, I69. Water 2021, 13, 370 26 of 26 26 of 26 52. Facchinei, F.; Pang, J.S. Finite-Dimensional Variational Inequalities and Complementarity Problems; Springer Series in Operations Research; Springer: New York, NY, USA, 2003; Volume II, pp. 1–34, 625–1234, II1–II57. p g pp 53. Bethke, C.M. Geochemical and Biogeochemical Reaction Modeling, 2nd ed.; Cambridge University Press: Cambridge, UK, 2008. 54 C J L ki f f l ti f th ti t t b h k f M M S ith SPECY C t G i p g pp 53. Bethke, C.M. Geochemical and Biogeochemical Reaction Modeling, 2nd ed.; Cambridge University Press: Cambridge, UK, 2008. 54. Carrayrou, J. Looking for some reference solutions for the reactive transport benchmark of MoMaS with SPECY. Comput. Geosci. f [ ] 55. Leal, A.M.; Blunt, M.J.; LaForce, T.C. Efﬁcient chemical equilibrium calculations for geochemical speciation and reactive transport modelling. Geochim. Cosmochim. Acta 2014, 131, 301–322. [CrossRef] 56. Eymard, R.; Guichard, C.; Herbin, R.; Masson, R. Vertex centred discretization of two-phase Darcy ﬂows on general meshes. References In Congrès National de Mathématiques Appliquées et Industrielles; ESAIM Proc.; EDP Sciences: Les Ulis, France, 2011; Volume 35, pp. 59–78. [CrossRef] 57. Eymard, R.; Guichard, C.; Herbin, R.; Masson, R. Vertex-centred discretization of multiphase co general meshes. Comput. Geosci. 2012, 16, 987–1005. [CrossRef] 58. Dalissier, E.; Guichard, C.; Havé, P.; Masson, R.; Yang, C. ComPASS: A tool for distributed parallel ﬁnite volume discretizations on general unstructured polyhedral meshes. In CEMRACS 2012; ESAIM Proc.; EDP Sciences: Les Ulis, France, 2013; Volume 43, pp. 147–163. [CrossRef] [ ] van Loan, C.F. Matrix Computations, 4th ed.; Johns Hopkins University Press: Baltimore, MD, USA, 2013. pp 59. Golub, G.H.; van Loan, C.F. Matrix Computations, 4th ed.; Johns Hopkins University Press: Baltimore, M 60. Brookins, D.G. Eh-pH Diagrams for Geochemistry; Springer: Berlin/Heidelberg, Germany, 1988. 61. Carrayrou, J. Modélisation du Transport de Solutes Réactifs en Milieu Poreux Saturé. Thèse de Doctorat, Université Louis Pasteur, Strasbourg, France, 2001. g 62. Ogata, A.; Banks, R.B. A Solution of the Differential Equation of Longitudinal Dispersion in Porous Media; Technical Report Prof. Pap., 411-A; USGS: Reston, VA, USA, 1961. 63. Feike, L.; Dane, J.H. Analytical solution of the one dimensional advection equation equation and two- or tree-dimentional dispersion equation. Water Ressources Res. 1990, 26, 1475–1482. 64. Haeberlein, F. Time Space Domain Decomposition Methods for Reactive Transport—Application to CO2 Geological Storage. Thèse de Doctorat, Université Paris 13, Paris, France, 2011. e de Doctorat, Université Paris 13, Paris, France, 2011. 65. Ahusborde, E.; El Ossmani, E. A sequential approach for numerical simulation of two-phase multicomponent ﬂow with reactive transport in porous media. Math. Comput. Simul. 2017, 137, 71–89. [CrossRef]
https://openalex.org/W2739753320	https://europepmc.org/articles/pmc5524810?pdf=render	English	null	Recent Progress in Vascular Aging: Mechanisms and Its Role in Age-related Diseases	Aging and disease	2,017	cc-by	15,611	Xianglai Xu1,2, , Brian Wang2, Changhong Ren2,4, Jiangnan Hu2, David A. Greenberg5, Tianxiang Chen6, Liping Xie3, , Kunlin Jin2, * 1Zhongshan Hospital, Fudan University, Shanghai 200032, China. 1Zhongshan Hospital, Fudan University, Shanghai 200032, China. 2Department of Pharmacology and Neuroscience, University of North Texas Health Science Center at Fort Worth, TX 76107, USA. g p , y, g , 2Department of Pharmacology and Neuroscience, University of North Texas Health Scien TX 76107, USA. 3Department of Urology, the First Affiliated Hospital, Zhejiang University, Zhejiang Province, China. 4Institute of Hypoxia Medicine, Xuanwu Hospital, Capital Medical University. Beijing, China. 5Buck Institute for Research on Aging, Novato, CA 94945, USA. 6Department of Thoracic Surgery, Shanghai Chest Hospital, Shanghai Jiaotong University, Shanghai, China. yp , p , p y j g, 5Buck Institute for Research on Aging, Novato, CA 94945, USA. 6Department of Thoracic Surgery, Shanghai Chest Hospital, Shanghai Jiaotong University, Shanghai, China. g g, , , Department of Thoracic Surgery, Shanghai Chest Hospital, Shanghai Jiaotong University, g g cic Surgery, Shanghai Chest Hospital, Shanghai Jiaotong University, Shanghai, China. [Received December 7, 2016; Revised April 15, 2017; Accepted May 7, 2017] [Received December 7, 2016; Revised April 15, 2017; Accepted May 7, 2017] ABSTRACT: As with many age-related diseases including vascular dysfunction, age is considered an independent and crucial risk factor. Complicated alterations of structure and function in the vasculature are linked with aging hence, understanding the underlying mechanisms of age-induced vascular pathophysiological changes holds possibilities for developing clinical diagnostic methods and new therapeutic strategies. Here, we discuss the underlying molecular mediators that could be involved in vascular aging, e.g., the renin-angiotensin system and pro-inflammatory factors, metalloproteinases, calpain-1, monocyte chemoattractant protein-1 (MCP-1) and TGFβ-1 as well as the potential roles of testosterone and estrogen. We then relate all of these to clinical manifestations such as vascular dementia and stroke in addition to reviewing the existing clinical measurements and potential interventions for age-related vascular dysfunction. Key words: vascular aging, stroke, dementia, arterial stiffness, endothelial dysfunction Correspondence should be addressed to: Dr. Xianglai Xu, Zhongshan Hospital, Fudan University, Shanghai, China. Email: xuxianglai@163.com; Dr. Liping Xie, the First Affiliated Hospital, Zhejiang University, China. Email: xielp@zjuem.zju.edu.cn; Dr. Kunlin Jin, University of North Texas Health Science Center, Fort Worth, TX 76107, USA. Email: Kunlin.Jin@unthsc.edu Copyright: © 2017. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Correspondence should be addressed to: Dr. Xianglai Xu, Zhongshan Hospital, Fudan University, Shanghai, China. Email: xuxianglai@163.com; Dr. Liping Xie, the First Affiliated Hospital, Zhejiang University, China. Email: xielp@zjuem.zju.edu.cn; Dr. Kunlin Jin, University of North Texas Health Science Center, Fort Worth, TX 76107, USA. Email: Kunlin.Jin@unthsc.edu Copyright: © 2017. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. http://dx.doi.org/10.14336/AD.2017.0507 Volume 8, Number 4; 486-505, August 2017 Review Introduction Cerebrovascular diseases, if considered separately, accounts for 12% of all deaths in Europe annually. Moreover, cerebrovascular diseases such as vascular cognitive impairment dramatically decrease the quality of life, even though they do not cause death directly. Therefore, aging research is of great importance, especially in the context of cerebrovascular diseases, from bench to bedside. Understanding the mechanisms underlying age-induced vascular pathophysiological alterations also holds possibilities for developing clinical diagnostic methods and new therapeutic strategies. Among aging-related diseases, age-specific mortality rates from cardiovascular diseases (CVDs) rise exponentially from more than 40% of all deaths at age 65– 74 to almost 60% in the 85 and over age group to become the leading cause of death [1]. The Global Burden of Disease Study reported that 15.6 million people died from CVDs worldwide in 2010, accounting for 29.6% of all deaths, or twice as many as the deaths from cancers [2]. This proportion is even higher in Europe, where CVDs are responsible for nearly half of all deaths [3]. The current review will discuss the molecular 486 ISSN: 2152-5250 Vascular aging: mechanisms and disease Xu X., et al expression and activity of the angiotensin-converting enzyme-1 (ACE-1) significantly increase in both ECs and VSMCs during aging [6-8]. Chymase, an alternative angiotensin convertase, also shows increased expression within the aging arterial wall [7, 8]. Consequently, the cleaved product, Ang II, is markedly upregulated in rats, monkeys and humans [7-9]. AT1 levels also increase within the aged arterial wall. In contrast, the expression of ACE2 decreases with age, reducing its inhibitory effects on the RAS, which might lead to higher expression levels of Ang II and more Ang II-related vascular alteration [10]. Thus, the net activity of the RAS is increased in the elderly. Under normal physiological conditions, the RAS is well regulated. However, increased activity is associated with several vascular dysfunctions (Fig. 1) [11, 12]. alterations of aging vessels and their associated clinical conditions, especially in cerebral vessel-related diseases. We will also address possible therapeutic approaches, which may potentially improve cardiovascular and cerebrovascular health in the aged. MFG-E8 MFG-E8 is a major component of the human milk fat globule and a multifunctional glycoprotein found in mammary epithelial cells, ECs, VSMCs, dendritic cells, blood cells and fetal hematopoietic stem cells [32-34]. Abundant MFG-E8 expression can be found in advanced human atherosclerotic plaques [35] and certain tumors [36], where it functions in accumulating apoptotic cells, accelerating atherosclerosis and inhibiting neutrophil migration through αVβ3-integrin-dependent MAP kinase activation [37]. MFG-E8 is N-linked glycosylated and is overexpressed in the intima and aortic wall media of aged rodents, nonhuman primates and humans. Angiotensin II induces dysfunction and ROS in human brain microvascular endothelial cells, promotes cerebrovascular remodeling, vascular inflammation and results in the impairment of regulation of cerebral blood flow [14]. Comparing with placebo, clinical studies have shown that some RAS-inhibiting agents improve arterial stiffness (as evidenced by a reduction in PWV), endothelial dysfunction or number of EPCs independently of blood pressure reduction. Although data relating the RAS activity to these vascular functions are inconclusive, there are more consistent studies that show that drugs interfering with Ang II reduce arterial stiffness and even decrease arterial-stiffness-associated diseases such as cardiovascular, metabolic and renal disorders [15-18]. Double-label immunofluorescence staining show that MFG-E8 co-localizes with a marker of VSMCs, α- smooth muscle actin. It co-localizes with Ang II and MCP-1 as well and may take part in developmental homeostatic processes, showing protective effects in various models of organ injury [38, 39]. MFG-E8 is known to support both cell-cell and cell-lipid interactions. It also mediates tissue remodeling of different organs including the vasculature, under pathological conditions [39, 40]. Treating early passage VSMCs from the young aorta with Ang II can enhance the expression of both MFG-E8 and MCP-1, which are required for increased VSMC invasion. MFG-E8 siRNA can inhibit Ang II induced MCP-1 protein expression, whereas exposure to MFG-E8 increases MCP-1 expression level and invasion capacity of VSMCs; these effects, in turn, are inhibited by a MCP-1 receptor blocker, the poxvirus-derived viral CC- chemokine inhibitor protein, vCCI. Conversely, knockdown of MFG-E8 by siRNA also substantially reduces VSMCs invasion capacity. These data suggest that Ang II is upstream of MFG-E8 while MCP-1 is a downstream molecule within the Ang II signaling cascade that regulates VSMC invasion. MFG-E8 is a pivotal relay element within the Ang II/MCP-1/VSMC invasion signaling cascade [40]. , [ ] Moreover, Ang II directly regulates permeability of BBB ECs via oxidative stress [19, 20]. Renin/angiotensin system and pro-inflammatory factors The renin/angiotensin system (RAS) is of great importance in vascular biology, mediating normal vessel function and participating in pathogenesis. Ang II acts through the AT1 and AT2 receptors and Ang II signaling plays a crucial role in the vascular remodeling process [4, 5] The majority of Ang II is the product of Ang I cleavage by angiotensin converting enzyme (ACE). The Figure 1. Angiotensin II signaling pathway underlying vascular aging. Figure 1. Angiotensin II signaling pathway underlying vascular aging. Aging and Disease • Volume 8, Number 4, August 2017 487 Xu X., et al Vascular aging: mechanisms and disease incidence rate of cognitive impairment with age, with ARB being relatively more effective [31]. Ang II is an inflammatory trigger that increases the activation of inflammatory factors in the arterial wall with advancing age. It activates nuclear transcription factor- kappa B (NF-κB), a key inflammatory marker, through the AT1 and AT2 receptors and then initiates an inflammatory signaling loop [13]. Induction of NF-κB upregulates the activities of MMP-2/-9, calpain-1, MCP- 1, TGF-β1 and ROS, which may contribute to structural and functional vascular remodeling. Aging and Disease • Volume 8, Number 4, August 2017 MFG-E8 Collectively, an increase in MFG-E8 signaling may be a mechanism underlying the age-associated increase in aortic VSMC proliferation and invasion. regulated by PDGF and its receptor-signaling cascade as well [42]. Expression of the PDGF receptor-α and PDGF receptor-β is increased in aged VSMCs. MFG-E8 treatment dose- and time-dependently increases expressions of PDGF-A, PDGFR-α and PDGFR-β, in young and old VSMCs while MFG-E8 siRNA decreases expressions of PDGFR-α and PDGFR-β [41]. Collectively, an increase in MFG-E8 signaling may be a mechanism underlying the age-associated increase in aortic VSMC proliferation and invasion. which leads to vessel and perivascular pathophysiological alterations. Additionally, chronic cerebral hypoperfusion associated with vascular aging upregulates MMP-2 expression in microglia and the vascular endothelium of white matter [51]. Aging and Disease • Volume 8, Number 4, August 2017 MMPs The extracellular matrix (ECM) maintains vascular structural stability and is essential for the mechanical and biological properties of vessel walls. The ECM is degraded with advancing age, which facilitates VSMC migration. MEROPS (https://merops.sanger.ac.uk/), an online peptidases database, classifies metalloprotease families into 15 different clans such as MA, MC and 7 unassigned families. MMPs comprise the M10 family of the zinc- containing MA clan of metallopeptidases. They are ubiquitous enzymes, with an active site where a Zn atom, coordinated by three histidines, can play a catalytic role to degrade the vascular ECM [44]. MMPs involved in vascular ECM remodeling consist of collagenases (MMP- 1 and MMP-8, expressed by ECs and SMCs), gelatinases (MMP-2, expressed by SMCs and macrophages; MMP-9, expressed in macrophages and vascular cells), elastases (MMP-7, expressed at a low level in the vascular wall, MMP-12, synthesized by macrophages), and stromelysins and membrane-type metalloproteinases (MT-MMPs) [45- 47]. In addition, MMP-8 can promote the migration and proliferation of VSMCs [47]. Calpain-1 Calpain-1 is a ubiquitous cytosolic Ca2+-activated neutral protease. As discussed in above, increased MMP-2 can be found in aging vessels and MMP-2 activation is implicated in age-associated VSMC migration, elastin degradation and collagen deposition. Calpain-1 regulates MMP-2 activity in VSMCs, facilitating vascular calcification and fibrosis [48] and promoting the invasion of fibroblasts and leukemic cells [52]. Endothelial integrity is of great importance to a normal vessel. However, aging enhances endothelial cell (EC) susceptibility to apoptosis. The overexpression of MFG-E8 in ECs triggers apoptosis through increasing the Bax/Bcl-2 ratio and caspase-9 and caspase-3 activation. Knocking down MFG-E8 by siRNA decreases both caspase-3 activity and the phosphorylation of glycogen synthase kinase 3β, thereby suppressing advanced glycation end product (AGE)-induced apoptosis [43]. Ang II signaling mediates the age-related increase in MMP-2 activity in the vascular wall. This can be blocked using an exogenous or endogenous calpain inhibitor (calpastatin), indicating that calpain activity is required for Ang II-associated activation of MMP-2 [48]. Calpain- 1 is also linked to cytoskeleton remodeling within VSMCs and VSMC migration as migratory capacity can be stimulated by overexpression of calpain-1 in young VSMCs. Furthermore, calpain-1 has higher mRNA transcript and protein abundance and activity in the aged rat aorta and is regulated by Ang II signaling. Collectively, these findings indicate that increased calpain-1 activity is a central mechanism for the exaggerated Ang II signaling involved in age-related arterial remodeling. MFG-E8 Studies have shown that RAS blockade improves cerebrovascular structure, normalizes CBF autoregulation and reduces cerebral ischemia [21-23] while improving vascular compliance and endothelial function in healthy normotensive elderly individuals [24]. Some evidence indicate that RAS blockade therapy may be effective in the prevention of cognitive decline and dementia [25, 26], which may be independent of its blood pressure lowering function [27]. Drugs that inhibit the RAS include angiotensin converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB). ACEI blocks ACE and decreases Ang II production while ARB blocks the angiotensin type 1 receptor rather than type 2. The RAS is involved in the production and clearance of amyloid-β (Aβ) and vascular and inflammatory factors, which may contribute to AD [28, 29]. Thus, ACEI and ARB may be protective against cognitive impairment in hypertensive patients. However, according to several clinical studies, the effects of RAS blockade on cognitive impairment and dementia have been controversial [30]. A meta-analysis that combined randomized clinical trials and observational studies, suggested that the use of RAS blockade was associated with a 35% reduction in cognitive impairment incidence and 20% reduction in AD incidence. Among RAS blockade, ACEI and ARB have 13% and 31% association with the decrease in the incidence of AD, respectively, while no significant difference was seen between ACEI and ARB in AD risk. RAS blockade provided a significant reduction in the MFG-E8 affects VSMC proliferation as well. Exogenously overexpressing MFG-E8 in young VSMCs increases their proliferation rate. Treatment with MFG-E8 triggers a dose-dependent activation of ERK1/2, which are important modulators of mitosis, in both young and aged VSMCs. Exposure to MFG-E8 also enhances levels of PCNA and CDK4, which can be blocked by the ERK1/2 phosphorylation inhibitor, U126, and promotes VSMC proliferation via αVβ5 integrins. In contrast, knocking down MFG-E8 by stealth siRNA dramatically reduces PCNA and CDK4 levels and slows down proliferation [41]. Proliferation of VSMC can be Aging and Disease • Volume 8, Number 4, August 2017 488 Vascular aging: mechanisms and disease Xu X., et al regulated by PDGF and its receptor-signaling cascade as well [42]. Expression of the PDGF receptor-α and PDGF receptor-β is increased in aged VSMCs. MFG-E8 treatment dose- and time-dependently increases expressions of PDGF-A, PDGFR-α and PDGFR-β, in young and old VSMCs while MFG-E8 siRNA decreases expressions of PDGFR-α and PDGFR-β [41]. TGF-β1 TGF-β1 is a widely expressed cytokine, playing a critical role in the processes of proliferation, wound healing, synthesis of ECM molecules and inflammation. It also has multiple functions in both vascular development and remodeling [60]. Ang II can enhance oxidative stress as well as impair endothelial function [70]. Increased Ang II signaling activity can promote the generation of cellular ROS, together with the activation of redox-sensitive signaling cascades [71]. Therefore, ROS may be essential signaling molecules for maintaining vascular homeostasis. TGF-β1 is synthesized as a latent precursor molecule (LTGF-β) that contains an amino-terminal hydrophobic signal peptide region, named the latency-associated peptide (LAP). LAP is cleaved from the complex in the Golgi apparatus, but remains non-covalently bonded with TGF-β1. LTGF-β is secreted as a small latent complex (SLC) or as a large latent complex that is covalently bonded with the LTGF-β binding protein (LTBP) [60]. Inflammatory stimuli, MMP-2 and MMP-9 promote the release of TGF-β1 from this complex in its mature form, which binds to its receptor and then activates downstream signaling pathways. Superoxide molecules are produced by transferring an electron to oxygen; superoxide concentrations can be reduced to the picomolar levels by superoxide dismutase (SOD). However, superoxide molecules react with nitric oxide (NO) at least 10 times faster than SOD can scavenge NO [72]. This reaction may have some biological significance when the concentration of superoxide molecules rises in blood vessels with advancing age [73, 74]. The higher level of superoxide inhibits the generation of NO by vascular cells, resulting in the impairment of endothelium-dependent relaxation. On the other hand, both eNOS activation and NO bioavailability are decreased with age [75, 76]; lower NO levels further increase ROS production. Moreover, peroxynitrite (ONOO–), a highly reactive molecule produced by the reaction of NO with superoxide, has been implicated in impaired EC function and vascular aging [74]. Several studies reported the importance of age-related mitochondrial oxidative stress as a characteristic of endothelial dysfunction. It is associated with the over- activation of NADPH oxidase, an enzyme localized in the cytoplasm and membranes. The NADPH oxidases are considered to be enzymatic sources of ROS production in vascular cells, while over-activated NADPH oxidase mediated ROS production is regulated by various factors including hypoperfusion and cytokines and hormones such as Ang II, platelet-derived growth factor, TGF-β as well. TGF-β1 ROS promote age-related vascular dysfunction not only by compromising EC function, but also inducing the expression of factors involved in vascular structural remodeling such as MMPs [77, 78]. The active receptor complex phosphorylates the receptor-activated Smads (R-Smads), Smad2 and Smad3, which propagate the signal. R-Smads form complexes with Smad4 and translocate to the nucleus, where they recruit co-activators or co-repressors into transcriptional complexes and regulate the transcriptional activity of various genes [61]. The signal transduction pathway has its own endogenous regulators as well. An inhibitory Smad, Smad7, blocks TGF-β1 signaling by physically interacting with the activated TGFBR1 receptor and prevents the phosphorylation of Smad2/3 [62]. Importantly, upregulated Ang II and MMP2/MMP9 increase the activation of TGF-β1 [63, 64]. Signaling molecules downstream of activated TGF-β1 (p-Smad-2/- 3/-4) are increased while inhibitory Smad (Smad-7) is decreased in the arterial wall with aging [65]. TGF-β1 signaling affects ECs function and causes arterial stiffening by increasing collagen types I and III in vessel walls [66, 67]. Aging and Disease • Volume 8, Number 4, August 2017 MCP-1 MCP-1 is a 76-amino acid member of the C-C subfamily of chemokines [53]. Through the activation of CCR2, a 7- transmembrane G protein–coupled receptor, MCP-1 can promote the migration of monocytes, lymphocytes, ECs, VSMCs and human fibroblasts as well as induce MMP-1 synthesis [54-56]. MCP-1 also mediates TGF-β-induced angiogenesis by stimulating VSMC migration. The upregulated expression of MCP-1 and CCR2 could be found in the intima of the aorta of aged rats [57]. MCP-1 expression by vascular cells is regulated by Ang II via the NF-κB-dependent pathway and Ang II levels are upregulated with aging [58, 59]. Interestingly, an age- dependent increase of circulating MCP-1 levels was found in healthy aged people without known diseases. MCP-1 plasma level is increased in older individuals. This means that MCP-1 has the potential to serve as an aging biomarker, although its specific biological significance and functions still need to be further investigated. Among the family of MMPs, increased MMP-2 and MMP-9 expressions and activities were found to be associated with enhanced Ang II signaling of aged arteries in various species including humans [8, 9, 48, 49]. MMP- 2 activation is increased during aging, localized to ECs and VSMCs and was detected in the intima, internal elastic lamina and elastin fibers of the inner tunica media [50]. MMP-9 activity is increased in situ during aging in grossly normal aortic segments. Increased MMP-9 levels reduce vascular permeability and induce inflammation, Aging and Disease • Volume 8, Number 4, August 2017 489 Xu X., et al Vascular aging: mechanisms and disease Imbalance in the regulation of oxidative stress contributes to vascular alterations characterized by mitochondrial dysfunction and increased ROS production, and, eventually, leads to the development of cardiovascular pathological alterations, such as hypertension and stroke. TGF-β1 ROS and NO bioavailability Oxidative stress is widely acknowledged as an important factor associated with aging and disease, which increases in the arterial system of humans and experimental models. It has been linked to the development of age-related pathogenesis, leading to arterial dysfunction. Reactive oxygen species also play a physiological role in the vascular wall, including endothelium-dependent functions, smooth muscle and endothelial cell growth and survival, and regulation of remodeling of the vascular wall [68]. Importantly, oxidative stress can be altered by the imbalance between antioxidant defenses and reactive oxygen species (ROS) that are produced in vessel walls and regulate cell functions and cellular senescence [69]. g Mitochondrial DNA (mtDNA) is a double-stranded DNA circular genome that encodes 37 genes including tRNA, rRNA and respiratory chain enzymes. MtDNA has a higher mutation rate than nuclear DNA due to the lack of protective histones and the close proximity of the mitochondrial genome to the inner membrane where ROS are continually generated. The amount of mtDNA mutations have been observed to increase as humans age [79] e.g., increased deletions in mtDNA have been reported in the aged human central nervous system and cardiac muscle [80-82]. Mutations of mtDNA may accumulate as a result of unrepaired DNA damage caused 490 Vascular aging: mechanisms and disease Xu X., et al by ROS or by replication errors that undergo clonal expansion throughtout the adult life [83]. The replication of mtDNA occurs randomly and independently of the cell cycle [84]. Furthermore, a single cell may contain more than one type of mutated mtDNA, namely heteroplasmy. Once it reaches a certain threshold, respiratory chain dysfunction may occur. amyotrophic lateral sclerosis [88-91]. Furthermore, mtDNA from atherosclerotic plaques or circulating cells from CVDs patients show increased frequency of mtDNA adducts or the common mitochondrial deletion mtDNA4977, compared with normal vessels or patients without CVDs, respectively [92, 93]. However, the direct correlation between mtDNA damage and neurological diseases via vascular aging remains unclear. Mitochondria oxidative stress increases with age and is a strong trigger of age-related ECs and VSMCs function. Antioxidant enzymes knockout animal models with ablated mitochondrial aldehyde dehydrogenase or mitochondrial superoxide dismutase showed increased mitochondrial ROS formation and oxidative mtDNA lesions. Meanwhile, endothelium-dependent acetylcholine-induced relaxation of aortas was significantly deteriorated. Due to the correlation analysis, mitochondrial ROS formation is directly related to endothelial function while the correlation between mtDNA damage and endothelial dysfunction is relatively less striking but still efficient [85]. ROS and NO bioavailability Studies showed that H2O2 is able to induce mtDNA damage in ECs and VSMCs. H2O2 treatment leads to mtDNA damage within 10 and 15 minutes in ECs and VSMCs, respectively. Apart from H2O2, either low or high ONOO− concentration causes mtDNA damage in ECs while only high ONOO− concentration in VSMCs causes mtDNA damage, indicating that ECs are more sensitive to ROS than VSMCs [86]. A mouse model with a homozygous mutation in the exonuclease encoding domain of the mtDNA polymerase gamma (Polgm/m) that is prone to age- dependent accumulation of mtDNA mutations was reported to have increased mitochondrial ROS levels, which resulted in an accelerated form of age-dependent cardiomyopathy. Oxidative stress and respiratory chain dysfunction due to mtDNA point mutations accumulation in protein-coding mitochondrial genes are considered to contribute to the premature aging phenotype of the Polgm/m mice. The susceptibility of mtDNA to oxidative damage may be considered a combination of factors besides the higher superoxide formation rate in the mitochondrial matrix, which is different from nuclear DNA, in that the mtDNA has no protective histones and possesses a low level of DNA repair activity. Therefore, the mitochondrial 8-oxo-deoxyguanosine DNA lesion may be the underlying burden of oxidative stress during the aging process in the heart and brain [87]. Aging and Disease • Volume 8, Number 4, August 2017 Testosterone The majority of testosterone is produced by the testis while small amounts are produced by the adrenal glands. Testosterone plays an essential role in the development of the reproductive system, promoting male secondary sexual characteristics. Testosterone can freely diffuse across the plasma membrane, enter the cytoplasm and bind to the androgen receptor (AR). The hormone-bound AR acts as a transcription regulatory element that binds to specific DNA response elements in target gene promoters, thereby modulating transcription and protein synthesis [94]. Simultaneously, there are other non-classical and non- genomic mechanisms of testosterone action. Testosterone can activate intracellular signaling molecules such as the mitogen-activated protein kinase (MAPK) family, ERK1/2, protein kinase A and protein kinase C pathways [95]. ARs are widely distributed in cells and organs including ECs and VSMCs [96]. Thus, testosterone levels may influence vascular function in health and disease. Circulating testosterone in men declines progressively with aging, starting in the early years of adulthood. The decline is paralleled by a number of pathophysiological alterations leading to cardiovascular and cerebrovascular diseases, metabolic syndrome and insulin resistance [97]. Testosterone is involved in vasorelaxation via endothelium-dependent mechanisms, ion channel modulation, and vascular structural remodeling. g Testosterone can modulate endothelial function through NO release [98]. Physiological concentration of testosterone increases NO synthesis by ECs through rapid recruitment of the ERK1/2 and PI3-kinase/Akt signaling pathways [99, 100]. Moreover, testosterone stimulates NO synthesis, which increases cyclic guanosine monophosphate (cGMP) formation in VSMCs to induce vasorelaxation [101]. A low plasma testosterone level is associated with endothelial dysfunction in men independent of other risk factors, suggesting a protective effect of testosterone on the endothelium [102]. In addition, testosterone is metabolized by aromatase to form Human clinical studies have shown that the mtDNA4977 deletion levels in the putamen, cortex and cerebellum accumulates with advancing age. The levels of mtDNA damage in the cortex and putamen of the aged brain is much higher than the aged heart [81]. Numerous studies indicated that mtDNA deletion is related to neurodegenerative disorders such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease and 491 Xu X., et al Vascular aging: mechanisms and disease 17β-estradiol, which stimulates NO release. in older men with low or low-to-normal testosterone concentrations treatment did not improve cognitive function [116]. This may indicate that testosterone treatment could be used as maintenance therapy in that it lowers the risk of AD, but does not necessarily improve brain function. Estrogen Estrogen is a protective factor for the cardiovascular and cerebrovascular systems [118]. Both types of estrogen receptors (ERs), ERα and ERβ, can be found in the vascular endothelium, SMCs and adventitia. Like testosterone, estrogens bind to nuclear receptors. Activation of ERs regulates gene expression through estrogen-response elements to mediate transcription. Estrogen can also bind to plasma membrane sites thus inhibiting Ca2+ L-type voltage-gated channels in VSMCs and activating several signaling pathways including the PI3K pathway in ECs [119, 120]. Testosterone induces relaxation of isolated blood vessels in vitro. Testosterone levels in men are positively associated with flow-mediated dilatation (FMD) as an independent determinant of endothelial vasomotor function in men [102]. Short-term intracoronary testosterone administration has been reported to induce vasodilation that can increase blood flow in men [111]. A supraphysiological concentration of testosterone was able to trigger direct vasodilator actions via membrane ion channels in isolated blood vessels rather than through the endothelium and AR-dependent pathway [112]. On the other hand, oral supplementation of testosterone in men enhanced FMD without affecting the basal diameter of the brachial artery, suggesting improved endothelial function in men [113]. Progressively lower concentrations of total or free testosterone would be associated with increasing incidence of cerebrovascular events such as incident stroke and transient ischemic attack [114]. However, the relationship between low testosterone level and AD is not well understood. A meta-analysis of seven prospective cohort studies with a total of 5251 elderly men and 240 cases of Alzheimer’s disease, supports the view that low testosterone level is significantly associated with increased risk of AD in elderly men, which may suggest that testosterone has important roles in regulating cognitive function in elderly men [115]. Nevertheless, although low testosterone level is a risk factor for AD, a long-term testosterone administration study of 36 months Estrogen is known to mediate vascular function by increasing NO bioavailability and suppressing ROS levels. Estrogen modulates vasoconstrictive factors or positively upregulates vasodilating factors such as prostacyclin (PGI2). Estrogen also inhibits the expression or function of cyclooxygenase-derived vasoconstrictors including prostaglandin H2 and thromboxane A2, which are mediators of vascular tone in females [121]. Furthermore, ACE activity can be decreased by estrogen therapy, which reduces Ang II production, seen in both animal models and postmenopausal females. Moreover, estrogen attenuates AT1 receptor expression and AT1-mediated responses in the aorta, heart, and kidneys [122]. Testosterone In addition, Jamadar et al. (117) reviewed 11 clinical studies examining the effects of androgen- deprivation therapy (ADT), a mainstay therapy for advanced prostate cancer (PCa), on cognition as measured by standardized tests in cognitive domains. They found that spatial memory and verbal memory may be especially sensitive to ADT, indicating that it is important to consider the benefits of ADT, especially in patients with early stage prostate cancer, which might contribute to vulnerability to negative cognitive effects. β , Testosterone also modulates ion fluxes in VSMCs by either activating K+ channel or inactivating Ca2+ channel. Testosterone within the physiological range is a selective and potent inhibitor of L-type voltage-operated Ca2+ channels, which results in vasodilation [103]. Testosterone also interacts with a large conductance voltage- and Ca2+-activated K+ (BKCa or MaxiK) channel, inducing vascular relaxation by increasing intracellular K+ efflux [104]. This channel regulates vascular tone and diameter and is a target for various vasoconstrictor and vasodilator agents under physiological conditions. Its expression is reduced in aged coronary arteries, suggesting decreased vasodilating capacity [105]. Long- term testosterone deprivation in rat VSMCs reduces expression of the Kv1.5 voltage-gated K+ channel protein whilst restoration of testosterone to physiological concentrations rescues the resulting impairment in Kv1.5 channel function [106]. Intimal-medial thickness and PWV can be reduced by long-term testosterone administration [107, 108], indicating that testosterone replacement can diminish vascular remodeling and arterial stiffness associated with male hypogonadism [109]. Furthermore, low testosterone levels give rise to higher responsiveness to vasoconstrictor agents [104, 110]. Aging and Disease • Volume 8, Number 4, August 2017 Diseases linked to vascular aging Blood pressure changes with advancing age. There is a linear rise in systolic blood pressure induced by arterial stiffness and a concurrent increase in diastolic blood pressure due to endothelial dysfunction-induced high peripheral vascular resistance until about the age of 50 [138]. Thereafter, high systolic blood pressure accelerates arterial stiffness, perpetuating a vicious cycle. In older individuals, increased arterial stiffness decreases diastolic blood pressure while systolic blood pressure keeps rising, which forms an age-related systolic-diastolic blood pressure divergence, resulting in a widen pulse pressure. Even in older normotensives, they have higher arterial stiffness, systolic blood pressure and pulse pressure vs. the young [139]. Increasing pulse pressure or hypertension increases cardiac afterload, then leads to left ventricular hypertrophy. In the brain, increased pulse pressure penetrates further into cerebral microcirculation, which may potentially expose capillaries to damaging levels of elevated pressure pulsatility. Thus, cerebral microvascular remodeling and dysfunction increase vascular resistance in response to elevated pulse pressure and limit the penetration of excessive pulsatility directly into the capillaries [140]. However, cerebral microvascular remodeling and dysfunction increases minimal resistance and reduces vasodilatory reserve. Accompanying age-related microvessel loss, these lead to an increased susceptibility to hypoperfusion and reduce the efficiency of oxygen and energy delivery, which can Several reports have shown that vascular aging is associated with both AD and VD. Hanon et al. [147] observed that PWV was elevated in AD (13.3±2.9 m/s) and VD (15.2±3.9 m/s) patients compared with controls (11.5±2.0 m/s), suggesting that arterial stiffness may be involved in the development of these disorders. Calik, et al. [148] pointed out that arterial diastolic dysfunction and increased arterial stiffness are detected in AD patients. Other studies have demonstrated that vascular stiffness is an independent risk factor for and has potential to be a strong predictor of cognitive loss [149, 150]. Interestingly, PWV is significantly higher in VD than in AD [151], which has a more complex etiology in which the contribution of vascular disease is less clear. Endothelial dysfunction may associate with cognitive impairment in the elderly population as well. It was reported that mildly cognitive impaired patients with prevalent amnestic multiple domains exhibited much worse brachial FMD [152]. Arterial stiffness is largely associated with Aβ deposition, either by baseline comparison or accumulating in the brain over the years in non-demented elderly adults; this association differs with different arterial bed measurements [153, 154]. Alzheimer’s disease and vascular cognitive impairment Dementia is a disorder of cognitive impairment that interferes with everyday life. Alzheimer's disease (AD) is the most common cause of dementia and is rare before 60 years of age [145]; vascular cognitive impairment (VCI) is a group of cognitive disorders with a presumed vascular cause and has generally superseded the term vascular dementia. In addition to AD and VCI’s adverse effects on affected individuals and their families, dementia greatly increases demands on the medical care system. VCI is a larger concept that includes VCI–no dementia, vascular dementia (VD) and cognitive impairment of mixed origin (Alzheimer's disease and vascular dementia) [146]. In this section, we focus on vascular-related cognitive impairment by reviewing recent studies. Estrogen In middle-aged women, subclinical vascular dysfunction may be aggravated by estrogen deficiency due to menopause or ovariectomy, which may be improved by estrogen replacement. However, some evidence suggest that estrogen can delay the progression of atherosclerotic lesions, measured by carotid artery intima-media thickness, in early postmenopausal women [123]. Beneficial effects of estrogen replacement therapy on mood and cognition in early but not late postmenopausal women have also been reported [124, 125]. Further research into sex hormone deficiency related to vascular dysfunction is still necessary to define Aging and Disease • Volume 8, Number 4, August 2017 492 Xu X., et al Vascular aging: mechanisms and disease optimal diagnostic and therapeutic strategies. promote cerebral microvascular disease [141, 142] and cognitive impairment [143, 144]. p g p g The menstrual cycle may be a great estrogen fluctuation cycle for studying the function of estrogen on vascular function. A comprehensive study reported that FMD, acetylcholine administration to skin microvessels via iontophoresis and arterial compliance, were all consistently reduced during the early luteal phase, while pulse wave velocity did not [126, 127], suggesting that vascular endothelial function possibly fluctuates with hormonal changes throughout the menstrual cycle although there was little alteration found in arterial stiffness. However, longer periods of estrogen deficiency may trigger changes in arterial stiffness. Several studies have reported the augmenting age-related increase in arterial stiffness in postmenopausal women [128-131]. Moreover, the evidence is inconclusive on the relationship between menopause and cognition. Some suggest that menopause at younger ages was associated with reduced cognitive performance in later years [132, 133], while others reported that no clinically significant changes were found with respect to cognitive function [134, 135] or that only short-lasting cognitive decline was observed in limited domains with menopause [136, 137]. Thus, vascular aging may be linked with neurological diseases in two parallel pathways: (1) reduction in brain microcirculation, arterial stiffening, endothelial dysfunction causing hypoperfusion and (2) blood-brain barrier breakdown associated with brain accumulation of serum proteins and several vasculotoxic and/or neurotoxic macromolecules ultimately leading to secondary neuronal degenerative changes. Aging and Disease • Volume 8, Number 4, August 2017 Diseases linked to vascular aging Peripheral rather than central vascular stiffness is relatively stronger in the correlation with the amount of Aβ being deposited whilst systemic arterial stiffness (a composite parameter of 493 Vascular aging: mechanisms and disease Xu X., et al subjects. During follow-up, 63 subjects developed a stroke (mean follow-up period, 3.2 years). Among this group, the risk of stroke increased with increasing aortic PWV index. Although other comorbidities of aging such as obesity and diabetes limit the predictive value of the conventional cardiovascular risk factors, PWV may be an independent predictor of coronary heart disease and stroke [167]. Willum-Hansen, et al. [168] also documented the prognostic value of PWV in a 1678-person population- based study. The meta-analysis consisted 10 studies that revealed that carotid stiffness is strongly associated with a higher stroke incidence, independent of aortic stiffness, supporting the idea that carotid stiffening is an important factor in the pathogenesis of stroke [169]. Indeed, the increased pulse wave is able to go through the carotid artery quickly and penetrate distally into the cerebral microcirculation, which may promote hypertrophic remodeling and induce chronic ischemia by hypoperfusion. central and peripheral arteries) or central arterial stiffness is strongly correlated with the progression of Aβ accumulation. Furthermore, as a result of arterial stiffness and endothelial dysfunction, pulse pressure increases and becomes a risk factor for AD and dementia in older adults [155]. High pulse pressure promotes vascular remodeling, which further impairs vascular function. Clinical data showed that increased pulse pressure is related to reduced cerebrospinal fluid Aβ, indicating that elevated pulse pressure may affect the clearance of Aβ [156]. This was confirmed by another group that pulse pressure elevation was significantly associated with reduced cerebrospinal fluid Aβ in the aged group [157]. An alternative mechanism may involve reduced cerebral blood flow (CBF). Decreased CBF was observed by arterial spin-labeling MRI in the frontal and parietal cortices of human subjects with either AD or VD [158]. Total (tCBF) and mean CBF decrease with age whereas pulsatile CBF velocity increases [159], perhaps due to the age-related increase in stiffness of elastic arteries and decreased compliance of muscular arteries and arterioles. Interestingly, females have relatively higher tCBF and lower cerebrovascular resistance than males [160]. In vivo studies demonstrated that subsequent hypoperfusion accelerates deposition of Aβ via increased generation and impaired clearance and could cause cognitive impairment [142, 161]. Diseases linked to vascular aging Greater Aβ concentration decreases the production of NO, thus causing a significant decrease in endothelial-dependent vasodilation along with the induction oendothelial cell apoptosis [162-164]. Arterial stiffness also serves as a predictor of stroke outcome. Low carotid–femoral PWV is significantly associated with excellent stroke outcome, even after adjustment for age, baseline NIHSS score on admission and stroke history [170]. In stroke patients, aortic PWV helps predict asymptomatic coronary artery disease beyond the predictive ability of classical risk factors. In fact, coronary artery disease is a significant cause of morbidity and mortality in patients who have had a stroke [171]. Further, arterial stiffness is independently related to cerebral microbleeds in patients with stroke [172]. Endothelial dysfunction is another potentially adverse alteration with advancing age in stroke development. Endothelial dysfunction reduces the tone of aging arteries and increases their sensitivity to vasoconstrictors or stimulus, which may result in a higher risk of acute or chronic ischemia. It has been reported that acute ischemic stroke patients have severe endothelial dysfunction [173]. More specifically, the relationship between endothelial dysfunction and stroke subtypes has been studied, which revealed that only lacunar stroke is closely related with endothelial dysfunction [174]. Endothelial dysfunction may be involved in the pathogenesis of lacunar stroke especially in those subjects with concomitant silent lacunar infarcts and ischemic white matter hyperintensities (WMH) [175]. Endothelial function measured by brachial FMD is impaired in stroke patients compared with healthy controls, which has also been observed in other studies [176, 177]. Selecting homocysteine, von Willebrand factor (vWF), E-selectin, P-selectin, intercellular adhesion molecule-1 (ICAM), and vascular cellular adhesion molecule-1 (VCAM) as blood markers of endothelial dysfunction, Wiseman et al. [178] found that these markers were significantly increased in lacunar stroke vs. the non-stroke group. On Aging and Disease • Volume 8, Number 4, August 2017 Stroke Stroke is the leading cause of death worldwide that accounts for about 1 in 10 deaths or 5.7 million deaths a year. Stroke incidence is doubled in low- and middle- income regions compared with those living in high- income regions [165]. It can be fatal or dramatically decrease the quality of life in survivors and often necessitates extremely costly medical care. Stroke incidence increases with advancing age, so research on the relationship between vascular aging and stroke could have significant clinical benefits. Arterial stiffness is an independent predictor for fatal stroke as demonstrated in a long-term clinical study of 1715 patients with essential hypertension whose carotid- femoral PWV was measured at entry. Importantly, the predictive value of PWV remained significant even after full adjustment for classical cardiovascular risk factors such as age, gender, cholesterol level, diabetes and smoking [166]. Another study comprising 2835 participants in the third examination phase of the Rotterdam Study, demonstrated that arterial stiffness is an independent predictor of stroke among otherwise healthy 494 Vascular aging: mechanisms and disease Xu X., et al arterial abnormalities were found. Thus, endothelial dysfunction of brain vessels can be an important determinant of lacunar stroke [179]. Overall, vascular aging may have important roles in stroke pathogenesis and could be a potential marker of stroke risk and outcome. the cerebrovascular level, by assessing cerebral vasoreactivity and extracerebral vascular functions in symptomatic lacunar stroke patients and non-stroke controls, symptomatic lacunar stroke patients had more severe endothelial dysfunction in the cerebral arteries while relatively smaller difference of extracerebral Table 1. Devices and methods used for evaluating arterial stiffness. Table 1. Devices and methods used for evaluating arterial stiffness. Device Manufacturer Method Region or arterial site Regional arterial stiffness Sphygmocor AtCor Medical, Au Pressure cfPWV Diasys Novacor, Fr Korotkov sounds & ECG baPWV Omron VP-1000 Omron Healthcare, Jp Pressure cfPWV, baPWV, faPWV Vasera system Fukuda Denshi, Jp Pressure & ECG Cardio-ankle vascular index Mobil-O-Graph I.E.M. Ge Pressure cfPWV PulsePen Diatechne, It Pressure & ECG cfPWV PulseTrace Micromedical, UK Doppler & ECG cfPWV, baPWV Vicorder SMT medical GmbH & Co, Ge Pressure cfPWV, baPWV pOpmètre Axelife SAS, Fr Photodiodes sensors Finger-toe PWV Local arterial stiffness Artlab system Esaote, It Ultrasound CCA, CFA, BA E-Tracking Aloka, Jp Ultrasound CCA, CFA, BA MRI - Cine-MRI Aorta ECG, electrocardiography; CCA, common carotid artery; CFA, common femoral artery; BA, branchial artery Aging and Disease • Volume 8, Number 4, August 2017 Other neurological diseases Taken together, vascular aging may affect the process and outcome of neurological diseases including dementia, cognitive impairment and stoke. Therefore, it needs to be carefully considered as an important detrimental factor in these diseases, especially in certain populations such as patients with hypertension. Cerebral small vessel disease (SVD) is common in aged people and can be potentiated by hypertension and diabetes mellitus [180]. SVD can be visualized by MRI as WMH, lacunar infarcts, and cerebral microbleeds [181]; it also contributes to the development of cognitive decline and dementia [182, 183]. Several studies have reported an association between arterial stiffness (measured based on brachial-ankle or carotid-femoral pulse wave velocity) and SVD [184-187]. Age is associated with WMH, although the relationship between age and lacunar infarction, in particular, differs among studies. Some studies report no statistically significant link [188, 189], but this may be explained by the relatively young population (mean age 46.9-58.2 years) and the low prevalence of lacunar infracts (4.3%). When older populations were examined, arterial stiffness was shown to be associated with lacunar infarcts [185, 190-192]. Brain atrophy is also related to vascular aging [193, 194]. Increased aorta or carotid stiffness is associated with larger WMH and increased carotid stiffness is associated with decreased total brain parenchyma, grey matter and especially, white matter volumes. These associations were even stronger in uncontrolled hypertension and aged patients [195]. Considering the importance of the cerebral blood supply for brain function, SVD, thus, has a great impact on cognitive function. The vascular alterations-induced vascular dysfunction that finally leads to cognitive impairment are diverse, which includes hemodynamic changes or alterations on cerebral blood vessels. General hypoperfusion can produce transient or permanent cognitive impairment [196], which is a result of structural and functional age-related changes. The most acknowledged vascular lesions promoting vascular cognitive impairment is SVD, especially in white matter [197]. These lesions often coexist in one single patient and contribute to the cognitive impairment. Lacunar infarcts and leukoaraiosis that are common in VCI are also associated with arterial stiffness. Patients with leukoaraiosis have higher PWV that transmits increased pulse pressure into the brain through the middle cerebral artery [198]. Tortuous arterioles, increased intima-media thickness, reduced capillaries number and length may be the underlying vascular alterations. Aging and Disease • Volume 8, Number 4, August 2017 Other neurological diseases Microbleeds is a Aging and Disease • Volume 8, Number 4, August 2017 495 Vascular aging: mechanisms and disease Xu X., et al predictor of cognitive dysfunction and arterial stiffness may be important in the development of microbleeds, especially in stroke patients [172, 199]. Furthermore, it is well known that Aβ deposition or cerebral amyloid angiopathy is associated with vascular cognitive impairment, although it seems to be a “marker” of AD, with it being present in over 90% of patients [200]; it can be promoted by arterial aging-induced arterial stiffness and endothelial dysfunction. AD-related cerebrovascular lesions and VCIs are similar. Data containing different average prevalence of leukoaraiosis and lacunar infarcts, microbleeds, microinfarcts, Aβ deposition as well as hypoperfusion in AD and VCI or VD could reveal the pattern of vascular alterations in these diseases. A previous review reported that the most remarkable vascular-related difference between AD and VCI is cerebral amyloid angiopathy and microinfarcts/lacunes [200]. Cerebral amyloid angiopathy can be found in almost 98% of AD patients vs. 30% of VCI patients and 23-45% of aged controls. Microinfarcts/lacunes were seen in 70% of VCI patients vs. 10-15% of AD patients and just over 10% of aged controls. (CCA) and the common femoral artery (cfPWV). cfPWV is often thought to be the gold standard for measuring arterial stiffness as it evaluates the arteries from the aortic through the aortoiliac pathway including big branches, which are essential in the hemodynamic load for the end organ [201]. The femoral–ankle PWV (faPWV) is considered a peripheral arterial stiffness index. Furthermore, brachial-ankle (baPWV) assesses the mechanical property of a large territory, integrating both the large-sized central elastic and medium-sized peripheral muscular arteries thus correlating better with left ventricular mass and diastolic function and other indices of arterial function than cf-PWV; it also can possibly be considered a relative systemic index of arterial stiffness [202]. For local arterial stiffness, superﬁcial arteries can be directly measured by ultrasound. The carotid may be a region with great interest to clinicians not only because it is a superficial artery, but it also is a high incidence region of atherosclerosis. All types of classic bi-dimensional vascular ultrasound devices can probably be used to examine parameters including local arterial distensibility, diameter and thickness. However, because of the limitation of ultrasound technology itself, the readings from devices always differ from clinician to clinician, even with the same device. Other neurological diseases Thus, there is increasing interest in measuring arterial stiffness through MRI. Interestingly, faster and more robust MRI sequences have enabled MRI-based PWV measurements. Nonetheless, the lack of compatible computing software limited the use of MRI-based PWV [203]. Because of its ease and wide clinical utility coupled with a much lower cost, the ultrasound technology is still the most prevalent method for local arterial stiffness assessment (Table 1). Table 2. Approaches for assessing endothelial function. Method Vascular site Invasive or Non-invasive Coronary epicardial vasoreactivity Coronary Invasive FMD Peripheral conduit artery Non-invasive EndoPAT Small arteries and finger microvasculature Non-invasive Intracoronary doppler wire Coronary microcirculation Invasive Venous occlusion plethysmography Forearm vein Invasive Table 2. Approaches for assessing endothelial function. Endothelial dysfunction is a pathological condition during aging. Similar to measurements for arterial stiffness, endothelial function assessment would require stimulating the human body by pacing or exercise, or stimulating the arteries locally by acetylcholine or reactive hyperemia, then investigate if the blood flow or dilation increases as healthy arteries are expected to. Despite its high cost and invasive process, coronary epicardial vasoreactivity is the gold standard for epicardial macrovascular endothelial dysfunction (Table 2). However, for an asymptomatic individual, non- invasive and cheaper measurements would be appropriate for screening. As a noninvasive approach, FMD has been the most commonly used measurement of endothelial function of the brachial or other peripheral conduit artery. It measures the artery’s response to endothelial NO during reactive hyperemia after a 5-minute occlusion of the brachial artery using a blood pressure cuff to estimates endothelial function. Measuring endothelial function with peripheral arterial tonometry (EndoPAT) is another non- invasive approach that evaluates changes in finger arterial Acknowledgements This work was supported by the China Scholarship Council (No. 201406320118) (to XX), American Heart Association Predoctoral Fellowship (15PRE25700416) (to BW), Sigma Xi Grants-in-Aid of Research (G201503151162287) (to BW), National Natural Science Foundation of China (81301996) (to TC), Ph.D. Programs Foundation of Ministry of Education of China (20130101120017) (to TC). National Institutes of Health grant NS084148-01A1 (to KJ), the National Science Foundation of Beijing (No. 7161014) (to KJ). Potential intervention measures concerted effort from researchers and clinicians to work together to advance the field of vascular aging and find interventions to promote cardiovascular and cerebrovascular health in the elderly. As life expectancy is expected to increase with the use of modern medical technologies, biomedical research is likely to focus increasingly on cerebrovascular and cardiovascular diseases. Consequently, numerous research groups consider vascular aging an important area of research (Table 3). However, even in the case of less invasive strategies such as exercise and diet, purported benefits of new interventions must be weighed against potential adverse effects. Measurements There are many methods that can be used for assessing vascular aging. As vascular aging is a ubiquitous phenomenon with both clinical and subclinical manifestations, large-scale application of invasive techniques may be inappropriate and can lead to overtreatment. Arterial stiffness can be assessed at different levels, either regional or local. For regional arterial stiffness assessment, PWV seems to be a widely used and reproducible approach. Central arterial stiffness can only be estimated by PWV between the common carotid artery Aging and Disease • Volume 8, Number 4, August 2017 496 Vascular aging: mechanisms and disease Xu X., et al internal control [204]. pulse wave amplitude under the stimulation of reactive hyperemia. The contralateral arm could serve as an internal control [204]. pulse wave amplitude under the stimulation of reactive hyperemia. The contralateral arm could serve as an internal control [204]. Table 3. Potential intervention measures for vascular aging. Intervention Measure Influence of intervention Underlying Mechanism Reference Exercise Decreases vessel tortuosity, cardiovascular morbidity and mortality Increases CBF, capillary vascularity Maintain endothelial function eNOS↑ ROS↓ [205-207] Caloric restriction Decreases serum cholesterol, arterial stiffness and attenuate vascular inflammation Protects endothelial function eNOS↑ ROS↓ Circulating C-reactive protein (CRP) ↓ TNFα↓ [208, 209] Statins Decreases PWV Protects endothelial function Anti-vascular inflammation eNOS↑ Serum CRP ↓ Activity of NF-kB↓ ROS↓ [210-212] RAS Drugs Decreases PWV Increases CBF Block Ang II signaling [15, 213, 214] Table 3. Potential intervention measures for vascular aging. Summary The molecular targets through which aging perturbs vascular homeostasis are manifold and appear to include increased expression or activation of TNF-α, IL-1β, IL-6 family members and CRP, which promote inflammation and endothelial dysfunction. Currently, there are many non-invasive ways to examine vascular aging through arterial stiffness and endothelial function assessments. New interventions such as the use of statins and RAS drugs to treat vascular aging have been suggested, but clinicians and patients should consider potential adverse effects. Aging and Disease • Volume 8, Number 4, August 2017 [1] Ungvari Z, Kaley G, de Cabo R, Sonntag WE, Csiszar A (2010). Mechanisms of vascular aging: new perspectives. J Gerontol A Biol Sci Med Sci, 65: 1028- 1041 [2] Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. (2012). Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet, 380: 2095- References Hypertension, 50: 219-227 22] Dupuis F, Atkinson J, Liminana P, Chillon JM (2005). Captopril improves cerebrovascular structure and function in old hypertensive rats. Br J Pharmacol, 144: 349-356 [9] [23] [9] Wang M, Zhang J, Spinetti G, Jiang LQ, Monticone R, Zhao D, et al. (2005). Angiotensin II activates matrix metalloproteinase type II and mimics age-associated carotid arterial remodeling in young rats. Am J Pathol, 167: 1429-1442 Estato V, Obadia N, Carvalho-Tavares J, Freitas FS, Reis P, Castro-Faria NH, et al. (2013). Blockade of the renin-angiotensin system improves cerebral microcirculatory perfusion in diabetic hypertensive rats. Microvasc Res, 87: 41-49 [10] [24] [10] Xie X, Chen J, Wang X, Zhang F, Liu Y (2006). Age- and gender-related difference of ACE2 expression in rat lung. Life Sci, 78: 2166-2171 [24] Rajagopalan S, Kariisa M, Dellegrottaglie S, Bard RL, Kehrer C, Matlow S, et al. (2006). Angiotensin receptor blockade improves vascular compliance in healthy normotensive elderly individuals: results from a randomized double-blind placebo-controlled trial. J Clin Hypertens (Greenwich), 8: 783-790 [11] Rajagopalan S, Brook R, Mehta RH, Supiano M, Pitt B (2002). Effect of losartan in aging-related endothelial impairment. Am J Cardiol, 89: 562-566 [12] [25] [12] Modrick ML, Didion SP, Sigmund CD, Faraci FM (2009). Role of oxidative stress and AT1 receptors in cerebral vascular dysfunction with aging. Am J Physiol Heart Circ Physiol, 296: H1914-H1919 Fournier A, Oprisiu-Fournier R, Serot JM, Godefroy O, Achard JM, Faure S, et al. (2009). Prevention of dementia by antihypertensive drugs: how AT1- receptor-blockers and dihydropyridines better prevent dementia in hypertensive patients than thiazides and ACE-inhibitors. Expert Rev Neurother, 9: 1413-1431 [13] [13] Wolf G, Wenzel U, Burns KD, Harris RC, Stahl RA, Thaiss F (2002). Angiotensin II activates nuclear transcription factor-kappaB through AT1 and AT2 receptors. Kidney Int, 61: 1986-1995 [26] Shah K, Qureshi SU, Johnson M, Parikh N, Schulz PE, Kunik ME (2009). Does use of antihypertensive drugs affect the incidence or progression of dementia? A systematic review. Am J Geriatr Pharmacother, 7: 250-261 [14] [14] Tota S, Goel R, Pachauri SD, Rajasekar N, Najmi AK, Hanif K, et al. (2013). Effect of angiotensin II on spatial memory, cerebral blood flow, cholinergic neurotransmission, and brain derived neurotrophic factor in rats. Psychopharmacology (Berl), 226: 357- 369 [27] McGuinness B, Todd S, Passmore P, Bullock R (2009). Blood pressure lowering in patients without prior cerebrovascular disease for prevention of cognitive impairment and dementia. References [1] Ungvari Z, Kaley G, de Cabo R, Sonntag WE, Csiszar A (2010). Mechanisms of vascular aging: new perspectives. J Gerontol A Biol Sci Med Sci, 65: 1028- 1041 [2] Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. (2012). Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet, 380: 2095- What may seem like an insurmountable task with regard to the treatment for vascular dysfunction may in fact hold much promise and excitement. It will take a Aging and Disease • Volume 8, Number 4, August 2017 497 Xu X., et al Vascular aging: mechanisms and disease Suppl, 27: S3-S7 2128 [3] [18] [3] Nichols M, Townsend N, Scarborough P, Rayner M (2014). Cardiovascular disease in Europe 2014: epidemiological update. Eur Heart J, 35: 2950-2959 [18] Sohn SI, Kim CJ (2014). Modulation of renin- angiotensin system and arterial stiffness: evidence from clinical trials. Curr Hypertens Rev, 10: 37-40 [4] [19] [4] Wang M, Jiang L, Monticone RE, Lakatta EG (2014). Proinflammation: the key to arterial aging. Trends Endocrinol Metab, 25: 72-79 [19] Fleegal-DeMotta MA, Doghu S, Banks WA (2009). Angiotensin II modulates BBB permeability via activation of the AT(1) receptor in brain endothelial cells. J Cereb Blood Flow Metab, 29: 640-647 [5] [5] Stegbauer J, Coffman TM (2011). New insights into angiotensin receptor actions: from blood pressure to aging. Curr Opin Nephrol Hypertens, 20: 84-88 [20] Zhang M, Mao Y, Ramirez SH, Tuma RF, Chabrashvili T (2010). Angiotensin II induced cerebral microvascular inflammation and increased blood-brain barrier permeability via oxidative stress. Neuroscience, 171: 852-858 [6] [6] Basso N, Cini R, Pietrelli A, Ferder L, Terragno NA, Inserra F (2007). Protective effect of long-term angiotensin II inhibition. Am J Physiol Heart Circ Physiol, 293: H1351-H1358 [21 [21] Nishimura Y, Ito T, Saavedra JM (2000). Angiotensin II AT(1) blockade normalizes cerebrovascular autoregulation and reduces cerebral ischemia in spontaneously hypertensive rats. Stroke, 31: 2478- 2486 [7] [7] Wang M, Takagi G, Asai K, Resuello RG, Natividad FF, Vatner DE, et al. (2003). Aging increases aortic MMP-2 activity and angiotensin II in nonhuman primates. Hypertension, 41: 1308-1316 [8] [22] [8] Wang M, Zhang J, Jiang LQ, Spinetti G, Pintus G, Monticone R, et al. (2007). Proinflammatory profile within the grossly normal aged human aortic wall. Aging and Disease • Volume 8, Number 4, August 2017 References Cochrane Database Syst Rev, D4034 [15] Topouchian J, El FR, Pannier B, Wang S, Zhao F, Smetana K, et al. (2007). Arterial stiffness and pharmacological interventions--the TRanscend arterial stiffNess Substudy (TRANS study). Vasc Health Risk Manag, 3: 381-387 [28] Marchesi VT (2011). Alzheimer's dementia begins as a disease of small blood vessels, damaged by oxidative-induced inflammation and dysregulated amyloid metabolism: implications for early detection and therapy. Faseb J, 25: 5-13 [16] Werner C, Baumhakel M, Teo KK, Schmieder R, Mann J, Unger T, et al. (2008). RAS blockade with ARB and ACE inhibitors: current perspective on rationale and patient selection. Clin Res Cardiol, 97: 418-431 [29] Hajjar I, Rodgers K (2013). Do angiotensin receptor blockers prevent Alzheimer's disease? Curr Opin Cardiol, 28: 417-425 [17] [30] Chang-Quan H, Hui W, Chao-Min W, Zheng-Rong W, Jun-Wen G, Yong-Hong L, et al. (2011). The association of antihypertensive medication use with Perret-Guillaume C, Joly L, Jankowski P, Benetos A (2009). Benefits of the RAS blockade: clinical evidence before the ONTARGET study. J Hypertens Aging and Disease • Volume 8, Number 4, August 2017 498 Xu X., et al Vascular aging: mechanisms and disease glycation end products and protective effects of grape seed procyanidin B2 and resveratrol. Apoptosis, 16: 732-745 risk of cognitive decline and dementia: a meta-analysis of longitudinal studies. Int J Clin Pract, 65: 1295-1305 [31] [31] Zhuang S, Wang HF, Wang X, Li J, Xing CM (2016). The association of renin-angiotensin system blockade use with the risks of cognitive impairment of aging and Alzheimer's disease: A meta-analysis. J Clin Neurosci, [44] Sbardella D, Fasciglione GF, Gioia M, Ciaccio C, Tundo GR, Marini S, et al. (2012). Human matrix metalloproteinases: an ubiquitarian class of enzymes involved in several pathological processes. Mol Aspects Med, 33: 119-208 [32] [32] Brissette MJ, Lepage S, Lamonde AS, Sirois I, Groleau J, Laurin LP, et al. (2012). MFG-E8 released by apoptotic endothelial cells triggers anti- inflammatory macrophage reprogramming. Plos One, 7: e36368 [45] Galis ZS, Khatri JJ (2002). Matrix metalloproteinases in vascular remodeling and atherogenesis: the good, the bad, and the ugly. Circ Res, 90: 251-262 [33] [46] [33] Lee J, Choi BI, Park SY, An SY, Han J, Kim JH (2015). Fetal hematopoietic stem cells express MFG-E8 during mouse embryogenesis. Exp Mol Med, 47: e174 [ Jacob MP (2003). Extracellular matrix remodeling and matrix metalloproteinases in the vascular wall during aging and in pathological conditions. References Biomed Pharmacother, 57: 195-202 [34] [34] Veron P, Segura E, Sugano G, Amigorena S, Thery C (2005). Accumulation of MFG-E8/lactadherin on exosomes from immature dendritic cells. Blood Cells Mol Dis, 35: 81-88 [47] [47] Xiao Q, Zhang F, Grassia G, Hu Y, Zhang Z, Xing Q, et al. (2014). Matrix metalloproteinase-8 promotes vascular smooth muscle cell proliferation and neointima formation. Arterioscler Thromb Vasc Biol, 34: 90-98 [35] [35] Ait-Oufella H, Kinugawa K, Zoll J, Simon T, Boddaert J, Heeneman S, et al. (2007). Lactadherin deficiency leads to apoptotic cell accumulation and accelerated atherosclerosis in mice. Circulation, 115: 2168-2177 [48] Jiang L, Zhang J, Monticone RE, Telljohann R, Wu J, Wang M, et al. (2012). Calpain-1 regulation of matrix metalloproteinase 2 activity in vascular smooth muscle cells facilitates age-associated aortic wall calcification and fibrosis. Hypertension, 60: 1192- 1199 [36] [36] Neutzner M, Lopez T, Feng X, Bergmann-Leitner ES, Leitner WW, Udey MC (2007). MFG-E8/lactadherin promotes tumor growth in an angiogenesis-dependent transgenic mouse model of multistage carcinogenesis. Cancer Res, 67: 6777-6785 [49] [49] Li Z, Froehlich J, Galis ZS, Lakatta EG (1999). Increased expression of matrix metalloproteinase-2 in the thickened intima of aged rats. Hypertension, 33: 116-123 [37] [37] Aziz M, Yang WL, Corbo LM, Chaung WW, Matsuo S, Wang P (2015). MFG-E8 inhibits neutrophil migration through alphavbeta(3)-integrin-dependent MAP kinase activation. Int J Mol Med, 36: 18-28 [50] [50] McNulty M, Spiers P, McGovern E, Feely J (2005). Aging is associated with increased matrix metalloproteinase-2 activity in the human aorta. Am J Hypertens, 18: 504-509 [38] [38] Cheyuo C, Jacob A, Wu R, Zhou M, Qi L, Dong W, et al. (2012). Recombinant human MFG-E8 attenuates cerebral ischemic injury: its role in anti-inflammation and anti-apoptosis. Neuropharmacology, 62: 890-900 [51] [51] Ihara M, Tomimoto H, Kinoshita M, Oh J, Noda M, Wakita H, et al. (2001). Chronic cerebral hypoperfusion induces MMP-2 but not MMP-9 expression in the microglia and vascular endothelium of white matter. J Cereb Blood Flow Metab, 21: 828- 834 [39] p p p gy [39] Raymond A, Ensslin MA, Shur BD (2009). SED1/MFG-E8: a bi-motif protein that orchestrates diverse cellular interactions. J Cell Biochem, 106: 957-966 [40] [52] Popp O, Heidinger M, Ruiz-Heinrich L, Ries C, Jochum M, Gil-Parrado S (2003). The calpastatin- derived calpain inhibitor CP1B reduces mRNA expression of matrix metalloproteinase-2 and -9 and invasion by leukemic THP-1 cells. References Biol Chem, 384: 951-958 [40] Fu Z, Wang M, Gucek M, Zhang J, Wu J, Jiang L, et al. (2009). Milk fat globule protein epidermal growth factor-8: a pivotal relay element within the angiotensin II and monocyte chemoattractant protein-1 signaling cascade mediating vascular smooth muscle cells invasion. Circ Res, 104: 1337-1346 [52] [41] [53] [53] Rollins BJ (1996). Monocyte chemoattractant protein 1: a potential regulator of monocyte recruitment in inflammatory disease. Mol Med Today, 2: 198-204 41] Wang M, Fu Z, Wu J, Zhang J, Jiang L, Khazan B, et al. (2012). MFG-E8 activates proliferation of vascular smooth muscle cells via integrin signaling. Aging Cell, 11: 500-508 [54] Ma J, Wang Q, Fei T, Han JD, Chen YG (2007). MCP- 1 mediates TGF-beta-induced angiogenesis by stimulating vascular smooth muscle cell migration. Blood, 109: 987-994 [42] [42] Marrero MB, Schieffer B, Li B, Sun J, Harp JB, Ling BN (1997). Role of Janus kinase/signal transducer and activator of transcription and mitogen-activated protein kinase cascades in angiotensin II- and platelet- derived growth factor-induced vascular smooth muscle cell proliferation. J Biol Chem, 272: 24684- 24690 [55] 5] Rossi D, Zlotnik A (2000). The biology of chemokines and their receptors. Annu Rev Immunol, 18: 217-242 [56] Salcedo R, Ponce ML, Young HA, Wasserman K, Ward JM, Kleinman HK, et al. (2000). Human endothelial cells express CCR2 and respond to MCP- 1: direct role of MCP-1 in angiogenesis and tumor progression. Blood, 96: 34-40 [43] Li BY, Li XL, Cai Q, Gao HQ, Cheng M, Zhang JH, et al. (2011). Induction of lactadherin mediates the apoptosis of endothelial cells in response to advanced Aging and Disease • Volume 8, Number 4, August 2017 499 Xu X., et al Vascular aging: mechanisms and disease [57] oxidase activation. Contribution to alterations of vasomotor tone. J Clin Invest, 97: 1916-1923 [57] Spinetti G, Wang M, Monticone R, Zhang J, Zhao D, Lakatta EG (2004). Rat aortic MCP-1 and its receptor CCR2 increase with age and alter vascular smooth muscle cell function. Arterioscler Thromb Vasc Biol, 24: 1397-1402 [71] Touyz RM (2003). Reactive oxygen species in vascular biology: role in arterial hypertension. Expert Rev Cardiovasc Ther, 1: 91-106 [58] [72] Beckman JS (2001). -OONO: rebounding from nitric oxide. Circ Res, 89: 295-297 [58] Tsuchiya K, Yoshimoto T, Hirono Y, Tateno T, Sugiyama T, Hirata Y (2006). References Angiotensin II induces monocyte chemoattractant protein-1 expression via a nuclear factor-kappaB-dependent pathway in rat preadipocytes. Am J Physiol Endocrinol Metab, 291: E771-E778 [73] Faraci FM, Didion SP (2004). Vascular protection: superoxide dismutase isoforms in the vessel wall. Arterioscler Thromb Vasc Biol, 24: 1367-1373 [74] Zhao H, Han Z, Ji X, Luo Y (2016). Epigenetic Regulation of Oxidative Stress in Ischemic Stroke. Aging Dis, 7: 295-306 [59] [59] Chen XL, Tummala PE, Olbrych MT, Alexander RW, Medford RM (1998). Angiotensin II induces monocyte chemoattractant protein-1 gene expression in rat vascular smooth muscle cells. Circ Res, 83: 952- 959 [75] Guzik TJ, Mussa S, Gastaldi D, Sadowski J, Ratnatunga C, Pillai R, et al. (2002). Mechanisms of increased vascular superoxide production in human diabetes mellitus: role of NAD(P)H oxidase and endothelial nitric oxide synthase. Circulation, 105: 1656-1662 [60] [60] Pohlers D, Brenmoehl J, Loffler I, Muller CK, Leipner C, Schultze-Mosgau S, et al. (2009). TGF-beta and fibrosis in different organs - molecular pathway imprints. Biochim Biophys Acta, 1792: 746-756 [76] Soucy KG, Ryoo S, Benjo A, Lim HK, Gupta G, Sohi JS, et al. (2006). Impaired shear stress-induced nitric oxide production through decreased NOS phosphorylation contributes to age-related vascular stiffness. J Appl Physiol (1985), 101: 1751-1759 [61] [61] Feng XH, Derynck R (2005). Specificity and versatility in tgf-beta signaling through Smads. Annu Rev Cell Dev Biol, 21: 659-693 [62] [62] Nakao A, Afrakhte M, Moren A, Nakayama T, Christian JL, Heuchel R, et al. (1997). Identification of Smad7, a TGFbeta-inducible antagonist of TGF- beta signalling. Nature, 389: 631-635 [77] Nelson KK, Melendez JA (2004). Mitochondrial redox control of matrix metalloproteinases. Free Radic Biol Med, 37: 768-784 [63] [78] [63] Weigert C, Brodbeck K, Klopfer K, Haring HU, Schleicher ED (2002). Angiotensin II induces human TGF-beta 1 promoter activation: similarity to hyperglycaemia. Diabetologia, 45: 890-898 [78] [63] Weigert C, Brodbeck K, Klopfer K, Haring HU, Schleicher ED (2002). Angiotensin II induces human TGF-beta 1 promoter activation: similarity to hyperglycaemia. Diabetologia, 45: 890-898 [64] Yu Q, Stamenkovic I (2000). Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis Genes Dev 14: 163 176 [78] [79] Rajagopalan S, Meng XP, Ramasamy S, Harrison DG, Galis ZS (1996). Reactive oxygen species produced by macrophage-derived foam cells regulate the activity of vascular matrix metalloproteinases in vitro. Implications for atherosclerotic plaque stability. J Clin Invest, 98: 2572-2579 [64] [64] Yu Q, Stamenkovic I (2000). Aging and Disease • Volume 8, Number 4, August 2017 References Cell surface-localized matrix metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor invasion and angiogenesis. Genes Dev, 14: 163-176 [79] [79] Krishnan KJ, Greaves LC, Reeve AK, Turnbull DM (2007). Mitochondrial DNA mutations and aging. Ann N Y Acad Sci, 1100: 227-240 [65] [65] Wang M, Zhao D, Spinetti G, Zhang J, Jiang LQ, Pintus G, et al. (2006). Matrix metalloproteinase 2 activation of transforming growth factor-beta1 (TGF- beta1) and TGF-beta1-type II receptor signaling within the aged arterial wall. Arterioscler Thromb Vasc Biol, 26: 1503-1509 [80] [80] Mecocci P, MacGarvey U, Kaufman AE, Koontz D, Shoffner JM, Wallace DC, et al. (1993). Oxidative damage to mitochondrial DNA shows marked age- dependent increases in human brain. Ann Neurol, 34: 609-616 [66] [81] [66] Ying WZ, Aaron KJ, Sanders PW (2013). Transforming growth factor-beta regulates endothelial function during high salt intake in rats. Hypertension, 62: 951-956 Corral-Debrinski M, Horton T, Lott MT, Shoffner JM, Beal MF, Wallace DC (1992). Mitochondrial DNA deletions in human brain: regional variability and increase with advanced age. Nat Genet, 2: 324-329 [67] [82] Pak JW, Vang F, Johnson C, McKenzie D, Aiken JM (2005). MtDNA point mutations are associated with deletion mutations in aged rat. Exp Gerontol, 40: 209- 218 [67] Fleenor BS, Marshall KD, Durrant JR, Lesniewski LA, Seals DR (2010). Arterial stiffening with ageing is associated with transforming growth factor-beta1- related changes in adventitial collagen: reversal by aerobic exercise. J Physiol, 588: 3971-3982 [82] [83] [83] Larsson NG (2010). Somatic mitochondrial DNA mutations in mammalian aging. Annu Rev Biochem, 79: 683-706 [68] [68] Fortuno A, San JG, Moreno MU, Diez J, Zalba G (2005). Oxidative stress and vascular remodelling. Exp Physiol, 90: 457-462 [84] [84] Clayton DA (1982). Replication of animal mitochondrial DNA. Cell, 28: 693-705 [69] [69] Erusalimsky JD (2009). Vascular endothelial senescence: from mechanisms to pathophysiology. J Appl Physiol (1985), 106: 326-332 [85] Wenzel P, Schuhmacher S, Kienhofer J, Muller J, Hortmann M, Oelze M, et al. (2008). Manganese superoxide dismutase and aldehyde dehydrogenase deficiency increase mitochondrial oxidative stress and aggravate age-dependent vascular dysfunction. Cardiovasc Res, 80: 280-289 [70] [70] Rajagopalan S, Kurz S, Munzel T, Tarpey M, Freeman BA, Griendling KK, et al. (1996). References Phys Chem Chem Phys, 15: 18353-18356 [105] [91] [106] 91] Banoei MM, Houshmand M, Panahi MS, Shariati P, Rostami M, Manshadi MD, et al. (2007). Huntington's disease and mitochondrial DNA deletions: event or regular mechanism for mutant huntingtin protein and CAG repeats expansion?!. Cell Mol Neurobiol, 27: 867-875 [106] [107] Zhou P, Fu L, Pan Z, Ma D, Zhang Y, Qu F, et al. (2008). Testosterone deprivation by castration impairs expression of voltage-dependent potassium channels in rat aorta. Eur J Pharmacol, 593: 87-91 [107] [107] Mathur A, Malkin C, Saeed B, Muthusamy R, Jones TH, Channer K (2009). Long-term benefits of testosterone replacement therapy on angina threshold and atheroma in men. Eur J Endocrinol, 161: 443-449 [92] [92] Botto N, Berti S, Manfredi S, Al-Jabri A, Federici C, Clerico A, et al. (2005). Detection of mtDNA with 4977 bp deletion in blood cells and atherosclerotic lesions of patients with coronary artery disease. Mutat Res, 570: 81-88 [108] Yaron M, Greenman Y, Rosenfeld JB, Izkhakov E, Limor R, Osher E, et al. (2009). Effect of testosterone replacement therapy on arterial stiffness in older hypogonadal men. Eur J Endocrinol, 160: 839-846 [93] [93] Yu E, Calvert PA, Mercer JR, Harrison J, Baker L, Figg NL, et al. (2013). Mitochondrial DNA damage can promote atherosclerosis independently of reactive oxygen species through effects on smooth muscle cells and monocytes and correlates with higher-risk plaques in humans. Circulation, 128: 702-712 [109] [109] Dockery F, Bulpitt CJ, Donaldson M, Fernandez S, Rajkumar C (2003). The relationship between androgens and arterial stiffness in older men. J Am Geriatr Soc, 51: 1627-1632 [110] [110] English KM, Jones RD, Jones TH, Morice AH, Channer KS (2000). Aging reduces the responsiveness of coronary arteries from male Wistar rats to the vasodilatory action of testosterone. Clin Sci (Lond), 99: 77-82 [94] [94] Roy AK, Lavrovsky Y, Song CS, Chen S, Jung MH, Velu NK, et al. (1999). Regulation of androgen action. Vitam Horm, 55: 309-352 [95] [95] Foradori CD, Weiser MJ, Handa RJ (2008). Non- genomic actions of androgens. Front Neuroendocrinol, 29: 169-181 [111] [111] Webb CM, McNeill JG, Hayward CS, de Zeigler D, Collins P (1999). Effects of testosterone on coronary vasomotor regulation in men with coronary heart disease. Circulation, 100: 1690-1696 [96] Wu FC, von Eckardstein A (2003). Androgens and coronary artery disease. Endocr Rev, 24: 183-217 [112] Jones RD, Pugh PJ, Jones TH, Channer KS (2003). References Angiotensin II- mediated hypertension in the rat increases vascular superoxide production via membrane NADH/NADPH Aging and Disease • Volume 8, Number 4, August 2017 500 Vascular aging: mechanisms and disease Xu X., et al [100] [86] [100] Goglia L, Tosi V, Sanchez AM, Flamini MI, Fu XD, Zullino S, et al. (2010). Endothelial regulation of eNOS, PAI-1 and t-PA by testosterone and dihydrotestosterone in vitro and in vivo. Mol Hum Reprod, 16: 761-769 [86] Ballinger SW, Patterson C, Yan CN, Doan R, Burow DL, Young CG, et al. (2000). Hydrogen peroxide- and peroxynitrite-induced mitochondrial DNA damage and dysfunction in vascular endothelial and smooth muscle cells. Circ Res, 86: 960-966 [87] [101] [101] Kelly DM, Jones TH (2013). Testosterone: a vascular hormone in health and disease. J Endocrinol, 217: R47-R71 [87] Barja G, Herrero A (2000). Oxidative damage to mitochondrial DNA is inversely related to maximum life span in the heart and brain of mammals. Faseb J, 14: 312-318 [102] [102] Akishita M, Hashimoto M, Ohike Y, Ogawa S, Iijima K, Eto M, et al. (2007). Low testosterone level is an independent determinant of endothelial dysfunction in men. Hypertens Res, 30: 1029-1034 [88] [88] Aliev G, Gasimov E, Obrenovich ME, Fischbach K, Shenk JC, Smith MA, et al. (2008). Atherosclerotic lesions and mitochondria DNA deletions in brain microvessels: implication in the pathogenesis of Alzheimer's disease. Vasc Health Risk Manag, 4: 721- 730 [103] Scragg JL, Jones RD, Channer KS, Jones TH, Peers C (2004). Testosterone is a potent inhibitor of L-type Ca(2+) channels. Biochem Biophys Res Commun, 318: 503-506 [89] 89] Bender A, Krishnan KJ, Morris CM, Taylor GA, Reeve AK, Perry RH, et al. (2006). High levels of mitochondrial DNA deletions in substantia nigra neurons in aging and Parkinson disease. Nat Genet, 38: 515-517 [104] [104] [104] Yildiz O, Seyrek M, Gul H, Un I, Yildirim V, Ozal E, et al. (2005). Testosterone relaxes human internal mammary artery in vitro. J Cardiovasc Pharmacol, 45: 580-585 [90] [105] Marijic J, Li Q, Song M, Nishimaru K, Stefani E, Toro L (2001). Decreased expression of voltage- and Ca(2+)-activated K(+) channels in coronary smooth muscle during aging. Circ Res, 88: 210-216 Kim Y, Yoo JM, Jeon HR, Hong BH (2013). Efficient n-doping of graphene films by APPE (aminophenyl propargyl ether): a substituent effect. Aging and Disease • Volume 8, Number 4, August 2017 [101] [112] References Asian J Androl, 14: 232-238 Huang G, Wharton W, Bhasin S, Harman SM, Pencina KM, Tsitouras P, et al. (2016). Effects of long-term testosterone administration on cognition in older men with low or low-to-normal testosterone concentrations: a prespecified secondary analysis of data from the randomised, double-blind, placebo-controlled TEAAM i l L Di b E d i l 4 657 665 [130] [130] Smulyan H, Asmar RG, Rudnicki A, London GM, Safar ME (2001). Comparative effects of aging in men and women on the properties of the arterial tree. J Am Coll Cardiol, 37: 1374-1380 [131] [131] Takahashi K, Miura S, Mori-Abe A, Kawagoe J, Takata K, Ohmichi M, et al. (2005). Impact of menopause on the augmentation of arterial stiffness with aging. Gynecol Obstet Invest, 60: 162-166 TEAAM trial. Lancet Diabetes Endocrinol, 4: 657-665 [117] Jamadar RJ, Winters MJ, Maki PM (2012). Cognitive changes associated with ADT: a review of the literature. Asian J Androl, 14: 232-238 [117] [132] [132] Ryan J, Scali J, Carriere I, Amieva H, Rouaud O, Berr C, et al. (2014). Impact of a premature menopause on cognitive function in later life. BJOG, 121: 1729-1739 [118] Mendelsohn ME (2002). Protective effects of estrogen on the cardiovascular system. Am J Cardiol, 89: 12E- 17E, 17E-18E [133] [133] Mitchell ES, Woods NF (2011). Cognitive symptoms during the menopausal transition and early postmenopause. Climacteric, 14: 252-261 [119] Hisamoto K, Ohmichi M, Kurachi H, Hayakawa J, Kanda Y, Nishio Y, et al. (2001). Estrogen induces the Akt-dependent activation of endothelial nitric-oxide synthase in vascular endothelial cells. J Biol Chem, 276: 3459-3467 [134] [134] Henderson VW, Sherwin BB (2007). Surgical versus natural menopause: cognitive issues. Menopause, 14: 572-579 [135] [120] [120] Nakajima T, Kitazawa T, Hamada E, Hazama H, Omata M, Kurachi Y (1995). 17beta-Estradiol inhibits the voltage-dependent L-type Ca2+ currents in aortic smooth muscle cells. Eur J Pharmacol, 294: 625-635 [135] [135] Henderson VW, Guthrie JR, Dudley EC, Burger HG, Dennerstein L (2003). Estrogen exposures and memory at midlife: a population-based study of women. Neurology, 60: 1369-1371 [136] [121] Novella S, Dantas AP, Segarra G, Medina P, Hermenegildo C (2012). Vascular Aging in Women: is Estrogen the Fountain of Youth? Front Physiol, 3: 165 [136] Fuh JL, Wang SJ, Lee SJ, Lu SR, Juang KD (2006). A longitudinal study of cognition change during early menopausal transition in a rural community. References The vasodilatory action of testosterone: a potassium- channel opening or a calcium antagonistic action? Br J Pharmacol, 138: 733-744 [97] Bhasin S, Buckwalter JG (2001). Testosterone supplementation in older men: a rational idea whose time has not yet come. J Androl, 22: 718-731 [98] Miller VM, Mulvagh SL (2007). Sex steroids and endothelial function: translating basic science to clinical practice. Trends Pharmacol Sci, 28: 263-270 [113] Kang SM, Jang Y, Kim J, Chung N, Cho SY, Chae JS, et al. (2002). Effect of oral administration of testosterone on brachial arterial vasoreactivity in men with coronary artery disease. Am J Cardiol, 89: 862- 864 [99] [99] Yu J, Akishita M, Eto M, Ogawa S, Son BK, Kato S, et al. (2010). Androgen receptor-dependent activation of endothelial nitric oxide synthase in vascular endothelial cells: role of phosphatidylinositol 3- kinase/akt pathway. Endocrinology, 151: 1822-1828 [114] Yeap BB, Hyde Z, Almeida OP, Norman PE, Chubb SA, Jamrozik K, et al. (2009). Lower testosterone Yeap BB, Hyde Z, Almeida OP, Norman PE, Chubb SA, Jamrozik K, et al. (2009). Lower testosterone Aging and Disease • Volume 8, Number 4, August 2017 501 Xu X., et al Vascular aging: mechanisms and disease levels predict incident stroke and transient ischemic attack in older men. J Clin Endocrinol Metab, 94: 2353-2359 [128] [128] Zaydun G, Tomiyama H, Hashimoto H, Arai T, Koji Y, Yambe M, et al. (2006). Menopause is an independent factor augmenting the age-related increase in arterial stiffness in the early postmenopausal phase. Atherosclerosis, 184: 137-142 [115] Lv W, Du N, Liu Y, Fan X, Wang Y, Jia X, et al. (2016). Low Testosterone Level and Risk of Alzheimer's Disease in the Elderly Men: a Systematic Review and Meta-Analysis. Mol Neurobiol, 53: 2679- 2684 [129] [129] Tang Y, Wang L, Wang J, Lin X, Wang Y, Jin K, et al. (2016). Ischemia-induced Angiogenesis is Attenuated in Aged Rats. Aging Dis, 7: 326-335 [116] [116] Huang G, Wharton W, Bhasin S, Harman SM, Pencina KM, Tsitouras P, et al. (2016). Effects of long-term testosterone administration on cognition in older men with low or low-to-normal testosterone concentrations: a prespecified secondary analysis of data from the randomised, double-blind, placebo-controlled TEAAM trial. Lancet Diabetes Endocrinol, 4: 657-665 [117] Jamadar RJ, Winters MJ, Maki PM (2012). Cognitive changes associated with ADT: a review of the literature. Aging and Disease • Volume 8, Number 4, August 2017 References Acta Neuropathol, 123: 381- Aging and Disease • Volume 8, Number 4, August 2017 502 Xu X., et al Vascular aging: mechanisms and disease 394 pressure is associated with Alzheimer biomarkers in cognitively normal older adults. Neurology, 81: 2024- 2027 [143] Luckhaus C, Flub MO, Wittsack HJ, Grass-Kapanke B, Janner M, Khalili-Amiri R, et al. (2008). Detection of changed regional cerebral blood flow in mild cognitive impairment and early Alzheimer's dementia by perfusion-weighted magnetic resonance imaging. Neuroimage, 40: 495-503 [157] Nation DA, Edmonds EC, Bangen KJ, Delano-Wood L, Scanlon BK, Han SD, et al. (2015). Pulse pressure in relation to tau-mediated neurodegeneration, cerebral amyloidosis, and progression to dementia in very old adults. Jama Neurol, 72: 546-553 [144] [144] Osawa A, Maeshima S, Shimamoto Y, Maeshima E, Sekiguchi E, Kakishita K, et al. (2004). Relationship between cognitive function and regional cerebral blood flow in different types of dementia. Disabil Rehabil, 26: 739-745 [158] y Schuff N, Matsumoto S, Kmiecik J, Studholme C, Du A, Ezekiel F, et al. (2009). Cerebral blood flow in ischemic vascular dementia and Alzheimer's disease, measured by arterial spin-labeling magnetic resonance imaging. Alzheimers Dement, 5: 454-462 [145] Roman GC (2003). Vascular dementia: distinguishing characteristics, treatment, and prevention. J Am Geriatr Soc, 51: S296-S304 [159] [159] Pase MP, Grima NA, Stough C, Scholey A, Pipingas A (2014). Association of pulsatile and mean cerebral blood flow velocity with age and neuropsychological performance. Physiol Behav, 130: 23-27 [146] Moorhouse P, Rockwood K (2008). Vascular cognitive impairment: current concepts and clinical developments. Lancet Neurol, 7: 246-255 [160] [160] Tarumi T, Ayaz KM, Liu J, Tseng BY, Parker R, Riley J, et al. (2014). Cerebral hemodynamics in normal aging: central artery stiffness, wave reflection, and pressure pulsatility. J Cereb Blood Flow Metab, 34: 971-978 [147] [147] Hanon O, Haulon S, Lenoir H, Seux ML, Rigaud AS, Safar M, et al. (2005). Relationship between arterial stiffness and cognitive function in elderly subjects with complaints of memory loss. Stroke, 36: 2193- 2197 [161] [161] Liu H, Xing A, Wang X, Liu G, Li L (2012). Regulation of beta-amyloid level in the brain of rats with cerebrovascular hypoperfusion. Neurobiol Aging, 33: 826-831 [148] Calik AN, Ozcan KS, Yuksel G, Gungor B, Arugarslan E, Varlibas F, et al. (2014). Altered diastolic function and aortic stiffness in Alzheimer's disease. References Clin Interv Aging, 9: 1115-1121 [162] [162] Niwa K, Porter VA, Kazama K, Cornfield D, Carlson GA, Iadecola C (2001). A beta-peptides enhance vasoconstriction in cerebral circulation. Am J Physiol Heart Circ Physiol, 281: H2417-H2424 [149] [149] Scuteri A, Tesauro M, Appolloni S, Preziosi F, Brancati AM, Volpe M (2007). Arterial stiffness as an independent predictor of longitudinal changes in cognitive function in the older individual. J Hypertens, 25: 1035-1040 [163] [163] Price JM, Chi X, Hellermann G, Sutton ET (2001). Physiological levels of beta-amyloid induce cerebral vessel dysfunction and reduce endothelial nitric oxide production. Neurol Res, 23: 506-512 [150] Scuteri A, Brancati AM, Gianni W, Assisi A, Volpe M (2005). Arterial stiffness is an independent risk factor for cognitive impairment in the elderly: a pilot study. J Hypertens, 23: 1211-1216 [164] [164] Paris D, Humphrey J, Quadros A, Patel N, Crescentini R, Crawford F, et al. (2003). Vasoactive effects of A beta in isolated human cerebrovessels and in a transgenic mouse model of Alzheimer's disease: role of inflammation. Neurol Res, 25: 642-651 [151] Rabkin SW, Jarvie G (2011). Comparison of vascular stiffness in vascular dementia, Alzheimer dementia and cognitive impairment. Blood Press, 20: 274-283 [165] [152] Vendemiale G, Romano AD, Dagostino M, de Matthaeis A, Serviddio G (2013). Endothelial dysfunction associated with mild cognitive impairment in elderly population. Aging Clin Exp Res, 25: 247-255 Kim AS, Johnston SC (2013). Temporal and geographic trends in the global stroke epidemic. Stroke, 44: S123-S125 [166] Laurent S, Katsahian S, Fassot C, Tropeano AI, Gautier I, Laloux B, et al. (2003). Aortic stiffness is an independent predictor of fatal stroke in essential hypertension. Stroke, 34: 1203-1206 [153] [153] Hughes TM, Kuller LH, Barinas-Mitchell EJ, Mackey RH, McDade EM, Klunk WE, et al. (2013). Pulse wave velocity is associated with beta-amyloid deposition in the brains of very elderly adults. Neurology, 81: 1711-1718 [167] Mattace-Raso FU, van der Cammen TJ, Hofman A, van Popele NM, Bos ML, Schalekamp MA, et al. (2006). Arterial stiffness and risk of coronary heart disease and stroke: the Rotterdam Study. Circulation, 113: 657-663 [154] gy [154] Hughes TM, Kuller LH, Barinas-Mitchell EJ, McDade EM, Klunk WE, Cohen AD, et al. (2014). Arterial stiffness and beta-amyloid progression in nondemented elderly adults. Jama Neurol, 71: 562- 568 [168] [168] Willum-Hansen T, Staessen JA, Torp-Pedersen C, Rasmussen S, Thijs L, Ibsen H, et al. (2006). References Maturitas, 53: 447-453 [137] [122] Wu Z, Maric C, Roesch DM, Zheng W, Verbalis JG, Sandberg K (2003). Estrogen regulates adrenal angiotensin AT1 receptors by modulating AT1 receptor translation. Endocrinology, 144: 3251-3261 [137] Greendale GA, Huang MH, Wight RG, Seeman T, Luetters C, Avis NE, et al. (2009). Effects of the menopause transition and hormone use on cognitive performance in midlife women. Neurology, 72: 1850- 1857 p gy [123] Hodis HN, Mack WJ, Henderson VW, Shoupe D, Budoff MJ, Hwang-Levine J, et al. (2016). Vascular Effects of Early versus Late Postmenopausal Treatment with Estradiol. N Engl J Med, 374: 1221- 1231 [138] [138] Franklin SS, Gustin WT, Wong ND, Larson MG, Weber MA, Kannel WB, et al. (1997). Hemodynamic patterns of age-related changes in blood pressure. The Framingham Heart Study. Circulation, 96: 308-315 [139] [124] McCarrey AC, Resnick SM (2015). Postmenopausal hormone therapy and cognition. Horm Behav, 74: 167- 172 [ Franklin SS, Lopez VA, Wong ND, Mitchell GF, Larson MG, Vasan RS, et al. (2009). Single versus combined blood pressure components and risk for cardiovascular disease: the Framingham Heart Study. Circulation, 119: 243-250 [125] Gleason CE, Dowling NM, Wharton W, Manson JE, Miller VM, Atwood CS, et al. (2015). Effects of Hormone Therapy on Cognition and Mood in Recently Postmenopausal Women: Findings from the Randomized, Controlled KEEPS-Cognitive and Affective Study. Plos Med, 12: e1001833, e1001833 [140] James MA, Watt PA, Potter JF, Thurston H, Swales JD (1995). Pulse pressure and resistance artery structure in the elderly. Hypertension, 26: 301-306 [141] Fernando MS, Simpson JE, Matthews F, Brayne C, Lewis CE, Barber R, et al. (2006). White matter lesions in an unselected cohort of the elderly: molecular pathology suggests origin from chronic hypoperfusion injury. Stroke, 37: 1391-1398 [126] Robb AO, Mills NL, Din JN, Smith IB, Paterson F, Newby DE, et al. (2009). Influence of the menstrual cycle, pregnancy, and preeclampsia on arterial stiffness. Hypertension, 53: 952-958 [127] [142] [127] Williams MR, Westerman RA, Kingwell BA, Paige J, Blombery PA, Sudhir K, et al. (2001). Variations in endothelial function and arterial compliance during the menstrual cycle. J Clin Endocrinol Metab, 86: [142] Okamoto Y, Yamamoto T, Kalaria RN, Senzaki H, Maki T, Hase Y, et al. (2012). Cerebral hypoperfusion accelerates cerebral amyloid angiopathy and promotes cortical microinfarcts. [165] Aging and Disease • Volume 8, Number 4, August 2017 [167] References J Vasc Interv Neurol, 5: 33-39 [188] [188] Poels MM, Zaccai K, Verwoert GC, Vernooij MW, Hofman A, van der Lugt A, et al. (2012). Arterial stiffness and cerebral small vessel disease: the Rotterdam Scan Study. Stroke, 43: 2637-2642 [174] Chen PL, Wang PY, Sheu WH, Chen YT, Ho YP, Hu HH, et al. (2006). Changes of brachial flow-mediated vasodilation in different ischemic stroke subtypes. Neurology, 67: 1056-1058 [189] [189] van Elderen SG, Brandts A, Westenberg JJ, van der Grond J, Tamsma JT, van Buchem MA, et al. (2010). Aortic stiffness is associated with cardiac function and cerebral small vessel disease in patients with type 1 diabetes mellitus: assessment by magnetic resonance imaging. Eur Radiol, 20: 1132-1138 [175] Knottnerus IL, Ten CH, Lodder J, Kessels F, van Oostenbrugge RJ (2009). Endothelial dysfunction in lacunar stroke: a systematic review. Cerebrovasc Dis, 27: 519-526 [176] Hassan A, Hunt BJ, O'Sullivan M, Parmar K, Bamford JM, Briley D, et al. (2003). Markers of endothelial dysfunction in lacunar infarction and ischaemic leukoaraiosis. Brain, 126: 424-432 [1 [190] [190] Brisset M, Boutouyrie P, Pico F, Zhu Y, Zureik M, Schilling S, et al. (2013). Large-vessel correlates of cerebral small-vessel disease. Neurology, 80: 662-669 [191] ] Hashimoto J, Aikawa T, Imai Y (2008). Large artery stiffening as a link between cerebral lacunar infarction and renal albuminuria. Am J Hypertens, 21: 1304- 1309 [177] Hassan A, Gormley K, O'Sullivan M, Knight J, Sham P, Vallance P, et al. (2004). Endothelial nitric oxide gene haplotypes and risk of cerebral small-vessel disease. Stroke, 35: 654-659 [191] [178] [192] Ochi N, Kohara K, Tabara Y, Nagai T, Kido T, Uetani E, et al. (2010). Association of central systolic blood pressure with intracerebral small vessel disease in Japanese. Am J Hypertens, 23: 889-894 [178] Wiseman S, Marlborough F, Doubal F, Webb DJ, Wardlaw J (2014). Blood markers of coagulation, fibrinolysis, endothelial dysfunction and inflammation in lacunar stroke versus non-lacunar stroke and non- stroke: systematic review and meta-analysis. Cerebrovasc Dis, 37: 64-75 [ [ [193] [193] Mitchell GF, van Buchem MA, Sigurdsson S, Gotal JD, Jonsdottir MK, Kjartansson O, et al. (2011). Arterial stiffness, pressure and flow pulsatility and brain structure and function: the Age, Gene/Environment Susceptibility--Reykjavik study. Brain, 134: 3398-3407 [179] [179] Deplanque D, Lavallee PC, Labreuche J, Gongora- Rivera F, Jaramillo A, Brenner D, et al. (2013). Cerebral and extracerebral vasoreactivity in symptomatic lacunar stroke patients: a case-control study. References Prognostic value of aortic pulse wave velocity as index of arterial stiffness in the general population. Circulation, 113: 664-670 [155] [155] Qiu C, Winblad B, Viitanen M, Fratiglioni L (2003). Pulse pressure and risk of Alzheimer disease in persons aged 75 years and older: a community-based, longitudinal study. Stroke, 34: 594-599 [169] van Sloten TT, Sedaghat S, Laurent S, London GM, Pannier B, Ikram MA, et al. (2015). Carotid stiffness is associated with incident stroke: a systematic review and individual participant data meta-analysis. J Am [156] [156] Nation DA, Edland SD, Bondi MW, Salmon DP, Delano-Wood L, Peskind ER, et al. (2013). Pulse Aging and Disease • Volume 8, Number 4, August 2017 503 Xu X., et al Vascular aging: mechanisms and disease Coll Cardiol, 66: 2116-2125 Functions in Older Parkinson's Disease Patients: a Cross-Sectional Study. Aging Dis, 7: 28-35 [170] [170] Gasecki D, Rojek A, Kwarciany M, Kubach M, Boutouyrie P, Nyka W, et al. (2012). Aortic stiffness predicts functional outcome in patients after ischemic stroke. Stroke, 43: 543-544 [184] Kuo HK, Chen CY, Liu HM, Yen CJ, Chang KJ, Chang CC, et al. (2010). Metabolic risks, white matter hyperintensities, and arterial stiffness in high- functioning healthy adults. Int J Cardiol, 143: 184-191 [171] [171] Calvet D, Touze E, Laurent S, Varenne O, Sablayrolles JL, Boutouyrie P, et al. (2014). Aortic stiffness measurement improves the prediction of asymptomatic coronary artery disease in stroke/transient ischemic attack patients. Int J Stroke, 9: 291-296 [185] [185] Hatanaka R, Obara T, Watabe D, Ishikawa T, Kondo T, Ishikura K, et al. (2011). Association of arterial stiffness with silent cerebrovascular lesions: the Ohasama study. Cerebrovasc Dis, 31: 329-337 [186] Matsumoto M, Inoue K, Moriki A (2007). Associations of brachial-ankle pulse wave velocity and carotid atherosclerotic lesions with silent cerebral lesions. Hypertens Res, 30: 767-773 [172] Seo WK, Lee JM, Park MH, Park KW, Lee DH (2008). Cerebral microbleeds are independently associated with arterial stiffness in stroke patients. Cerebrovasc Dis, 26: 618-623 [187] [187] Ochi N, Tabara Y, Igase M, Nagai T, Kido T, Miki T, et al. (2009). Silent cerebral microbleeds associated with arterial stiffness in an apparently healthy subject. Hypertens Res, 32: 255-260 [173] Blum A, Vaispapir V, Keinan-Boker L, Soboh S, Yehuda H, Tamir S (2012). Endothelial dysfunction and procoagulant activity in acute ischemic stroke. References Int J Stroke, 8: 413-421 [194] [194] Tsao CW, Seshadri S, Beiser AS, Westwood AJ, Decarli C, Au R, et al. (2013). Relations of arterial stiffness and endothelial function to brain aging in the community. Neurology, 81: 984-991 [180] [180] Khan U, Porteous L, Hassan A, Markus HS (2007). Risk factor profile of cerebral small vessel disease and its subtypes. J Neurol Neurosurg Psychiatry, 78: 702- 706 D s c [195] J [195] [195] Jochemsen HM, Muller M, Bots ML, Scheltens P, Vincken KL, Mali WP, et al. (2015). Arterial stiffness and progression of structural brain changes: The SMART-MR study. Neurology, 84: 448-455 [181] [181] Wardlaw JM, Smith EE, Biessels GJ, Cordonnier C, Fazekas F, Frayne R, et al. (2013). Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol, 12: 822-838 [19 [196] [196] Marshall RS (2012). Effects of altered cerebral hemodynamics on cognitive function. J Alzheimers Dis, 32: 633-642 [182] [182] Kalaria RN (2002). Small vessel disease and Alzheimer's dementia: pathological considerations. Cerebrovasc Dis, 13 Suppl 2: 48-52 Jellinger KA (2013). Pathology and pathogenesis of vascular cognitive impairment-a critical update. Front Aging Neurosci, 5: 17 [183] Yalcin A, Atmis V, Cengiz OK, Cinar E, Aras S, Varli M, et al. (2016). Evaluation of Cardiac Autonomic Aging N [198] Webb A M, et al. (2016). Evaluation of Cardiac Autonomic [198] Webb AJ, Simoni M, Mazzucco S, Kuker W, Schulz Webb AJ, Simoni M, Mazzucco S, Kuker W, Schulz Aging and Disease • Volume 8, Number 4, August 2017 504 Xu X., et al Vascular aging: mechanisms and disease cerebrovascular aging. Am J Physiol Heart Circ Physiol, 305: H620-H633 U, Rothwell PM (2012). Increased cerebral arterial pulsatility in patients with leukoaraiosis: arterial stiffness enhances transmission of aortic pulsatility. Stroke, 43: 2631-2636 U, Rothwell PM (2012). Increased cerebral arterial pulsatility in patients with leukoaraiosis: arterial stiffness enhances transmission of aortic pulsatility. Stroke, 43: 2631-2636 [207] Viboolvorakul S, Patumraj S (2014). Exercise training could improve age-related changes in cerebral blood flow and capillary vascularity through the upregulation of VEGF and eNOS. Biomed Res Int, 2014: 230791 [199] 199] Ding J, Mitchell GF, Bots ML, Sigurdsson S, Harris TB, Garcia M, et al. (2015). Carotid arterial stiffness and risk of incident cerebral microbleeds in older people: the Age, Gene/Environment Susceptibility (AGES)-Reykjavik study. References Arterioscler Thromb Vasc Biol, 35: 1889-1895 [208] Rippe C, Lesniewski L, Connell M, LaRocca T, Donato A, Seals D (2010). Short-term calorie restriction reverses vascular endothelial dysfunction in old mice by increasing nitric oxide and reducing oxidative stress. Aging Cell, 9: 304-312 [200] [200] Akinyemi RO, Mukaetova-Ladinska EB, Attems J, Ihara M, Kalaria RN (2013). Vascular risk factors and neurodegeneration in ageing related dementias: Alzheimer's disease and vascular dementia. Curr Alzheimer Res, 10: 642-653 [209] [209] Meyer TE, Kovacs SJ, Ehsani AA, Klein S, Holloszy JO, Fontana L (2006). Long-term caloric restriction ameliorates the decline in diastolic function in humans. J Am Coll Cardiol, 47: 398-402 [201] [201] Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, et al. (2006). Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J, 27: 2588-2605 [210] Laufs U, La Fata V, Plutzky J, Liao JK (1998). Upregulation of endothelial nitric oxide synthase by HMG CoA reductase inhibitors. Circulation, 97: 1129- 1135 [211] [202] [202] Yu WC, Chuang SY, Lin YP, Chen CH (2008). Brachial-ankle vs carotid-femoral pulse wave velocity as a determinant of cardiovascular structure and function. J Hum Hypertens, 22: 24-31 [211] Matsuo T, Iwade K, Hirata N, Yamashita M, Ikegami H, Tanaka N, et al. (2005). Improvement of arterial stiffness by the antioxidant and anti-inflammatory effects of short-term statin therapy in patients with hypercholesterolemia. Heart Vessels, 20: 8-12 [203] [203] Wentland AL, Grist TM, Wieben O (2014). Review of MRI-based measurements of pulse wave velocity: a biomarker of arterial stiffness. Cardiovasc Diagn Ther, 4: 193-206 [212] Lahera V, Goicoechea M, de Vinuesa SG, Miana M, de Las HN, Cachofeiro V, et al. (2007). Endothelial dysfunction, oxidative stress and inflammation in atherosclerosis: beneficial effects of statins. Curr Med Chem, 14: 243-248 [204] Kuvin JT, Patel AR, Sliney KA, Pandian NG, Sheffy J, Schnall RP, et al. (2003). Assessment of peripheral vascular endothelial function with finger arterial pulse wave amplitude. Am Heart J, 146: 168-174 [213] [213] Mahmud A, Feely J (2004). Arterial stiffness and the renin-angiotensin-aldosterone system. J Renin Angiotensin Aldosterone Syst, 5: 102-108 [205] [205] Bullitt E, Rahman FN, Smith JK, Kim E, Zeng D, Katz LM, et al. (2009). The effect of exercise on the cerebral vasculature of healthy aged subjects as visualized by MR angiography. Aging and Disease • Volume 8, Number 4, August 2017 [214] [210] References AJNR Am J Neuroradiol, 30: 1857-1863 Ito T, Yamakawa H, Bregonzio C, Terron JA, Falcon- Neri A, Saavedra JM (2002). Protection against ischemia and improvement of cerebral blood flow in genetically hypertensive rats by chronic pretreatment with an angiotensin II AT1 antagonist. Stroke, 33: 2297-2303 [206] [206] Bolduc V, Thorin-Trescases N, Thorin E (2013). Endothelium-dependent control of cerebrovascular functions through age: exercise for healthy Aging and Disease • Volume 8, Number 4, August 2017 505
https://openalex.org/W2169669898	https://europepmc.org/articles/pmc4825486?pdf=render	English	null	Ranking the Effects of Urban Development Projects on Social Determinants of Health: Health Impact Assessment	Global journal of health science	2,014	cc-by	7,140	Global Journal of Health Science; Vol. 6, No. 5; 2014 ISSN 1916-9736 E-ISSN 1916-9744 Published by Canadian Center of Science and Education Global Journal of Health Science; Vol. 6, No. 5; 2014 ISSN 1916-9736 E-ISSN 1916-9744 Published by Canadian Center of Science and Education Ranking the Effects of Urban Development Projects on Social Determinants of Health: Health Impact Assessment Parisa Shojaei1, Masoud Karimlou1, Jafar Nouri2, Farahnaz Mohammadi 1, Hosein Malek Afzali2 & Ameneh Setareh Forouzan1 Parisa Shojaei1, Masoud Karimlou1, Jafar Nouri2, Farahnaz Mohammadi 1, Hosein Malek Afzali2 & Ameneh Setareh Forouzan1 2 The School of Public Health, Tehran University of Medical Sciences, Tehran, Iran Correspondence: Masoud Karimlou, PhD of Biostatistics, Associated Professor, Social Determinant of Health Research Center, University of Social Welfare & Rehabilitation Sciences, Tehran, Iran. Tel: 98-21-22-180-146. E-mail: mkarimlo@yahoo.com Received: April 5, 2014 Accepted: May 8, 2014 Online Published: May 30, 2014 doi:10.5539/gjhs.v6n5p183 URL: http://dx.doi.org/10.5539/gjhs.v6n5p183 Received: April 5, 2014 Accepted: May 8, 2014 Online Published: May 30, 2014 doi:10.5539/gjhs.v6n5p183 URL: http://dx.doi.org/10.5539/gjhs.v6n5p183 Abstract Background and Objective: Health impact assessment (HIA) offer a very logical and interesting approach for those aiming to integrate health issues into planning processes. With a lot of works and plans waiting to be done (e.g., developing and updating plans, counseling planning commissions, cooperation with other organizations), planners find it difficult to prioritize health among a variety of possible issues and solutions they confront. Method: In the present article, first, the list of social determinants of health associated with Chitgar man-made lake was extracted out using a qualitative method and with content analysis approach, and then they were prioritized using analytic hierarchy process. Results: 28 social determinants of health including “intermediary” and “structural” determinants were extracted out. Regarding positive effects of lake on these determinants, “recreational services” and “traffic” received the highest and the lowest weights with 0.895 and 0.638 respectively among structural determinants and with consideration to “construction” option. Furthermore, among intermediary determinants for “construction” option, sub-criteria of both “physical activity” and “air quality” received the final highest weight (0.889) and “pathogenesis” indicated the lowest weight with 0.617. Moreover, lake demonstrated the highest negative effects on “housing” among “structural” determinants which it takes the highest weight (0.476) in “non-construction” option. Additionally, lake had the highest negative effects on “noise pollution” among “intermediary determinants” and it takes the highest weight (0.467) in “non-construction” option. Conclusion: It has been shown that urban development projects such as green spaces, man-made lakes … have a huge range of effects on community’s health, and having not considered these effects by urban planners and mangers is going to confront urban health with many challenges. Keywords: Analytic Hierarchy Process, social determinants of health, man-made lake, health impact assessment 1. Introduction 1. Introduction In most of the cases, urban planning entities are not the only elements or even, particularly, the main responsible institutions for health-related goals in the urban planning process, but the cooperation between these entities with planning process is quite necessary. There are some other social, economic, and environmental organizations involved in this process. For such conditions, we need a cooperative approach, i.e. a common activity towards reaching pre-agreed goals. This cooperation in urban planning gets both approaches together in order to have consciously decisions and efforts for society interference. Some part of this process requires identifying interest groups in the society who owns the benefits of adversaries in the decisions (Barati, 2011). The meeting point between health and urban planning has engrossed much interest so far. Practitioners are looking for clear-cut, systematic, and wide-ranging means to incorporate health concerns into planning processes. There is a substantial discussion about such matters, and a wide range of literature points to associations between the two disciplines. While the practice is rhetorically long, it is unluckily short when it comes to execution. However, perfectly and entirely accounting for health in plans and policies remains a key challenge. Health, however, is a 183 Global Journal of Health Science Vol. 6, No. 5; 2014 www.ccsenet.org/gjhs vital issue in urban areas, and as the world goes towards urbanization, the link between cities and health will become gradually more important. HIAs are one means to help these challenges out, providing a rubric to bring health into planning. They can assist planners to systematically consider a variety of health concerns, serving as a tool to evaluate plans, policies, and development proposals. However, considering the current status of HIAs, there is still substantial uncertainty about how or when they should be used and what they say in the end. In many cases, the applicability of planning does not immediately become obvious. The urban planning community requires clear-cut yet comprehensive tools that help planners go forward through this cause in a way that is also nonthreatening and easy to utilize (“Design for Health. Building public understanding: The link between health and planning,” 2007). An HIA is a systematic appraisal that amalgams scientific data, professional proficiency, and stakeholder participation to find out the effects that a potential policy, plan, program, or project might have on the health of a particular population. 1. Introduction HIAs offer information to decision-makers that can help minimize the expected adverse health effects and maximize positive health outcomes. It can be used for a variety number of sectors and can be employed at the federal, state, tribal, and local levels (“National Research Council of the National Academies (NRCNA). Improving Health in the United States: The Role of Health Impact Assessment.”, 2011). There exists a wide range of studies about evaluation of health effects of urban development projects which most of them are about green spaces such as parks, and a few have particularly focused on urban man-made lakes. As defined by Green Space Scotland (2008), the term “green space” can be applied to any vegetated land or surface water body within or adjacent to an urban area, including natural and semi-natural habitats; countryside immediately adjoining a town which people can access; green corridors—paths, rivers and canals; amenity grassland, parks and gardens; outdoor sports facilities, and playing fields (Greenspace Scotland, 2008). Therefore, the man-made lakes are green space and thus have the effects of a green space on heath.In a study carried out by Santana et al namely “evaluation of health effects of walkable urban green spaces”, they stated that these spaces have both direct and indirect impacts on health, and not only are connected with good health condition of habitats but also with environmental quality improvement(Santana, Santos, & Costa, 2009)Results of using HIA in urban planning process of Manukau city in Tailand indicated that it reinforces the intersection, especially structural one, between city council, interested people, and organizations out of the council which we do not see much of such cooperation at the time of urban planning (Field, September 2011). Therefore, considering health issues in urban plans and projects is really important. Creating man-made lakes in developing countries is followed by a lot of dangers. Stakeholders need to actively integrate in face-to-face cooperation in order to fully utilize, improve, and maintain water sources and lakes projects. As it looks, it is essential to have health professionals especially biologists, sociologists, and economists during the planning and implementation processes of development projects besides engineers (Araoye, 2002). Constructing the biggest man-made lake in north west of Tehran in district 22 is one of these under-construction projects in Iran. A-Qualitative Method: A-Qualitative Method: This qualitative study was carried out with a content analysis approach. Data were collected by individual interviews as well as focus group discussions (FGDs) from September to March 2012 respectively. Individual interviews were done on 30 informants in three groups: 1- people living in the 22 District, 2-Members involved in the man made lake project from the Municipal District 22 with following characteristics: educated individuals with at least B.A degree, experts on urban development, water sources, environment, sociology and urban planning, civil engineering; managers and administrative-executive experts in region 22 of municipality of Tehran, experts of Civil engineering consulting firms with at least one-year work experience, lake surroundings residents, and local authorities( managers of local health houses) and 3-third group to avoid information bias include Members from outside the municipal District 22 and experts in the field of health determinants and environment with at least a bachelor's degree. The focus group discussions were carried out with 2 experts group that participated in individual interviews. Sampling was conducted using purposive in this study and until data saturation and new ideas and insights was driven. The focus group discussion have ability to explore people’s ideas, worries, attitudes and experiences of individuals regarding a specific subject matter (Barbour & Kitzinger, 1998; Dahlgren, Emmelin, & Winkvist, 2007). Individual interviews lasted for 50 minutes. For implementation of Interviews were used of focus group discussion guiding questionnaires and in-depth individual interviews. These procedures were designed based on the research team’s perspectives and experts who were familiar with health and urban development issues as well as library sources and objectives of the study. The concepts discussed in this questionnaire includes various health dimensions of man made lake construction. Validity of questionnaires was conducted through a pilot study, and necessary modifications in content, sequence and timing done. We arranged sessions with municipality coordination to make them aware of the study and asked for any permission required. In addition, at the beginning of interview informants were justified about the whole research, from its objectives to the methods and confidentiality of the data, and all agreed to take part voluntarily. All interviews were guided by the principal investigator. Focus Group discussion was held with the coordinating investigator for one and a half hours on Location with Participants agreed in the office of the honourable Deputy Mayor of Tehran 22. 1. Introduction Chitgar man made lake, located in the urban District-22 Municipality of Tehran northwest of Tehran, Iran is by volume l0 Million Cubic Meter (MCM), by depth average 10 meters (m.) and by area 225 hectares (ha) as the Iran’s largest man-made lake. Approximately 80% of its water is supplied by Can river and the rest by run-off of Intermediate and surface area of the region (Nouri, Rahimipour, & Nezakati, 2004). History of construction the lake project goes back to develop the first comprehensive plan of Tehran city in 1986 which anticipated construction a lake in west of Tehran. But construction of the lake due to technical restrictions and budget remain dormant for years. Finally, in the background studies, since 1382-1389, by the consultant detailed studies were done at different times and uncertainties survey was completed. on October 2010, Catchment area of operations and on July 2012 coastal zones operations were began by the supervision of Sabir Company according to study and design of Tunnel Pars engineering company (Eshtuky Pars Company). The first phase of the project opened named Martyrs Persian Gulf on May 15th 2013 (Chitgar man made lake, 2013). It includes three Islands as well as green, recreational, playing and sport spaces. Although the objective of constructing man-made lakes is to create areas for attracting tourists, improve recreational capacity of area, and bring people joy and happiness morale, not considering environmental and health issues in designing, implementation, and utilization of such areas can have negative environmental and health consequences. With consideration to what mentioned above and a need to evaluate health effects of urban development projects, a study namely “Health impact assessment of constructing man-made lake in city development strategy on health and social determinants of health” is under implementation. The current article is taken from the “identification” step of above study which aims to identify and prioritize social determinants of health in the existing study. Based on the health impact assessment guidelines, multi-criteria decision-making model was used as a quantitative method in the present article (Harris, Harris-Roxas, Harris, & Kemp, 2007). Additionally, the Analytic Hierarchy Process (AHP) is a structured method for organizing and analyzing complex decisions, based on mathematics and psychology. It was 184 Global Journal of Health Science Vol. 6, No. 1. Introduction 5; 2014 www.ccsenet.org/gjhs provides the treatment for complicated problems with multiple criteria, stakeholders, and decision makers in a scenario with a high uncertainty and high risk. It offers a good compromise between the target, understanding, and objectivity to the extent that is a tool supported by the basic mathematics letting ordinary people arrange tangible and intangible factors in a process of conflict resolution or order of priorities (Parra-Lopez, Calatrava-Requena, & de-Haro-Gimenez, 2007). Despite the employment of specialists for the realization of judgments, the AHP provides the likelihood that such judgments may remain inconsistent. In this sense, the AHP allows to assess the consistency and to find out the degree of inconsistency in a matrix of parity judgment (Thomas L Saaty, 2005). At this point, it is significant to consider that AHP is a technique for selecting the best alternative that integrates qualitative concerns and quantitative factors to the subjective process of decision-making, i.e., it lets the decision maker deal with the intuitive, rational, and irrational aspects simultaneously. This model has four stages ranging from creation of the decision problem, measurement and data collection, establishment of standard weights to presentation of summaries of findings solutions to the problem (Rosa & Haddad, 2013). 2. Methods This study was two step that step A is Qualitative method for Extracting social determinants associated with artificial lakes and step B was weighting and prioritizing of its determinants. B: AHP and Matrix Construction In this step, after analyzing content analysis methodology by the help of OpenCode software, list of all health determinants associated with man-made lake was extracted out. AHP is a structured method for organizing and analyzing complex decisions, based on mathematics and psychology. It was developed by Thomas L. Saaty in the 1970s (T.L Saaty, 1980). Its indispensable feature is that it changes individual preferences into ratio scale weights so providing the opportunity for effective comparison and ranking of the decision aspects. It has been utilized in various areas ranging from key factors of urbanization (Thapa & Murayama, 2010), fire risk analysis, engineering decision making, project selection, evaluation and management, vendor selection, housing sector, banking to marketing .The key features of AHP include hierarchical structure of complexity, pair-wise comparisons, unneeded judgments, an eigenvector method for providing weights and steadiness considerations (Bhatta & Doppler, 2010). As owned by many other methods, AHP provides opportunities for decision makers to build a model of a complicated problem with the goal at the top and criteria, sub-criteria (factors) and alternatives at levels in drop-down manner(Thomas L Saaty, 2008). Figure 1 shows the general framework of AHP used in the study. Since seven criteria were chosen in order to find their impact on the predominance of a particular farm practice, the dimension of the matrix, therefore, is 7x7. In this sense, one column and one row correspond to each factor and there were 21 pair-wise comparisons using the formula {n(n-1)/2}. If this matrix is denoted as A={aij} where aij is the element of ith row and jth column of the matrix, all it’s entries are obtained by marking the relative importance of each criterion over another with consideration to the goal. Pair-wise comparison can be done by adopting an integer ranging from 1 to 9 or the mutual of such an integer to each cell of the matrix to gauge the relative importance of the factors that characterize the cell (T. L. Saaty, 1980). According to this scale, the available values for the pair-wise comparisons are members of the set: {9, 8, 7, 6, 5, 4, 3, 2, 1, 1/2, 1/3, 1/4, 1/5, 1/6, 1/7, 1/8, 1/9} (see also Table 1). According to the AHP approach assumption, each of the factors under assessment is independent of another. It makes some small variation in the judgment since human responses are not always constant. A-Qualitative Method: In meetings of individual interviews and focus groups were used of public and comprehensive questions and it progressed to detailed questions by the time. Interviews and group discussions were recorded and verbatim was rewritten. We used Graneheim approach for data analysis (Graneheim & Lundman, 2004). Therefore rewritten texts several times reread and meaning units was identified, and were coded finally. Next, codes based on the meaning units represented by informants and their similarities and discrepancies were extracted and grouped, then from these groupings, themes were identified. At the beginning of the data analysis, according to the list of common social determinants of and core categories, data analysis was conducted using content analysis matrix (Averill, 2002). Then, with analysis continued, new categories were extracted as and they were put in the body of the total data analysis. In order to increase reliability all codes and themes were checked mutually by research team, and its summary was given to the informants in the end. The objectivity of the data, coded by two researchers were also and codes and categories were compared together. To trustworthiness of findings, data were collected based on the age, gender, and education of each informant separately. Additionally, other methods for assuring data trustworthiness include: Combining the methods of data collection and 185 Global Journal of Health Science Vol. 6, No. 5; 2014 www.ccsenet.org/gjhs individual interviews as well as FGDs with each other, long-standing relationship with participants and research topic, extensive and detailed research report writing, reviews by participants and observers. In order to uphold the ethics of research, these points were met: Written and informed consent, permission to collect and record the data, maintaining the anonymity of the participants, the right to withdraw from the study, send results to all stakeholders and obtaining approval of the ethics committee. B: AHP and Matrix Construction In practice, we can hardly achieve 100% consistency, but the method can still be employed when there is some degree of interdependence. In order to calculate the index of inconsistency, AHP makes use of consistency ratio (CR). Values of inconsistency index lower than 10% are acceptable, especially if matrix is 4 by 4 or above. In some cases, although higher value of inconsistency index asks for re-evaluation of pair-wise comparisons, decisions acquired in such circumstances could also be taken as ‘the best alternative’. Consistency index (CI) is calculated as (Yu, Wang, & Gong, 2013): At the beginning, a hierarchical tree with its aim, criterion, sub-criterion, and options was formed for executing hierarchical analysis. Then all determinants extracted from qualitative method were put in a checklist in the form of pair-wise comparisons matrix in order to weigh them. In the next step, weights were assigned to each of them by the same participants participated in interviews (experts of second group and third group) using Saaty method. In order to facilitate the completion of this checklist by participants, it was completed once by having a positive approach on the effect of lake on determinants and again with consideration to its negative effects on health determinants. In this way, it might be possible to understand that which determinant is affected the most by the positive effects and the negative effects of lake construction. Indeed, everyone completed both checklists thereby helping us to have their both negative and positive points of view. Table 1. Scale of relative importances (according to Saaty (1980)) Intensity of Importance 1 3 5 7 9 2,4,6,8 Definition Equal importance Weak importance of one over another Essential or strong importance Demonstrated importance Absolute importance Intermediate values between the two adjacent judgments After having checklists completed, comparative tables of each respondent were combined together using Table 1. Scale of relative importances (according to Saaty (1980)) Intensity of Importance 1 3 5 7 9 2,4,6,8 Definition Equal importance Weak importance of one over another Essential or strong importance Demonstrated importance Absolute importance Intermediate values between the two adjacent judgments After having checklists completed, comparative tables of each respondent were combined together using 186 Global Journal of Health Science www.ccsenet.org/gjhs Vol. 6, No. 5; 2014 geometric mean. Since pair-wise scales create data in the form of ratio, therefore, geometric mean seemed suitable among other methods. B: AHP and Matrix Construction Since the interview subjects that completed this matrix were two groups of experts, so after completing paired matrix, a geometric mean was calculated for each group. Then the average of the two groups were evaluated by a panel of experts on health determinants. Finally, the review panel was formed matrix showed that the overall mean And the matrix for analysis and measurement of weights as raw data was used into the Expert choice software. Figure 1. Different criteria and sub-criteria under investigation and their hierarchy Figure 1. Different criteria and sub-criteria under investigation and their hierarchy 3. Results This study was conducted in district 22 of Tehran municipality in 2012. Mean age of participants was 30, and they were experts in urban development, environmental, water source, civil engineering, sociology, urbanization, and health determinants areas. Their education ranged from B.S to P.H.D, and four of them were indigenous to the region and the others were from other parts of the city. Qualitative analysis of data provided us with 261 codes. Furthermore, conceptual framework structure developed by commission on social determinants of health along with their sub-groups was considered as data analysis matrix (O. Solar & A. Irwin, 2010). Thus, two “structural” and “intermediary” groups were formed, and 14 sub-groups were determined in each stratum. Then these 28 determinants were put in a checklist and were given to participants through focus group discussion and interviews in order to implement AHP process. Therefore, we had one goal, two criteria, and 28 sub-criteria as well as two options (construction and non-construction of man-made lake) for depicting their hierarchy. Findings regarding positive effects of lake construction on social determinants of health: According to the findings, the maximum relative weight among structural determinants was calculated for income and tourism. 187 www.ccsenet.org/gjhs Global Journal of Health Science Vol. 6, No. 5; 2014 Moreover, “air quality” and “physical and social development of children’ had the highest weight among intermediary determinants. Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs Moreover, “air quality” and “physical and social development of children’ had the highest weight among intermediary determinants. 188 Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs As seen in Table 2, regarding “construction” option, “recreational services” and “traffic” received the highest and the lowest weights with 0.895 and 0.638 respectively. These criteria, for sure, had the lowest and the highest weight for “non-construction” option with 0.105 and 0.362 respectively. About intermediary determinants for “construction” option, sub-criteria of both “physical activity” and “air quality” received the final highest weight (0.889) and sub-criterion of “pathogenesis” indicated the lowest weight with (0.617). For “non-construction” option, the above sub-criteria had the lowest and the highest final weights with (0.111) and (0.375) respectively. 189 Table 2. 3. Results Weightings of assessment criteria and sub-criteria in positive impact Criterion/ Weight Sub criteria Options Final Weight Construction Non Construction Structural Determinants Construction: 0.636 Non Construction: 0.364 Income 0.889 0.111 Culture 0.857 0.143 Education 0.873 0.127 Job 0.833 0.167 Limitation of available resources 0.750 0.250 Traffic 0.638 0.362* Ethnicity 0.804 0.196 Tourism 0.889 0.111 Region Popularity 0.891 0.109 Recreational Services 0.895* 0.105* political participation 0.863 0.132 Spirituality 0.860 0.140 Housing 0.667 0.333 Social Security 0.681 0.319 Intermediary determinants Construction: 0.602 Non Construction: 0.398 Physical and Social Development of Children 0.877 0.123 Physical Activity 0.889* 0.111* Quality of Life 0.857 0.143 Nutrition 0.783 0.217 Unhealthy Behavior 0.750 0.250 Unintentional Injuries 0.786 0.214 Pathogenesis 0.617* 0.333 Air Quality 0.889* 0.111* Mental Health 0.869 0.131 Aesthetic 0.857 0.143 Social Capital 0.872 0.128 Weather Changes 0.836 0.164 Noise Pollution 0.667 0.333 Environmental Health 0.625 0.375 Table 2. Weightings of assessment criteria and sub-criteria in positive impact 189 Global Journal of Health Science Vol. 6, No. 5; 2014 www.ccsenet.org/gjhs Findings regarding negative effects of lake construction on social determinants of health: Findings regarding negative effects of lake construction on social determinants of health: Findings regarding negative effects of lake construction on social determinants of health: Findings regarding negative effects of lake construction on social determinants of health: Findings indicated that, among structural determinants, “traffic” and “housing” had higher weight compared to other elements. Findings indicated that, among structural determinants, “traffic” and “housing” had higher weight compared to other elements. Moreover, “air pollution” and “pathogenesis” received relatively the highest weights among intermediary determinants. Moreover, “air pollution” and “pathogenesis” received relatively the highest weights among intermediary determinants. 190 Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs Final weights of criteria and sun-criteria for two “construction” and “non-construction” options are shown in table 3 with regard to their negative effects. As stated, lake has the highest negative effects on “housing” among structural determinants which it takes the highest weight (0.476) in “non-construction” option. Additionally, it has the highest negative effects on “noise pollution” among intermediary determinants and it takes the highest weight (0.467) in “non-construction” option. Table 3. 4. Discussion and Conclusion As broadcast in the findings, “recreational services” and “traffic” demonstrated the highest and the lowest final weights with regard to positive effects of lake on social determinants of health. It means that with consideration to sub-criterion of “recreational services,” “construction” option should be selected. Furthermore, as participants declared their points in qualitative study, Chitgar man-made lake and green spaces surrounding it provide a wide range of recreational facilities for people living there. King County’s Open Space System (2004) specified that everyone enjoy recreational activities. Parks, recreational and open spaces “provide places to exercise, take part in competitive sports, socialize with others, and a space for people to stand far from development and experience the natural environment,” which all play roles in physical, mental, and emotional health of people. Recreation paves the ways for us to learn, explore, and challenge things that contribute to human growth. Moreover, cultural and historic places play roles in shaping our sense of community identity and relationship with others (King County’s Open Space System: Parks, Trails, Natural Areas, and Working Resource Lands, 2004). A community’s park system can offer passive and active recreational chances close to home for a variety of residents and tourists. Neighborhood and community parks meet urgent urban area needs (Enger, 2005). Urban regeneration is defined as creating sustainable places, such as green spaces. These places should offer high-quality recreational services for urban residents (Arnberger, 2012). With consideration to “traffic” sub-criterion, “non-construction” option would be the priority. According to the participants, lake construction brings in a lot of visitors and tourists and will increase the traffic in the region which will have negative impacts on people’s health. Findings of researches done by South Carolina Institute of Medicine and Public Health (2013) demonstrated that the potential creation of a park could bring in visitors from outside the community, which could raise the traffic and reduce safety for pedestrians and cyclists. Crashes between motor vehicles and pedestrians or cyclists can cause injury or death. These incidents of injuries happen more often in areas with large amounts of pedestrians and motor vehicles, and children and elderly are most likely the victims of the related accidents (South Carolina Institute of Medicine and Public Health (IMPH). A Health Impact Assessment (HIA) of Park, Trail, and Green Space Planning in the West Side of Greenville, South Carolina, 2013). 4. Discussion and Conclusion In addition, from the aspect of positive effects of lake on intermediary determinants, findings indicated that “physical activity” and “quality of air” received the highest weights among these elements. The lake and surrounding green spaces with recreational areas and sports fields - for hiking, rafting, ... provides the opportunity for individuals to exercise, do physical activity and lessens obesity and its related diseases. Findings of an urban heart study in 2012 in Tehran showed that the percentage of overweight and obese men and women was high in one of the neighborhoods near the lake in District 22 of Tehran which the lake might have effects on decreasing those. Adjustment of the built environment makes green spaces augment chances for advantageous ‘green exercise’ such as walking (J. Pretty, M. Griffin, M. Sellens et al., 2003). According to Godbey et al., there exists possible spare-time physical activity in a variety of community environments such as local parks (Godbey & AJ, 2003). Several studies go with this attitude that ‘the built environment can facilitate or constrain physical activity (Bedimo-Rung, Mowen, & Cohen, 2005)’. In a European study carried out in eight cities, high levels of greenery caused people to be physically active three times more than the others and the chance of being overweight and obese was about 40% lower for these people than for those living in similar areas with lower levels of greenery. On the other hand, people living in areas with high levels of anti-social activities were less likely to be physically active and more likely to be overweight or obese (Ellaway, Macintyre, & Bonnefoy, 2005). In a systematic review of 50 quantitative studies investigating the associations between green space access and physical activity, 20 reported positive links (higher physical activity with increased green space access), 15 were weak or mixed, 2 were negative and 13 found no evidence of any association(Lachowycz, Jones, Page, Wheeler, & Cooper, 2012). Result of present study indicated that Chitgar man-made lake with its surrounding green spaces as new ecological zones helped to soften the air in the area. Since the particular heat capacity of water is high, the lake water can accumulate large amounts of heat. This sometimes causes a decline in the magnitude of temperature alterations. 3. Results Weightings of assessment criteria and sub-criteria in negative impact Criterion/ Weight Sub criteria Options Final Weight Construction Non Construction Structural determinants Construction: 0.856 Non Construction: 0.135 Income 0.889 0.111 Culture 0.813 0.187 Education 0.833 0.167 Job 0.817 0.183 Limitation of available resources 0.609 0.391 Traffic 0.573 0.427 Ethnicity 0.805 0.195 Tourism 0.896 0.104 Region Popularity 0.889 0.111 Recreational Services 0.895 0.105 political participation 0.834 0.166 Spirituality 0.811 0.189 Housing 0.524 0.476* Social Security 0.667 0.333 Intermediary determinants Construction: 0.856 Non Construction: 0.135 Physical and Social Development of Children 0.875 0.125 Physical Activity 0.890 0.110 Quality of Life 0.875 0.125 Nutrition 0.676 0.324 Unhealthy Behavior 0.571 0.429 Unintentional Injuries 0.643 0.357 Pathogenesis 0.539 0.461 Air Quality 0.900 0.100 Mental Health 0.892 0.108 Aesthetic 0.857 0.143 Social Capital 0.860 0.140 Weather Changes 0.820 0.180 Environmental Health 0.564 0.436 Noise Pollution 0.533 0.467* Table 3. Weightings of assessment criteria and sub-criteria in negative impact 191 Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs 4. Discussion and Conclusion South Carolina Institute of Medicine and Public Health (IMPH) assessed the health effects of park, forest paths, and green spaces plans in west part of Greenville, and found out “quality of air” as one of the important determinants connected to these spaces (South Carolina Institute of Medicine and Public Health (IMPH). A Health Impact Assessment (HIA) of Park, Trail, and Green Space Planning in the West Side of Greenville, South Carolina, 2013). Moreover, Faculty of Public Health of London found improvement of welfare and mental health for all, better physical activity, declined violence, and anti-social behaviors, reduced health disparities, reduced mortality and cardiovascular diseases, enhancement of air quality, solving noise pollution, and fiscal profits as health determinants connected to green spaces (Faculty of Public Health. Great Outdoors: How Our Natural Health Service Uses Green Space To Improve Wellbeing 2010) According to the Heinze healthy and appropriately maintained green space offers 192 Global Journal of Health Science Vol. 6, No. 5; 2014 www.ccsenet.org/gjhs (Heinze, 2011). “Non-construction” option was also chosen with regard to sub-criterion of pathogenesis. Hunter et al. pointed out that construction of man-made lakes and dams as well as development of irrigation projects in tropical areas are especially implicated because, besides the obvious profits brought to the economy of a country, they have a tendency to ruin the environment by devastation of forests, increasing soil erosion, and the production of more biotypes than before to intermediate hosts or vectors of parasitic or infectious diseases. Such parasitic and infectious diseases transmission cycles profoundly affect series of ecosystem (Hunter, Rey, & Scott, 1983). A policy position is recommended whereby disease prevention procedures are incorporated with development projects from the scratch along with identification and inclusion of infrastructural investment and operational costs for health maintenance in the total benefit-cost analysis (Hunter, Rey, & Scott, 1982). “Noise pollution” came to have the highest weight with regard to negative effects of lake on intermediary determinants. Potchter stated that urban areas suffer from a variety of environmental troubles because of human activities which boost the development of urban heat islands, augment air pollution and noise. This study examines the potential of vegetation to improve local air quality, mitigate air temperature, and reduce noise levels. 4. Discussion and Conclusion Also, in a study carried out in the city of Tel Aviv, Israel, they investigated the total environmental consequences of diverse green urban spaces on climate, noise and air pollution (Potchter, Shashua-Bar, Choen, Boltansky, & Yaakov, 2009). Most European cities are now facing with major socioeconomic problems such as interrelated issues of urban sprawl, traffic congestion, noise, and air pollution. A methodology is now being introduced for assessing the role of green space and urban form in mitigating the adverse effects of urban station, with a focus on the environment but also considering socioeconomic aspects. Of other negative effects of lake construction was on “housing”. Participants believed that lake construction raises the price of housing but it may decrease its quality. Such a rise in property values is considered as a potential power for more displacement and even homelessness because of high rents. The potential gentrification could also disable new people of alike economic conditions to come into the neighborhood. Furthermore, property taxes for homeowners may be increased as well. If they lack abilities to afford to stay in their homes, they could also be displaced. A rise in property values could result in the possibility for displacement due to gentrification. Gentrification points to the boost in property values as a result of renovation and redevelopment in poor areas (Huestis, 2005). This shift can have a significant influence on health differences, especially for the poor, women, and children, the elderly and racial minorities. Studies show that these vulnerable groups have proclivities for higher rates of asthma, diabetes, and cardiovascular disease. Those residents who are influenced by such shift can experience an alteration in stress levels, crime, and/or mental health. Other health related impacts of displacement can comprise a lack of access to healthy food choices, transport, and quality of schools, bicycle, and walking pathways as well as affordable housing (Center for Disease Control and Prevention (CDC), Health Effects of Gentrification, 2012). Klessig believed that sustainability should start from needs. What are the needs of community and which of them can be met by lakes are important, and through this way, they can play their roles in sustainability issues. He pointed out eleven needs essential for a community to consider. 4. Discussion and Conclusion A sustainable social structure should provide measures to reach them, and so the functions of lakes can be defined accordingly as provision of aesthetic, cultural, economic, educational, and recreational opportunities as well as collective, emotional, environmental, spiritual, and individual safety (Klessig, 2001). Parks and urban green spaces as one of the most important public areas in modern cities play a very crucial role in meeting various social, cultural, and psychological needs of citizens. It is expected that with an increase in number of studies done on evaluation of health effects in Iran, the adverse effects of implementing projects, designs, and policies on people’s health would be recognized, prevented, and eradicated in coming futures. Acknowledgements This research is derived from the PhD thesis in specialized research social determinant of health approved by Research Council Session in Welfare and Rehabilitation Sciences University and the meeting of Ethics Committee of that university. We appreciate the Cooperation by honorable research deputy of university and honorable Municipality of District 22 of Tehran. Arnberger, A. (2012). Urban Densification and Recreational Quality of Public Urban Green Spaces—A Viennese Case Study .Sustainability, 4(4), 703-720. http://dx.doi.org/10.3390/su4040703 Araoye, P. (2002). Man-made lakes, ecological studies and conservation needs in Nigeria. Revista de biología tropical, 50(3-4), 857-864. Averill, J. B. (2002). Matrix analysis as a complementary analytic strategy in qualitative inquiry. Qualitative tropical, 50(3-4), 857-864. Arnberger, A. (2012). Urban Densification and Recreational Quality of Public Urban Green Spaces—A Viennese Case Study .Sustainability, 4(4), 703-720. http://dx.doi.org/10.3390/su4040703 Averill, J. B. (2002). Matrix analysis as a complementary analytic strategy in qualitative inquiry. Qualitative References Araoye, P. (2002). Man-made lakes, ecological studies and conservation needs in Nigeria. Revista de biología tropical, 50(3-4), 857-864. Arnberger, A. (2012). Urban Densification and Recreational Quality of Public Urban Green Spaces—A Viennese Case Study .Sustainability, 4(4), 703-720. http://dx.doi.org/10.3390/su4040703 Arnberger, A. (2012). Urban Densification and Recreational Quality of Public Urban Green Spaces—A Viennese Case Study .Sustainability, 4(4), 703-720. http://dx.doi.org/10.3390/su4040703 Averill, J. B. (2002). Matrix analysis as a complementary analytic strategy in qualitative inquiry. Qualitative 193 Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs Health Research, 12(6), 855-866. http://dx.doi.org/10.1177/104973230201200611 Health Research, 12(6), 855-866. http://dx.doi.org/10.1177/104973230201200611 Barati, N. (2011). Roykardhaye jadid dar barnamerizi va modiriate shahri: Barnamereizie shahre salem. Tehran: Urban Planning and Research Center of Tehran: dsneshe shahr. Barbour, R., & Kitzinger, J. (1998). Developing focus group research: Politics, theory and practice. Sage Publications Limited. Bedimo-Rung, A. L., Mowen, A. J., & Cohen, D. A. (2005). The significance of parks to physical activity and public health: a conceptual model. American journal of preventive medicine, 28(2), 159-168. http://dx.doi.org/10.1016/j.amepre.2004.10.024 Bhatta, G. D., & Doppler, W. (2010). Farming differentiation in the rural-urban interface of the middle mountains, Nepal: Application of analytic hierarchy process (AHP) modeling. Journal of Agricultural Science, 2(4), p37. http://dx.doi.org/10.5539/jas.v2n4p37 Center for Disease Control and Prevention (CDC). (2012). Health Effects of Gentrification. Dahlgren, L., Emmelin, M., & Winkvist, A. (2007). Qualitative methodology for international publ Design for Health. Building public understanding: The link between health and planning. (2007). Ellaway, A., Macintyre, S., & Bonnefoy, X. (2005). Graffiti, greenery, and obesity in adults: secondary analysis of European cross sectional survey. BMJ: British Medical Journal, 331(7517), 611. http://dx.doi.org/10.1136/bmj.38575.664549.F7 Enger, S. C. (2005). Planning for Parks, Recreation, and Open Space in Your Community. Washington: Washington State Department of Community, Trade and Economic Development. Municipal Research & Services Center Seattle Faculty of Public Health. Great Outdoors: How Our Natural Health Service Uses Green Space To Improve Wellbeing. (2010). Retrieved from http://www.fph.org.uk/uploads/bs_great_outdoors.pdf Field, A. (2011). Integrating Health Impact Assessment in Urban Design and Planning: The Manukau Experience. Ministry of Health. Godbey, G. C., & AJ, M. (2003). The role of parks and recreation in promoting physical activity and health. Franklin County Healthier Communities. Graneheim, U. H., & Lundman, B. (2004). Qualitative content analysis in nursing research: concepts, procedures and measures to achieve trustworthiness. Nurse education today, 24(2), 105-112. http://dx.doi.org/10.1016/j.nedt.2003.10.001 Scotland, G. (2008). Health Impact Assessment of Greenspace, A Guide.Greenspace Scotland. Retrieved form www. green spacescotland. org. References New York, NY, U.S.A.: McGraw-Hill International. http://dx.doi.org/10.1002/0470011815.b2a4a002 Saaty, T. L. (2005). Analytic hierarchy process. Encyclopedia of Biostatistics. http://dx.doi.org/10.1504/IJSSCI.2008.017590 Saaty, T. L. (2008). Decision making with the analytic hierarchy process. International Journal of Services Sciences, 1(1), 83-98. Santana, P., Santos, R., & Costa, C. (2009). Walkable Urban Green Spaces: Health Impact Assessment in Amadora, Portugal'. REAL CORP 2009: CITIES 3.0–Smart, Sustainable, Integrative Strategies, Concepts and Technologies for Planning the Urban Future, 579-5 85 . Solar, O, & Irwin, A. (2010.). A conceptual framework for action on the social determinants of health. South Carolina Institute of Medicine and Public Health (IMPH). (2013). A Health Impact Assessment (HIA) of Park, Trail, and Green Space Planning in the West Side of Greenville, South Carolina. Thapa, R. B., & Murayama, Y. (2010). Drivers of urban growth in the Kathmandu Valley, Nepal: Examining the efficacy of the analytic hierarchy process. Applied Geography, 30(1), 70-83. http://dx.doi.org/10.1016/j.apgeog.2009.10.002 Through Active Living Summit, Kauffman Station PA. (2003). Through Active Living Summit, Kauffman Station PA. (2003). Yu, J., Wang, L., & Gong, X. (2013). Study on the Status Evaluation of Urban Road Intersections Traffic Congestion Base on AHP-TOPSIS Modal. Procedia-Social and Behavioral Sciences, 96, 609-616. http://dx.doi.org/10.1016/j.sbspro.2013.08.071 References uk/upload/File/Greenspace% 20HIA. pdf Harris, P., Harris-Roxas, B., Harris, E., & Kemp, L. (2007). Health impact assessment: a practical guid Heinze, J. (2011). Benefits of Green Space – Recent Research: A nonprofit Research Foundation Specializing in Health and EnvironmentalL Science. Huestis, J. (2005). Gentrification: Will the University of Illinois at Chicago Threaten Surrounding Communities. Hunter, J. M., Rey, L., & Scott, D. (1982). Man-made lakes and man-made diseases: towards a policy resolution. Social science & medicine, 16(11), 1127-1145. http://dx.doi.org/10.1016/0277-9536(82)90115-0 Hunter, J. M., Rey, L., & Scott, D. (1983). Man-made lakes- man-made diseases. In World Health Forum (Vol. 4, No. 2, pp. 177-182). King County’s Open Space System: Parks, Trails, Natural Areas, and Working Resource Lands. (2004). King County. Klessig, L. L. (2001). Lakes and society: The contribution of lakes to sustainable societies. Lakes & Reservoirs: Research & Management, 6(2), 95-101. http://dx.doi.org/10.1046/j.1440-1770.2001.00131.x Lachowycz, K., Jones, A. P., Page, A. S., Wheeler, B. W., & Cooper, A. R. (2012). What can global positioning systems tell us about the contribution of different types of urban greenspace to children's physical activity? Health & Place, 18(3), 586-594. National Research Council of the National Academies (NRCNA). Improving Health in the United States: The Role of Health Impact Assessment. (2011). Retrieved from http://www.nap.edu/catalog.php?record_id=13229, Retrieved June 26, 2012 194 Vol. 6, No. 5; 2014 Global Journal of Health Science www.ccsenet.org/gjhs Nouri, J., Rahimipour, H., & Nezakati, R. (2004). Simulation of Thermal Stratification in Urban Man-made Lakes. American Journal of Applied Sciences, 1(4), 309. http://dx.doi.org/10.3844/ajassp.2004.309.315 Parra-Loper, C., Calatrava-Requena, J., & de-Haro-Gimenez, T. (2007). A multi-criteria evaluation of the environmental performances of conventional, organic and integrated olive-growing systems in the south of Spain based on experts' knowledge. Renewable Agriculture and Food Systems, 22(3), 189-203. http://dx.doi.org/10.1017/S1742170507001731 Potchter, O., Shashua-Bar, L., Choen, P., Boltansky, D., & Yaakov, Y. (2009). The Effect of Green Urban Spaces on Climate, Noise and Air Pollution in the city of Tel Aviv, Israel. Paper presented at the The seventh International Conference on Urban Climate . Pretty, J., Griffin, M., Sellens, M., et al. (2003). Green exercise: complementary roles of nature, exercise and diet in physical and emotional well-being and implications for public health policy. (University of Essex, CES Occasional Paper). Rosa, L. V., & Haddad, A. N. (2013). Assessing the sustainability of existing buildings using the analytic hierarchy process. American Journal of Civil Engineering, 1(1), 24-30. http://dx.doi.org/10.11648/j.ajce.20130101.14 Saaty, T. L. (1980). The Analytic Hierarchy Process. Copyrights py g Copyright for this article is retained by the author(s), with first publication rights granted to the journa Copyright for this article is retained by the author(s), with first publication rights granted to the journal. py g y ( ), p g g j This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). 195
https://openalex.org/W4308566630	https://hal.science/hal-04010465/file/aa44522-22.pdf	English	null	Ion irradiation triggers the formation of the precursors of complex organics in space	Astronomy & astrophysics	2,022	cc-by	15,188	To cite this version: R. Urso, E. Hénault, R. Brunetto, D. Baklouti, G. Baratta, et al.. Ion irradiation triggers the formation of the precursors of complex organics in space. Astronomy and Astrophysics - A&A, 2022, 668, pp.A169. 10.1051/0004-6361/202244522. hal-04010465 Ion irradiation triggers the formation of the precursors of complex organics in space R. Urso, E. Hénault, R. Brunetto, D. Baklouti, G. Baratta, Z. Djouadi, A. Elsaesser, C. Scirè, G. Strazzulla, M. Palumbo Distributed under a Creative Commons Attribution 4.0 International License ABSTRACT Among other molecules, we observe the formation of H2CO and CH3CHO, and we find that their abundance depends on the dose and on the stoichiometry of the mixtures. We find that the H2CO abundance reaches the highest value after a dose of 10 eV/16u and then it decreases as the dose increases. Results. Among other molecules, we observe the formation of H2CO and CH3CHO, and we find that their abundance depends on the dose and on the stoichiometry of the mixtures. We find that the H2CO abundance reaches the highest value after a dose of 10 eV/16u and then it decreases as the dose increases. Conclusions. The data suggest that surfaces exposed to high doses are depleted in H2CO. This explains why the amount of HMT in organic residues and that formed after irradiation of ices depends on the dose deposited in the ice. Because the H2CO abundance decreases at doses higher than 10 eV/16u, a lower quantity of H2CO is available to form HMT during the subsequent warm-up. The H2CO abundances caused by ion bombardment are insufficient to explain the ISM abundances, but ion bombardment can account for the abundance of CH3CHO towards the ISM and comets. s. astrochemistry – ISM: molecules – ISM: abundances – methods: laboratory: solid state – techniques: spectroscopi understanding of the processes that contribute to the formation of complex organics. Experiments allowed studying how ener- getically charged particles in space, such as cosmic rays (CR) and solar particles, affect the spectral properties of frozen sur- faces in the ISM and in the Solar System, such as the appearance of red slopes in irradiated ices (e.g. Strazzulla et al. 1983; Moore et al. 1996; Brunetto et al. 2006; Palumbo 2006; Dartois et al. 2015; Rothard et al. 2017; Poston et al. 2018). Changes primar- ily depend on the initial composition of the ice, the source of processing, and the irradiation dose (e.g. Rothard et al. 2017). Simple and complex compounds are formed after irradiation of mixtures that simulate the pristine composition of astrophysical ices (e.g. Moore et al. 1996; Palumbo et al. 1997; Kaˇnuchová et al. 2016; Urso et al. 2019). Among them, vibrational mode bands attributed to aldehydes are found in the mid-IR spectra of samples exposed to UV photons and ion irradiation. R. G. Urso1,2 , E. Hénault3, R. Brunetto3, D. Baklouti3, G. A. Baratta2, Z. Djouadi3, A. Elsaesser1, C. Scirè2, G. Strazzulla2, and M. E. Palumbo2 1 Experimental Biophysics and Space Sciences, Institute of Experimental Physics, Freie Universitaet, Arnimallee 14, 14195 Berlin, Germany e mail: riccardo urso@inaf it e-mail: riccardo.urso@inaf.it 2i e-mail: riccardo.urso@inaf.it 2i 2 INAF–Osservatorio Astrofisico di Catania, via Santa Sofia 78, 95123 Catania, Italy 3 Université Paris-Saclay, CNRS, Institut d’Astrophysique Spatiale, 91405 Orsay, France 2 INAF–Osservatorio Astrofisico di Catania, via Santa Sofia 78, 95123 Catania, Italy 3 Université Paris-Saclay, CNRS, Institut d’Astrophysique Spatiale, 91405 Orsay, France Received 16 July 2022 / Accepted 19 October 2022 Received 16 July 2022 / Accepted 19 October 2022 ABSTRACT In partic- ular, formaldehyde (H2CO) and acetaldehyde (CH3CHO) have been reported among the byproducts of irradiation of frozen mixtures containing H-, C-, and O-bearing compounds (e.g. Khare et al. 1989; Moore & Hudson 1998; Palumbo et al. 1999). In space, solid-phase H2CO is likely identified toward massive Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication. Ion irradiation triggers the formation of the precursors of complex organics in space The case of formaldehyde and acetaldehyde R. G. Urso1,2 , E. Hénault3, R. Brunetto3, D. Baklouti3, G. A. Baratta2, Z. Djouadi3, A. Elsaesser1, C. Scirè2, G. Strazzulla2, and M. E. Palumbo2 HAL Id: hal-04010465 https://hal.science/hal-04010465v1 Submitted on 3 Mar 2023 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License Astronomy & Astrophysics Astronomy & Astrophysics A&A 668, A169 (2022) https://doi.org/10.1051/0004-6361/202244522 © R. G. Urso et al. 2022 A169, page 1 of 10 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication. ABSTRACT Context. Cosmic rays and solar energetic particles induce changes in the composition of compounds frozen onto dust grains in the interstellar medium (ISM), in comets, and on the surfaces of atmosphere-less small bodies in the outer Solar System. This induces the destruction of pristine compounds and triggers the formation of various species, including the precursors of complex organics. Context. Cosmic rays and solar energetic particles induce changes in the composition of compounds frozen onto dust grains in the interstellar medium (ISM), in comets, and on the surfaces of atmosphere-less small bodies in the outer Solar System. This induces the destruction of pristine compounds and triggers the formation of various species, including the precursors of complex organics. Aims. We investigate the role of energetic ions in the formation of formaldehyde (H2CO) and acetaldehyde (CH3CHO), which are observed in the ISM and in comets and which are thought to be the precursors of more complex compounds such as hexamethylenete Context. Cosmic rays and solar energetic particles induce changes in the composition of compounds frozen onto dust grains in the interstellar medium (ISM), in comets, and on the surfaces of atmosphere-less small bodies in the outer Solar System. This induces the destruction of pristine compounds and triggers the formation of various species, including the precursors of complex organics. Aims. We investigate the role of energetic ions in the formation of formaldehyde (H2CO) and acetaldehyde (CH3CHO), which are observed in the ISM and in comets, and which are thought to be the precursors of more complex compounds such as hexamethylenete- tramine (HMT), which is found in carbonaceous chondrites and in laboratory samples produced after the irradiation and warm-up of astrophysical ices. Aims. We investigate the role of energetic ions in the formation of formaldehyde (H2CO) and acetaldehyde (CH3CHO), which are observed in the ISM and in comets, and which are thought to be the precursors of more complex compounds such as hexamethylenete- tramine (HMT), which is found in carbonaceous chondrites and in laboratory samples produced after the irradiation and warm-up of astrophysical ices.i p y Methods. We performed ion irradiation of water, methanol, and ammonia mixtures at 14–18 K. We bombarded frozen films with 40–200 keV H+ that simulate solar energetic particles and low-energy cosmic rays. Samples were analysed by infrared transmission spectroscopy.i p py Results. 1. Introduction young stellar objects (MYSOs) and low-mass YSOs (LYSOs), with abundances of 0.06–6.00 relative to CH3OH (Bottinelli et al. 2007; Boogert et al. 2015). In the Solar System, H2CO is detected in comets (e.g. Mitchell et al. 1987; Altwegg et al. 2019) with abundances that span between 0.02 and 5 with respect to CH3OH (e.g. le Roy et al. 2015; Boogert et al. 2015; Schuhmann et al. 2019). Solid-phase CH3CHO is only tentatively detected in the ISM and in comets, while it is observed in the gas-phase in both environments (e.g. Cazaux et al. 2003; Crovisier et al. 2004; Öberg et al. 2010; Jaber et al. 2014; Codella et al. 2015; Schuhmann et al. 2019). Information on solid-phase CH3CHO is lacking, possibly because its main vibrational mode bands over- lap with those of other species. In particular, HCOO−shows a vibrational feature at the same wavelength and can contribute to the 1350 cm−1 feature (e.g. Schutte et al. 1999). ( g ) In space, aldehydes are among the frozen compounds that can react to form more complex species. Reactions are also favoured by thermal processing that increases the diffusion and reactivity of molecules (e.g. Mispelaer et al. 2013; Theulé et al. 2013; Vinogradoff et al. 2013). When irradiation and thermal processing of astrophysical ices is simulated in the laboratory, a complex organic refractory material forms. Organic refrac- tory residues are laboratory analogues of cometary refractory organics (e.g. Strazzulla & Johnson 1991; Baratta et al. 2019), of the organics in extraterrestral samples (e.g. Nuevo et al. 2008; de Marcellus et al. 2017), and of organic materials that can form on the surface of KBOs and TNOs (e.g. Strazzulla et al. 2003; Urso et al. 2020b). The characterization of residues reveals various astrobiologically relevant compounds, including amino acids, nucleobases, and sugars (Danger et al. 2013; Fresneau et al. 2017; Meinert et al. 2016; Nuevo et al. 2018; Urso et al. 2020b). According to Vinogradoff et al. (2012b), together with NH3 and HCOOH, which act as a catalyst, H2CO is one of the precursor of hexamethylenetetramine (HMT). This is one of the most abundant organics detected in residues and forms at relatively high temperatures (>280 K) during the warm-up of ices. HMT was recently found in the carbonaceous chon- drites Murchison, Murray, and Tagish Lake (Oba et al. 2. Experimental methods The ion irradiation experiments were performed at the Intitut d’Astrophysique Spatiale, Irène Joliot-Curie Laboratoire (IAS- IJCLab Orsay, France) and at the Laboratory of Experimental Astrophysics (LASp) at the Istituto Nazionale di Astrofisica INAF-Osservatorio Astrofisico di Catania (Catania, Italy). We used the IrradiatioN de Glaces et Météorites Analysées par Réflectance VIS-IR (INGMAR) setup at IAS-IJCLab to irradi- ate ices with 40 keV H+ by means of the SIDONIE accelerator (Chauvin et al. 2004). The setup at the LASp allows accelerating ions up to 200–400 keV (Fulvio et al. 2009; Urso et al. 2016). For this work, we used 200 keV H+. Although the two experimen- tal setups show distinct characteristics, they both consist of high vacuum chambers (P ∼10−8–10−9 mbar), where He cryocool- ers allow cooling substrates down to ∼15 K. H2O:CH3OH:NH3 gaseous mixtures were injected in the vacuum chambers, where they condensed on the cold substrates. The resulting frozen films were then exposed to H+ beams that were rastered on samples, in order to ensure a homogeneous covering of their surfaces. The energy of impinging ions determines their penetration depth in the frozen mixture. According to simulations per- formed with the SRIM software (Ziegler et al. 2008), 40 keV and 200 keV H+ implant in H2O:CH3OH:NH3 mixtures within 1 µm and 3.3 µm, respectively. We therefore deposited frozen films with thicknesses that were lower than the H+ implanta- tion depth to ensure that the impinging particles released their energy homogeneously within the ice, and to limit their implan- tation. In particular, for the experiments with 40 keV H+, we deposited films between 0.4 and 0.6 µm, while in the case of irradiation with 200 keV, we deposited 0.5 and 1.2 µm thick films. During ion irradiation, we integrated the ion current to estimate the fluence, that is, the number of impinging ions cm−2. We then estimated the dose, that is, the energy deposited per molecule in the sample by incident particles, by multiply- ing the fluence by the stopping power (eV cm2/16u) calculated with SRIM. We give the dose in eV/16u, where 16u is the atomic mass unit of a CH4 molecule that we take as a reference (Strazzulla & Johnson 1991). CH3CHO is also invoked as a precursor of various complex compounds. Laboratory experiments that simulated the thermal processing of CH3CHO-rich ices showed that CH3CHO reacts with NH3 to form α-aminoethanol (NH2CH(CH3)OH) above 100 K (Duvernay et al. 2010). 1. Introduction Frozen compounds are present on dust grains in the interstel- lar medium (ISM; e.g. Boogert et al. 2015) and on the surface of small bodies in the Solar System, such as comets (e.g. Bockelée-Morvan et al. 2000; Altwegg et al. 2017), satellites of giant planets (e.g. Johnson 2005; Spencer & Nimmo 2013; Dougherty & Spilker 2018), Kuiper-belt (KBOs), and trans- neptunian objects (TNOs; e.g. Cruikshank et al. 1998; Brown et al. 2011; Grundy et al. 2020; Barucci & Merlin 2020). Var- ious KBOs and TNOs exhibit frozen compounds together with red slopes in the visible and near-infrared (NIR) spectra that could be attributed to complex C-rich materials (Brunetto et al. 2006; Brown et al. 2011; Dalle Ore et al. 2011). Complex organ- ics are also found in materials sampled from asteroids (e.g. Chan et al. 2021; Pilorget et al. 2021; Parker et al. 2022), in interplanetary dust particles (IDPs; e.g. Chan et al. 2020) and carbonaceous chondrites (CC; e.g. Sephton 2002; Busemann et al. 2006; Schmitt-Kopplin et al. 2010; Oba et al. 2020, 2022). pp Laboratory experiments allow us to shed light on the physi- cal and chemical properties of small bodies and to improve our A&A 668, A169 (2022) compounds observed in YSOs (e.g. Caselli & Ceccarelli 2012; Boogert et al. 2015) and in small bodies in the outer Solar System (e.g. Cruikshank et al. 1998; Delsanti et al. 2010; Grundy et al. 2016; Barucci & Merlin 2020). Furthermore, organic refrac- tory residues produced after irradiation of H2O:CH3OH:NH3 mixtures are well characterized and show complex organics, including HMT (e.g. Bernstein et al. 1995; Danger et al. 2016; Urso et al. 2020b). We can thus interpret our results by taking into account similar experiments reported in literature. The irra- diation with 40 and 200 keV H+ allows us to simulate the effects of solar charged particles and CR on icy surfaces in the Solar System and on icy grain mantles. As discussed by Mewaldt et al. (2007) and Urso et al. (2020a), protons in the range of tens of keV are abundant in the spectrum of solar energetic particles (SEP) impinging on frozen surfaces in the outer Solar System. Furthermore, several previous studies used hundreds ok keV H+ to simulate the effects of low-energy CR (e.g. Palumbo 2006; Islam et al. 2014; Urso et al. 2019). 1. Introduction 2020), and in the laboratory, it forms after UV and ion irradiation at 10–20 K and subsequent warm-up to room temperature of mix- tures containing water (H2O), methanol (CH3OH), and ammonia (NH3, Hulett et al. 1971; Bernstein et al. 1995; Cottin et al. 2001; Vinogradoff et al. 2013; Danger et al. 2016; Urso et al. 2020b). Urso et al. (2020b) reported that the composition of organic refractory residues depends on the extent of irradiation experienced by ices. They also found that the HMT abundance decreased when the dose was increased. 2. Experimental methods For replicate experiments, the last letter represents the range of dose used; L: low dose; H: high dose. µ Fig. 1. Mid-IR spectra of a H2O:CH3OH:NH3 = 1:1:1 mixture as deposited (dashed blue line) and after irradiation with 200 keV H+ (solid red line). Bottom axis: wavenumber (cm−1). Top axis: wavelength (µm). Spectra are on the same optical depth scale with an offset to avoid superposition. µ the arrival of two ions on the same target area. Ions travelling through the target cause ionizations and excitations of molecules and the formation of molecular fragments and radicals, whose recombination forms species that were not present in the origi- nal sample (e.g. Rothard et al. 2017). The whole process, from the excitation to the recombination and thus relaxation, requires about 10−12 s or less. The mixtures we deposited at 14–18 K had a density of about 0.9 g cm−3. In these samples, keV pro- tons create a radial struggle of r = 1.7×10−5 cm as calculated by means of the SRIM software. The ion track within the ice can be approximated to a cylinder with a section area a = πr2, equal to 9.1×10−10 cm2, which we can approximate to 10−9 cm2. Thus, it is necessary to reach a fluence of 109 ions cm−2 to fully cover the surface of a sample of 1 cm2. Taking into account a constant flux value as that used in our experiment, i.e. about 6.25×1012 ions cm−2 s−1, a fluence of 109 ions cm−2 is reached after 1.6×10−2 s. This is the time required for an impinging ion to hit the same target area a already excited by a former ion in our experiments. This time is about 1010 times longer than the relaxation time along the ion track (about 10−12 s). As a consequence, impinging ions always travel through volumes of ices that are stable and relaxed. This allows our laboratory experiments to be representative of the particle irradiation taking place in astrophysical ices exposed to cosmic and solar charged particles. Fig. 1. Mid-IR spectra of a H2O:CH3OH:NH3 = 1:1:1 mixture as deposited (dashed blue line) and after irradiation with 200 keV H+ (solid red line). Bottom axis: wavenumber (cm−1). Top axis: wavelength (µm). Spectra are on the same optical depth scale with an offset to avoid superposition. and quantitative information that show consistent trends, thus confirming the reliability of our results. 2. Experimental methods Furthermore, above 70 K and in presence of HCOOH, the reaction between CH3CHO and NH3 leads to the CH3CHO-NH3 trimer after the formation of the reac- tion intermediates α-aminoethanol and ethanimine (Vinogradoff et al. 2012a). We here report new experiments that we performed to inves- tigate the role of ion irradiation in the production of H2CO and CH3CHO. We study how their abundance is affected by the dose and we discuss how the formation of HMT can be affected by the destruction of H2CO, its precursor. For this study, we deposited H2O:CH3OH:NH3 mixtures at 14–18 K and exposed them to 40 and 200 keV H+. H2O, CH3OH, and NH3 mixtures were chosen because they are among the most abundant solid-phase To properly simulate the ion irradiation of frozen surfaces, we kept the ion current in the range of some hundred nA and below 1 µA. For an ion current density of 1 µA cm−2, the flux is equal to 6.25×1012 ions cm−2 s−1. This value prevents the macroscopic heating of the samples. Furthermore, the low current is required to ensure that enough time passes between A169, page 2 of 10 R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space Table 1. List of all the experiments. Frozen mixtures are deposited and irradiated at 14–18 K. Table 1. List of all the experiments. Frozen mixtures are deposited and irradiated at 14–18 K. Experiment H2O:CH3OH:NH3 H+ energy Dose min. Dose max. label mixture ratio (keV) (eV/16u) 3_200_L 3:1:1 200 2 23 3_200_H 3:1:1 200 16 80 1_200_L 1:1:1 200 2 27 1_200_H 1:1:1 200 19 93 3_40 3:1:1 40 7 54 1_40 1:1:1 40 26 103 Notes. Experiments are labelled as follows: the first number gives the amount of H2O with respect to CH3OH and NH3 in the mixture; the second number gives the energy of the H+ beam used for irradiation. For replicate experiments, the last letter represents the range of dose used; L: low dose; H: high dose. Experiment H2O:CH3OH:NH3 H+ energy Dose min. Dose max. Notes. Experiments are labelled as follows: the first number gives the amount of H2O with respect to CH3OH and NH3 in the mixture; the second number gives the energy of the H+ beam used for irradiation. 2. Experimental methods Samples are analysed in situ by means of transmission Fourier-Transform infrared spectroscopy (FT-IR) with a resolu- tion of 1 cm−1. In the INGMAR setup, the IR beam arrives on the sample with an angle of 10◦with respect to the surface nor- mal, while at LASp the IR beam arrives with an incident angle of 45◦with respect to the surface normal. 3. Results Table 1 groups all the experiments taken into account in this work. Experiments are labelled according to (i) the amount of H2O with respect to CH3OH and NH3 in the mixture and (ii) the energy of the H+ beam used to bombard the samples. In the case of the experiments performed with 200 keV H+, we deposit the same mixture ratio twice, in the same experimental conditions, and we perform irradiation in two ranges of dose that overlap with each other. These experiments are labelled according to the dose range we investigate (L for the low dose range and H for the high dose range). H2O, CH3OH, and NH3 are deposited at 14–18 K on IR- transparent substrates (ZnSe or KBr) and bombarded with 40 and 200 keV H+. Spectra are acquired after the deposition and after each irradiation step, in order to monitor the sample spec- tral evolution as a function of irradiation dose. Spectra are then analysed to obtain qualitative and quantitative estimations on the newly-formed compounds. We give the abundance of the newly- formed compounds N(X) by estimating the column density, i.e. N(X) = R τ(ν)dν A , where τ(ν) is the optical depth (–ln T, trans- mittance) and A is the band strength (cm molecule−1) of the vibrational mode band taken into account. 3.1. Methanol destruction and methane formation 9 1 C H 3 O H i n H 2 O : C H 3 O H : N H 3 m i x t u r e s + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 N ( C H 3 O H ) / N ( C H 3 O H ) i D o s e ( e V / 1 6 u ) Experiment σ (16u/eV) y∞ 3_200_L 0.035 ± 0.011 0.045 ± 0.009 3_200_H 0.048 ± 0.005 0.042 ±0.001 1_200_L 0.053 ± 0.004 0.037 ± 0.002 1_200_H 0.036 ± 0.001 0.048 ± 0.001 3_40 0.051 ± 0.005 0.035 ± 0.0014 1_40 0.043 ± 0.006 0.046 ± 0.002 Notes. Data from the analysis of the 1305 cm−1 band. Notes. Data from the analysis of the 1305 cm−1 band. Notes. Data from the analysis of the 1305 cm−1 band. 3.1. Methanol destruction and methane formation Figure 1 shows the spectra of a H2O:CH3OH:NH3 = 1:1:1 mixture before and after irradiation with 200 keV H+. Before irradiation we observe the N-H stretching mode band of NH3 at 3375 cm−1, the O-H stretching mode band in water and CH3OH at about 3270 cm−1, the C-H stretching mode bands of CH3OH The comparison between experimental results obtained in different laboratories is not trivial. We thus carefully plan exper- iments and we use the same procedures to acquire, reduce, and analyse data so to minimise the differences that arise when com- paring different experiments. This allow us to obtain qualitative A169, page 3 of 10 A&A 668, A169 (2022) Table 2. CH4 formation cross section in H2O:CH3OH:NH3 mixtures after H+ bombardment. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 0 . 7 0 . 8 0 . 9 1 C H 3 O H i n H 2 O : C H 3 O H : N H 3 m i x t u r e s + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 N ( C H 3 O H ) / N ( C H 3 O H ) i D o s e ( e V / 1 6 u ) Fig. 2. Column density of CH3OH during irradiation with 40–200 keV H+ in H2O:CH3OH:NH3 normalized to the initial amount of methanol, (CH3OH)i. The CH3OH column density is estimated from the analysis of the 1025 cm−1 band. Experimental data are fitted with an exponential curve y = e−σD (dashed grey line). 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 0 . 6 0 . 7 0 . 8 0 . 3.2. Formaldehyde and acetaldehyde formation D o s e ( e V / 1 6 u ) Various vibrational mode bands attributed to H2CO (1720, 1500, 1250 cm−1) and CH3CHO (1720, 1382, 1350 cm−1, e.g. Moore & Hudson 1998; Urso et al. 2020b) are found in both 3:1:1 and 1:1:1 mixtures after irradiation. We perform quantitative analysis to study the evolution of their abundance with increasing irradi- ation dose. For the analysis, we do not consider the 1720 cm−1 band as it contains contributions from all the –C=O bearing species in the samples. Furthermore, cautions has to be taken because of other features in the same spectral range (NH2CHO at 1385 cm−1, NH+ 4 at 1480 and 1460 cm−1, HCOO−at 1385 and 1350 cm−1; e.g. Brucato et al. 2006; Raunier et al. 2004; Schutte et al. 1999). Fig. 2. Column density of CH3OH during irradiation with 40–200 keV H+ in H2O:CH3OH:NH3 normalized to the initial amount of methanol, (CH3OH)i. The CH3OH column density is estimated from the analysis of the 1025 cm−1 band. Experimental data are fitted with an exponential curve y = e−σD (dashed grey line). between about 2980 and 2830 cm−1, two broad features between 1720 and 1370 cm−1 attributed to bending modes in water and NH3 molecules, a feature at 1125 attributed to the wagging mode in NH3 and to the rocking mode in CH3OH, the C-O stretching mode of CH3OH at 1025 cm−1, and the water libration mode band between about 990 and 500 cm−1 (e.g. Urso et al. 2020b). Irradiation determines evident changes in the spectra. The fea- tures at about 3300 and 1025 cm−1 decrease and new bands appear. In particular, the CH3OH band at 1025 cm−1 decreases with increasing dose, following an exponential decay, because of the ion-induced destruction of CH3OH. Figure 2 shows the vari- ation in the CH3OH abundance as a function of the dose. We fit the data with an exponential curve (Eq. (1)), between about 2980 and 2830 cm−1, two broad features between 1720 and 1370 cm−1 attributed to bending modes in water and NH3 molecules, a feature at 1125 attributed to the wagging mode in NH3 and to the rocking mode in CH3OH, the C-O stretching mode of CH3OH at 1025 cm−1, and the water libration mode band between about 990 and 500 cm−1 (e.g. Urso et al. 2020b). Irradiation determines evident changes in the spectra. 3.2. Formaldehyde and acetaldehyde formation The fea- tures at about 3300 and 1025 cm−1 decrease and new bands appear. In particular, the CH3OH band at 1025 cm−1 decreases with increasing dose, following an exponential decay, because of the ion-induced destruction of CH3OH. Figure 2 shows the vari- ation in the CH3OH abundance as a function of the dose. We fit the data with an exponential curve (Eq. (1)), ) H2CO column densities were estimated by analysing the CH2 rocking mode band at 1250 cm−1 using the band strength value given by Bouilloud et al. (2015), that is, 1.5×10−18 cm molecule−1. Panel A of Fig. 5 shows the H2CO abundance with respect to the initial abundance of CH3OH as a function of the dose. In all the experiments, the H2CO abun- dance increases rapidly at low doses and the highest abundance is observed at about 10 eV/16u. At this dose, the irradiation of 3:1:1 mixtures produces 6.2% of H2CO, and the irradiation of 1:1:1 mixtures produces 2.4%. At higher irradiation doses the H2CO abundance starts to decrease, following an exponential decay. At about 30 eV/16u, the 1:1:1 and 3:1:1 mixtures show comparable amounts of H2CO. At the highest irradiation dose we investigated, the H2CO band is at the detection limit and the abundance is about 1% with respect to the initial amount of CH3OH. In Fig. 5 we also show the abundance of H2CO with respect to the amount of CH3OH at the same dose. Again, the ratio increases at low doses, with higher abundances in the case of the 3:1:1 mixtures. At doses higher than 50 eV/16u both H2CO and CH3OH are efficiently destroyed and the observed trend is based on features that are at the detection limits. y = e−σD, (1) y = e−σD, (1) CH4 abundance relative to the initial amount of methanol, (CH3OH)i with increasing irradiation dose. The CH4 column density is estimated from the analysis of the 1305 cm−1 band; the CH3OH col- umn density is estimated from the 1025 cm−1 band. of about the 4% of CH3CHO. In 1:1:1 mixtures, the highest abundance is about 3%. We also performed multi-Gaussian fits between 1500 and 1300 cm−1 by simulating a Gaussian curve to fit the 1350 cm−1 band (Fig. B.2, panel B). This method estimates the total area of the 1350 cm−1 band from the baseline at opti- cal depth zero. The estimated of the CH3CHO abundances we obtained through the multi-Gaussian fits are reported in Fig. 6, panel B. CH3CHO abundances reach 12% in 3:1:1 mixtures and 6% in 1:1:1 mixtures. We fitted the data in Fig. 6 with Eq. (2) to obtain the cross sections that are given in Table 3. The val- ues of y∞vary according to the method we used to estimate the 1350 cm−1 band area. Figure 6, panel C, shows the ratio of the CH3CHO abundance and the CH3OH abundance at the same dose. The exponential trend of the data has to be attributed to the formation of CH3CHO and the destruction of CH3OH. 1580/1350 band ratios that are as low as about 50% of the ratio in Vinogradoff et al. (2011). This preliminary analysis gives us con- fidence in attributing the 1350 cm−1 band to CH3CHO, although dedicated studies with more sensitive techniques are needed to estimate the relative amounts of HCOO−and CH3CHO in irradi- ated mixtures, such as in situ mass spectrometry currently under implementation at LASp. 1580/1350 band ratios that are as low as about 50% of the ratio in Vinogradoff et al. (2011). This preliminary analysis gives us con- fidence in attributing the 1350 cm−1 band to CH3CHO, although dedicated studies with more sensitive techniques are needed to estimate the relative amounts of HCOO−and CH3CHO in irradi- ated mixtures, such as in situ mass spectrometry currently under implementation at LASp. For our estimations of the CH3CHO abundance, we took into account a 1350 cm−1 band strength of 7.11×10−18 cm molecule−1 (Hudson & Ferrante 2020). This value is higher than the one given by Schutte et al. (1999), that is, 1.5×10−18 cm molecule−1, and thus allows a more conservative estimation of the CH3CHO abundance. (1) Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space 1 5 5 0 1 5 0 0 1 4 5 0 1 4 0 0 1 3 5 0 1 3 0 0 1 2 5 0 1 5 5 0 1 5 0 0 1 4 5 0 1 4 0 0 1 3 5 0 1 3 0 0 1 2 5 0 O p t i c a l d e p t h ( a . u . ) W a v e n u m b e r ( c m - 1 ) H 2 O : C H 3 O H : N H 3 m i x t u r e s a s d e p o s i t e d ( n o i r r a d i a t i o n ) 3 _ 2 0 0 _ L , 1 8 K 3 _ 2 0 0 _ H , 1 8 K 1 _ 2 0 0 _ L , 1 7 K 1 _ 2 0 0 _ H , 1 7 K 3 _ 4 0 , 1 5 K 1 _ 4 0 , 1 4 K A * O p t i c a l d e p t h ( a . u . ) W a v e n u m b e r ( c m - 1 ) + 2 0 0 k e V H + ( L A S p ) + 1 0 . 7 e V / 1 6 u + 1 5 . 9 e V / 1 6 u + 6 . 2 e V / 1 6 u + 1 8 . 5 e V / 1 6 u 1 2 5 0 c m - 1 ρ ( C H 2 ) H 2 C O 1 3 5 0 c m - 1 δa s ( C H ) C H 3 C H O 1 3 0 5 c m - 1 δa s ( C H ) C H 4 B * * + 4 0 k e V H + ( I A S ) + 6 . 8 e V / 1 6 u + 2 5 . 7 e V / 1 6 u Fig. (1) The CH4 column density is estimated from the analysis of the 1305 cm−1 band; the CH3OH col- umn density is estimated from the 1025 cm−1 band. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 1 0 . 0 2 0 . 0 3 0 . 0 4 0 . 0 5 C H 4 i n H 2 O : C H 3 O H : N H 3 m i x t u r e s + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 N ( C H 4 ) / N ( C H 3 O H ) i 0 10 20 30 40 50 60 70 80 90 100 0 0.01 0.02 0.03 0.04 0.05 CH4 in H2O:CH3OH:NH3 mixtures + 200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N(CH4)/N(CH3OH)i Dose (eV/16u) Model BoxLucas1 Equation y = a(1 - exp(-bx)) Plot N(CH4)/CH3OHi a 0.045 ± 0.00949 b 0.0357 ± 0.01046 Reduced Chi-Sqr 7.37408E-7 R-Square (COD) 0.99356 Adj. R-Square 0.99195 Table 3. CH3CHO formation cross section in H2O:CH3OH:NH3 mix- tures after ion bombardment. Experiment σ (16u/eV) y∞ [gauss] (linear) [gauss] (linear) 3_200_L , 3_200_H [0.037] (0.038) [0.104] (0.041) 1_200_L , 1_200_H [0.061] (0.060) [0.040] (0.022) 3_40 [0.055] (0.052) [0.115] (0.044) 1_40 [0.055] (0.058) [0.051] (0.026) Notes. [gauss] refer to cross section and asymptotic values estimated from the analysis of the 1350 cm−1 band area through gaussian fit; (lin- ear) refers to cross section and asymptotic values estimated from the analysis of the 1350 cm−1 band area after the subtraction of the linear continuum below the band (see Fig. B.2). Table 3. CH3CHO formation cross section in H2O:CH3OH:NH3 mix- tures after ion bombardment. Notes. [gauss] refer to cross section and asymptotic values estimated from the analysis of the 1350 cm−1 band area through gaussian fit; (lin- ear) refers to cross section and asymptotic values estimated from the analysis of the 1350 cm−1 band area after the subtraction of the linear continuum below the band (see Fig. B.2). ( ) Fig. 4. (1) 5 e V / 1 6 u 1 2 5 0 c m - 1 ρ ( C H 2 ) H 2 C O 1 3 5 0 c m - 1 δa s ( C H ) C H 3 C H O 1 3 0 5 c m - 1 δa s ( C H ) C H 4 B * * + 4 0 k e V H + ( I A S ) + 6 . 8 e V / 1 6 u + 2 5 . 7 e V / 1 6 u Fig. 3. 1600–1200 cm−1 spectra of H2O:CH3OH:NH3 mixtures as deposited (panel A) and after irradiation with 40 or 200 keV H+. For a better comparison, spectra were normalized to the most intense peak (marked with an asterisk) between 1478 and 1462 cm−1. In the spectra in panel A the spectral continuum at about 1550 cm−1 is affected by the amount of water in the mixture. The continuum is higher in 3:1:1 mixtures because of the higher amount of water with respect to 1:1:1 mixtures. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 1 0 . 0 2 0 . 0 3 0 . 0 4 0 . 0 5 C H 4 i n H 2 O : C H 3 O H : N H 3 m i x t u r e s + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 N ( C H 4 ) / N ( C H 3 O H ) i D o s e ( e V / 1 6 u ) 0 10 20 30 40 50 60 70 80 90 100 0 0.01 0.02 0.03 0.04 0.05 CH4 in H2O:CH3OH:NH3 mixtures + 200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N(CH4)/N(CH3OH)i Dose (eV/16u) Fig. 4. CH4 abundance relative to the initial amount of methanol, (CH3OH)i with increasing irradiation dose. (1) 3. 1600–1200 cm−1 spectra of H2O:CH3OH:NH3 mixtures as deposited (panel A) and after irradiation with 40 or 200 keV H+. For a better comparison, spectra were normalized to the most intense peak (marked with an asterisk) between 1478 and 1462 cm−1. In the spectra in panel A the spectral continuum at about 1550 cm−1 is affected by the amount of water in the mixture. The continuum is higher in 3:1:1 mixtures because of the higher amount of water with respect to 1:1:1 mixtures. 1 5 5 0 1 5 0 0 1 4 5 0 1 4 0 0 1 3 5 0 1 3 0 0 1 2 5 0 O p t i c a l d e p t h ( a . u . ) W a v e n u m b e r ( c m - 1 ) H 2 O : C H 3 O H : N H 3 m i x t u r e s a s d e p o s i t e d ( n o i r r a d i a t i o n ) 3 _ 2 0 0 _ L , 1 8 K 3 _ 2 0 0 _ H , 1 8 K 1 _ 2 0 0 _ L , 1 7 K 1 _ 2 0 0 _ H , 1 7 K 3 _ 4 0 , 1 5 K 1 _ 4 0 , 1 4 K A * 1 5 5 0 1 5 0 0 1 4 5 0 1 4 0 0 1 3 5 0 1 3 0 0 1 2 5 0 W a v e n u m b e r ( c m - 1 ) + 2 0 0 k e V H + ( L A S p ) + 1 0 . 7 e V / 1 6 u + 1 5 . 9 e V / 1 6 u + 6 . 2 e V / 1 6 u + 1 8 . (1) We used two methods to estimate the 1350 cm−1 band area, as shown in Fig. B.2. We first subtracted a linear continuum between 1330 and 1360 cm−1 and then integrated the band intensity (Fig. B.2, panel A). This method tends to under- estimate the band area because a relevant portion of it is found beneath the linear baseline that is subtracted to exclude other components. Panel A of Fig. 6 shows the results of the estima- tions. As in the case of H2CO, the irradiation of more diluted mixtures results in a higher abundance of CH3CHO. The irradia- tion of 3:1:1 mixtures up to 50 eV/16u determines the production (1) (1) where σ is the cross section, which is found to be 0.34 16u/eV, and D is the dose in eV/16u. About 50% of CH3OH is destroyed after a dose of about 20 eV/16u, and only 10% of CH3OH is left at about 70 eV/16u. The analysis also shows that about 3% of methanol survives the highest irradiation dose. The most intense new features are those attributed to CO2 (2345 cm−1), OCN− (2165 cm−1), and CO (2139 cm−1). Furthermore, Fig. 3 shows a zoom in the 1600–1200 cm−1 range before (panel A) and after irradiation with 40 and 200 keV H+ (panel B). In this range, several new features appear after irradiation. The band centred at 1305 cm−1 is attributed to CH4 that formed after irradiation. Figure 4 shows the CH4 abundance as a function of irradiation dose. For our estimations, we used the CH4 1300 cm−1 band strength given by Mulas et al. (1998). The highest column den- sity (4.6%) is reached after about 50 eV/16u. Figure 4 also shows the CH4 formation σ that we estimate by means of Eq. (2), We estimated the CH3CHO abundance by analysing the 1350 cm−1 band. Various features overlap with this band, in particular the νC-O of HCOO−(Schutte et al. 1999). We there- fore also analysed the 1580 cm−1 band, which is attributed to HCOO−alone (Schutte et al. 1999), and we calculated the 1580/1350 band area ratio in our experiments and in the spec- trum of a H2CO:NH3:HCOOH mixture reported by Vinogradoff et al. (2011), in order to check the relative intensity of the two bands and to verify that the 1350 cm−1 band is not mainly due to HCOO−. Vinogradoff et al. (2011) attributed both bands to HCOO−because their mixtures did not contain CH3CHO. When we use their 1580/1350 ratio as a reference, we would find a similar 1580/1350 band ratio if our mixtures were also to con- tain manly HCOO−. However, our mixtures show averages of the y = y∞(1 −e−σD), (2) y = y∞(1 −e−σD), (2) where y∞is the asymptotic value of the N(CH4)/N(CH3OH)i ratio. Table 2 reports the values of y∞and σ for each experiment. The analysis of the CH3OH destruction and of the CH4 forma- tion does not show differences attributable to the H+ energy used to irradiate samples. A169, page 4 of 10 R. G. 4.1. Chemical changes induced by ion bombardment The irradiation of H2O:CH3OH:NH3 mixtures at 14–18 K deter- mines the formation of CH4 (Fig. 4), a compound observed in various star-forming regions and on frozen surfaces in the outer Solar System. The quantitative analyses show that cross sections and asymptotic values reported in Table 2 are comparable to those reported in literature (Islam et al. 2014). The CH4 abun- dance reaches a plateau above 50 eV/16u, and at these doses, we find 4.5% of CH4 with respect to the initial CH3OH. This abundance represents the highest amount of CH4 that can form on a CH3OH-rich body exposed to irradiation. However, other processes could contribute to form CH4 and thus increase its abundance with respect to CH3OH. In the ISM, hydrogenation of C atoms on the dust grains forms CH4 (e.g. Allen & Robinson 1977). On frozen small bodies, higher CH4 abundances could be attributed to the accretion of pristine CH4 from the preso- lar cloud, during the bodies formation, to the irradiation of other CH4 precursors, or to the more efficient migration of CH4 towards the surface as it is more volatile than CH3OH (e.g. Fray & Schmitt 2009). g Aldehydes are invoked as precursors of complex organics that can form in ices exposed to ion irradiation and thermal pro- cessing in space. In particular, H2CO is a precursor of HMT. Due to the decreasing H2CO abundance that we observe at doses higher than 10 eV/16u, a lower quantity of H2CO would be available in the ice to form HMT during the subsequent warm-up, explaining why the dose affects the final amount of HMT in residues, as reported by Urso et al. (2020b). By warm- ing up the frozen mixtures with a constant heating rate of 3 K min−1, as reported in the literature for similar experiments (e.g. Urso et al. 2020b), after their irradiation up to doses well above this 10 eV/16u value, we do not detect any feature that would be attributable to HMT in the IR spectra. We thus expect astrophysical ices exposed to doses well above the threshold value of 10 eV/16u to be depleted in H2CO and thus they would be less likely to lead to the formation of HMT in space. We also detect H2CO and CH3CHO. 4. Discussion The reactions induced by ion irradiation in ices are due to the transfer of kinetic energy between ions and the target molecules, resulting in their excitation and ionization along the ion track (e.g. Rothard et al. 2017). Figure B.1 shows that for both 40 and 200 keV H+, the interaction occurs mainly through inelas- tic collisions between the impinging keV H+ and the electronic cloud of the target molecules, while the contribution from elastic collisions, that is, the nuclear stopping power S n, can be neglected. The similar values of inelastic stopping power of A169, page 5 of 10 A169, page 5 of 10 A&A 668, A169 (2022) 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 2 0 . 0 4 0 . 0 6 0 . 0 8 0 . 1 0 . 1 2 0 . 1 4 0 . 1 6 N ( H 2 C O ) / N ( C H 3 O H ) D o s e ( e V / 1 6 u ) B 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 1 0 . 0 2 0 . 0 3 0 . 0 4 0 . 0 5 0 . 0 6 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 2 0 . 0 4 0 . 0 6 0 . 0 8 0 . 1 0 . 1 2 0 . 1 4 0 . 4. Discussion 1 6 N ( H 2 C O ) / N ( C H 3 O H ) i D o s e ( e V / 1 6 u ) H 2 C O i n H 2 O : C H 3 O H : N H 3 + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 3:1:1 mixtures: 6.2% 1:1:1 mixtures: 2.4% A N ( H 2 C O ) / N ( C H 3 O H ) D o s e ( e V / 1 6 u ) B Fig. 5. H2CO in H2O:CH3OH:NH3 mixtures exposed to ion bombardment. Estimates are based on the analysis of the 1250 cm−1 band for H2CO and 1025 cm−1 for CH3OH and are normalized to the initial methanol abundance, (CH3OH)i. Panel A: H2CO as a function of dose, with respect to the amount of deposited CH3OH. The grey area shows the range in which the H2CO abundance was found to increase with increasing the dose. The horizontal dashed line indicates the maximum amount of H2CO found in 1:1:1 mixtures (dashed black line) and 3:1:1 mixtures (dashed grey line). Panel B: H2CO with respect to the CH3OH at the same irradiation dose. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 0 1 0 . 0 2 0 . 0 3 0 . 0 4 0 . 0 5 0 . 4. Discussion 0 6 N ( H 2 C O ) / N ( C H 3 O H ) i D o s e ( e V / 1 6 u ) H 2 C O i n H 2 O : C H 3 O H : N H 3 + 2 0 0 k e V H + 3 _ 2 0 0 _ L 3 _ 2 0 0 _ H 1 _ 2 0 0 _ L 1 _ 2 0 0 _ H + 4 0 k e V H + 3 _ 4 0 1 _ 4 0 3:1:1 mixtures: 6.2% 1:1:1 mixtures: 2.4% A A Fig. 5. H2CO in H2O:CH3OH:NH3 mixtures exposed to ion bombardment. Estimates are based on the analysis of the 1250 cm−1 band for H2CO and 1025 cm−1 for CH3OH and are normalized to the initial methanol abundance, (CH3OH)i. Panel A: H2CO as a function of dose, with respect to the amount of deposited CH3OH. The grey area shows the range in which the H2CO abundance was found to increase with increasing the dose. The horizontal dashed line indicates the maximum amount of H2CO found in 1:1:1 mixtures (dashed black line) and 3:1:1 mixtures (dashed grey line). Panel B: H2CO with respect to the CH3OH at the same irradiation dose. ratio because of the efficient formation of H2CO at low dose. At higher dose, the ratio decreases because H2CO is destroyed, while the amount of CH3CHO reaches a plateau. In our exper- iments, the formation of H2CO could be explained by taking the dehydrogenation of CH3OH induced by the impinging H+ into account, in a mechanism similar to that of secondary elec- trons in CH3OH-rich mixtures (Ciaravella et al. 2010). CH3CHO could be formed through the reaction path reported by Bennett et al. (2005), which involves solid-phase CO and CH4, which are both abundant byproducts of CH3OH irradiation. The cleavage of a C-H bond in CH4 would form CH3 and a suprathermal H that would then react with CO, forming the formyl radical HCO. Neighbouring CH3 and HCO would then recombine to form CH3CHO. This mechanism involving CH4 also explains why more quantities of CH3CHO form during irradiation, as the CH4 abundance also increases with the dose. We also find that both H2CO and CH3CHO are formed more efficiently in mixtures richer in H2O. 4. Discussion Although dedicated studies are needed to explain the role of H2O in the production of aldehydes, it is possible that on one hand, the aldehyde formation is independent of the abun- dance of H2O and during irradiation aldehydes would destroy more efficiently simply because they are more highly concen- trated in the samples; on the other hand, H2O could take part in the reactions leading to aldehydes, favouring their formation through a reaction network that includes OH radicals and H. 40 and 200 keV H+ could explain the consistent results we obtain in all the experiments, with changes that can accordingly be attributed to the irradiation dose. However, the H+ energy deter- mines the H+ penetration depth before their implantation in the ice (e.g. Urso et al. 2020a). The higher the H+ energy, the higher the penetration depth. R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space Estimates are based on the analysis of the H2CO 1250 cm−1 and of the CH3CHO 1350 cm−1 bands. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 4 6 1 2 1 4 1 6 1 8 H2CO/CH3CHO in H2O:CH3OH:NH3 +200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N ( H 2 C O ) / N ( C H 3 C H O ) 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 14 16 18 H2CO/CH3CHO in H2O:CH3OH:NH3 +200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N(H2CO)/N(CH3CHO) Dose (eV/16u) 0 10 20 30 40 50 60 70 80 90 100 0 0.01 0.02 0.03 0.04 0.05 0.06 CH3CHO in H2O:CH3OH:NH3 + 200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H + 40 keV H+ 3_40 1_40 N(CH3CHO)/N(CH3OH)i Dose (eV/16u) A Integration from linear baseline Dose (eV/16u) Dose (eV/16u) 0 10 20 30 40 50 60 70 80 90 100 0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 N(CH3CHO)/N(CH3OH)i Dose (eV/16 u) B Integration from gaussian fit 0 5 0 6 0 Dose (eV/16u) D o s e ( e V / 1 6 u ) Fig. 7. H2CO over CH3CHO with respect to dose in H2O:CH3OH:NH3 mixtures. Estimates are based on the analysis of the H2CO 1250 cm−1 and of the CH3CHO 1350 cm−1 bands. 1.3×10−15 s−1 and 1.3×10−17 s−1. With this ζ, icy grain man- tles would accumulate 10 eV/16u within 1.1×106 and 1.1×108 yr, respectively. At this dose, we find the highest abundance of H2CO (Fig. 5), while the CH3OH abundance decreases to about 40% of its initial value. Higher doses determine the destruc- tion of H2CO and at 60 eV/16u its abundance drops to 1% with respect to CH3OH. Considering our most conservative estimates, up to 4% of CH3CHO relative to the initial CH3OH can form within ices exposed to irradiation. This abundance is reached at about 50 eV/16u, which corresponds to 5.5×106–5.5×108 yr. These timescales are comparable to the expected lifetimes of icy mantles in dense molecular clouds and in protoplanetary disks (e.g. Greenberg 1982; Larson 2003). However, when we take the data in Fig. 5 into account, the column density of H2CO detected in icy grain mantles (e.g. Boogert et al. R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space 2015) is higher than the amount that can form after CR irradiation of CH3OH-rich ices, suggesting that other processes contribute to the formation of solid H2CO in icy grain mantles, such as hydrogenation of solid CO (e.g. Watanabe & Kouchi 2002). So far, CH3CHO is only tentatively detected in the line of sight of the high-mass YSO W33A (Schutte et al. 1999). In the near future, this tentative detection could be confirmed by James Webb Space Telescope (JWST) observations that may also include information on the CH3CHO relative abundances and in turn contribute to constrain its formation route (e.g. McClure et al. 2017). 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 N ( C H 3 C H O ) / N ( C H 3 O H ) D o s e ( e V / 1 6 u ) C D o s e ( e V / 1 6 u ) D o s e ( e V / 1 6 u ) Fig. 6. CH3CHO column density with respect to the initial methanol abundance ((CH3OH)i, panels A and B) or during irradiation (CH3OH, panel C). Estimates are based on the analysis of the 1350 cm−1 band for CH3CHO and on the 1025 cm−1 band for CH3OH. Panel A: estimates obtained after the subtraction of a linear baseline between 1360 and 1330 cm−1 and integration of the 1350 cm−1 band intensity. Panel B: estimates obtained from the integration of Gaussian curves centred at about 1350 cm−1 obtained after a multi-Gaussian fit performed in the range 1500–1300 cm−1. Panel C: CH3CHO with respect to the amount of CH3OH during irradiation. The ratio is calculated using the abun- dances of CH3CHO obtained with the method of the integration from linear baseline. ( g ) During the star-formation process, icy mantles sublimate and thus inject volatile species into the gas phase (e.g. Caselli & Ceccarelli 2012). Before sublimation, frozen species can react and form new compounds, however. In particular, the reaction between H2CO and NH3 is the first step of a more complex reaction leading to HMT (e.g. Bernstein et al. 1995; Vinogradoff et al. 2012b). R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space 0 10 20 30 40 50 60 70 80 90 100 0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0 10 20 30 40 50 60 70 80 90 100 0 0.01 0.02 0.03 0.04 0.05 0.06 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 0 0 . 1 0 . 2 0 . 3 0 . 4 0 . 5 N(CH3CHO)/N(CH3OH)i Dose (eV/16 u) B Integration from gaussian fit CH3CHO in H2O:CH3OH:NH3 + 200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H + 40 keV H+ 3_40 1_40 N(CH3CHO)/N(CH3OH)i Dose (eV/16u) A Integration from linear baseline N ( C H 3 C H O ) / N ( C H 3 O H ) D o s e ( e V / 1 6 u ) C Fig. 6. CH3CHO column density with respect to the initial methanol abundance ((CH3OH)i, panels A and B) or during irradiation (CH3OH, panel C). Estimates are based on the analysis of the 1350 cm−1 band for CH3CHO and on the 1025 cm−1 band for CH3OH. Panel A: estimates obtained after the subtraction of a linear baseline between 1360 and 1330 cm−1 and integration of the 1350 cm−1 band intensity. Panel B: estimates obtained from the integration of Gaussian curves centred at about 1350 cm−1 obtained after a multi-Gaussian fit performed in the range 1500–1300 cm−1. Panel C: CH3CHO with respect to the amount of CH3OH during irradiation. The ratio is calculated using the abun- dances of CH3CHO obtained with the method of the integration from linear baseline. 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 4 6 1 2 1 4 1 6 1 8 H2CO/CH3CHO in H2O:CH3OH:NH3 +200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N ( H 2 C O ) / N ( C H 3 C H O ) D o s e ( e V / 1 6 u ) 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 14 16 18 H2CO/CH3CHO in H2O:CH3OH:NH3 +200 keV H+ 3_200_L 3_200_H 1_200_L 1_200_H +40 keV H+ 3_40 1_40 N(H2CO)/N(CH3CHO) Dose (eV/16u) Fig. 7. H2CO over CH3CHO with respect to dose in H2O:CH3OH:NH3 mixtures. R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space The IR spectrum of HMT shows two intense and narrow bands at 1007 and 1234 cm−1 (Bernstein et al. 1994). Although the detection of the 1005 band would be hindered by the silicate feature at the same wavelength, the 1234 cm−1 band might be detected. However, even if the temperature would increase enough to allow the HMT forma- tion, we estimate that the 1234 cm−1 HMT band would show an optical depth peak of about 2×10−2–4×10−3 if we were to assume that 10–40% of NH3 observed in interstellar ices would react to form HMT and if we used the spectral param- eters reported by Bernstein et al. (1994, 1995). This value is below the detection limits of the instruments on board the JWST 4.1. Chemical changes induced by ion bombardment The highest amount of H2CO with respect to CH3OH, that is, 6%, is observed after irra- diation of 3:1:1 mixtures up to about 10 eV/16u, while only 2% of H2CO is found in 1:1:1 mixtures. In all mixtures, the H2CO abundance decreases at doses higher than 10 eV/16u, reveal- ing that ices exposed to a higher extent of irradiation will be depleted in H2CO. Taking into account the most conservative approach, we find up to 4% of CH3CHO with respect to the initial CH3OH abundance. The CH3CHO abundance could be as high as 12% by considering the data obtained through the multi-Gaussian fit of the 1350 cm−1 band. In Fig. 7 we show the ratio of the two aldehydes, H2CO/CH3CHO as a function of the dose. Ices exposed to a low extent of irradiation show a high A169, page 6 of 10 4.2. Astrophysical implications and timescales of irradiation in space In this section, we estimate the timescales necessary to form H2CO and CH3CHO in ices exposed to energetic H+ on icy grain mantles and on frozen surfaces in the outer Solar System. To obtain the timescale of irradiation on icy grain mantles, we followed the estimations reported by Kaˇnuchová et al. (2016) and the approximation for the energy of cosmic ions reported by Mennella et al. (2003). Timescales of irradiation in dense molecular clouds, protostars, and protoplanetary disks can be evaluated from the measured cosmic-ray ionization rate ζ (s−1). For our calculation, we considered the lowest and highest val- ues of ζ estimated in different sources in the ISM reported by Mennella et al. (2003), Podio et al. (2014), Woods et al. (2015) and Kaˇnuchová et al. (2016). These values vary between A169, page 7 of 10 A169, page 7 of 10 A&A 668, A169 (2022) (Wells et al. 2015). However, the presence of HMT in the early Solar System and its incorporation in the meteoritic par- ent bodies could explain the presence of such compounds in carbonaceous chondrites, although also hydrothermal aqueous alteration in asteroids may contribute to the formation of HMT Oba et al. (2020). the age of the Solar System. Deeper layers are less strongly irra- diated, but about 2 eV/16u can cumulate because radioactive elements decay within the nucleus (e.g. Draganic & Draganic 1984; Modica et al. 2012). Observational studies showed that the abundance of complex molecules relative to water in cometary comae is comparable (with differences smaller than one order of magnitude) to the abundances observed towards YSOs, supporting the hypothe- sis that comets inherited ices from the presolar cloud or that cometary ices were processed in the same physical and chem- ical ways as interstellar ices (e.g. Bockelée-Morvan et al. 2019, and references therein). Various comets show an abundance of H2CO that is comparable to that of CH3OH (e.g. le Roy et al. 2015; Schuhmann et al. 2019). Thus, other processes might con- tribute to the presence of H2CO in addition to its formation from the CH3OH-rich ices irradiation. On the other hand, the abundance ratio CH3CHO/CH3OH ranges between 0.01 and 0.2, which means that CH3CHO might form after cosmic ion irradia- tion of CH3OH-rich ices. However, CH3CHO has been detected only towards a limited number of comets (e.g. le Roy et al. 2015; Bockelée-Morvan et al. 2019). 4.2. Astrophysical implications and timescales of irradiation in space The capabilities of the JWST will enable future observations to contribute to better constrain the origin of complex molecules in comets as well as on other frozen bodies located in the outer Solar System. Since their formation, frozen bodies in the outer Solar Sys- tem are exposed to SEP and to CR, although the shielding of the heliosphere strongly limits the flux of CR with energies lower than a few hundred MeV (e.g. Langner et al. 2003). CH3OH is detected on 5145 Pholus (semimajor axis, S.A. 20.36 au), 2002 VE95 (S.A. 39.18 au), and 486958 Arrokoth (S.A. 44.58 au), but neither solid-phase H2CO nor CH3CHO are detected on frozen surfaces in the Solar System. By scaling the values reported by Strazzulla et al. (2003), who estimated the timescales of irradia- tion at various locations in the outer Solar System, we find that a frozen body located at 40 au accumulated about 65 eV/16u at surface depth of 1 µm and about 46 eV/16u at surface depth of 1 cm since its formation 4.6 billion years ago. Deeper layers, however, experience decreasing dose values. p g The missing detection of HMT in CH3OH-rich bodies could be explained by the fact that IR photons only probe limited thick- nesses of the surface and HMT would only form in deeper layers, where the lower dose allows its precursors to form and survive. In the outer Solar System, we estimate that energetic ions can induce the formation of the highest abundance of H2CO and CH3CHO on timescales of about 106 yr (Urso et al. 2020b). These timescales imply a fast and efficient conversion of CH3OH into its byproducts, including CH4 and aldehydes. However, bod- ies in the Kuiper belt are thought to have remained largely undisturbed since their formation (e.g. Grundy et al. 2020). Dur- ing their lifetime, these bodies might have accumulated doses up to 40–70 eV/16u by CR and much higher doses at smaller surface depth by SEP. These doses would destroy not only the complex organic precursors, but also volatiles species that are observed on frozen surfaces, however, including CH3OH (e.g. Barucci et al. 2011). Nevertheless, our analysis on the irradiation of CH3OH shows that its destruction reaches a plateau and that about 3% of CH3OH survives at 100 eV/16u. 5. Conclusions We exposed H2O, CH3OH, and NH3 to 40 and 200 keV H+ to simulate the irradiation of frozen surfaces in both the ISM and in the Solar System. We followed the destruction of CH3OH and the formation of its byproducts CH4, H2CO, and CH3CHO. We find that the composition of ices depends on the irradia- tion dose and that the abundance of aldehydes is affected by the amount of water in mixtures. In all mixtures, the H2CO abundance decreased at doses higher than 10 eV/16u, thus ices exposed to a higher dose would be depleted in such a compound. In turn, lower abundances of HMT might form in the refractory organic material that is produced after irradiation and warm-up of ices due to the role of H2CO as a precursor of HMT. Frozen volatile compounds such as CH3OH, aldehydes, and complex organics might be buried below the surface that is probed by astronomical observations. Volatile compounds might survive below surfaces and the irradiation by CR would induce the partial destruction of methanol and the formation of its byproducts, including aldehydes, which might then take part in reactions to form more complex compounds. p p Further data are needed to better constrain the fate of organ- ics after their formation, to understand how radiation affects their chemical properties, and to determine whether they can survive the harsh conditions on atmosphere-less surfaces. In this regard, exposure platforms on board space facilities allow the exposure to the full solar spectrum as well as to higher fluxes of galac- tic CR and solar particles (e.g. Baratta et al. 2015, 2019; Cottin et al. 2017; Bryson et al. 2015). The Organics Exposure in Orbit (OREOcube; Elsaesser et al. 2014) and Exocube (Sgambati et al. 2020) facilities will allow exposing various organic compounds of relevance for astrobiology as pure and in contact with min- eral surfaces and gaseous compounds, allowing us to simulate the physico-chemical conditions of various astrophysical sce- narios. Both platforms will allow the in situ analysis by means of UV-Visible (OREOcube) or mid-IR (Exocube) spectroscopy, allowing to monitor the sample evolution during flight. At the end of the mission, samples will be brought back to Earth allow- ing their further characterization. Both experiments will thus Cometary comae also show H2O, CH3OH, NH3, H2CO, and CH3CHO along with various other compounds (e.g. Bockelée-Morvan et al. 2000; le Roy et al. 2015, and references therein). 4.2. Astrophysical implications and timescales of irradiation in space Thus, lower quan- tities of CH3OH could survive the high extent of irradiation on small bodies, and aldehydes could still form on highly irradiated surfaces. Another consideration regards the fact that IR photons that are detected by remote-sensing observations and that inform on the composition of frozen surfaces can probe larger thick- nesses than the penetration depth of SEP (Urso et al. 2020a). Low-energy SEP alter a lower thickness than high-energy SEP and CR. IR photons might travel through both layers strongly altered by SEP and underlying layers that are only affected by CR and high-energy SEP. These buried layers, exposed to much lower doses, can still contain high quantities of CH3OH and its irradiation byproducts, such as aldehydes, which might then take part in reactions to form complex organics. Furthermore, reju- venating processes have to be taken into account as they cause the exposure of fresh and less altered materials buried below frozen surfaces. These unprocessed ices would then undergo ion bombardment and form byproducts within the short timescales reported above. 5. Conclusions MEP, GAB, and CS acknowledge the financial support by the Italian Ministero dell’Istruzione, Università e Ricerca through the grant Progetti Premiali – iALMA. Langner, U. W., Potgieter, M. S., & Webber, W. R. 2003, J. Geophys. Res. Space Phys., 108, 8039 Larson, R. B. 2003, Rep. Prog. Phys., 66, 1651 Le Roy, L., Altwegg, K., Balsiger, H., et al. 2015, A&A, 583, A2 Le Roy, L., Altwegg, K., Balsiger, H., et al. 2015 McClure, M., Bailey, J., Beck, T., et al. 2017, IceAge: Chemical Evolution of Ices during Star Formation, JWST Proposal ID 1309. Cycle 0 Early Release Science Meinert, C., Myrgorodska, I., de Marcellus, P., et al. 2016, Science, 352, 208 Mennella, V., Baratta, G. A., Esposito, A., Ferini, G., & Pendleton, Y. J. 2003, ApJ, 587, 727 g, 716, 825 Palumbo, M. E. 2006, A&A, 453, 903 Bockelée-Morvan, D., Leyrat, C., Erard, S., et al. 2019, A&A, 630, A22 Bockelée-Morvan, D., Leyrat, C., Erard, S., et al. 2019, A&A, 630, A22 Bockelée-Morvan D Lis D C Wink J E et al 2000 A&A 353 1101 Palumbo, M. E., Geballe, T. R., & Tielens, A. G. G. M. 1997, A Palumbo, M. E., Geballe, T. R., & Tielens, A. G. G. M. 1997, ApJ, 479, 839 Palumbo, M. E., Castorina, A. C., & Strazzulla, G. 1999, A&A, 342, 551 M. E., Geballe, T. R., & Tielens, A. G. G. M. 1997, ApJ, 479, 839 kelée-Morvan, D., Lis, D. C., Wink, J. E., et al. 2000, A&A, 353, Boogert, A. C. A., Gerakines, P. A., & Whittet, D. C. B. 2015, ARA&A, 53, 541l Palumbo, M. E., Castorina, A. C., & Strazzulla, G. 1999, A&A, 342, 551 Parker, E. T., Chan, Q. H. S., Glavin, D. P., & Dworkin, J. P. 2022, Meteor. Planet. Sci., 57, 776 Bottinelli, S., Ceccarelli, C., Williams, J. P., & Lefloch, B. 2007, A&A, 463, 601 ouilloud, M., Fray, N., Bénilan, Y., et al. 2015, MNRAS, 451, 2145 Pilorget, C., Okada, T., Hamm, V., et al. 2021, Nat. Astron., 6, 221l Brown, M. E., Schaller, E. L., & Fraser, W. C. 2011, ApJ, 739, L60 Brucato, J. R., Baratta, G. A., & Strazzulla, G. 2006, A&A, 455, 395 Podio, L., Lefloch, B., Ceccarelli, C., Codella, C., & Bachiller, R. 2014, A&A, 565, A64 Brunetto, R., Barucci, M. A., Dotto, E., & Strazzulla, G. 2006, ApJ, 644, 646 Poston, M. J., Mahjoub, A., Ehlmann, B. L., et al. 2018, ApJ, 856, 124 Poston, M. J., Mahjoub, A., Ehlmann, B. L., et al. 2018, ApJ, 856, 124 Raunier S Chiavassa T Duvernay F et al 2004 A&A 416 165 Bryson, K. L., Salama, F., Elsaesser, A., et al. 2015, Int. J. Astrobiol., 14, 55 Poston, M. J., Mahjoub, A., Ehlmann, B. L., et al. 2018, ApJ, 856, 124 Raunier, S., Chiavassa, T., Duvernay, F., et al. 2004, A&A, 416, 165 j p Raunier, S., Chiavassa, T., Duvernay, F., et al. 2004, A&A, 416, 165 Busemann, H., Young, A. F., O’D. Alexander, C. M., et al. 2006, Science, 312, 727 Rothard, H., Domaracka, A., Boduch, P., et al. 2017, J. Phys. B, 50, 062011 Caselli, P., & Ceccarelli, C. 2012, A&ARv, 20, 56 Schmitt-Kopplin, P., Gabelica, Z., Gougeon, R. References Mewaldt, R. A., Cohen, C. M. S., Mason, G. M., Haggerty, D. K., & Desai, M. I. 2007, Space Sci. Rev., 130, 323 p Mispelaer, F., Theulé, P., Aouididi, H., et al. 2013, A&A, 555, A13 Allen, M., & Robinson, G. W. 1977, ApJ, 212, 396 Mispelaer, F., Theulé, P., Aouididi, H., et al. 2013, A&A, 555, A13 Mitchell, D. L., Lin, R. P., Anderson, K. A., et al. 1987, Science, 237, 626 Mitchell, D. L., Lin, R. P., Anderson, K. A., et al. 1987, Science, 237, 626 Altwegg, K., Balsiger, H., Berthelier, J. J., et al. 2017, MNRAS, 469, S130 Altwegg, K., Balsiger, H., Berthelier, J. J., et al. 2017, MNRAS, 469, S130 Altwegg, K., Balsiger, H., & Fuselier, S. A. 2019, ARA&A, 57, 113 Altwegg, K., Balsiger, H., Berthelier, J. J., et al. 2017, MNRAS, 469, S130 Altwegg, K., Balsiger, H., & Fuselier, S. A. 2019, ARA&A, 57, 113 Modica, P., Palumbo, M. E., & Strazzulla, G. 2012, Planet. Space Sci., 73, 425 Modica, P., Palumbo, M. E., & Strazzulla, G. 2012, Planet. Space Sci., 73, 425 Moore, M. H., & Hudson, R. L. 1998, Icarus, 135, 518 Altwegg, K., Balsiger, H., & Fuselier, S. A. 2019, ARA&A, 57, 113 Moore, M. H., & Hudson, R. L. 1998, Icarus, 135, 518 Baratta, G. A., Chaput, D., Cottin, H., et al. 2015, Planet. Space Sci., 118, 211 Moore, M. H., Ferrante, R. F., & Nuth, J. A., I. 1996, Planet. Space Sci., 44, 927 Moore, M. H., Ferrante, R. F., & Nuth, J. A., I. 1996, Plan Baratta, G. A., Accolla, M., Chaput, D., et al. 2019, Astrobiology, 19, 1018 ci, M. A., & Merlin, F. 2020, in The Trans-Neptunian Solar Syste Mulas, G., Baratta, G. A., Palumbo, M. E., & Strazzulla, G. 1998, A&A, 333, 1025 Barucci, M. A., & Merlin, F. 2020, in The Trans-Neptunian Solar System, eds D. Prialnik, M. A. Barucci, & L. Young (Amsterdam: Elsevier), 109 D. Prialnik, M. A. Barucci, & L. Young (Amsterdam: Elsevier), 10 Nuevo, M., Auger, G., Blanot, D., & D’Hendecourt, L. 2008, Origins Life Evol. Biosph., 38, 37 g Barucci, M. A., Alvarez-Candal, A., Merlin, F., et al. 2011, Icarus, 214, 297 Bennett, C. J., Jamieson, C. S., Osamura, Y., & Kaiser, R. I. 2005, ApJ, 624, 1097 Nuevo, M., Cooper, G., & Sandford, S. A. 2018, Nat. Commun., 9, 5276 Oba, Y., Takano, Y., Naraoka, H., et al. Spectrochim. Acta A Mol. Biomo p p Greenberg, J. M. 1982, IAU Colloq., 61, 131 Wells, M., Pel, J. W., Glasse, A., et al. 2015, PASP, 127, 646 5. Conclusions Oort cloud comets are exposed to high fluxes of CR that induce the formation of organic refractory materials, and Strazzulla & Johnson (1991) estimated that organic refractory materials form in the upper 1–2 m of cometary surfaces within A169, page 8 of 10 . G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space R. G. Urso et al.: Ion irradiation triggers the formation of the precursors of complex organics in space provide information about the survival of organics and about the possibility that complex molecules formed after the processing of solid-phase simple volatiles can be trapped in small bodies and delivered on planetary surfaces. Hulett, H. R., Wolman, Y., Miller, S. L., et al. 1971, Science, 174, 1038 Islam, F., Baratta, G. A., & Palumbo, M. E. 2014, A&A, 561, A73 Hulett, H. R., Wolman, Y., Miller, S. L., et al. 1971, Science, 174, 1038 Islam, F., Baratta, G. A., & Palumbo, M. E. 2014, A&A, 561, A73 Jaber, A. A., Ceccarelli, C., Kahane, C., & Caux, E. 2014, ApJ, 791, 29 Jaber, A. A., Ceccarelli, C., Kahane, C., & Caux, E. 2014, ApJ, 79 Johnson, T. V. 2005, Space Sci. Rev., 116, 401 , , p , , Kaˇnuchová, Z., Urso, R. G., Baratta, G. A., et al. 2016, A&A, 585, A155 Khare, B. N., Thompson, W. R., Chyba, C. F., Arakawa, E. T., & Sagan, C. 1989, Adv. Space Res., 9, 41 Acknowledgements. We thank the referee, Dr. V. Vinogradoff for the useful comments to improve the manuscript. AE and RGU acknowledges the Einstein Foundation Berlin (grant IPF-2018-469). INGMAR is a IAS-IJCLab facility funded by the French Programme National de Planétologie (PNP), Faculté des Sciences d’Orsay, Université Paris-Sud (Attractivité 2012), P2IO LabEx (ANR-10-LABX-0038) in the framework Investissements d’Avenir (ANR-11- IDEX-0003-01). MEP, GAB, and CS acknowledge the financial support by the Italian Ministero dell’Istruzione, Università e Ricerca through the grant Progetti Premiali – iALMA. Acknowledgements. We thank the referee, Dr. V. Vinogradoff for the useful comments to improve the manuscript. AE and RGU acknowledges the Einstein Foundation Berlin (grant IPF-2018-469). INGMAR is a IAS-IJCLab facility funded by the French Programme National de Planétologie (PNP), Faculté des Sciences d’Orsay, Université Paris-Sud (Attractivité 2012), P2IO LabEx (ANR-10-LABX-0038) in the framework Investissements d’Avenir (ANR-11- IDEX-0003-01). y Phys. Res. A, 521, 149 2015, A&A, 576, A125 de Marcellus, P., Fresneau, A., Brunetto, R., et al. 2017, MNRAS, 464, 114 Delsanti A Merlin F Guilbert-Lepoutre A et al 2010 A&A 520 A40 j ApJ, 894, L3 g de Marcellus, P., Fresneau, A., Brunetto, R., et al. 2017, MNRAS, 464, Urso, R. G., Vuitton, V., Danger, G., et al. 2020b, A&A, 644, A115 Delsanti, A., Merlin, F., Guilbert-Lepoutre, A., et al. 2010, A&A, 520, A40 Dougherty, M. K., & Spilker, L. J. 2018, Rep. Prog. Phys., 81, 065901 Vinogradoff, V., Duvernay, F., Danger, G., Theulé, P., & Chiavassa, T. 2011, A&A, 530, A128 Dougherty, M. K., & Spilker, L. J. 2018, Rep. Prog. Phys., 81, 065901 Draganic, I. G., & Draganic, Z. D. 1984, Adv. Space Res., 4, 115 Duvernay F Dufauret V Danger G et al 2010 A&A 523 A79 g y p p g y Draganic, I. G., & Draganic, Z. D. 1984, Adv. Space Res., 4, 115 Vinogradoff, V., Duvernay, F., Farabet, M., et al. 2012a, J. Phys. Chem. A, 116, 2225 Duvernay, F., Dufauret, V., Danger, G., et al. 2010, A&A, 523, A79 Elsaesser, A., Quinn, R. C., Ehrenfreund, P., et al. 2014, Langmuir, 30, 13217 Fray, N., & Schmitt, B. 2009, Planet. Space Sci., 57, 2053 Elsaesser, A., Quinn, R. C., Ehrenfreund, P., et al. 2014, Langmuir, 30, 13217 Elsaesser, A., Quinn, R. C., Ehrenfreund, P., et al. 2014, Langmuir, 30 Fray, N., & Schmitt, B. 2009, Planet. Space Sci., 57, 2053 Vinogradoff, V., Rimola, A., Duvernay, F., et al. 2012b, Phys. Chem. Chem. Phys Incorporating Faraday Transac 14 12309 Vinogradoff, V., Rimola, A., Duvernay, F., et al. 2012b, Phys. Chem. Chem. Vinogradoff, V., Rimola, A., Duvernay, F., et al. 2012b, Phys. Chem. Chem. Phys. Incorporating Faraday Transac., 14, 12309 Fray, N., & Schmitt, B. 2009, Planet. Space Sci., 57, 2053 Fresneau, A., Abou Mrad, N., d’Hendecourt, L. L., et al. 201 Fresneau, A., Abou Mrad, N., d’Hendecourt, L. L., et al. 2017, ApJ, 837, 168 Phys. Incorporating Faraday Transac., 14, 12309 Vinogradoff, V., Fray, N., Duvernay, F., et al. 2013, A&A, 551, A12 Fulvio, D., Sivaraman, B., Baratta, G. A., Palumbo, M. E., & , , , , , , , , Spectrochim. Acta A Mol. Biomol. Spectros., 72, 1007 Watanabe, N., & Kouchi, A. 2002, ApJ, 571, L173 g, 716, 825 D., et al. 2010, Proc. Natl. Acad. Sci., 107, 2763 Cazaux, S., Tielens, A. G. G. M., Ceccarelli, C., et al. 2003, ApJ, 593, L51 Ch Q H S F hi I A Zh X l 2020 M Pl S i 55 Cazaux, S., Tielens, A. G. G. M., Ceccarelli, C., et al. 2003, ApJ, Cazaux, S., Tielens, A. G. G. M., Ceccarelli, C., et al. 2003, ApJ, 593, L51 Chan, Q. H. S., Franchi, I. A., Zhao, X., et al. 2020, Meteor. Planet. Sci., 55, 1320 Schuhmann, M., Altwegg, K., Balsiger, H., et al. 2019, ACS Earth Space Chem., 3, 1854 Chan, Q. H. S., Franchi, I. A., Zhao, X., et al. 2020, Meteor. Planet. Sci., 55, 1320 Schutte, W. A., Boogert, A. C. A., Tielens, A. G. G. M., et al. 1999, A&A, 343, 966 Chan, Q. H. S., Stephant, A., Franchi, I. A., et al. 2021, Sci. Rep., 11, 5125 Chauvin, N., Dayras, F., Le Du, D., & Meunier, R. 2004, Nucl. Instrum. Methods Phys. Res. A, 521, 149 Sephton, M. A. 2002, Natural Product Rep., 19, 292 References 2020, Nat. Commun., 11, 6243 Nuevo, M., Cooper, G., & Sandford, S. A. 2018, Nat. Commun., 9, 5276 Oba, Y., Takano, Y., Naraoka, H., et al. 2020, Nat. Commun., 11, 6243 Bernstein, M. P., Sandford, S. A., Allamandola, L. J., & Chang, S. 1994, J. Phys. Chem., 98, 12206 Oba, Y., Takano, Y., Furukawa, Y., et al. 2022, Nat. Commun., 13, 2008 Öberg, K. I., Bottinelli, S., Jørgensen, J. K., & van Dishoeck, E. F. 2010, ApJ, 716, 825 Bernstein, M. P., Sandford, S. A., Allamandola, L. J., Chang, S., & Scharberg, M. A. 1995, ApJ, 454, 327 y Phys. Res. A, 521, 149 p p Sgambati, A., Deiml, M., Stettner, A., et al. 2020, Acta Astron., 166, 377 y Ciaravella, A., Muñoz Caro, G., Jiménez Escobar, A., et al. 2010, ApJ, 722, L45 Sgambati, A., Deiml, M., Stettner, A., et al. 2020, Ac Spencer, J. R., & Nimmo, F. 2013, Ann. Rev. Earth Planet. Sci., 41, 693 Codella, C., Fontani, F., Ceccarelli, C., et al. 2015, MNRAS, 449, L11 Strazzulla, G., & Johnson, R. E. 1991, Irradiation Effects on Comets and Cometary Debris, Comets in the Post-Halley Era (Dordrecht: Springer Netherlands) Cottin, H., Szopa, C., & Moore, M. H. 2001, ApJ, 561, L139 Cottin, H., Kotler, J. M., Billi, D., et al. 2017, Space Sci. Rev., 209, 83 Crovisier, J., Bockelée-Morvan, D., Colom, P., et al. 2004, A&A, 418, 1141 Strazzulla, G., Pirronello, V., & Foti, G. 1983, A&A, 123, 93 S ll G C Ch i i & h ruikshank, D. P., Roush, T. L., Bartholomew, M. J., et al. 1998, Icar Dalle Ore, C. M., Fulchignoni, M., Cruikshank, D. P., et al. 2011, A&A, 533, A98 Strazzulla, G., Cooper, J. F., Christian, E. R., & Johnson, R. E. 2003, Comptes Rendus Phys., 4, 791 Danger, G., Orthous-Daunay, F. R., de Marcellus, P., et al. 2013, Geochim. Cosmochim. Acta, 118, 184 Theulé, P., Duvernay, F., Danger, G., et al. 2013, Adv. Space Res., 52, 1567 Urso, R. G., Scirè, C., Baratta, G. A., Compagnini, G., & Palumbo, M. E. 2016, Theulé, P., Duvernay, F., Danger, G., et al. 2013, Adv. Space Res., 52, 1567 Urso, R. G., Scirè, C., Baratta, G. A., Compagnini, G., & Palumbo, M. E. 2016, A&A, 594, A80 Urso, R. G., Scirè, C., Baratta, G. A., Compagnini, G., & Palumbo, M. E. 2016, A&A, 594, A80 , , , , , , p g , , , A&A, 594, A80 Danger, G., Fresneau, A., Abou Mrad, N., et al. 2016, Geochim. Cosmochim. Acta, 189, 184 Urso, R. G., Palumbo, M. E., Ceccarelli, C., et al. 2019, A&A, 628, A72 Dartois, E., Augé, B., Boduch, P., et al. 2015, A&A, 576, A125 Urso, R. G., Baklouti, D., Djouadi, Z., Pinilla-Alonso, N., & Brunetto, R. 2020a, ApJ, 894, L3 Urso, R. G., Baklouti, D., Djouadi, Z., Pinilla-Alonso, N., & Brunetto, R. 2020a A J 894 L3 Urso, R. G., Baklouti, D., Djouadi, Z., Pinilla-Alonso, N., & Brunetto, R. 2 ApJ, 894, L3 Dartois, E., Augé, B., Boduch, P., et al. p p Greenberg, J. M. 1982, IAU Colloq., 61, 131 Woods, P., Occhiogrosso, A., Viti, S., et al. 2015, MNRAS, 450, 1256 Ziegler, J. F., Biersack, J. P., & Ziegler, M. D. 2008, The Stopping and Range of Ions in Solids (New York: Pergamon Press) Woods, P., Occhiogrosso, A., Viti, S., et al. 2015, MNRAS, 450, Woods, P., Occhiogrosso, A., Viti, S., et al. 2015, MNRAS, 450, 1256 Ziegler, J. F., Biersack, J. P., & Ziegler, M. D. 2008, The Stopping and Range of Grundy, W. M., Binzel, R. P., Buratti, B. J., et al. 2016, Science, 351, aad9189 Grundy, W. M., Bird, M., Britt, D., et al. 2020, Science, 367, eaay3705 Woods, P., Occhiogrosso, A., Viti, S., et al. 2015, MNRAS, 450, 1256 Ziegler, J. F., Biersack, J. P., & Ziegler, M. D. 2008, The Stopping and Range of y, , , , , , , , , Grundy, W. M., Bird, M., Britt, D., et al. 2020, Science, 367, eaay3705 Ziegler, J. F., Biersack, J. P., & Ziegler, M. D. 2008, The Stopping and Range of Ions in Solids (New York: Pergamon Press) iegler, J. F., Biersack, J. P., & Ziegler, M. D. 2008, The Stopping Ions in Solids (New York: Pergamon Press) Hudson, R. L., & Ferrante, R. F. 2020, MNRAS, 492, 283 A169, page 9 of 10 A169, page 9 of 10 A169, page 9 of 10 A&A 668, A169 (2022) 1400 1375 1350 1325 1300 1275 0 0.02 0.04 1400 1375 1350 1325 1300 1275 0 0.02 0.04 Optical depth Wavenumber (cm-1) A 1350 H2O:CH3OH:NH3=1:1:1 +200 keV H+ up to 18.5 eV/16u Integration in linear fit mode Optical depth Wavenumber (cm-1) B 1350 Integration of the gaussian curve Fig. B.2. Estimates of the 1350 cm−1 band area in a H2O:CH3OH:NH3=1:1:1 mixture bombarded with 200 keV H+. Panel A: Integration of the optical depth after the subtraction of a linear continuum between about 1360 and 1330 cm−1; Panel B: Multi-Gaussian fit between 1450-1250 cm−1 and estimate of the 1350 cm−1 band area from the integration of the optical depth of the synthetic Gaussian curve. 1400 1375 1350 1325 1300 1275 0 0.02 0.04 Optical depth Wavenumber (cm-1) A 1350 H2O:CH3OH:NH3=1:1:1 +200 keV H+ up to 18.5 eV/16u Integration in linear fit mode Appendix A: H+ stopping power H2O:CH3OH:NH3=1:1:1 +200 keV H+ up to 18.5 eV/16u Figure B.1 shows the simulations we ran with SRIM to estimate the stopping power, that is, the energy released by the impinging ion per unit path length, of 40 and 200 keV H+ impinging on H2O:CH3OH:NH3 mixtures. H+ release their energy via inelas- tic collisions with the electronic cloud of the target molecules (electronic stopping power, Se), while the energy released by elastic collisions with the nuclei of the target molecules (nuclear stopping power, Sn) is negligible. Appendix B: Estimates of the 1350 cm−1 band We here give an example of the two methods we used to estimate the 1350 cm−1 band area in the spectra of H2O:CH3OH:NH3 mixtures after irradiation with 40 or 200 keV H+. Panel A shows the estimate of the 1350 cm−1 band area after the subtraction of a linear continuum between about 1360 and 1330 cm−1. Panel B shows the multi-Gaussian fit we used to model the same spectrum as reported in panel A. The dotted red line shows the modelled spectrum. In both panels, the grey area represents the integrated optical depth of the 1350 cm−1 band taken into account for our estimates. Wavenumber (cm-1) 1400 1375 1350 1325 1300 1275 0 0.02 0.04 Optical depth Wavenumber (cm-1) B 1350 Integration of the gaussian curve B 1 10 100 1000 0 20 40 60 80 Stopping power (keV µm-1) H+ energy (keV) Electronic and nuclear stopping power of H+ in H2O:CH3OH:NH3 mixtures Se in 3:1:1 Sn in 3:1:1 Se in 1:1:1 Sn in 1:1:1 40 keV 200 keV Fig. B.1. Elastic and nuclear stopping power (Bragg curves) of protons in H2O:CH3OH:NH3=3:1:1 and 1:1:1 frozen mixtures as calculated by means of the SRIM software (Ziegler et al. 2008). The Sn curves overlap and cannot be distinguished on the given scale. 1400 1375 1350 1325 1300 1275 0 0.02 Optical depth Wavenumber (cm-1) 1350 Fig. B.2. Estimates of the 1350 cm−1 band area in a H2O:CH3OH:NH3=1:1:1 mixture bombarded with 200 keV H+. Panel A: Integration of the optical depth after the subtraction of a linear continuum between about 1360 and 1330 cm−1; Panel B: Multi-Gaussian fit between 1450-1250 cm−1 and estimate of the 1350 cm−1 band area from the integration of the optical depth of the synthetic Gaussian curve. 1 10 100 1000 0 20 40 60 80 Stopping power (keV µm-1) H+ energy (keV) Electronic and nuclear stopping power of H+ in H2O:CH3OH:NH3 mixtures Se in 3:1:1 Sn in 3:1:1 Se in 1:1:1 Sn in 1:1:1 40 keV 200 keV Fig. B.1. Elastic and nuclear stopping power (Bragg curves) of protons in H2O:CH3OH:NH3=3:1:1 and 1:1:1 frozen mixtures as calculated by means of the SRIM software (Ziegler et al. 2008). The Sn curves overlap and cannot be distinguished on the given scale. Appendix B: Estimates of the 1350 cm−1 band 1 10 100 1000 0 20 40 60 80 Stopping power (keV µm-1) H+ energy (keV) Electronic and nuclear stopping power of H+ in H2O:CH3OH:NH3 mixtures Se in 3:1:1 Sn in 3:1:1 Se in 1:1:1 Sn in 1:1:1 40 keV 200 keV Wavenumber (cm-1) Fig. B.2. Estimates of the 1350 cm−1 band area in a H2O:CH3OH:NH3=1:1:1 mixture bombarded with 200 keV H+. Panel A: Integration of the optical depth after the subtraction of a linear continuum between about 1360 and 1330 cm−1; Panel B: Multi-Gaussian fit between 1450-1250 cm−1 and estimate of the 1350 cm−1 band area from the integration of the optical depth of the synthetic Gaussian curve. H+ energy (keV) H+ energy (keV) Fig. B.1. Elastic and nuclear stopping power (Bragg curves) of protons in H2O:CH3OH:NH3=3:1:1 and 1:1:1 frozen mixtures as calculated by means of the SRIM software (Ziegler et al. 2008). The Sn curves overlap and cannot be distinguished on the given scale. A169, page 10 of 10 A169, page 10 of 10
https://openalex.org/W3107809640	https://link.springer.com/content/pdf/10.1007/s00441-020-03329-z.pdf	English	null	Plasticity and modulation of olfactory circuits in insects	Cell & tissue research/Cell and tissue research	2,020	cc-by	14,575	ERROR: type should be string, got "https://doi.org/10.1007/s00441-020-03329-z\nCell and Tissue Research (2021) 383:149–164 https://doi.org/10.1007/s00441-020-03329-z\nCell and Tissue Research (2021) 383:149–164 REVIEW REVIEW Keywords Antenna . Antennal lobe . Mushroom body . Neuromodulation . Structural synaptic plasticity Keywords Antenna . Antennal lobe . Mushroom body . Neuromodulation . Structural synaptic plasticity Sylvia Anton1\n& Wolfgang Rössler2 Sylvia Anton1\n& Wolfgang Rössler2 Received: 12 August 2020 /Accepted: 27 October 2020\n# The Author(s) 2020\n/ Published online: 4 December 2020 Abstract Olfactory circuits change structurally and physiologically during development and adult life. This allows insects to respond to\nolfactory cues in an appropriate and adaptive way according to their physiological and behavioral state, and to adapt to their\nspecific abiotic and biotic natural environment. We highlight here findings on olfactory plasticity and modulation in various\nmodel and non-model insects with an emphasis on moths and social Hymenoptera. Different categories of plasticity occur in the\nolfactory systems of insects. One type relates to the reproductive or feeding state, as well as to adult age. Another type of plasticity\nis context-dependent and includes influences of the immediate sensory and abiotic environment, but also environmental condi-\ntions during postembryonic development, periods of adult behavioral maturation, and short- and long-term sensory experience. Finally, plasticity in olfactory circuits is linked to associative learning and memory formation. The vast majority of the available\nliterature summarized here deals with plasticity in primary and secondary olfactory brain centers, but also peripheral modulation\nis treated. The described molecular, physiological, and structural neuronal changes occur under the influence of neuromodulators\nsuch as biogenic amines, neuropeptides, and hormones, but the mechanisms through which they act are only beginning to be\nanalyzed. 2\nBehavioral Physiology and Sociobiology (Zoology II), Biozentrum,\nUniversity of Würzburg, Am Hubland, 97074 Würzburg, Germany 1\nIGEPP, INRAE, Institut Agro, Univ Rennes, INRAE,\n49045 Angers, France * Sylvia Anton\nsylvia.anton@inrae.fr\n* Wolfgang Rössler\nroessler@biozentrum.uni-wuerzburg.de Introduction This finally promotes fitness of an insect in a given ecological\nand evolutionary context (Agrawal 2001). Many insect species predominantly rely on olfaction for intra-\nand interspecific communication and searching food. Olfaction is\nof particularly high importance in night- or dim-light active spe-\ncies and for social communication as in social insects. The mul-\ntitude of available olfactory cues in the natural environment com-\nbined with limited size of the nervous system and the resulting\nneuronal processing capacities render neuronal plasticity and\nmodulation as major factors to optimize the use of neural sub-\nstrate (Dukas 2008; Gadenne et al. 2016; Groh and Rössler\n2020). However, the complexity of the nervous system can also\nset limits for behavioral and ecological plasticity (Bernays 2001). Plasticity in insect olfactory systems occurs at multiple levels,\nfor example as a function of physiological state, in response to\nenvironmental factors, social interactions, and experience. Whereas most of the literature on the mechanisms of olfactory\nplasticity and modulation in insects concentrated on the central\nnervous system (CNS), recent work has also shown modulation\nalready at the olfactory receptor level within olfactory sensory\nneurons (OSNs) on the antennae, for example, as a function of\nodor exposure (Tsitoura and Iatrou 2016; Guo et al. 2017; Wicher\n2018). Furthermore, modulation of OSN sensitivity due to expe-\nrience has been described, for example, in male moths (Guerrieri\net al. 2012) and due to feeding and maturity in female mosquitoes\nand blood-feeding bugs (Gadenne et al. 2016 and references there-\nin (Davis 1984; Grant and O’Connell 2007; Siju et al. 2008;\nReisenman 2014)). Within the CNS, several levels of plasticity\nhave been identified. We will summarize here mainly the most\nrecent results from studies on plasticity and modulation within the\nprimary and secondary olfactory centers in the brain—the anten-\nnal lobes (ALs) and the mushroom bodies (MBs). * Sylvia Anton\nsylvia.anton@inrae.fr\n* Wolfgang Rössler\nroessler@biozentrum.uni-wuerzburg.de Similar to other sensory systems, various mechanisms are\ninvolved in olfactory plasticity, and recent methodological\nadvances provide increasing access to study these 150 Cell Tissue Res (2021) 383:149–164 mechanisms. At the molecular level, the expression of genes\nassociated with olfactory reception and genes coding for\nneuromodulators and hormones and their receptors can vary\nin status- or context-dependent manners (Gadenne et al. 2016). Introduction At the cellular level, neuronal elements have been iden-\ntified physiologically and anatomically, and various parame-\nters of their activity were monitored to show modulation (e.g.,\nNeupert et al. 2018). New technology allows to detect the\npresence of neuromodulators within individual neurons or\nsmall populations of neurons, such as biogenic amines, neu-\nropeptides, and hormones (Ly et al. 2019). moth Agrotis ipsilon, behavioral responses to the female-\nemitted sex pheromone are inhibited transiently after mat-\ning. This plasticity seems to originate from a decrease in\nsensitivity of AL neurons to the sex pheromone, probably\nthrough the implication of ecdysteroids, whereas antennal\ndetection of the sex pheromone does not change after\nmating (for review see Gadenne et al. (2016)). More re-\ncently, differences in the occurrence of a few neuropep-\ntides, such as insulin-like peptides, have been found be-\ntween brains and more specifically ALs of mated and\nunmated male A. ipsilon, indicating a potential role in\npost-mating sex pheromone response inhibition (Diesner\net al. 2018). In another noctuid moth, Spodoptera\nlittoralis, behavioral inhibition of male sex pheromone\nresponses after mating rather originates from modulation\nin OSNs (Kromann et al. 2014). In the same species, an\nincrease in antennal sensitivity to host plant volatiles has\nbeen shown in females after mating (Martel et al. 2009;\nSaveer et al. 2012). Mating-dependent plasticity of pe-\nripheral sensitivity to fruit odors and sex pheromones\nhas also been investigated in different fruit fly species. In female Drosophila suzukii, mating causes strong up-\nand downregulation of olfactory genes within the antenna. In parallel, female antennae increased their sensitivity to\nisoamyl acetate significantly after mating, which is coher-\nent with the attractant role of this compound emitted by\nfresh fruit to mated females (Revadi et al. 2015; Crava\net al. 2019). In Ceratitis capitata, sensitivity of antennae\nand palps to pheromone components emitted by sexually\nmature males, which attract both males and females, de-\ncreases after mating in both sexes (Sollai et al. 2018). However, the mechanisms leading to mating-induced\nchanges in antennal sensitivity are unknown so far. In\nthe hymenopteran parasitoid wasp Nasonia vitripennis,\nfemales are attracted to a male-emitted sex pheromone,\nand the mating-induced lack of behavioral pheromone re-\nsponses seems to be mediated by dopamine, as virgin\nfemales injected with dopamine did not respond any more\nto the pheromone and mated females injected with a do-\npamine antagonist continued to respond (Lenschow et al. Introduction 2018). However, appetitive learning leads to recovery of\nsex pheromone attraction in these females (Lenschow\net al. 2018). Concerning changes in olfactory sensitivity\ndepending on the reproductive state in social insects, very\nlittle information is available. In the ant Harpegnathos\nsaltator, a reduction in antennal sensitivity to queen-\nproduced cuticular hydrocarbons, involved in inhibiting\nworkers from reproduction, has been revealed in female\nworkers becoming reproductive substitute queens called\ngamergates (Ghaninia et al. 2017). However, this repre-\nsents a special case within social Hymenoptera that may\nbe limited to ponerine ants in which mature workers retain\nthe potential to mate and reproduce Mechanisms of olfactory plasticity have specifically been\nstudied in Drosophila melanogaster, due to the available ge-\nnetic tools. As literature on D. melanogaster has been recently\nreviewed (e.g., Sayin et al. 2018; Amin and Lin 2019; Boto\net al. 2020), we concentrate here on other experimental insect\nmodels, mainly moths and social Hymenoptera, with a focus\non their specific ecological context, because a vast amount of\nliterature is available in these two classical models for olfac-\ntory plasticity. We do, however, also include occasional ref-\nerences to further insect species such as locusts, blood-feeding\ninsects, and aphids, because we would like to emphasize and\npromote the importance of comparative investigations in this\nfield. In order to illustrate neuronal mechanisms, we will pro-\nvide here an integrative view of olfactory plasticity in a be-\nhavioral and ecological context emphasizing the importance\nof structural neuronal plasticity and neuromodulation in\nolfaction. State-dependent plasticity and modulation Responses to intra- and interspecific volatile olfactory stimuli can\nbe modulated as a function of the physiological state. Depending\non the role of an olfactory cue or signal, the age, reproductive or\nfeeding state, but also the circadian rhythm can influence the\nsensitivity of the olfactory system to certain olfactory stimuli. Such modulation is mostly caused by an interplay between hor-\nmones, neuropeptides, and biogenic amines, acting at the periph-\neral or the central olfactory levels (for review see Gadenne et al. (2016)). As an example, the titer of the biogenic amine serotonin\n(5HT) within the AL varies in a circadian fashion in male moths,\nwhichis correlated withAL neuronand behavioral responsiveness\nto sex pheromone or host plant volatiles (Kloppenburg et al. 1999;\nGatellier et al. 2004). However, circadian modulation of olfactory\nsensitivity seems to be primarily modulated at the peripheral level\nand has been reviewed earlier (Gadenne et al. 2016). Feeding state In the mosquito Aedes aegypti, the abundance of two\npeptides within the ALs, short neuropeptide F2 (sNPF-2) and\nallatostatin-A-5 (AstA-5), increased 24 and 48 h after a blood\nmeal and systemic injection of both neuropeptides mimicked\nthe host-seeking inhibition effect of a blood meal in unfed\nfemales, thus downregulating responses to food odor (Christ\net al. 2017). In the oriental fruit fly, Bactrocera dorsalis, sNPF\nhas been shown to be involved in feeding state-dependent\nantennal sensitivity to a host plant odor, but in this case up-\nmodulating responses to food odor. When sNPF gene expres-\nsion was inhibited via RNAi, the sensitivity to the odor de-\ncreased in starved flies, which normally exhibit a high sensi-\ntivity (Jiang et al. 2017). In D. melanogaster, sNPF also con-\ntributes to starving-induced improved responses to food odor\nat the receptor neuron level and changes odor representation in\nthe AL, resulting in more robust food-search behavior (Root\net al. 2011). A more general nutritional effect on olfactory\nsensitivity has been identified in D. melanogaster: when fed\nwith a high fat diet, antennal sensitivity to various odors de-\ncreased. This was correlated with a decreased expression of\nthe olfactory co-receptor Orco (Jung et al. 2018). A big chal-\nlenge now is to unravel how peripheral and central nervous\nmodulation interact in feeding state-dependent changes of ol-\nfactory sensitivity. Many aphid species change their dispersal capacities by\nproducing winged morphs when population density increases,\nplant quality decreases, or stress factors such as enemy attacks\noccur (Braendle et al. 2006). The formation of wings in par-\nthenogenetic aphid females improves their dispersal capacities\nand allows them to colonize new habitats more easily than\nwingless females. It is known for several aphid species that\nthe sensory equipment of winged individuals is more elabo-\nrate than that of wingless aphids: besides differences in eye\nmorphology (Ishikawa and Miura 2007; Kollmann et al. 2010), they possess longer antennae and more olfactory or-\ngans, so-called rhinaria, on their antennae (Shambaugh et al. 1978; Miyazaki 1987). A recent study has found evidence that\nalso primary sensory centers in the brain, i.e., visual neuropils\nand ALs, are larger in winged females than in wingless indi-\nviduals of the pea aphid, Acyrthosiphon pisum (Gadenne et al. 2019). Feeding state threshold, the insects change from a solitary to a gregarious\nlifestyle (Simpson and Sword 2008). This phase change\ncauses not only behavioral but also morphological and phys-\niological modifications. Among others, the olfactory system is\nstrongly modified: gregarious locusts have less olfactory sen-\nsilla on the antennae than solitary locusts, along with a lower\ndiscrimination ability for food sources (Greenwood and\nChapman 1984; Ochieng et al. 1998). This correlates with a\nsmaller relative size of the ALs in relation to the volume of the\nentire brain and to the midbrain in gregarious compared to\nsolitary locusts, even though the total brain size is much larger\nin gregarious locusts (Ott and Rogers 2010). Whereas the\nanatomy of AL projection neurons did not show any obvious\ndifferences between the two phases, solitary adult females\npossessed a higher proportion of AL neurons responding to\ntwo components of the egg-laying aggregation pheromone\n(Anton et al. 2002). In addition, AL neurons in solitary third\ninstar nymphs responded more frequently to phenylacetonitril,\nthe major component of the adult aggregation pheromone\n(Ignell et al. 1999). Interestingly, opposing attraction\n(gregarious) and repulsion (solitary) behavior of the aggrega-\ntion pheromone are mediated by octopamine and tyramine,\nrespectively (Ma et al. 2015). Phase switch in locusts does\nnot only modify the olfactory system and its sensitivity but\nalso influences associative food odor learning. Gregarious lo-\ncusts do not acquire new olfactory aversions, contrary to sol-\nitary locusts (Simoes et al. 2016). The effect of the feeding state on olfactory sensitivity has\nmainly been investigated in blood-feeding insects. The ex-\npression of olfactory genes in the antennae and responses of\nOSNs are modulated after a blood meal in, e.g., mosquitoes,\ntsetse flies, and triatomine bugs. This leads to reduced behav-\nioral responses to host cues, but increased responses to alter-\nnative signals, such as odors emitted by oviposition sites, or\naggregation pheromones in bugs (e.g., Rinker et al. 2013;\nTaparia et al. 2017) (for review see also Gadenne et al. (2016)). Recently, also a role of neuropeptides in the brain\nhas been revealed to contribute to feeding-dependent modula-\ntion of olfactory sensitivity in blood-feeding and herbivorous\ndiptera. Environmental conditions during development in\nnon-social insects Several insect species change their lifestyle and phenotype\ndepending on environmental conditions during\npostembryonic development and accordingly modify their ol-\nfactory communication skills. Locusts, for example, strongly\nchange their lifestyle as a function of population densities\nduring development. When densities exceed a certain Feeding state The available genome for this aphid species should\nallow in the future to pinpoint neuromodulators and their re-\nceptors involved in the structural (and probably physiological)\nchanges between winged and wingless females. Reproductive state The reproductive state, i.e., either the mating state or the\ncapacity to reproduce, has important effects on the re-\nsponses to pheromones or volatile host cues. In the male Cell Tissue Res (2021) 383:149–164 151 Variations of postembryonic brood care in social\ninsects For example, honeybee queens do not respond to their\nown mandibular pheromone bouquet, and in sterile workers,\nthe response to the queen pheromone is both age- and stage-\ndependent (Vergoz et al. 2009). Interestingly, the effects of\nqueen mandibular pheromone are mediated by a single com-\nponent (homovanillyl alcohol) that has high chemical similar-\nity with dopamine and acts on brain dopamine receptors that\nmodulate aversive olfactory learning (Vergoz et al. 2009). The\npostembryonic pupal development in the two female castes\nshows marked differences. The ALs develop much faster\n(by about 4 days) in queens compared to workers, and the\nsame applies to synaptogenesis in olfactory sub-regions with-\nin secondary olfactory centers in the MBs (Groh and Rössler\n2011). Whereas the number of olfactory glomeruli in the adult\nAL is only slightly smaller in queens, the spatial arrangement\nand sizes of individual glomeruli show marked differences in\nqueens compared to workers. Differences in the AL and MB\nphenotypes are even more pronounced in ants comprising\npermanent worker castes, especially in leaf-cutting ants\n(Kelber et al. 2010; Groh et al. 2014). For example, in Atta\nvollenweideri, the development of trail pheromone–specific\nmacroglomeruli is worker size-dependent, and the overall\nnumber of glomeruli in a specific AL glomerular cluster (T4\ncluster) may differ by more than 50 glomeruli in minor vs. major workers (Kelber et al. 2010). Whether this marked AL\npolyphenism in the female worker castes is a sole effect of\ndifferential feeding still needs to be investigated. Whereas\nminor workers engage as fungus gardeners inside the nest,\nlarge workers leave the nest as foragers and search for profit-\nable food sources using primarily olfactory cues. Consequently, the behavioral responses to trail pheromone\nwere also shown to be worker size-dependent (Kleineidam\net al. 2007). Variations of postembryonic brood care in social\ninsects Cooperative brood care is a hallmark feature of insect socie-\nties, and differential conditions during postembryonic brood 152 Cell Tissue Res (2021) 383:149–164 regions of the MBs. The resulting synaptic changes correlate\nwith inferior olfactory learning and memory capabilities or\nchanges in the timing of foraging (Tautz et al. 2003; Jones\net al. 2005; Becher et al. 2009). Furthermore, bees raised at\nlower temperatures performed less well in associative olfacto-\nry memory tasks, and they differed in dance-communication\nperformance and undertaking behavior compared to bees\nraised at higher temperatures within the range of naturally\noccurring temperatures in the brood area. Similarly, in\nCamponotus ants, workers control the temperature of pupae\nto specific temperature ranges during postembryonic meta-\nmorphic development by brood carrying behavior. Ant nurses\nrespond to changes in the ambient temperature by placing the\nbrood to nest compartments with the appropriate temperatures\nfollowing circadian rhythms (Roces and Núñez 1989;\nFalibene et al. 2016). Also in ants, suboptimal temperature\nregimes affect proper development of olfactory synapses in\nthe MBs (Falibene et al. 2016). Most interestingly, ants that\nhad experienced diverging brood temperature regimes exhibit\ndifferences in their stimulus response thresholds for adult\nbrood carrying behavior, most likely due to changes in senso-\nry thresholds (Weidenmüller et al. 2009). Taken together, dif-\nferential environmental influences caused by variations of\nbrood care conditions during postembryonic metamorphic de-\nvelopment affect olfactory circuits in the brain and have con-\nsequences for a range of adult olfaction-related behaviors. The\nabove forms of plasticity, therefore, represent interesting cases\nof metaplasticity (Abraham 2008), meaning that olfactory\nplasticity induced by brood care conditions affects adult be-\nhavioral plasticity in social insect colonies. However, how\nexactly changes in olfactory circuits are causally linked to\nchanges in complex olfactory behavior (both at the individual\nand colony levels) still requires further investigations. development may affect the adult phenotype (Weaver 1957). For example, in the honeybee, the reproductive status and\ndevelopment of the female castes (queen-worker\npolymorphism) are induced by differential larval feeding and\nmediated via an epigenetic mechanism involving royal jelly\nproduced in the hypopharyngeal glands (Kucharski et al. 2008). Queens develop from fertilized eggs that are genetical-\nly not different from those that develop into workers, but they\ndevelop faster, are larger, live much longer, and differ mark-\nedly in their adult behavior, including olfactory-guided behav-\niors. Variations of postembryonic brood care in social\ninsects For example, honeybee queens do not respond to their\nown mandibular pheromone bouquet, and in sterile workers,\nthe response to the queen pheromone is both age- and stage-\ndependent (Vergoz et al. 2009). Interestingly, the effects of\nqueen mandibular pheromone are mediated by a single com-\nponent (homovanillyl alcohol) that has high chemical similar-\nity with dopamine and acts on brain dopamine receptors that\nmodulate aversive olfactory learning (Vergoz et al. 2009). The\npostembryonic pupal development in the two female castes\nshows marked differences. The ALs develop much faster\n(by about 4 days) in queens compared to workers, and the\nsame applies to synaptogenesis in olfactory sub-regions with-\nin secondary olfactory centers in the MBs (Groh and Rössler\n2011). Whereas the number of olfactory glomeruli in the adult\nAL is only slightly smaller in queens, the spatial arrangement\nand sizes of individual glomeruli show marked differences in\nqueens compared to workers. Differences in the AL and MB\nphenotypes are even more pronounced in ants comprising\npermanent worker castes, especially in leaf-cutting ants\n(Kelber et al. 2010; Groh et al. 2014). For example, in Atta\nvollenweideri, the development of trail pheromone–specific\nmacroglomeruli is worker size-dependent, and the overall\nnumber of glomeruli in a specific AL glomerular cluster (T4\ncluster) may differ by more than 50 glomeruli in minor vs. major workers (Kelber et al. 2010). Whether this marked AL\npolyphenism in the female worker castes is a sole effect of\ndifferential feeding still needs to be investigated. Whereas\nminor workers engage as fungus gardeners inside the nest,\nlarge workers leave the nest as foragers and search for profit-\nable food sources using primarily olfactory cues. Consequently, the behavioral responses to trail pheromone\nwere also shown to be worker size-dependent (Kleineidam\net al. 2007). In addition to controlled feeding of larvae, many social\ni\nt\ni\nid\nt ll d li\nt\nditi\n(\ni\nd development may affect the adult phenotype (Weaver 1957). For example, in the honeybee, the reproductive status and\ndevelopment of the female castes (queen-worker\npolymorphism) are induced by differential larval feeding and\nmediated via an epigenetic mechanism involving royal jelly\nproduced in the hypopharyngeal glands (Kucharski et al. 2008). Queens develop from fertilized eggs that are genetical-\nly not different from those that develop into workers, but they\ndevelop faster, are larger, live much longer, and differ mark-\nedly in their adult behavior, including olfactory-guided behav-\niors. Adult maturation and polyethism Some of the related\nchanges in olfactory circuits were assigned to age, but the\ntemporal flexibility of task-related changes in adult behavior\n(adult polyethism) adds another level of complexity of olfac-\ntory plasticity in social insects that needs to be studied in more\ndetail in the future. In addition to the AL, robust structural changes associated\nwith adult behavioral maturation were observed in olfactory\ninput regions of the MBs, as reported by several studies in\nsocial Hymenoptera (e.g., honeybee (Withers et al. 1993;\nDurst et al. 1994; Fahrbach et al. 1998; Groh et al. 2012;\nScholl et al. 2014; Muenz et al. 2015), ants (Gronenberg\net al. 1996; Kühn-Bühlmann and Wehner 2006; Stieb et al. 2010, 2012)). The cellular processes underlying these volume\nchanges involve massive outgrowth of Kenyon cell (KC) den-\ndrites and, at the same time, pruning of presynaptic boutons\nwithin microglomerular synaptic complexes (Farris et al. 2001; Stieb et al. 2010; Groh et al. 2012; Muenz et al. 2015). Dendritic expansion is the main cause for the volume\nincrease in the MB calyx during the transition from nursing to\nforaging. The overall result of this structural plasticity is an\nincrease in the olfactory projection neuron to KC synaptic\ndivergence of olfactory circuits by about 33% (the number\nof KC dendritic profiles forming synapses with one\nprojection neuron bouton; Groh et al. 2012). Pharmacological stimulation suggests that the underlying pro-\ncesses are promoted by activity in muscarinic cholinergic\ntransmission during foraging experience (Ismail et al. 2006). Sensory exposure was also shown to play an important role in\nthis olfactory plasticity in leaf-cutting ants (Falibene et al. 2015). A combined anatomical and patch-clamp study in\nD. melanogaster confirmed that structural plasticity of olfac-\ntory MB input synapses is induced by sensory activation\n(Kremer et al. 2010). Interestingly, aged honeybee queens\nexhibit an increase in the relation of olfactory versus visual\ninput synapses in the MB calyx (Groh et al. 2006). More\nrecent studies revealed that social experience influences the The rather drastic interior-forager transition in social in-\nsects correlates with changes in diverse neuromodulators and\nhormones (Hamilton et al. 2016). For example, variations\nwere found in biogenic amine levels (reviewed in Kamhi\nand Traniello (2013)), juvenile hormone (Robinson 1987;\nBloch et al. 2002; Dolezal et al. 2012), and vitellogenin\n(e.g., Amdam and Omholt 2003). Adult maturation and polyethism Future cohort experiments using more\ntightly controlled sensory manipulations and high-resolution\nanatomical and behavioral analyses are needed to dissect the\nchanges in olfactory circuits caused by differences in sensory\nor social experience in order to find the mechanisms how they\naffect adult olfactory behavior (Groh and Rössler 2020). Winter bees (the last generation of bees in fall) might be a\nvaluable model for studying adult olfactory plasticity in the\nfuture, as they live much longer than summer bees and start to\nresume foraging in the next spring after staying in the hive\nduring the entire winter. The winter bee model may help to\ndissect more clearly effects of age and sensory experience. The molecular mechanisms underlying structural neuronal\nplasticity of olfactory circuits are still unknown. A gene ex-\npression study (Becker et al. 2016) revealed several genes that\nmight be associated with epigenetic regulation of neuronal\nplasticity during behavioral maturation of the honeybee, and\nGTPase activities were correlated with the nurse-forager tran-\nsition (Dobrin and Fahrbach 2012). However, in both cases, it\nremained unclear how the molecular changes causally link to\nstructural plasticity in olfactory circuits, which opens an im-\nportant field for future studies. Recent studies on changes in\nthe activity of immediate early genes following odorant expo-\nsure are highly promising in this respect (reviewed in\nSommerlandt et al. (2019)). Schachtner 2005; Tomé et al. 2014). These age-dependent\nchanges in olfactory sensitivity, at least in moths, have been\nshown to be independent of experience. Adult behavioral maturation and the associated changes in\nsensory experience affect the olfactory system in social in-\nsects. Various studies revealed substantial effects of sensory\nexperience on the development of the AL particularly mor-\nphological aspects of individual olfactory glomeruli and their\nresponses to odorants in the AL of honeybees (Winnington\net al. 1996; Jernigan et al. 2020). Calcium-imaging experi-\nments suggest that the odor responsiveness of AL glomeruli\nin honeybee workers increases during the first days of adult\nlife (Wang et al. 2005). Studies in D. melanogaster indicate\nthat activity-related volume increases in olfactory glomeruli\nare mainly caused by an increase in synaptic density within\nthe glomeruli, most likely mediated via local interneurons\n(Devaud et al. 2003; Sachse et al. 2007). Adult maturation and polyethism There is ample evidence that early adult development modu-\nlates behavioral, peripheral, and central nervous olfactory sen-\nsitivity to pheromones in various insects, but also to non-\npheromonal odors. During early adult life, increasing antennal\nresponses to pheromones or kairomones have been shown to\ncorrelate, for example, with increased odorant receptor ex-\npression in mosquitoes or increased hormone receptor expres-\nsion in a noctuid moth (Bigot et al. 2012; Bohbot et al. 2013). At the CNS level, the age-dependent modulation of attraction\nbehavior and AL neuron sensitivity to sex or aggregation\npheromones in moths and locusts has been shown to depend,\namong others, on juvenile hormone titers (for review see\nGadenne et al. (2016)). In various insects, morphological\nchanges have been observed in primary and secondary olfac-\ntory centers associated with age-dependent increases in olfac-\ntory sensitivity to specific cues or changes in olfactory learn-\ning and memory performance (e.g., Huetteroth and In addition to controlled feeding of larvae, many social\ninsect species provide controlled climate conditions (reviewed\nby, e.g., Seeley and Heinrich (1981)), which have conse-\nquences for metamorphic development including the forma-\ntion of olfactory centers in the brain. Experimental manipula-\ntions in the honeybee have shown that accurate temperature\ncontrol is required for proper development of olfactory sub-\nregions in the MBs (Groh et al. 2004, 2006). Slight deviations\n(1 °C) from the optimal temperature range (36 °C ± 0.5 °C)\nlead to deficits in synaptic maturation in olfactory input 153 Cell Tissue Res (2021) 383:149–164 number of MB olfactory input synapses in worker bees\n(Cabirol et al. 2017, 2018). However, a major problem with\nmanipulations of the social environment is that too many var-\niables (e.g., pheromonal, tactile, visual) may change at the\nsame time and, in most cases, are difficult to control. This\nproblem, for example, became evident while studying the in-\nfluence of the primer pheromone ethyl oleate on maturation of\nthe olfactory circuits in the honeybee brain (Muenz et al. 2015). Ethyl oleate is present at high concentrations on the\ncuticle of experienced foragers, sensed by OSNs on the anten-\nnae of nurse bees, processed in the AL (Muenz et al. 2012),\nand finally causes a delay in adult behavioral maturation\n(Leoncini et al. 2004). Adult maturation and polyethism However, the causal links\nof these modulators and hormones, especially how they affect\nspecific sensory pathways, including the olfactory pathways,\nand/or individual behavioral modules, are still discussed con-\ntroversially (reviewed in Hamilton et al. (2016)). In recent\nyears, studies on social Hymenoptera began to focus on the\nlarge and diverse group of neuropeptides as potential modu-\nlators of behavioral pattern transitions (Takeuchi et al. 2003;\nBrockmann et al. 2009; Pratavieira et al. 2014; Schmitt et al. 2015, 2017; Han et al. 2015; Gospocic et al. 2017). For Cell Tissue Res (2021) 383:149–164 154 example, in the desert ant Cataglyphis fortis, tachykinin was\nshown to express age- and behavioral state-related changes\nassociated with task transitions (Schmitt et al. 2017). In the\nponerine ant Harpegnathos saltator, corazonin was identified\nas an important driver of behavioral changes (e.g., worker-\nspecific hunting behavior) associated with the transition of\nfemale workers into reproductive substitute queens\n(Gospocic et al. 2017). Neuropeptides have a specifically high\npotential to mediate a variety of specific or highly localized\nmodulatory actions on neuronal circuits associated with dif-\nferent behavioral patterns, as they represent a very large and\ndiverse group of messenger molecules that may act both as\nneurohormones or neuromodulators (e.g., reviewed by\nSchoofs et al. (2017); Nässel and Zandawala (2019)). Future\nlocalization analyses of stage-specific changes in the spatial\ndistribution of individual neuropeptides within primary and\nsecondary olfactory centers combined with functional analy-\nses using manipulation experiments appear highly promising\nin understanding the role of neuropeptides in age- and stage-\nspecific plasticity of olfactory behaviors and circuits. simultaneous application (Dupuy et al. 2017), which corre-\nlates with delayed behavioral responses. In another noctuid\nmoth, Helicoverpa armigera, calcium imaging revealed a re-\nduced increase of intracellular calcium levels when stimulated\nwith a blend of sex pheromone and complex plant odors as\ncompared to individual odor application (Ian et al. 2017). On\nthe other hand, synergistic responses to a mixture of sex pher-\nomone and a volatile originating from the larval host plant,\npear ester, were reported in the AL of the codling moth Cydia\npomonella and well correlated with behavioral responses\n(Trona et al. 2013). In the noctuid moth S. littoralis, host plant\nvolatiles enhance the selectivity for conspecific pheromone\nblends (Borrero-Echeverry et al. 2018), but nothing is known\nso far about the underlying neural mechanisms. Context-dependent plasticity and modulation Behavioral responses to olfactory signals are modulated by\nvarious environmental factors, including different sensory\ncues emitted by conspecifics, for example, social interactions\n(see following paragraphs), or other organisms, as well as\nabiotic factors, such as climate, and pollutants. Such modula-\ntion and plasticity can occur at different levels within the ol-\nfactory system, starting from the periphery in OSNs, and, in\nmany cases, results in changes within the AL, and especially\nthe MBs (Fig. 1). Adult maturation and polyethism Non-host volatiles or herbivore feeding-induced volatiles\nhave been shown in several insects to reduce responses to\npheromones. One example is the response to aggregation\npheromones in bark beetles, which is inhibited by non-host\nvolatiles or volatiles emitted by attacked host trees, originating\nfrom inhibition within the OSNs on the antennae (Zhang et al. 1999; Jactel et al. 2001; Andersson et al. 2010). In several\nmoth species, non-host plant volatiles also modulate male\nsex pheromone responses, but again, interactions have only\nbeen investigated at the antennal level (Party et al. 2009, 2013;\nFaraone et al. 2013; Binyameen et al. 2013; Hatano et al. 2015; Wang et al. 2016). Signals emitted by herbivore-\nattacked plants can also modulate the attractiveness of host\nplants for female moths searching for an oviposition site. Females of the tobacco hawk moth, Manduca sexta, prefer\nundamaged host plants above herbivore-damaged plants, in\nwhich enhanced emission of (−) linalool renders the signal\nless attractive (Reisenman et al. 2013). This correlates with\ninhibitory interactions between two AL glomeruli specific for\nthe two linalool enantiomers in female M. sexta (Reisenman\n2005). How exactly odor responses are modulated in the pres-\nence of other volatiles at the different levels of the olfactory\npathway is still a matter of debate. In the peripheral system,\ndirect chemical interactions, competition for binding sites, and\ninteractions within co-localized neurons might be possible\n(Renou 2014). At the CNS level, additional interactions via\nseparate input channels have to be considered (Renou and\nAnton 2020). Immediate sensory environment The blue pathway depicts influences of ascending and protocerebral neu-\nronal systems mediating associative influences (e.g., octopaminergic, do-\npaminergic systems). G, olfactory glomerulus; KC, Kenyon cell; LN,\nlocal interneuron; OSN, olfactory sensory neurons; MBON, mushroom\nbody output neuron; PN, projection neuron\n155\nCell Tissue Res (2021) 383:149–164 indicate sites that have been shown to express structural plasticity in\nolfactory neuronal circuits (structural synaptic changes, changes in\naxonal/dendritic structure and connectivity, neuropil volume changes). The blue pathway depicts influences of ascending and protocerebral neu-\nronal systems mediating associative influences (e.g., octopaminergic, do-\npaminergic systems). G, olfactory glomerulus; KC, Kenyon cell; LN,\nlocal interneuron; OSN, olfactory sensory neurons; MBON, mushroom\nbody output neuron; PN, projection neuron Fig. 1 Schematic view of the insect olfactory pathway—from the sensory\nstructures on the antenna to primary (antennal lobe, AL) and secondary\nolfactory centers (mushroom body, MB and lateral horn, LH) in the brain,\nindicating factors inducing plasticity and modulation at various process-\ning levels. The blue asterisks indicate sites of action of neuromodulators\n(biogenic amines, neuropeptides) or hormones and sites for associated\nphysiological and molecular changes (spontaneous activity, response\nthreshold, changes in the expression of odorant receptors, changes in\nthe expression of receptors for modulators or hormones). Red asterisks (Skals et al. 2005). Other sensory modalities, such as vision\nand taste, are also known to modulate/modify olfactory-\nguided behavior, but these interactions rather occur within\nsecondary olfactory centers such as the MBs, for example,\nas shown in moths and honeybees (Balkenius and Balkenius\n2016; Strube-Bloss and Rössler 2018). impaired short-term memory, probably due to increased ex-\npression of nicotinic acetylcholine receptor expression and\nincreased neural sensitivity to acetylcholine (Desneux et al. 2007; Wright et al. 2015; Cabirol and Haase 2019). Opposite to decreased olfactory responses, sugar sensitivity\nin honeybees increased after treatments with sublethal doses\nof the neonicotinoid acetamiprid (El Hassani et al. 2008). Inversely, in the noctuid moth A. ipsilon, different sublethal\ndoses of the neonicotinoid insecticide clothianidin were\nshown to up- or downregulate the sensitivity of AL neurons\nto the sex pheromone, depending on the dose, and in parallel\nincreased or decreased the behavioral response probability to\nthe sex pheromone (Rabhi et al. 2014, 2016). In another noc-\ntuid moth, S. littoralis, peripheral and behavioral modulation\nof sex pheromone responses was caused by sublethal doses of Immediate sensory environment The presence of different sensory stimuli in the immediate\nenvironment of an insect can alter responses to a given olfac-\ntory stimulus through interactions of the odorants at the pe-\nripheral and/or central level. Even though this type of interac-\ntion needs not necessarily fall into the category of modulation,\nwe would like to include them here, because they might inter-\nfere or provide the basis for some cases of experience-\ndependent plasticity. A prominent example is the interaction\nbetween sex pheromones and plant-emitted volatiles in male\nmoths. A flower volatile, heptanal, for example, reduces re-\nsponses to the sex pheromone within the macroglomerular\ncomplex of the AL in the noctuid moth A. ipsilon both at the\ninput and output level (Deisig et al. 2012) but also results in an\nimproved temporal resolution of pheromone pulses by AL\noutput neurons (Chaffiol et al. 2014). When the two odors\nare presented with a time shift, the responses of AL neurons\nto the sex pheromone are delayed as compared to a There is also evidence for the modulation of sex phero-\nmone responses within the AL by mechanical stimulation of\nthe antennae in the noctuid moth S. littoralis with a clean air\npuff (Han et al. 2005). This indicates that modulation of ol-\nfactory responses occurs as a function of antennal\nmechanosensory detection, which can result from air move-\nments in the environment or from feedback of flight activity. Auditory and olfactory inputs also interact in the case of sex\npheromone responses in moths when predatory bats emit ul-\ntrasound signals. Behavioral responses to the ultrasound sig-\nnals depend on the quality of the sex pheromone stimulus 155 Cell Tissue Res (2021) 383:149–164 Fig. 1 Schematic view of the insect olfactory pathway—from the sensory\nstructures on the antenna to primary (antennal lobe, AL) and secondary\nolfactory centers (mushroom body, MB and lateral horn, LH) in the brain,\nindicating factors inducing plasticity and modulation at various process-\ning levels. The blue asterisks indicate sites of action of neuromodulators\n(biogenic amines, neuropeptides) or hormones and sites for associated\nphysiological and molecular changes (spontaneous activity, response\nthreshold, changes in the expression of odorant receptors, changes in\nthe expression of receptors for modulators or hormones). Red asterisks\nindicate sites that have been shown to express structural plasticity in\nolfactory neuronal circuits (structural synaptic changes, changes in\naxonal/dendritic structure and connectivity, neuropil volume changes). Abiotic environmental factors A major anthropogenic factor influencing the insect olfactory\nsystem are insecticides remaining in the environment for a\nlong time. Especially sublethal doses of neonicotinoid insec-\nticides were shown to have negative effects on pollinating\ninsects, such as honeybees and bumblebees, including de-\ncreased behavioral responses to attractive olfactory cues and Cell Tissue Res (2021) 383:149–164 156 another insecticide, deltamethrin, a widely used pyrethroid\n(Lalouette et al. 2016). So far, it is, however, not known if\nthe observed modulatory effects of insecticides are caused\ndirectly by receptor-ligand interactions, or if insecticides cause\nmodifications of neuromodulator levels or expression of their\nreceptors (Abrieux et al. 2013, 2014, 2016). In addition, py-\nrethroid insecticides were shown to disturb the wiring of ol-\nfactory glomeruli during postembryonic metamorphic devel-\nopment in M. sexta (Wegerhoff et al. 2001). MB calyces increased in size after brief pre-exposure to these\nsame stimuli (Guerrieri et al. 2012; Anton et al. 2015). Attractive and repellent gustatory stimuli also improved sub-\nsequent behavioral responses to the sex pheromone, but nei-\nther modified AL neuron responses to the sex pheromone nor\nthe volume of MGC glomeruli or the MB calyces, indicating\nthat the behavioral effects might originate from neural modi-\nfications in higher brain centers (Minoli et al. 2012; Anton\net al. 2015). Olfactory plasticity involving learning\nand memory Associative learning is very common in a large variety of\ninsects. As associative learning and memory represent a very\nlarge and rapidly expanding research field, we here mostly\nfocus on plasticity associated with stable olfactory long-term\nmemory (LTM), as it has the potential to affect insect behavior\nover extended time. The wealth of literature in the field of\nassociative learning and memory, especially in\nD. melanogaster, is beyond the scope of this review (for re-\ncent reviews, see, e.g., Kahsai and Zars (2011); Guven-Ozkan\nand Davis (2014); Sugie et al. (2018); Kacsoh et al. (2019);\nBoto et al. (2020)). Even though behavioral and molecular\nstudies of learning in parasitoid wasps are numerous (for re-\nview see Hoedjes et al. (2011); Smid and Vet (2016)), neuro-\nbiological studies besides D. melanogaster have largely fo-\ncused on social insects, which shall be the main topic here. Nevertheless, associative olfactory learning has been evi-\ndenced in various other insects, such as moths, locusts,\ncrickets, and parasitic wasps (e.g., Fan et al. (1997); Hartlieb\net al. (1999); Daly and Smith (2000); Meiners et al. (2003);\nSkiri et al. (2005); Costa et al. (2010); Simoes et al. (2016)),\nbut the underlying neurobiological mechanisms are unex-\nplored except for a few rare cases (Cayre et al. 2007;\nCassenaer and Laurent 2012). Among the social insects, the\nhoneybee has proven a very valuable model for the study of\nplasticity related to long-term memory (> 24 h) (e.g., Menzel\n(1999); Müller (2000); Menzel and Giurfa (2001); Menzel\net al. (2007)). Experience-related changes in the activity of\nglomeruli were described in the AL of the honeybee using\ncalcium-imaging techniques indicating that changes in olfac-\ntory responses persist over extended time periods after asso-\nciative learning at this early sensory processing level (Rath\net al. 2011). The robust and well-studied proboscis extension\nresponse is a favorable behavioral paradigm for classical con-\nditioning to study olfactory LTM in detail. Using sequential\nassociative conditioning, bees can be trained to memorize the\nassociation between a sugar reward and an odor over extended\ntime (> 3 days up to weeks, months, or lifetime). The reward\nor punishment pathways for appetitive and aversive olfactory\nlearning have been linked to ascending and brain\noctopaminergic and dopaminergic modulatory systems—\nmainly via their influences on odor responses at the level of Non-associative experience Multiple-\ntrial conditioning leading to LTM has previously been shown\nto depend on intracellular calcium levels, which indicates a\nrole of calcium in structural plasticity associated with stable\nLTM (Perisse et al. 2009). In the same line (Scholl et al. 2015),\nusing RNAi knockdown and pharmacological manipulation\nin the MBs showed that CaMKII is required for the formation\nof both early and late olfactory LTM, indicating that the\ncalcium-dependent “learning protein” might be involved in\ntriggering structural synaptic plasticity. The above studies\nsuggest that olfactory LTM is associated with structural A similar effect was observed in leaf-cutting ants, in that\ncase after aversive olfactory learning of odors associated with\nunsuitable plant material for cultivating the underground sym-\nbiotic fungus maintained by the ants (Falibene et al. 2015). The formation of an aversive olfactory LTM leads to an in-\ncrease of the synaptic densities in olfactory (not visual) cir-\ncuits of the MBs, whereas pure sensory exposure resulted in\nsynaptic pruning. Whereas the increase of synaptic boutons\nmay also represent a form of Hebbian plasticity, pruning of\nsynapses after pure sensory exposure may lead to adjustments\nin MB circuits resulting in homeostatic regulation to a drasti-\ncally changing olfactory sensory input. Physiological access to plasticity of olfactory circuits in the\nMBs is sparse, except for few calcium-imaging studies sug-\ngesting physiological plasticity at the olfactory projection\nneuron-to-KC synapses and electrophysiological recordings\nrevealing spike-timing-dependent plasticity at mushroom\nbody output neuron synapses (Faber et al. 1999; Szyszka\net al. 2008; Cassenaer and Laurent 2012). Learning-related\nolfactory plasticity was also revealed by intracellular record-\nings and calcium imaging of GABAergic neurons in the hon-\neybee forming recurrent circuits from the MB output to the\ninput (Grünewald 1999; Haenicke et al. 2018). Similarly, re-\ncordings revealed olfactory plasticity in another type of MB\nextrinsic neurons (Haehnel and Menzel 2012). However, as\nintracellular recordings and calcium imaging are limited to\nshort-term recording times, it is difficult or rather impossible\nto monitor changes over extended periods, for example, after\nassociative conditioning. More recent studies employing long-\nterm recordings (over several hours to days) of MB extrinsic\nor MB output neurons via multiple thin wire tetrodes emerged\nas a feasible approach to monitor learning- and memory-\nrelated long-term changes in olfactory circuits. In the honey-\nbee, multi-unit recordings can even be combined with olfac-\ntory conditioning experiments using the proboscis extension\nresponse (Strube-Bloss et al. 2012). Non-associative experience This technique also opens\nup possibilities to look into multimodal (olfactory-visual) in-\nteractions and their role in context-specific influences on ol-\nfactory perception (Strube-Bloss and Rössler 2018). Non-associative experience Experience has long been known to modify behavioral re-\nsponses to olfactory stimuli, and the neuronal and molecular\nmechanisms underlying these modifications have been inves-\ntigated for many years. Here we will review only recent data\non the role of physiological mechanisms and anatomical long-\nterm modifications that occur within the olfactory pathway as\na consequence of different forms of learning in insects. As an\nextreme case, experience can be acquired during early devel-\nopment and influence larval or adult behavior, or, more fre-\nquently, during the adult stage that may lead to long-lasting\nadaptive changes in olfactory behavior. Even though there are\nindications, that larval host plant experience in moths modu-\nlates female oviposition behavior and even male partner\nchoice, so far the neural substrate concerning the transfer of\nmemories from the larval to the adult stage remains largely\nspeculative (Anderson and Anton 2014). Simple forms of non-associative experience modulating\nolfactory-guided behavior, such as sensitization and habitua-\ntion, have been revealed in many insects. Nevertheless, very\nlittle is known about the neural mechanisms underlying these\nforms of learning. Brief exposure to a behaviorally relevant\ndose of the sex pheromone in the male moth S. littoralis, on\nthe other hand, has been shown to modify the expression of an\nodorant-binding protein in the antenna and leads to stronger\nsubsequent OSN responses to the same signal (Guerrieri et al. 2012). However, as physiological and anatomical changes\noccur in the AL, too, upon pheromone exposure, we cannot\nexclude a feedback to the peripheral system causing this form\nof sensitization (Anderson et al. 2007; Guerrieri et al. 2012). In addition, brief exposure to various behaviorally active sen-\nsory signals, including predator sound and different olfactory\nstimuli, improved behavioral responses and increased the sen-\nsitivity of neurons within the AL to the sex pheromone, rather\nthan in the antennae in the same moth species (Anton et al. 2011; Minoli et al. 2012). At the same time, the volume of the\nmacroglomerular complex (MGC) glomerulus, processing in-\nformation on the major sex pheromone component, and of the Cell Tissue Res (2021) 383:149–164 157 to keep in mind that structural changes in olfactory synaptic\ncircuits themselves may be part of a memory trace, but wheth-\ner they are actually required for memory storage and retrieval\nremains to be determined. Non-associative experience the MBs, but also at the levels of the AL and lateral horn (LH)\n(Mauelshagen 1993; Hammer and Menzel 1995; Okada et al. 2007; Tedjakumala et al. 2014; Jarriault et al. 2018). Whereas\nthe majority of modifications due to associative LTM are lo-\ncalized at the CNS level, physiological plasticity associated\nwith LTM has nevertheless also been evidenced at the anten-\nnal level. Expression of olfactory receptors known to bind the\nlearned odor compounds was significantly downregulated af-\nter associative learning, and electroantennogram responses\nwere significantly reduced in honeybees which had formed a\nLTM, compared to control bees (Claudianos et al. 2014). The\nfeedback mechanism towards the CNS, however, remained\nunclear in this case. At the CNS level, the formation of a stable\nolfactory LTM was shown to be transcription-dependent and\nto involve structural synaptic changes in olfactory circuits at\nthe input of the MBs (Hourcade et al. 2010). Only bees that\nhad received paired stimulation of the conditioned (odor\npulse, CS) and unconditioned stimulus (sugar water, US),\nand that were not injected with the transcription inhibitor ac-\ntinomycin D (ActD) after training, had retained a stable LTM\nwhen they were tested with the CS after 3 days. Most inter-\nestingly, stable LTM formation after 3 days was associated\nwith an increase in synaptic complexes within olfactory com-\npartments of the MB calyces. This effect was absent in neigh-\nboring visual input regions. Naïve bees, i.e., bees that had\nreceived unpaired stimulation and paired stimulated bees that\nhad received ActD, were unable of memory retrieval and did\nnot show any changes in synaptic densities. The authors con-\nclude that growth of new synapses may be involved in stable\nLTM in the insect brain, similar to what has been found in the\nmammalian brain (Abraham et al. 2019). Compared with syn-\naptic pruning following sensory exposure as described above,\nassociative olfactory learning and the formation of\ntranscription-dependent stable LTM resulted in a volume-\nindependent increase of synaptic complexes in olfactory com-\npartments of the honey bee MBs (Groh and Rössler 2020). This suggests that the increases in densities of synaptic\nboutons after associative LTM formation may represent a\nform of learning-related (Hebbian) structural plasticity in\nMB-calyx microcircuits. Transcription-independent memo-\nries, such as early long-term memory, did not lead to any\ndetectable structural changes in olfactory circuits. Outlook Pre-adaptations for high levels of olfactory plasticity\nmay allow species more easily to invade new habitats in the course\nof climate change. Olfactory plasticity is also an important feature\nfrom an ecological point of view. We should investigate how\ndifferent lifestyles and interactions within trophic networks as well\nas with the abiotic environment influence plasticity. Studies on the\nmechanistic nature and role of such differences between closely\nand distantly related insect species with similar or different life-\nstyles, habitat preferences, and olfactory behaviors provide a rich\nground for future comparative research on the causes and conse-\nquences of olfactory plasticity. There is still a long way to go until we fully understand the\npowerful mechanisms and influences of olfactory plasticity\nand modulation on insect behavior and their ecological con-\nsequences. Both the experimental accessibility and rich diver-\nsity of insects clearly promise exciting future advances in this\nimportant field of research. Another important perspective is to investigate the role of\nmultimodal interactions, aiming towards understanding multi-\nsensory, context-dependent plasticity influencing olfactory per-\nception. Here the role of the MBs has been highlighted, but the\nfunction of other protocerebral neuropils, like the function and\npotential interactions with lateral horn neurons, is still largely\nunexplored in most insects. Recent advances in high-resolution\ninsect neuronal brain atlases that started in D. melanogaster\n(Dolan et al. 2019) will help to explore plasticity in these brain\nareas. The potential role of the lateral horn in memory formation\nshould be explored in future studies, as another recent study in\nD. melanogaster already showed that specifically context-\ndependent LTM appears to be mediated by lateral horn neurons\nafter only single trial conditioning (Zhao et al. 2019). Funding Open Access funding enabled and organized by Projekt DEAL. The authors have been supported by grants from the French National\nFunding Agency (ANR), the Region “Pays de la Loire” and the French\nInstitute of Agricultural Research (INRAE) to SA, and the German\nResearch Foundation (DFG) grants SPP 1392 (Ro1177/5-2) and SFB\n1047 (B6) and the University of Würzburg to WR. Open Access This article is licensed under a Creative Commons\nAttribution 4.0 International License, which permits use, sharing, adap-\ntation, distribution and reproduction in any medium or format, as long as\nyou give appropriate credit to the original author(s) and the source, pro-\nvide a link to the Creative Commons licence, and indicate if changes were\nmade. Outlook A major conclusion from previous studies is that plasticity and\nmodulation occur at all levels of the insect olfactory pathway. Whereas some drivers of plasticity like internal programs,\nage- and status-/stage-specific causes of plasticity, seem to\nact at both peripheral and central levels, experience-\ndependent plasticity like learning and memory as well as Cell Tissue Res (2021) 383:149–164 158 multimodal interactions preferentially, but not exclusively,\noccur at higher central levels, particularly the MB. The mech-\nanisms by which sensory and modulatory influences target the\ndifferent levels of the olfactory pathway are comparably well\nunderstood for learning and memory in the MBs (especially\nfrom work in D. melanogaster and the honeybee on dopami-\nnergic or octopaminergic modulation). Much less, however, is\nknown for other modes of plasticity including bottom-up and\ntop-down influences of olfactory memory. This clearly needs\nmore intense investigations in the future, for example, efforts\nto understand distributed forms of plasticity and to identify the\nmajor neuromodulators, for example, within the large family\nof neuropeptides. Furthermore, we need to find causal links\nbetween changes in gene expression or epigenetic regulation,\nmessenger molecules, and their action on identified neuronal\ncircuits all the way up to how plasticity in these circuits mod-\nulates behavior. In addition, we need more information on\nlocal modulatory interactions, such as between different olfac-\ntory glomeruli in the AL, recurrent pathways within the MBs,\nor interactions (bottom up and top down) between primary\nand secondary olfactory centers (MB, LH, and AL). In that\nrespect, MB output neurons (MBONs) might play a key role\nin mediating such interactions. Integrative and multidisciplin-\nary approaches at different levels are necessary to fully under-\nstand the mechanisms underlying age-, status-, and state-\nspecific changes in olfactory processing and perception. moths and social Hymenoptera, with rich knowledge on their\nolfactory systems, behaviors, and their plasticity, are specifically\nimportant from an applied point of view, because they include\nbothimportant pest species, but also beneficial (pollinator) species. Understanding olfactory plasticity in these insects will largely\ncontribute to efforts of environmentally acceptable control of pest\ninsects and to improve protection of beneficial species. To study\nnon-model insects, novel tools like CRISPR/Cas9 manipulation of\ngene expression already started to become very helpful. Comparative mechanistic approaches are highly important in fu-\nture research aimed at understanding the role of olfactory plasticity\nin the dynamics of adaptation of insect species under global\nchange. Abraham WC (2008) Metaplasticity: tuning synapses and networks for\nplasticity. Nat Rev Neurosci 9:387–387. https://doi.org/10.1038/\nnrn2356 Outlook The images or other third party material in this article are included\nin the article's Creative Commons licence, unless indicated otherwise in a\ncredit line to the material. If material is not included in the article's\nCreative Commons licence and your intended use is not permitted by\nstatutory regulation or exceeds the permitted use, you will need to obtain\npermission directly from the copyright holder. To view a copy of this\nlicence, visit http://creativecommons.org/licenses/by/4.0/. In evolutionary terms, variations in olfactory plasticity between\ndifferent insect species provide a promising source of knowledge\nto understand their efficient adaptation to the environment. Insects\nrepresent by far the most abundant group of animal species with\nhighly diverse lifestyles. Because of this rich species diversity and\nthe multitude of evolutionary adaptations across insect taxa, it will\nbe most important to promote comparative research on plasticity\nin the olfactory system of diverse insect species. This includes\nclassical model insects like D. melanogaster, using the powerful\ngenetic manipulations available, but equally important, non-model\ninsect species should be investigated to reveal insight into novel\nmodes of plasticity in their olfactory systems and behaviors. Further investigations on the two focus groups of this review, References J Neurogenet 34:36–\n46. https://doi.org/10.1080/01677063.2020.1715971 Amin H, Lin AC (2019) Neuronal mechanisms underlying innate and\nlearned olfactory processing in Drosophila. Curr Opin Insect Sci\n36:9–17. https://doi.org/10.1016/j.cois.2019.06.003 Braendle C, Davis GK, Brisson JA, Stern DL (2006) Wing dimorphism\nin aphids. Heredity 97:192–199. https://doi.org/10.1038/sj.hdy. 6800863 Anderson P, Anton S (2014) Experience-based modulation of behaviour-\nal responses to plant volatiles and other sensory cues in insect her-\nbivores. Plant Cell Environ 37:1826–1835. https://doi.org/10.1111/\npce.12342 Brockmann A, Annangudi S, Richmond T, Ament SA, Xie F, Southey\nBR, Rodriguez-Zas SR, Robinson GE, Sweedler JV (2009)\nQuantitative peptidomics reveal brain peptide signatures of behav-\nior. Proc Natl Acad Sci U S A 106:2383–2388. https://doi.org/10. 1073/pnas.0813021106 Anderson P, Hansson BS, Nilsson U, Han Q, Sjöholm M, Skals N, Anton\nS (2007) Increased behavioral and neuronal sensitivity to sex pher-\nomone after brief odor experience in a moth. Chem Senses 32:483–\n491. https://doi.org/10.1093/chemse/bjm017 Cabirol A, Brooks R, Groh C, Barron AB, Devaud J-M (2017)\nExperience during early adulthood shapes the learning capacities\nand the number of synaptic boutons in the mushroom bodies of\nhoney bees ( Apis mellifera ). Learn Mem 24:557–562. https://doi. org/10.1101/lm.045492.117 Andersson MN, Larsson MC, Blazenec M, Jakus R, Zhang Q-H, Schlyter\nF (2010) Peripheral modulation of pheromone response by inhibito-\nry host compound in a beetle. J Exp Biol 213:3332–3339. https://\ndoi.org/10.1242/jeb.044396 Anton S, Chabaud M-A, Schmidt-Büsser D, Gadenne B, Iqbal J, Juchaux\nM, List O, Gaertner C, Devaud J-M (2015) Brief sensory experience\ndifferentially affects the volume of olfactory brain centres in a moth. Cell Tissue Res 364:59–65. https://doi.org/10.1007/s00441-015-\n2299-0 Cabirol A, Cope AJ, Barron AB, Devaud J-M (2018) Relationship be-\ntween brain plasticity, learning and foraging performance in honey\nbees. PLoS One 13:e0196749. https://doi.org/10.1371/journal.pone. 0196749 Cabirol A, Haase A (2019) The neurophysiological bases of the impact of\nneonicotinoid pesticides on the behaviour of honeybees. Insects 10:\n344–312. https://doi.org/10.3390/insects10100344 Anton S, Evengaard K, Barrozo RB, Anderson P, Skals N (2011) Brief\npredator sound exposure elicits behavioral and neuronal long-term\nsensitization in the olfactory system of an insect. Proc Natl Acad Sci\n108:3401–3405. https://doi.org/10.1073/pnas.1008840108 Cassenaer S, Laurent G (2012) Conditional modulation of spike-timing-\ndependent plasticity for olfactory learning. Nature 482:47–52. https://doi.org/10.1038/nature10776 Anton S, Ignell R, Hansson BS (2002) Developmental changes in the\nstructure and function of the central olfactory system in gregarious\nand solitary desert locusts. Microsc Res Tech 56:281–291. References Abraham WC (2008) Metaplasticity: tuning synapses and networks for\nplasticity. Nat Rev Neurosci 9:387–387. https://doi.org/10.1038/\nnrn2356 Cell Tissue Res (2021) 383:149–164 159 Abraham WC, Jones OD, Glanzman DL (2019) Is plasticity of synapses\nthe mechanism of long-term memory storage? Npj Sci Learn 1–10. https://doi.org/10.1038/s41539-019-0048-y Bernays EA (2001) Neural limitations in phytophagous insects: implica-\ntions for diet breadth and evolution of host affiliation. Annu Rev\nEntomol 46:703–727. https://doi.org/10.1146/annurev.ento.46.1. 703 Abrieux A, Debernard S, Maria A, Gaertner C, Anton S, Gadenne C,\nDuportets L (2013) Involvement of the G-protein-coupled\ndopamine/ecdysteroid receptor DopEcR in the behavioral response\nto sex pheromone in an insect. PLoS One 8:e72785. https://doi.org/\n10.1371/journal.pone.0072785.s001 Bigot L, Shaik HA, Bozzolan F, Party V, Lucas P, Debernard S, Siaussat\nD (2012) Peripheral regulation by ecdysteroids of olfactory respon-\nsiveness in male Egyptian cotton leaf worms, Spodoptera littoralis. Insect Biochem Mol Biol 42:22–31. https://doi.org/10.1016/j.ibmb. 2011.10.003 Abrieux A, Duportets L, Debernard S, Gadenne C, Anton A (2014) The\nGPCR membrane receptor, DopEcR, mediates the actions of both\ndopamine and ecdysone to control sex pheromone perception in an\ninsect. Front Behav Neurosci 8:312. https://doi.org/10.3389/fnbeh. 2014.00312 Binyameen M, Hussain A, Yousefi F, Birgersson G, Schlyter F (2013)\nModulation of reproductive behaviors by non-host volatiles in the\npolyphagous Egyptian cotton leafworm, Spodoptera littoralis. J\nChem Ecol 39:1273–1283. https://doi.org/10.1007/s10886-013-\n0354-4 Abrieux A, Mhamdi A, Rabhi KK, Egon J, Debernard S, Duportets L,\nTricoire-Leignel H, Anton S, Gadenne C (2016) An insecticide fur-\nther enhances experience-dependent increased behavioural re-\nsponses to sex pheromone in a pest insect. PLoS One 11:\ne0167469–e0167415. https://doi.org/10.1371/journal.pone. 0167469 Bloch G, Wheeler DE, Robinson GE (2002) Endocrine influences on the\norganization of insect societies. In: Pfaff DW, Arnold AP, Fahrbach\nSE et al (eds) Hormones, brain and behavior. Academic Press, San\nDiego, pp 195–235 Bohbot JD, Durand NF, Vinyard BT, Dickens JC (2013) Functional\ndevelopment of the octenol response in Aedes aegypti. Front\nPhysiol 4:39. https://doi.org/10.3389/fphys.2013.00039 Agrawal AA (2001) Phenotypic plasticity in the interactions and evolu-\ntion of species. Science 294:321–326. https://doi.org/10.1126/\nscience.1060701 Borrero-Echeverry F, Bengtsson M, Nakamuta K, Witzgall P (2018)\nPlant odor and sex pheromone are integral elements of specific mate\nrecognition in an insect herbivore. Evolution 72:2225–2233. https://\ndoi.org/10.1111/evo.13571 Amdam GV, Omholt SW (2003) The hive bee to forager transition in\nhoneybee colonies: the double repressor hypothesis. J Theor Biol\n223:451–464. https://doi.org/10.1016/s0022-5193(03)00121-8 Boto T, Stahl A, Tomchik S (2020) Cellular and circuit mechanisms of\nolfactory associative learning in Drosophila. References 00073 Daly KC, Smith BH (2000) Associative olfactory learning in the moth\nManduca sexta. J Exp Biol 203:2025–2038 Davis E (1984) Regulation of sensitivity in the peripheral chemoreceptor\nsystems for host-seeking behaviour by a hemolymph-borne factor in\nAedes aegypti. J Insect Physiol 30:179–183. https://doi.org/10.1016/\n0022-1910(84)90124-0 Fan R, Anderson P, Hansson B (1997) Behavioural analysis of olfactory\nconditioning in the moth Spodoptera littoralis (Boisd.) (Lepidoptera\n: Noctuidae). J Exp Biol 200:2969–2976 Deisig N, Kropf J, Vitecek S, Pevergne D, Rouyar A, Sandoz J-C, Lucas\nP, Gadenne P, Anton S, Barrozo R (2012) Differential interactions\nof sex pheromone and plant odour in the olfactory pathway of a male\nmoth. PLoS One 7:e33159. https://doi.org/10.1371/journal.pone. 0033159.g006 Faraone N, D’Errico G, Caleca V, de Cristofaro A, Trimble RM (2013)\nElectrophysiological and behavioral responses of oriental fruit moth\nto the monoterpenoid citral alone and in combination with sex pher-\nomone. Environ Entomol 42:314–322. https://doi.org/10.1603/\nEN12205 Desneux N, Decourtye A, Delpuech J-M (2007) The sublethal effects of\npesticides on beneficial arthropods. Annu Rev Entomol 52:81–106. https://doi.org/10.1146/annurev.ento.52.110405.091440 Farris SM, Robinson GE, Fahrbach SE (2001) Experience- and age-\nRrelated outgrowth of intrinsic neurons in the mushroom bodies of\nthe adult worker honeybee. J Neurosci 21:6395–6404. https://doi. /10 1523/JNEUROSCI 21 16 06395 2001 Devaud J-M, Acebes A, Ramaswami M, Ferrús A (2003) Structural and\nfunctional changes in the olfactory pathway of adult Drosophila take\nplace at a critical age. J Neurobiol 56:13–23. https://doi.org/10. 1002/neu.10215 org/10.1523/JNEUROSCI.21-16-06395.2001 Gadenne C, Barrozo RB, Anton S (2016) Plasticity in insect olfaction: to\nsmell or not to smell? Annu Rev Entomol 61:317–333. https://doi. org/10.1146/annurev-ento-010715-023523 Diesner M, Gallot A, Binz H, Gaertner C, Vitecek S, Kahnt J, Schachtner\nJ, Jacquin-Joly E, Gadenne C (2018) Mating-induced differential\npeptidomics of neuropeptides and protein hormones in Agrotis\nipsilon moths. J Proteome Res 17:1397–1414. https://doi.org/10. 1021/acs.jproteome.7b00779 Gadenne C, Groh C, Grübel K, Joschinski J, Krauss J, Krieger J, Rössler\nW, Anton S (2019) Neuroanatomical correlates of mobility: sensory\nbrain centres are bigger in winged than in wingless parthenogenetic\npea aphid females. Arthropod Struct Dev 52:100883. https://doi.org/\n10.1016/j.asd.2019.100883 Dobrin SE, Fahrbach SE (2012) Rho GTPase activity in the honey bee\nmushroom bodies is correlated with age and foraging experience. J\nInsect Physiol 58:228–234. https://doi.org/10.1016/j.jinsphys.2011. 11.009 Gatellier L, Nagao T, Kanzaki R (2004) Serotonin modifies the sensitivity\nof the male silkmoth to pheromone. J Exp Biol 207:2487–2496. References https://doi.org/10.1242/jeb.01035 Ghaninia M, Haight K, Berger SL, Reinberg D, Zwiebel LJ, Ray A,\nLiebig J (2017) Chemosensory sensitivity reflects reproductive sta-\ntus in the ant Harpegnathos saltator. Sci Rep 7:373–379. https://doi. org/10.1038/s41598-017-03964-7 Dolan M-J, Frechter S, Bates AS, Dan C, Huoviala P (2019)\nNeurogenetic dissection of the Drosophila lateral horn reveals major\noutputs, diverse behavioural functions, and interactions with the\nmushroom body. eLife 8:e43079. https://doi.org/10.7554/eLife. 43079 Gospocic J, Shields EJ, Glastad KM, Lin Y, Penick CA, Yan H,\nMikheyev AS, Linksvayer TA, Garcia BA, Berger SL, Liebig J,\nReinberg D, Bonasio R (2017) The neuropeptide corazonin controls\nsocial behavior and caste identity in ants. Cell 170:748–752.e12. https://doi.org/10.1016/j.cell.2017.07.014 Dolezal AG, Brent CS, Hölldobler B, Amdam GV (2012) Worker divi-\nsion of labor and endocrine physiology are associated in the harvest-\ner ant, Pogonomyrmex californicus. J Exp Biol 215:454–460. https://doi.org/10.1242/jeb.060822 Grant AJ, O’Connell RJ (2007) Age-related changes in female mosquito\ncarbon dioxide detection. J Med Entomol 44:617–623. https://doi. org/10.1093/jmedent/44.4.617 Dukas R (2008) Evolutionary biology of insect learning. Annu Rev\nEntomol 53:145–160. https://doi.org/10.1146/annurev.ento.53. 103106.093343 Greenwood M, Chapman R (1984) Differences in numbers of sensilla on\nthe antennae of solitarious and gregarious Locusta migratoria L. (Orthoptera: Acrididae). Int J Insect Morphol Embryol 13:295–\n301. https://doi.org/10.1016/0020-7322(84)90004-7 Dupuy F, Rouyar A, Deisig N, Bourgeois T, Limousin D, Wycke M-A,\nAnton S, Renou M (2017) A background of a volatile plant com-\npound alters neural and behavioral responses to the sex pheromone\nblend in a moth. Front Physiol 8:79. https://doi.org/10.3389/fphys. 2017.00079 Groh C, Ahrens D, Rössler W (2006) Environment- and age-dependent\nplasticity of synaptic complexes in the mushroom bodies of honey-\nbee queens. Brain Behav Evol 68:1–14. https://doi.org/10.1159/\n000092309 Durst C, Eichmüller S, Menzel R (1994) Development and experience\nlead to increased volume of subcompartments of the honeybee\nmushroom body. Behav Neural Biol 62:259–263. https://doi.org/\n10.1016/S0163-1047(05)80025-1 Groh C, Kelber C, Grübel K, Rössler W (2014) Density of mushroom\nbody synaptic complexes limits intraspecies brain miniaturization in\nhighly polymorphic leaf-cutting ant workers. Proc R Soc B Biol Sci\n281:20140432. https://doi.org/10.1098/rspb.2014.0432 El Hassani AK, Dacher M, Gary V, Lambin M, Gauthier M, Armengaud\nC (2008) Effects of sublethal doses of acetamiprid and\nthiamethoxam on the behavior of the honeybee (Apis mellifera). Arch Environ Contam Toxicol 54:653–661. https://doi.org/10. 1007/s00244-007-9071-8 81:20140432. https://doi.org/10.1098/rspb.2014.0432 Groh C, Lu Z, Meinertzhagen IA, Rössler W (2012) Age-related plastic-\nity in the synaptic ultrastructure of neurons in the mushroom body\ncalyx of the adult honeybee Apis mellifera. References https://\ndoi.org/10.1002/jemt.10032 Cayre M, Scotto-Lomassese S, Malaterre J, Strambi C, Strambi A (2007)\nUnderstanding the regulation and function of adult neurogenesis:\ncontribution from an insect model, the house cricket. Chem Senses\n32:385–395. https://doi.org/10.1093/chemse/bjm010 Balkenius A, Balkenius C (2016) Multimodal interaction in the insect\nbrain. BMC Neurosci 17:29. https://doi.org/10.1186/s12868-016-\n0258-7 Chaffiol A, Dupuy F, Barrozo RB, Kropf J, Renou M, Rospars J-P, Anton\nS (2014) Pheromone modulates plant odor responses in the antennal\nlobe of a moth. Chem Senses 39:451–463. https://doi.org/10.1093/\nchemse/bju017 Becher MA, Scharpenberg H, Moritz RFA (2009) Pupal developmental\ntemperature and behavioral specialization of honeybee workers\n(Apis mellifera L.). J Comp Physiol 195:673–679. https://doi.org/\n10.1007/s00359-009-0442-7 Christ P, Reifenrath A, Kahnt J, Hauser F, Hill SR, Schachtner J, Ignell R\n(2017) Feeding-induced changes in allatostatin-A and short neuro-\npeptide F in the antennal lobes affect odor-mediated host seeking in\nthe yellow fever mosquito, Aedes aegypti. PLoS One 12:e0188243–\ne0188215. https://doi.org/10.1371/journal.pone.0188243 Becker N, Kucharski R, Rössler W, Maleszka R (2016) Age-dependent\ntranscriptional and epigenomic responses to light exposure in the\nhoney bee brain. FEBS Open Bio 6:622–639. https://doi.org/10. 1002/2211-5463.12084@10.1002/(ISSN)2211-5463(CAT)\nVirtualIssues(VI)FEBSPressHighlights2016 Claudianos C, Lim J, Young M, Yan S, Cristino AS, Newcomb RD,\nGunasekaran N, Reinhard J (2014) Odor memories regulate 160 Cell Tissue Res (2021) 383:149–164 olfactory receptor expression in the sensory periphery. Eur J\nNeurosci 39:1642–1654. https://doi.org/10.1111/ejn.12539 Faber T, Joerges J, Menzel R (1999) Associative learning modifies neural\nrepresentations of odors in the insect brain. Nat Neurosci 2:74–78. https://doi.org/10.1038/4576 Costa A, Ricard I, Davison AC, Turlings TCJ (2010) Effects of rewarding\nand unrewarding experiences on the response to host-induced plant\nodors of the generalist parasitoid Cotesia marginiventris\n(Hymenoptera: Braconidae). J Insect Behav 23:303–318. https://\ndoi.org/10.1007/s10905-010-9215-y Fahrbach SE, Moore D, Capaldi EA, Farris SM, Robinson GE (1998)\nExperience-expectant plasticity in the mushroom bodies of the hon-\neybee. Learn Mem 5:115–123. https://doi.org/10.1101/lm.5.1.115 Falibene A, Roces F, Rössler W (2015) Long-term avoidance memory\nformation is associated with a transient increase in mushroom body\nsynaptic complexes in leaf-cutting ants. Front Behav Neurosci 9:84. https://doi.org/10.3389/fnbeh.2015.00084 Crava CM, Sassù F, Tait G, Becher P, Anfora G (2019) Functional tran-\nscriptome analyses of Drosophila suzukii antennae reveal mating-\ndependent olfaction plasticity in females. Insect Biochem Mol Biol\n105:51–59. https://doi.org/10.1016/j.ibmb.2018.12.012 Falibene A, Roces F, Rössler W, Groh C (2016) Daily thermal fluctua-\ntions experienced by pupae via rhythmic nursing behavior increase\nnumbers of mushroom body microglomeruli in the adult ant brain. Front Behav Neurosci 10:73. https://doi.org/10.3389/fnbeh.2016. References 1523/JNEUROSCI.1573-17.2017 Jactel H, Van Halder I, Menassieu P, Zhang QH, Schlyter F (2001) Non-\nhost volatiles disrupt the response of the stenographer bark beetle,\nIps sexdentatus (Coleoptera: Scolytidae), to pheromone-baited traps\nand maritime pine logs. Integr Pest Manag Rev 6:197–207. https://\ndoi.org/10.1023/A:1025775419193 Guven-Ozkan T, Davis RL (2014) Functional neuroanatomy of\nDrosophila olfactory memory formation. Learn Mem 21:519–526. https://doi.org/10.1101/lm.034363.114 Haehnel M, Menzel R (2012) Long-term memory and response general-\nization in mushroom body extrinsic neurons in the honeybee Apis\nmellifera. J Exp Biol 215:559–565. https://doi.org/10.1242/jeb. 059626 Jarriault D, Fuller J, Hyland BI, Mercer AR (2018) Dopamine release in\nmushroom bodies of the honey bee (Apis mellifera L.) in response to\naversive stimulation. Sci Rep 1–12. https://doi.org/10.1038/s41598-\n018-34460-1 Haenicke J, Yamagata N, Zwaka H, Nawrot M, Menzel R (2018) Neural\ncorrelates of odor learning in the presynaptic microglomerular cir-\ncuitry in the honeybee mushroom body calyx. eneuro 5:e.0128–\ne.0118. https://doi.org/10.1523/ENEURO.0128-18.2018 Jernigan CM, Halby R, Gerkin RC, Sinakevitch I, Locatelli F, Smith BH\n(2020) Experience-dependent tuning of early olfactory processing in\nthe adult honey bee, Apis mellifera. J Exp Biol 223:jeb206748. https://doi.org/10.1242/jeb.206748 Hamilton AR, Shpigler H, Bloch G, Wheeler DE, Robinson GE (2016)\nEndocrine influences on insect societies. In: Hormones, brain and\nbehavior. Elsevier, pp. 421–451 Jiang H-B, Gui S-H, Xu L, Pei Y-X, Smagghe G, Wang J-J (2017) The\nshort neuropeptide F modulates olfactory sensitivity of Bactrocera\ndorsalis upon starvation. J Insect Physiol 99:78–85. https://doi.org/\n10.1016/j.jinsphys.2017.03.012 Hammer M, Menzel R (1995) Learning and memory in the honeybee. J\nNeurosci 15:1617–1630. https://doi.org/10.1523/JNEUROSCI.15-\n03-01617 Jones JC, Helliwell P, Beekman M, Malezska R, Oldroyd BP (2005) The\neffects of rearing temperature on developmental stability and learn-\ning and memory in the honey bee, Apis mellifera. J Comp Physiol A\n191:1121–1129. https://doi.org/10.1007/s00359-005-0035-z Han B, Fang Y, Feng M, Hu H, Qi Y, Huo X, Meng L, Wu B, Li J (2015)\nQuantitative neuropeptidome analysis reveals neuropeptides are cor-\nrelated with social behavior regulation of the honeybee workers. J\nProteome Res 14:4382–4393. https://doi.org/10.1021/acs. jproteome.5b00632 Jung J, Kim D-I, Han G-Y, Kwon H (2018) The effects of high fat diet-\ninduced stress on olfactory sensitivity, behaviors, and transcriptional\nprofiling in Drosophila melanogaster. Int J Mol Sci 19:2855–2818. https://doi.org/10.3390/ijms19102855 Han Q, Hansson BS, Anton S (2005) Interactions of mechanical stimuli\nand sex pheromone information in antennal lobe neurons of a male\nmoth, Spodoptera littoralis. J Comp Physiol A 191:521–528. References J Comp Neurol 520:\n3509–3527. https://doi.org/10.1002/cne.23102 Cell Tissue Res (2021) 383:149–164 161 neuroscience. Proc R Soc B Biol Sci 278:889–897. https://doi.org/\n10.1093/molbev/msp188 Groh C, Rössler W (2020) Analysis of synaptic microcircuits in the\nmushroom bodies of the honeybee. Insects 11:43–17. https://doi. org/10.3390/insects11010043 Hourcade B, Muenz TS, Sandoz J-C, Rössler W, Devaud J-M (2010)\nLong-term memory leads to synaptic reorganization in the mush-\nroom bodies: a memory trace in the insect brain? J Neurosci 30:\n6461–6465. https://doi.org/10.1523/JNEUROSCI.0841-10.2010 Groh C, Rössler W (2011) Comparison of microglomerular structures in\nthe mushroom body calyx of neopteran insects. Arthropod Struct\nDev 40:358–367. https://doi.org/10.1016/j.asd.2010.12.002 Groh C, Tautz J, Rössler W (2004) Synaptic organization in the adult\nhoney bee brain is influenced by brood-temperature control during\npupal development. Proc Natl Acad Sci U S A 101:4268–4273. https://doi.org/10.1073/pnas.0400773101 Huetteroth W, Schachtner J (2005) Standard three-dimensional glomeruli\nof the Manduca sexta antennal lobe: a tool to study both develop-\nmental and adult neuronal plasticity. Cell Tissue Res 319:513–524. https://doi.org/10.1007/s00441-004-1016-1 Gronenberg W, Heeren S, Hölldobler B (1996) Age-dependent and task-\nrelated morphological changes in the brain and the mushroom bod-\nies of the ant Camponotus floridanus. J Exp Biol 199:2011–2019. https://doi.org/10.2307/2937655 Ian E, Kirkerud NH, Galizia CG, Berg BG (2017) Coincidence of pher-\nomone and plant odor leads to sensory plasticity in the heliothine\nolfactory system. PLoS One 12:e0175513. https://doi.org/10.1371/\njournal.pone.0175513 Grünewald B (1999) Physiological properties and response modulations\nof mushroom body feedback neurons during olfactory learning in\nthe honeybee, Apis mellifera. J Comp Physiol A Neuroethol Sens\nNeural Behav Physiol 185:565–576. https://doi.org/10.1007/\ns003590050417 Ignell R, Anton S, Hansson BS (1999) Integration of behaviourally rele-\nvant odours at the central nervous level in solitary and gregarious\nthird instar locusts, Schistocerca gregaria. J Insect Physiol 45:993–\n1000. https://doi.org/10.1016/s0022-1910(99)00080-3 Ishikawa A, Miura T (2007) Morphological differences between wing\nmorphs of two Macrosiphini aphid species Acyrthosiphon pisum\nand Megoura crassicauda (Hemiptera, Aphididae). Sociobiology\n50:881–893 Guerrieri F, Gemeno C, Monsempes C, Anton S, Jacquin-Joly E, Lucas\nP, Devaud J-M (2012) Experience-dependent modulation of anten-\nnal sensitivity and input to antennal lobes in male moths\n(Spodoptera littoralis) pre-exposed to sex pheromone. J Exp Biol\n215:2334–2341. https://doi.org/10.1242/jeb.060988 Ismail N, Robinson GE, Fahrbach SE (2006) Stimulation of muscarinic\nreceptors mimics experience-dependent plasticity in the honey bee\nbrain. Proc Natl Acad Sci 103:207–211. https://doi.org/10.1073/\npnas.0508318102 Guo H, Kunwar K, Smith D (2017) Odorant receptor sensitivity modu-\nlation in Drosophila. J Neurosci 37:9465–9473. https://doi.org/10. References https://doi.org/10.1007/s00359-005-0618-8 Kacsoh BZ, Barton S, Jiang Y, Zhou N, Mooney SD, Friedberg I,\nRadivojac P, Greene CS, Bosco G (2019) New Drosophila long-\nterm memory genes revealed by assessing computational function\nprediction methods. G3-Genes Genomes Genet 9:251–267. https://\ndoi.org/10.1534/g3.118.200867 Hartlieb E, Anderson P, Hansson BS (1999) Appetitive learning of\nodours with different behavioural meaning in moths. Physiol\nBehav 67:671–677. https://doi.org/10.1016/S0031-9384(99)\n00124-9 Kahsai L, Zars T (2011) Learning and memory in Drosophila: behavior,\ngenetics, and neural systems. Int Rev Neurobiol 99:139–167. https://\ndoi.org/10.1016/B978-0-12-387003-2.00006-9 Hatano E, Saveer AM, Borrero-Echeverry F, Strauch M, Zakir A,\nBengtsson M, Ignell R, Anderson P, Becher PG, Witzgall P,\nDekker T (2015) A herbivore-induced plant volatile interferes with\nhost plant and mate location in moths through suppression of olfac-\ntory signalling pathways. BMC Biol 13:75. https://doi.org/10.1186/\ns12915-015-0188-3 Kamhi JF, Traniello JFA (2013) Biogenic amines and collective organi-\nzation in a superorganism: neuromodulation of social behavior in\nants. Brain Behav Evol 82:220–236. https://doi.org/10.1159/\n000356091 Hoedjes KM, Kruidhof HM, Huigens ME, Dicke M, Vet LEM, Smid\nHM (2011) Natural variation in learning rate and memory dynamics\nin parasitoid wasps: opportunities for converging ecology and Kelber C, Rössler W, Kleineidam CJ (2010) Phenotypic plasticity in\nnumber of glomeruli and sensory innervation of the antennal lobe 162 Cell Tissue Res (2021) 383:149–164 in leaf-cutting ant workers (A. vollenweideri). Dev Neurobiol 70:\n222–234. https://doi.org/10.1002/dneu.20782 Menzel R (1999) Memory dynamics in the honeybee. J Comp Physiol A\n185:323–340. https://doi.org/10.1007/s003590050392 Kleineidam CJ, Rössler W, Hölldobler B, Roces F (2007) Perceptual\ndifferences in trail-following leaf-cutting ants relate to body size. J\nInsect Physiol 53:1233–1241. https://doi.org/10.1016/j.jinsphys. 2007.06.015 Menzel R, Brembs B, Giurfa M (2007) Cognition in invertebrates. In:\nKaas J (ed) Evolution of nervous systems, Vol.II: evolution of ner-\nvous systems in invertebrates. Oxford: Academic Press, Oxford, pp\n403–422 Kloppenburg P, Ferns D, Mercer AR (1999) Serotonin enhances central\nolfactory neuron responses to female sex pheromone in the male\nsphinx moth Manduca sexta. J Neurosci 19:8172–8181. https://\ndoi.org/10.1523/JNEUROSCI.19-19-08172.1999 Menzel R, Giurfa M (2001) Cognitive architecture of a mini-brain: the\nhoneybee. Trends Cogn Sci 5:62–71. https://doi.org/10.1016/s1364-\n6613(00)01601-6 Minoli S, Kauer I, Colson V, Party V, Renou M, Anderson P, Gadenne C,\nMarion-Poll F, Anton S (2012) Brief exposure to sensory cues elicits\nstimulus-nonspecific general sensitization in an insect. References https://doi.org/\n10.1007/s11356-015-5923-3 Ochieng SA, Hallberg E, Hansson BS (1998) Fine structure and distribu-\ntion of antennal sensilla of the desert locust, Schistocerca gregaria\n(Orthoptera: Acrididae). Cell Tissue Res 291:525–536. https://doi. org/10.1007/s004410051022 Lenschow M, Cordel M, Pokorny T, Mair MM, Hofferberth J, Ruther J\n(2018) The post-mating switch in the pheromone response of\nNasonia females is mediated by dopamine and can be reversed by\nappetitive learning. Front Behav Neurosci 12:14. https://doi.org/10. 3389/fnbeh.2018.00014 Okada R, Rybak J, Manz G, Menzel R (2007) Learning-related plasticity\nin PE1 and other mushroom body-extrinsic neurons in the honeybee\nbrain. J Neurosci 27:11736–11747. https://doi.org/10.1523/\nJNEUROSCI.2216-07.2007 Leoncini I, Le Conte Y, Costagliola G, Plettner E, Toth AL, Wang M,\nHuang Z, Bécard J-M, Crauser D, Slessor KN, Robinson GE (2004)\nRegulation of behavioral maturation by a primer pheromone pro-\nduced by adult worker honey bees. Proc Natl Acad Sci U S A 101:\n17559–17564. https://doi.org/10.1073/pnas.0407652101 Ott SR, Rogers SM (2010) Gregarious desert locusts have substantially\nlarger brains with altered proportions compared with the solitarious\nphase. Proc R Soc B Biol Sci R Soc 277:3087–3096. https://doi.org/\n10.1098/rspb.2010.0694 Party V, Hanot C, Büsser DS, Rochat D, Renou M (2013) Changes in\nodor background affect the locomotory response to pheromone in\nmoths. PLoS One 8:e52897. https://doi.org/10.1371/journal.pone. 0052897 Ly A, Ragionieri L, Liessem S, Becker M, Deininger S-O, Neupert S,\nPredel R (2019) Enhanced coverage of insect neuropeptides in tissue\nsections by an optimized mass-spectrometry-imaging protocol. Anal\nChem 91:1980–1988. https://doi.org/10.1021/acs.analchem. 8b04304 Party V, Hanot C, Said I, Rochat D, Renou M (2009) Plant terpenes affect\nintensity and temporal parameters of pheromone detection in a\nmoth. Chem Senses 34:763–774. https://doi.org/10.1093/chemse/\nbjp060 Ma Z, Guo X, Lei H, Li T, Hao S, Kang L (2015) Octopamine and\ntyramine respectively regulate attractive and repulsive behavior in\nlocust phase changes. Sci Rep 5:593–511. https://doi.org/10.1038/\nsrep08036 Perisse E, Raymond-Delpech V, Néant I, Matsumoto Y, Leclerc C,\nMoreau M, Sandoz J-C (2009) Early calcium increase triggers the\nformation of olfactory long-term memory in honeybees. BMC Biol\n7:30. https://doi.org/10.1186/1741-7007-7-30 Martel V, Anderson P, Hansson BS, Schlyter F (2009) Peripheral mod-\nulation of olfaction by physiological state in the Egyptian leaf worm\nSpodoptera littoralis (Lepidoptera: Noctuidae). J Insect Physiol 55:\n793–797. https://doi.org/10.1016/j.jinsphys.2009.04.012 Pratavieira M, da Silva Menegasso AR, Garcia AMC, dos Santons DS,\nGomes PC, Malaspina O, Palma MS (2014) MALDI imaging anal-\nysis of neuropeptides in the africanized honeybee (Apis mellifera)\nbrain: effect of ontogeny. J Proteome Res 13:3054–3064. References PLoS One 7:\ne34141 Kollmann M, Minoli S, Bonhomme J, Homberg U, Schachtner J, Tagu D,\nAnton S (2010) Revisiting the anatomy of the central nervous sys-\ntem of a hemimetabolous model insect species: the pea aphid\nAcyrthosiphon pisum. Cell Tissue Res 343:343–355. https://doi. org/10.1007/s00441-010-1099-9 Miyazaki M (1987) Morphology of aphids. In: Minks A, Harrewijn P\n(eds) Aphids, their biology, natural enemies and control. Elsevier,\nAmsterdam, pp 1–25 Kremer MC, Christiansen F, Leiss F, Paehler M, Knapek S, Andlauer\nTFM, Förstner F, Kloppenburg P, Sigrist SJ, Tavosanis G (2010)\nStructural long-term changes at mushroom body input synapses. Curr Biol 20:1938–1944. https://doi.org/10.1016/j.cub.2010.09.060 Muenz TS, Groh C, Maisonnasse A, Le Conte Y, Plettner E, Rössler W\n(2015) Neuronal plasticity in the mushroom body calyx during adult\nmaturation in the honeybee and possible pheromonal influences. Dev Neurobiol 75:1368–1384. https://doi.org/10.1002/dneu.22290 Kromann SH, Saveer AM, Binyameen M, Bengtsson M, Birgersson G,\nHansson BS, Schlyter F, Witzgall P, Ignell R, Becher PG (2014)\nConcurrent modulation of neuronal and behavioural olfactory re-\nsponses to sex and host plant cues in a male moth. Proc R Soc B\nBiol Sci 282:20141884. https://doi.org/10.1093/chemse/27.1.45 Muenz TS, Maisonnasse A, Plettner E, Le Conte Y, Rössler W (2012)\nSensory reception of the primer pheromone ethyl oleate. Sci Nat 99:\n421–425. https://doi.org/10.1007/s00114-012-0909-1 Müller U (2000) Signal transduction pathways in well-defined models of\nlearning and memory. In: Reith M (ed) Cerebral signal transduction. Humana Press, Totowa, pp 73–103 Kucharski R, Maleszka J, Foret S, Maleszka R (2008) Nutritional control\nof reproductive status in honeybees via DNA methylation. Science\n319:1827–1830. https://doi.org/10.1126/science.1153069 Nässel DR, Zandawala M (2019) Recent advances in neuropeptide sig-\nnaling in Drosophila, from genes to physiology and behavior. Prog\nNeurobiol 179:101607. https://doi.org/10.1016/j.pneurobio.2019. 02.003 Kühn-Bühlmann S, Wehner R (2006) Age-dependent and task-related\nvolume changes in the mushroom bodies of visually guided desert\nants, Cataglyphis bicolor. J Neurobiol 66:511–521. https://doi.org/\n10.1002/neu.20235 Neupert S, Fusca D, Kloppenburg P, Predel R (2018) Analysis of single\nneurons by perforated patch clamp recordings and MALDI-TOF\nmass spectrometry. ACS Chem Neurosci 9:2089–2096. https://doi. org/10.1021/acschemneuro.8b00163 Lalouette L, Pottier M-A, Wycke M-A, Boitard C, Bozzolan F, Maria A,\nDemondion E, Chertemps T, Lucas P, Renault D, Maibeche M,\nSiaussat D (2016) Unexpected effects of sublethal doses of insecti-\ncide on the peripheral olfactory response and sexual behavior in a\npest insect. Environ Sci Pollut Res 23:3073–3085. p\ny\nS0020-7322(78)80001-4 Reisenman CE, Riffell JA, Duffy K, Pesque A, Mikles D, Goodwin B\n(2013) Species-specific effects of herbivory on the oviposition be-\nhavior of the moth Manduca sexta. J Chem Ecol 39:76–89. https://\ndoi.org/10.1007/s10886-012-0228-1 Siju KP, Hansson BS, Ignell R (2008) Immunocytochemical localization\nof serotonin in the central and peripheral chemosensory system of\nmosquitoes. Arthropod Struct Dev 37:248–259. https://doi.org/10. 1016/j.asd.2007.12.001 Renou M (2014) Pheromones and general odor perception in insects. In:\nMucignat-Caretta C (ed) Neurobiology of chemical communication. CRC Press, Taylor and Francis, Boca Raton, London, New York, pp\n23–56 Simoes PMV, Ott SR, Niven JE (2016) Environmental adaptation, phe-\nnotypic plasticity, and associative learning in insects: the desert lo-\ncust as a case study. Integr Comp Biol 56:914–924. https://doi.org/\n10.1093/icb/icw100 Renou M, Anton S (2020) Insect olfactory communication in a complex\nand changing world. Curr Opin Insect Sci 42:1–7. https://doi.org/10. 1016/j.cois.2020.04.004 Simpson SJ, Sword GA (2008) Locusts. Curr Biol 18:R364–R366. https://doi.org/10.1016/j.cub.2008.02.029 Revadi S, Lebreton S, Witzgall P, Anfora G, Dekker T, Becher PG (2015)\nSexual behavior of Drosophila suzukii. Insects 6:183–196. https://\ndoi.org/10.3390/insects6010183 Skals N, Anderson P, Kanneworff M, Lfstedt C, Surlykke A (2005) Her\nodours make him deaf: crossmodal modulation of olfaction and\nhearing in a male moth. J Exp Biol 208:595–601. https://doi.org/\n10.1242/jeb.01400 Rinker DC, Pitts RJ, Zhou X, Suh E, Rokas A, Zwiebel LJ (2013) Blood\nmeal-induced changes to antennal transcriptome profiles reveal\nshifts in odor sensitivities in Anopheles gambiae. Proc Natl Acad\nSci U S A 110:8260–8265. https://doi.org/10.1073/pnas. 1302562110 Skiri HT, Stranden M, Sandoz J-C, Menzel R, Mustaparta H (2005)\nAssociative learning of plant odorants activating the same or differ-\nent receptor neurones in the moth Heliothis virescens. J Exp Biol\n208:787–796. https://doi.org/10.1242/jeb.01431 Smid HM, Vet LE (2016) The complexity of learning, memory and\nneural processes in an evolutionary ecological context. Curr Opin\nInsect Sci 15:61–69. https://doi.org/10.1016/j.cois.2016.03.008 Robinson GE (1987) Regulation of honey bee age polyethism by juvenile\nhormone. Behav Ecol Sociobiol 20:329–338. https://doi.org/10. 1007/BF00300679 Roces F, Núñez JA (1989) Brood translocation and circadian variation of\ntemperature preference in the ant Camponotus mus. Oecologia 81:\n33–37. https://doi.org/10.1007/BF00377006 Sollai G, Solari P, Crnjar R (2018) Olfactory sensitivity to major, inter-\nmediate and trace components of sex pheromone in Ceratitis\ncapitata is related to mating and circadian rhythm. J Insect Physiol\n110:23–33. https://doi.org/10.1016/j.jinsphys.2018.08.007 Root CM, Ko KI, Jafari A, Wang JW (2011) Presynaptic facilitation by\nneuropeptide signaling mediates odor-driven food search. Cell 145:\n133–144. References https://doi. org/10.1021/pr500224b Mauelshagen J (1993) Neural correlates of olfactory learning-paradigms\nin an identified neuron in the honeybee brain. J Neurophysiol 69:\n609–625. https://doi.org/10.1152/jn.1993.69.2.609 Rabhi KK, Deisig N, Demondion E, Le Corre J, Robert G, Tricoire-\nLeignel H, Lucas P, Gadenne C, Anton S (2016) Low doses of a\nneonicotinoid insecticide modify pheromone response thresholds of Meiners T, Wäckers F, Lewis WJ (2003) Associative learning of complex\nodours in parasitoid host location. Chem Senses 28:231–236. https://doi.org/10.1093/chemse/28.3.231 Cell Tissue Res (2021) 383:149–164 163 central but not peripheral olfactory neurons in a pest insect. Proc R\nSoc B Biol Sci 283:20152987. https://doi.org/10.1152/jn.00126. 2011 spectrometric analysis, localization, and age-related changes. J\nComp Neurol 525:901–918. https://doi.org/10.1002/cne.24109 central but not peripheral olfactory neurons in a pest insect. Proc R\nSoc B Biol Sci 283:20152987. https://doi.org/10.1152/jn.00126. 2011 spectrometric analysis, localization, and age-related changes. J\nComp Neurol 525:901–918. https://doi.org/10.1002/cne.24109 Scholl C, Kübert N, Muenz TS, Rössler W (2015) CaMKII knockdown\naffects both early and late phases of olfactory long-term memory in\nthe honeybee. J Exp Biol 218:3788–3796. https://doi.org/10.1242/\njeb.124859 Rabhi KK, Esancy K, Voisin A, Crespin L, Le Corre J, Tricoire-Leignel\nH, Anton S, Gadenne C (2014) Unexpected effects of low doses of a\nneonicotinoid insecticide on behavioral responses to sex pheromone\nin a pest insect. PLoS One 9:e114411. https://doi.org/10.1371/\njournal.pone.0114411 Scholl C, Wang Y, Krischke M, Mueller MJ, Amdam GV, Rössler W\n(2014) Light exposure leads to reorganization of microglomeruli in\nthe mushroom bodies and influences juvenile hormone levels in the\nhoneybee. Dev Neurobiol 74:1141–1153. https://doi.org/10.1002/\ndneu.22195 Rath L, Giovanni Galizia C, Szyszka P (2011) Multiple memory traces\nafter associative learning in the honey bee antennal lobe: memory\ntraces in the honey bee antennal lobe. Eur J Neurosci 34:352–360. https://doi.org/10.1111/j.1460-9568.2011.07753.x Schoofs L, De Loof A, Van Hiel MB (2017) Neuropeptides as regulators\nof behavior in insects. Annu Rev Entomol 62:35–52. https://doi.org/\n10.1146/annurev-ento-031616-035500 Reisenman CE (2005) Chemosensory selectivity of output neurons inner-\nvating an identified, sexually isomorphic olfactory glomerulus. J\nNeurosci 25:8017–8026. https://doi.org/10.1523/JNEUROSCI. 1314-05.2005 Seeley T, Heinrich B (1981) Regulation of temperature in the nests of\nsocial insects. In: Heinrich B (ed) Insect thermoregulation. Wiley,\nNew York, pp 160–234 Reisenman CE (2014) Hunger is the best spice: effects of starvation in the\nantennal responses of the blood-sucking bug Rhodnius prolixus. J\nInsect Physiol 71:8–13. https://doi.org/10.1016/j.jinsphys.2014.09. 009 Shambaugh GF, Frazier JL, Castell AEM, Coons LB (1978) Antennal\nsensilla of seventeen aphid species (homoptera: Aphidinae). References Int J\nInsect Morphol Embryol 7:389–404. https://doi.org/10.1016/\nS0020-7322(78)80001-4 doi.org/10.1186/s13064-018-0111-z 244–1254. https://doi.org/10.1016/j.jinsphys.2005.07.003 Szyszka P, Galkin A, Menzel R (2008) Associative and non-associative\nplasticity in Kenyon cells of the honeybee mushroom body. Front\nSyst Neurosci 2:3. https://doi.org/10.3389/neuro.06.003.2008 Weaver N (1957) Effects of larval age on dimorphic differentiation of the\nfemale honey bee. Ann Entomol Soc Am 50:283–294. https://doi. org/10.1093/aesa/50.3.283 Takeuchi H, Yasuda A, Yasuda-Kamatani YY, Kubo T, Nakajima T\n(2003) Identification of a tachykinin-related neuropeptide from the\nhoneybee brain using direct MALDI-TOF MS and its gene expres-\nsion in worker, queen and drone heads. Insect Mol Biol 12:291–298. https://doi.org/10.1046/j.1365-2583.2003.00414.x Wegerhoff R, Rössler W, Higgins M, Oland LA, Tolbert LP (2001)\nFenvalerate treatment affects development of olfactory glomeruli\nin Manduca sexta. J Comp Neurol 430:533–541. https://doi.org/\n10.1002/1096-9861(20010219)430:4<533::aid-cne1048>3.0.co;2-\nd Taparia T, Ignell R, Hill SR (2017) Blood meal induced regulation of the\nchemosensory gene repertoire in the southern house mosquito. BMC Genomics 18:393. https://doi.org/10.1186/s12864-017-\n3779-2 Weidenmüller A, Mayr C, Kleineidam CJ, Roces F (2009) Preimaginal\nand adult experience modulates the thermal response behavior of\nants. Curr Biol 19:1897–1902. https://doi.org/10.1016/j.cub.2009. 08.059 Tautz J, Maier S, Groh C, Rössler W, Brockmann A (2003) Behavioral\nperformance in adult honey bees is influenced by the temperature\nexperienced during their pupal development. Proc Natl Acad Sci U\nS A 100:7343–7347. https://doi.org/10.1073/pnas.1232346100 Wicher D (2018) Tuning insect odorant receptors. Front Cell Neurosci\n12:94. https://doi.org/10.3389/fncel.2018.00094 Winnington AP, Napper KM, Mercer AR (1996) Structural plasticity of\nidentified glomeruli in the antennal lobes of the adult worker honey\nbee. J Comp Neurol 365:479–490. https://doi.org/10.1002/(SICI)\n1096-9861(19960212)365:3<479::AID-CNE10>3.0.CO;2-M 100:7343–7347. https://doi.org/10.1073/pnas.1232346100 Tedjakumala SR, Aimable M, Giurfa M (2014) Pharmacological modu-\nlation of aversive responsiveness in honey bees. Front Behav\nNeurosci 7:221. https://doi.org/10.3389/fnbeh.2013.00221/abstract Withers GS, Fahrbach SE, Robinson GE (1993) Selective neuroanatom-\nical plasticity and division of labour in the honeybee. Nature 364:\n238–240. https://doi.org/10.1038/364238a0 Tomé HVV, Rosi-Denadai CA, Pimenta JFN, Guedes RNC, Martins G\n(2014) Age-mediated and environmentally mediated brain and be-\nhavior plasticity in the stingless bee Melipona quadrifasciata\nanthidioides. Apidologie 45:557–567. https://doi.org/10.1007/\ns13592-014-0272-7 Wright GA, Softley S, Earnshaw H (2015) Low doses of neonicotinoid\npesticides in food rewards impair short-term olfactory memory in\nforaging-age honeybees. Sci Rep 5:15322. https://doi.org/10.1038/\nsrep15322 Trona F, Anfora G, Balkenius A, Bengtsson M, Tasin M, Knight A, Janz\nN, Witzgall P, Ignell R (2013) Neural coding merges sex and habitat\nchemosensory signals in an insect herbivore. Proc R Soc B Biol Sci\n280:20130267. p\ny\nS0020-7322(78)80001-4 https://doi.org/10.1016/j.cell.2011.02.008 Sommerlandt FMJ, Brockmann A, Rössler W, Spaethe J (2019)\nImmediate early genes in social insects: a tool to identify brain\nregions involved in complex behaviors and molecular processes\nunderlying neuroplasticity. Cell Mol Life Sci 76:637–651. https://\ndoi.org/10.1007/s00018-018-2948-z Sachse S, Rueckert E, Keller A, Okada R, Tanaka NK, Ito K, Vosshall\nLB (2007) Activity-dependent plasticity in an olfactory circuit. Neuron 56:838–850. https://doi.org/10.1016/j.neuron.2007.10.035 Stieb SM, Hellwig A, Wehner R, Rössler W (2012) Visual experience\naffects both behavioral and neuronal aspects in the individual life\nhistory of the desert ant Cataglyphis fortis. Dev Neurobiol 72:729–\n742. https://doi.org/10.1002/dneu.20982 Saveer AM, Kromann SH, Birgersson G, Bengtsson M, Lindblom T,\nBalkenius A, Hansson BS, Witzgall P, Becher PG, Ignell R (2012)\nFloral to green: mating switches moth olfactory coding and prefer-\nence. Proc R Soc B Biol Sci 279:2314–2322. https://doi.org/10. 1098/rspb.2011.2710 Stieb SM, Muenz TS, Wehner R, Rössler W (2010) Visual experience\nand age affect synaptic organization in the mushroom bodies of the\ndesert ant Cataglyphis fortis. Dev Neurobiol 70:408–423. https://\ndoi.org/10.1002/dneu.20785 Sayin S, Boehm AC, Kobler JM, De Backer J-F, Grunwald Kadow IC\n(2018) Internal state dependent odor processing and perception—\nthe role of neuromodulation in the fly olfactory system. Front Cell\nNeurosci 12:11. https://doi.org/10.3389/fncel.2018.00011 Strube-Bloss MF, Herrera-Valdez MA, Smith BH (2012) Ensemble re-\nsponse in mushroom body outputneurons of the honey bee outpaces\nspatiotemporal odor processing two synapses earlier in the antennal\nlobe. PLoS One 7:e50322. https://doi.org/10.1371/journal.pone. 0050322 Schmitt F, Vanselow JT, Schlosser A, Kahnt J, Rössler W, Wegener C\n(2015) Neuropeptidomics of the carpenter ant Camponotus\nfloridanus. ACS Publ 14:1504–1514. https://doi.org/10.1021/\npr5011636 Strube-Bloss MF, Rössler W (2018) Multimodal integration and stimulus\ncategorization in putative mushroom body output neurons of the Strube-Bloss MF, Rössler W (2018) Multimodal integration and stimulus\ncategorization in putative mushroom body output neurons of the Schmitt F, Vanselow JT, Schlosser A, Wegener C, Rössler W (2017)\nNeuropeptides in the desert ant Cataglyphis fortis: mass 164 Cell Tissue Res (2021) 383:149–164 communication in a specialist herbivore. Sci Rep 6:32666. https://\ndoi.org/10.1038/srep32666 honeybee. R Soc Open Sci 5:171785. https://doi.org/10.1098/rsos. 171785 Sugie A, Marchetti G, Tavosanis G (2018) Structural aspects of plasticity\nin the nervous system of Drosophila. Neural Develop 13:14. https:// Sugie A, Marchetti G, Tavosanis G (2018) Structural aspects of plasticity\nin the nervous system of Drosophila. Neural Develop 13:14. https://\ndoi.org/10.1186/s13064-018-0111-z Wang S, Zhang S, Sato K, Srinivasan M (2005) Maturation of odor\nrepresentation in the honeybee antennal lobe. p\ny\nS0020-7322(78)80001-4 J Insect Physiol 51:\n1244–1254. https://doi.org/10.1016/j.jinsphys.2005.07.003 doi.org/10.1186/s13064-018-0111-z https://doi.org/10.1098/rspb.2013.0267 Zhang Q, Schlyter F, Anderson P (1999) Green leaf volatiles interrupt\npheromone response of spruce bark beetle, Ips typographus. J Chem\nEcol 25:2847–2861. https://doi.org/10.1023/A:1020816011131 Tsitoura P, Iatrou K (2016) Positive allosteric modulation of insect olfac-\ntory receptor function by ORco agonists. Front Cell Neurosci 10:\n275. https://doi.org/10.3389/fncel.2016.00275 Zhao B, Sun J, Zhang X, Mo H, Niu Y (2019) Long-term memory is\nformed immediately without the need for protein synthesis-\ndependent consolidation in Drosophila. Nat Commun 10:1–11. https://doi.org/10.1038/s41467-019-12436-7 Vergoz V, McQuillan HJ, Geddes LH, Pullar K, Nicholson BJ, Paulin\nMG, Mercer AR (2009) Peripheral modulation of worker bee re-\nsponses to queen mandibular pheromone. Proc Natl Acad Sci U S A\n106:20930–20935. https://doi.org/10.1073/pnas.0907563106 Publisher’s note Springer Nature remains neutral with regard to jurisdic-\ntional claims in published maps and institutional affiliations. Wang F, Deng J, Schal C, Lou Y, Zhou G, Ye B, Yin X, Xu Z, Schen L\n(2016) Non-host plant volatiles disrupt sex pheromone"
https://openalex.org/W3154532589	https://link.springer.com/content/pdf/10.1007/JHEP06(2021)129.pdf	English	null	Chemical-potential-assisted particle production in FRW spacetimes	The Journal of high energy physics/The journal of high energy physics	2,021	cc-by	30,931	Published for SISSA by Springer Received: April 27, 2021 Revised: June 7, 2021 Accepted: June 8, 2021 Published: June 21, 2021 Received: April 27, 2021 Revised: June 7, 2021 Accepted: June 8, 2021 Published: June 21, 2021 Received: April 27, 2021 Revised: June 7, 2021 Accepted: June 8, 2021 Published: June 21, 2021 JHEP06(2021)129 Open Access, c⃝The Authors. Article funded by SCOAP3. Keywords: Cosmology of Theories beyond the SM, Beyond Standard Model Chemical-potential-assisted particle production in FRW spacetimes Chon Man Sou, Xi Tong and Yi Wang Department of Physics, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P.R. China The HKUST Jockey Club Institute for Advanced Study, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P.R. China E-mail: cmsou@connect.ust.hk, xtongac@connect.ust.hk, phyw@ust.hk Chon Man Sou, Xi Tong and Yi Wang Department of Physics, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P.R. China The HKUST Jockey Club Institute for Advanced Study, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P.R. China E-mail: cmsou@connect.ust.hk, xtongac@connect.ust.hk, phyw@ust.hk E-mail: cmsou@connect.ust.hk, xtongac@connect.ust.hk, phyw@ust.hk Abstract: We analyze gravitational particle production assisted by chemical potential. By utilizing the uniformly smoothed Stokes-line method and Borel summation, we gain in- sight into the ﬁne-grained history of enhanced particle production. Analytic/semi-analytic formulae describing the production amount, time and width are obtained for both spin-1 and spin-1/2 particles in various FRW spacetimes. Our work also serves as a concrete demonstration of the uniformly smoothed Stokes-line method applied to cosmology. Keywords: Cosmology of Theories beyond the SM, Beyond Standard Model ArXiv ePrint: 2104.08772 Open Access, c⃝The Authors. Article funded by SCOAP3. https://doi.org/10.1007/JHEP06(2021)129 Contents 1 Introduction 1 2 Chemical potential: generalities and dS solutions 3 2.1 Spin-0 4 2.2 Spin- 1 2 6 2.3 Spin-1 8 3 The uniformly smoothed Stokes-line method 10 3.1 Bosonic case 10 3.2 Fermionic case 18 4 Analysis of particle production in various spacetimes 26 4.1 dS 27 4.2 Deviation from dS of the ϵ-type 34 4.3 Deviation from dS of the η-type 38 4.4 Radiation-domination era 46 4.5 Matter-dominated era 51 5 Summary and outlook 56 A A checklist of results 58 B The 1/4 puzzle 58 Contents 1 Introduction 1 2 Chemical potential: generalities and dS solutions 3 2.1 Spin-0 4 2.2 Spin- 1 2 6 2.3 Spin-1 8 3 The uniformly smoothed Stokes-line method 10 3.1 Bosonic case 10 3.2 Fermionic case 18 4 Analysis of particle production in various spacetimes 26 4.1 dS 27 4.2 Deviation from dS of the ϵ-type 34 4.3 Deviation from dS of the η-type 38 4.4 Radiation-domination era 46 4.5 Matter-dominated era 51 5 Summary and outlook 56 A A checklist of results 58 B The 1/4 puzzle 58 JHEP06(2021)129 1 Introduction The study of particle production in cosmology is, in a sense, the study of the origin of mat- ter itself. Arguably, the most natural choice of the initial state of the universe is the one in which strict spatial homogeneity is maintained, if space is to retain its physical meaning at all. Such a perfectly homogeneous state must also be devoid of any ordinary matter, which we view as a collection of fundamental particles. In addition, on the classical level, this strict homogeneity will be kept throughout the cosmic evolution if the Lagrangian respects translational symmetry. This would naively suggest a dull universe with no matter but zero modes only. However, this is not the whole story. The dynamics of the background spacetime can bring the vacuum ﬂuctuations to reality, breaking the spatial homogeneity on a quantum level. This spontaneous breaking of spatial translational invariance is accom- panied with the appearance of perturbative modes in spacetime, which upon quantization, become real particles. – 1 – Such is the scenario for the paradigm of inﬂation [1–4]. The initial condition of the inﬂationary universe is chosen as a Bunch-Davis (BD) vacuum, i.e., a coherent state of the inﬂaton ﬁeld that is annihilated by the positive-frequency part of all ﬁelds with non-zero momenta. However, the expansion of the spacetime stretches the wavelength of diﬀerent modes φk and produces a squeezed state with occupation number \|β(k)\|2 ̸= 0. For φ being the inﬂaton or the graviton, this occupation number is exponentially large so that these particles decohere and become essentially classical waves, which source the primor- dial inhomogeneity for the later cosmic evolutions. For φ being heavier degrees of freedom, the occupation number is ﬁnite and typically suppressed by an exponentially small factor \|β(k)\|2 ∝e−2πm/H, where H ≲1013GeV is the Hubble parameter during inﬂation. These gravitationally produced particles are the very ﬁrst matter emergent from the BD vacuum in the inﬂationary universe and their interactions leave characteristic non-Gaussian im- prints on the Cosmic Microwave Background (CMB) as well as the Large Scale Structure (LSS). This recently thriving ﬁeld known as cosmological collider physics [5–10] thus has an intimate relation with the phenomenon of particle production from the vacuum. For example, the signal strength is directly proportional to the square root of the production amount, S ∝\|β(k)\|. 1 Introduction Thus heavy particles with m ≫H are extremely diﬃcult to probe if they are produced in the purely gravitational way. JHEP06(2021)129 This problem motivates several proposals where the exponential suppression can be alleviated. One possibility is to consider special inﬂation models. For instance, in axion- monodromy inﬂation [11, 12], time-dependent mass terms violate adiabaticity and lead to a dramatic ampliﬁcation of particle number density and thus the size of non-Gaussianity [13]. In warm inﬂation [14, 15], thermally produced heavy particles are Boltzmann-suppressed by an alternative temperature much higher than the Hubble scale, also giving rise to enhanced cosmological collider signals [16]. Alternatively, one can consider the interesting possibility of introducing a chemical potential. This can be naturally achieved via a rolling scalar ﬁeld φ coupled to massive ﬁelds through operators of the form ∂µφJµ, where Jµ is a certain current made from massive ﬁelds. For massive spin-1 particles, choosing the Chern-Simons current JCS results in an enhanced production of vector boson [17–21]. For massive spin-1/2 fermions, choosing the chiral current J5 leads to a natural ampliﬁcation of fermion production rate [22–25]. Similar ampliﬁcation is also found for charged scalar particles if we generalize the chemical potential operator to κµJµ, and reinterpret the chemical potential κµ as a background gauge ﬁeld creating Schwinger pairs [26–28]. These mechanisms of chemical-potential- assisted particle production typically break parity or rotational invariance, hence leaving sizable and distinctive signatures in cosmological observables. In addition to physics during inﬂation, chemical potential can also play a role in the late universe. The aforementioned chemical potential introduced by a rolling Axion-Like Particle (ALP) generates a tachyonic instability in the gauge boson sector. This can, for ex- ample, eﬃciently convert the ALP to Dark Photon Dark Matter (DPDM) [29], and produce chiral gravitational waves [30–32]. In the fermion sector, chemical potential also sources the helicity asymmetry of fermion numbers, which can be important for baryogenesis [22, 33]. – 2 – Thus, chemical-potential-assisted particle production is a generic phenomenon that appears in many diﬀerent contexts and setups. A general and systematic investigation of particle production in the presence of chemical potential is necessary. Thus, chemical-potential-assisted particle production is a generic phenomenon that appears in many diﬀerent contexts and setups. A general and systematic investigation of particle production in the presence of chemical potential is necessary. 1 Introduction In this work, we re-derive Berry’s uniformly smoothed Stokes-line method [34–37] and apply it to analyze the ﬁne-grained production history of particles of mass m with chemical potential κ. By ﬁne-grained production history, we mean the production amount, time and width (duration) exact to the leading order in the super-adiabatic expansion. We found that for a constant chemical potential κ = const, spin-1/2 fermions and spin-1 vector bosons share a similar production history, with a simple yet subtle replacement rule m2 ↔m2+κ2. We also derive analytic/semi-analytic formulae for the production history in ﬁve common FRW spacetimes. JHEP06(2021)129 We note that there are many past studies in the literature that utilize the Stokes phe- nomenon to study particle production. For example, it is applied to the Sauter-Schwinger eﬀect [38–43], to Hawking radiation [44], to the adiabatic particle number in global de Sitter (dS) spacetime [45–47], to dark matter production at the end of inﬂation [48], to preheating [49, 50], and to particle production triggered by vacuum decay [51]. An ex- cellent review is recently given in [52]. However, we point out that most of them (with the exception of [40, 47, 51]) focus on the asymptotic production amount far away from the Stokes line. To our knowledge, the analytic calculation of production time and width presented in this work is a new ingredient in this area, with or without chemical potential. Understanding these ﬁne production details is useful, for example, in the loop-level estima- tion of cosmological collider signals, or in estimation of backreaction time scales. We hope this work also serves to demonstrate the application of the uniformly smoothed Stokes-line method applied to cosmology. This paper is organized as follows. In section 2, we ﬁrst discuss the generalities of chemical potential and deﬁne the model we work with. Then in section 3, we re-derive the uniformly smoothed Stokes phenomenon for both bosonic and fermionic systems and justify its validity for the case with signiﬁcant particle production. In section 4, we move on to work out the production details for spin-1,1/2 particles in ﬁve common FRW spacetimes. At last, we summarize and give outlooks in section 5. For readers who wish to skip the detailed analysis to directly look up the results, we assemble our formulae into a checklist in appendix A. 2 Chemical potential: generalities and dS solutions In this section, loosely following [53], we give some general discussions on chemical po- tential in cosmology. In thermodynamics, chemical potential is originally introduced by Gibbs to describe the change of internal energy of a system with respect to the change of particle numbers when the entropy and volume are held ﬁxed. More formally in statistical mechanics, it is identiﬁed with the Lagrange multiplier κ of total particle number when the systems in a grand canonical ensemble are allowed to exchange particles with each other. Starting from the partition function for a grand canonical ensemble, Z = e−(H−κN)/T , we can straightforwardly generalize it to the ﬁeld theory context as a path integral in the – 3 – phase space of a ﬁeld φ, Z = Z DφDπe−iR dt(H[φ,π]−κN[φ,π]) , (2.1) (2.1) where H is the Hamiltonian and N is the particle number operator associated with a certain symmetry (which may or may not be exact). Throughout this work, we will assume a spatially ﬂat FRW spacetime background with ds2 = −dt2 + a(t)2dx2. Going into the ﬁeld conﬁguration space and write the chemical potential term as the integral over a local density, we have Z = Z DφeiR dtd3x√ \|g\|(L(φ,∂φ)+κJ0(φ,∂φ)) , (2.2) (2.2) JHEP06(2021)129 where N ≡ R d3xa3J0. Now we can turn on the spacetime dependence of κ and interpret it as the zeroth component of a local vector ﬁeld κµ(t, x). Thus the general form of a chemical potential as a background ﬁeld coupled to the matter ﬁeld is ∆Lchem ≡κµ(x)Jµ(x) . (2.3) (2.3) Now if one inspects the eﬀect of introducing such a chemical potential term into the matter Lagrangian, the result will depend on two aspects. First, in the absence of κµ, if the matter current is derived from an exact symmetry, it is conserved as an operator identity, i.e., ⟨∇µJµ⟩= 0, where ∇is a covariant derivative with respect to the metric g. Then one can always consistently gauge this symmetry by minimally coupling the conserved current to a vector potential. We are free to choose κµ(x) to be such a background gauge ﬁeld. If the chemical potential κ = κµdxµ is closed in the 1-form sense, dDκ = dκ + κ ∧κ = 0, where D is compatible with the gauge ﬁeld connection κ, the background gauge ﬁeld then has no ﬁeld strength and is gauge-equivalent to vacuum. 2 Chemical potential: generalities and dS solutions Thus we can perform a gauge transformation to eliminate κ locally. Such is the case if κ = κ(t) is Abelian and spatially homogeneous. However, we point out that the elimination of the chemical potential term is not completely trivial in the scalar case, as we will see below, since it shifts the scalar mass in a quadratic way. The second possibility is that if the matter current is not built from an exact symmetry and is hence not conserved. This suggests that there is no consistent way of coupling it to a background gauge ﬁeld and thus interpreting it as the chemical potential. In summary, we give the following necessary condition of a chemical potential term that has physical eﬀects other than quadratically shifting the mass of the matter particle, either dDκ = dκ + κ ∧κ ̸= 0, or ∇· J ̸= 0 . (2.4) (2.4) either dDκ = dκ + κ ∧κ ̸= 0, or ∇· J ̸= 0 . (2 In the following discussion, we provide several examples that satisfy the above criterion and have interesting particle production features. We will start the general discussions in ﬂat FRW spacetime and retreat to exact dS spacetime when solving the Equations of Motion (EoMs). The general FRW EoMs will be discussed in later sections. In the following discussion, we provide several examples that satisfy the above criterion and have interesting particle production features. We will start the general discussions in ﬂat FRW spacetime and retreat to exact dS spacetime when solving the Equations of Motion (EoMs). The general FRW EoMs will be discussed in later sections. 2.1 Spin-0 Consider a complex scalar ﬁeld σ, the only non-trivial chemical potential term we can ﬁnd at quadratic level is with Jµ = i(σ∗∂µσ −σ∂µσ∗). If the original Lagrangian is U(1)- symmetric, this current is conserved. Therefore, according to (2.4), we have to go for the – 4 – ﬁrst possibility.1 We can write the following action in the FRW background, S0 = Z d4x√−g h −gµν∂µσ∗∂νσ −m2σ∗σ + κµJµi , (2.5) (2.5) where gµν ≡a(τ)2ηµν using comoving coordinates. Absorbing this chemical potential term into the derivative term and shifting the mass term accordingly, we have S0 = Z d4x√−g h −gµν(∂µ −iκµ)σ∗(∂ν + iκν)σ −(m2 −κµκµ)σ∗σ i . (2.6) (2.6) As mentioned above, if κµ=κ0(τ)δ0 µ, the whole system can be considered as a charged scalar moving in the vacuum with a new mass M2(τ) ≡m2 −κ0(τ)2 a(τ)2 . There is no enhancement of particle production unless M2(τ) becomes negative or its time dependence violates adiabatic condition. JHEP06(2021)129 If the ﬁrst criterion is satisﬁed, then there is a non-zero ﬁeld strength Fµν = ∂µκν − ∂νκµ. For instance, with κ0 = 0, κi = κi(τ), there is a uniform (time-dependent) electric ﬁeld Fi0 = −κ′ i(τ), where prime denotes a derivative with respect to the conformal time τ. The enhancement of particle production is exactly the Schwinger eﬀect of this electric ﬁeld [28]. The EoM of σ in momentum space reads (aσk)′′ + k2 −2k · κa + m2a2 −a′′ a (aσk) = 0 , (2.7) (2.7) where κ ≡κi a ˆei. Clearly the second term in the square bracket breaks rotational symmetry and stands for the eﬀect of the background electric ﬁeld. It introduces an angular depen- dence in the eﬀective mass of diﬀerent modes. Those with lighter eﬀective mass tend to get produced more easily, especially for the mode traveling at the same direction as κ. We shall see that this term linear in momentum is typical in the presence of chemical potentials. They represent the bias on the eﬀective mass of diﬀerent modes introduced by the chemical potential. For large enough \|κ\|, there is even a tachyonic instability and the pair creation rate becomes exponentially large. To be more quantitative, we set \|κ\| = const and limit ourselves to dS by taking a = −1 Hτ . 2.1 Spin-0 Then the solution to the EoM is given by σk = −eπˆk·˜κ/2 √ 2kHτ W−iˆk·˜κ,iµ(2ikτ) , µ ≡ r ˜m2 −9 4 , (2.8) (2.8) where ˆk = k \|k\|, ˜κ ≡κ H , ˜m ≡m H and W is the Whittaker function that matches the BD initial condition at τ →−∞. The late-time expansion reveals an angular-dependent production amount \|β(k)\|2 = e2π(µ+ˆk·˜κ) + 1 e4πµ −1 . (2.9) (2.9) 1In fact, even if the U(1) symmetry is explicitly broken with ∇· J ̸= 0, it can be shown that there is only a quadratic mass-shift and no enhancement particle production is present [53]. This is why (2.4) is only a necessary condition, rather than being suﬃcient. However, we note that a diﬀerent opinion is given in [54], where it is argued that a scalar chemical potential can also bring isotropic enhancement to particle production. – 5 – – 5 – To gain an intuitive understanding of the enhancement, we can go to the large mass limit m ≫H. Then the leading order particle number is To gain an intuitive understanding of the enhancement, we can go to the large mass limit m ≫H. Then the leading order particle number is \|β(k)\|2 µ≫1 −−−→e−2π(µ−ˆk·˜κ) . (2.10) (2.10) Therefore, the direct consequence of a chemical potential is a linear and biased (rather than quadratic and un-biased) shift of the eﬀective mass of the particle modes, making them easier or harder to produce.2 Therefore, the direct consequence of a chemical potential is a linear and biased (rather than quadratic and un-biased) shift of the eﬀective mass of the particle modes, making them easier or harder to produce.2 2See an alternative understanding of chemical potential as a non-trivial modiﬁcation of dispersion relation in [53]. 2.2 Spin- 1 2 JHEP06(2021)129 In the massive spin- 1 2 case, the second possibility of (2.4) can be satisﬁed by choosing the axial current Jµ 5 = eµ a ¯Ψγaγ5Ψ. Hence a time-like chemical potential κµ(τ) ∝δ0 µ has no interpretation as a trivial background pure-gauge and can become physically relevant. To illustrate how it assists gravitational particle production, we and choose a Majorana fermion model [23] written in a Weyl basis Ψ = ψ ψ† . The Dirac fermion case can be obtained by combing two Majorana fermions with analogous behaviors. The action of a Majorana fermion with chemical potential reads S1/2 = Z d4x√−g iψ†¯σaeµ aDµψ −1 2m(ψψ + ψ†ψ†) + κµeµ aψ†¯σaψ , (2.11) (2.11) where Dµ ≡∂µ −i 4ωµabσab, ωµab = eν b∇µe c ν ηac. This chemical potential term can arise from, for instance, a dimension-5 coupling to a rolling scalar, ∆Lchem = −∂µφJµ 5 2Λ = κµeµ aψ†¯σaψ, κµ ≡∂µφ Λ . (2.12) (2.12) After choosing a tetrad e a µ = a(τ)δa µ, eµ a = a(τ)−1δµ a and κµ = a(τ)κδ0 µ with κ = ˙φ Λ = const, the action simpliﬁes to After choosing a tetrad e a µ = a(τ)δa µ, eµ a = a(τ)−1δµ a and κµ = a(τ)κδ0 µ with κ = ˙φ Λ = const, the action simpliﬁes to S1/2 = Z dτd3x iψ†¯σaδµ a∂µψ −1 2am(ψψ + ψ†ψ†) + aκψ†¯σ0ψ , (2.13) (2.13) where for simplicity, we have rescaled ψ →a−3/2ψ. In momentum space, we can perform a standard decomposition into helicity eigenmodes, where for simplicity, we have rescaled ψ →a−3/2ψ. In momentum space, we can perform a standard decomposition into helicity eigenmodes, ψ(τ, x) = Z d3k (2π)3 X s=± h hs(ˆk)us(τ, k)eik·xbs k + iσ2hs∗(ˆk)vs(τ, k)∗e−ik·xbs† k i , (2.14) (2.14) where σ · ˆkhs(ˆk) = shs(ˆk). The EoM reads where σ · ˆkhs(ˆk) = shs(ˆk). The EoM reads iu′ ± = (∓k −aκ)u± + amv± iv′ ± = amu± + (±k + aκ)v± . (2.15) (2.15) 2See an alternative understanding of chemical potential as a non-trivial modiﬁcation of dispersion relation in [53]. – 6 – – 6 – This set of equations is exactly solvable in dS. 2.2 Spin- 1 2 With a BD initial condition, the mode functions take the form u+(τ, k) = ˜me−π˜κ/2 √ −2kτ W−1 2 +i˜κ, i √ ˜m2+˜κ2(2ikτ), u−(τ, k) = eπ˜κ/2 √ −2kτ W 1 2 −i˜κ, i √ ˜m2+˜κ2(2ikτ) v+(τ, k) = e−π˜κ/2 √ −2kτ W 1 2 +i˜κ, i √ ˜m2+˜κ2(2ikτ), v−(τ, k) = ˜meπ˜κ/2 √ −2kτ W−1 2 −i˜κ, i √ ˜m2+˜κ2(2ikτ) , (2.16) (2.16) where ˜m ≡m H and ˜κ = κ H . Similar to the scalar case, performing an IR expansion at τ = 0 and matching the Bogoliubov coeﬃcients, one can arrive at the production amount formula [22] JHEP06(2021)129 \|β±(k)\|2 = e2π( √ ˜m2+˜κ2∓˜κ) −1 e4π √ ˜m2+˜κ2 −1 . (2.17) (2.17) When the mass is large and chemical potential is small, the leading order particle number again takes the form of a Boltzmann factor with linearly biased eﬀective mass, When the mass is large and chemical potential is small, the leading order particle number again takes the form of a Boltzmann factor with linearly biased eﬀective mass, \|β±(k)\|2 ˜m≫\|˜κ\|, ˜m≫1 −−−−−−−−→e−2π( ˜m±˜κ) . (2.18) (2.18) For a positive chemical potential, the negative-helicity mode gets ampliﬁed whereas the positive-helicity mode is suppressed. However, when the chemical potential is larger than the mass scale, the enhancement in the negative-helicity particle production begins to saturate, For a positive chemical potential, the negative-helicity mode gets ampliﬁed whereas the positive-helicity mode is suppressed. However, when the chemical potential is larger than the mass scale, the enhancement in the negative-helicity particle production begins to saturate, \|β−(k)\|2 ˜κ≫˜m≫1 −−−−−→e−2π( √ ˜m2+˜κ2−˜κ) ≈e−π ˜m2/˜κ = e−πm2 κH . (2.19) (2.19) This is essentially the Pauli blocking phenomenon generic to all fermionic systems. The exclusion principle forbids any mode being occupied more than once. Actually, fermion production with constant chemical potential in general FRW space- times can be understood in an elegant way. In terms of the physical time t, the EoM is essentially a two-state system evolving according to a Schrödinger equation i ∂ ∂t u± v± ! = ∓k a −κ m m ±k a + κ ! u± v± ! . (2.20) (2.20) Without loss of generality, let us consider the negative helicity mode with s = −(the s = + helicity mode is obtained by k →−k). The unitary evolution governed by a Schrödinger equation (2.20) preserves the normalization \|u−\|2 + \|v−\|2 = 1. 2.2 Spin- 1 2 The BD initial condition selects the positive frequency mode u−in the early time limit t →−∞. If κ > 0, there can be a time when the physical wavelength k a of the mode is comparable to the chemical potential scale κ. In the language of quantum physics, we have an avoided crossing. Namely when the diagonal elements of the Hamiltonian vanish, its instantaneous eigenvalues approach each other, but a complete degeneracy is avoided due to the oﬀ- diagonal terms. If this process occurs adiabatically, according to the adiabatic theorem, the state smoothly maintains its positive-frequency trajectory and there is not much particle production. However, if κ is large, the avoided crossing becomes a non-adiabatic one, with – 7 – almost all positive-frequency crossed into negative-frequency part and thus nearly maximal particle production. Viewed in this way, Pauli exclusion principle is a built-in feature of fermions such that the evolution of (u, v) is unitary, as opposed to the symplectic evolution of bosons (σ, ˙σ), which can be unbounded from above and exponentiating (e.g., if κ ≫m in (2.10)). Interestingly, the large-chemical potential limit in the spin-1/2 fermion case enjoys a universal behavior in general FRW spacetimes. To see this more explicitly, we can assume κ ≫m and expand around the crossing time t∗, where κ = k/a(t∗), i ∂ ∂t u− v− ! ≈ κH(t∗)(t −t∗) m m −κH(t∗)(t −t∗) ! u− v− ! . (2.21) JHEP06(2021)129 (2.21) This is none other than a Landau-Zener (LZ) model with η = m and γ = 2κH(t∗) [55, 56]. The corresponding LZ parameter describing adiabaticity is z = η2 γ = m2 2κH(t∗), where z ≫1 corresponds to adiabatic transitions and 0 < z ≪1 corresponds to diabatic transi- tions. See ﬁgure 1 for justiﬁcation of approximation using an LZ transition in dS. Thus the crossing probability is given by the exponential factor This is none other than a Landau-Zener (LZ) model with η = m and γ = 2κH(t∗) [5 2 2 \|β−(k)\|2 = e−2πz = e− πm2 κH(t∗(k)) , (2.22) (2.22) in agreement with the exact dS result (2.19). Furthermore, we obtain the production time t∗as the solution to the equation κ = k a(t∗). Naively, we would expect the production width to be the time scale at which the LZ resonance happen, i.e., 4η γ = 2m κH(t∗). 2.2 Spin- 1 2 However, we shall see in section 3.2 that this is not the case. The actual particle production width is ∆t∗= q 2π γ = 1 √ κH(t∗)/π, which is typically shorter than the LZ time scale [36]. 2.3 Spin-1 One can also easily ﬁnd a non-conserved current in the Abelian massive vector boson case, namely the Chern-Simons current Jµ CS = EµνρσAνFρσ = 1 √−gϵµνρσAνFρσ. The chemical potential term looks like ∆Lchem = 1 2 Z d4x√−gκµJµ CS = 1 2 Z dτd3xκµϵµνρσAνFρσ . (2.23) (2.23) For a choice of κµ = θ′(τ)δ0 µ ≡a(τ)κδ0 µ, or κ ≡˙θ = const, an integration-by-part gives a time-dependent θ-term (a rolling axion), For a choice of κµ = θ′(τ)δ0 µ ≡a(τ)κδ0 µ, or κ ≡˙θ = const, an integration-by-part gives a time-dependent θ-term (a rolling axion), ∆Lchem = 1 2 Z d4x√−gκµJµ CS = −1 4 Z dτd3xθ(τ)ϵµνρσFµνFρσ . (2.24) (2.24) Thus the system is described by an axion electrodynamics Lagrangian [57] S1 = Z d4x√−g −1 4FµνF µν + 1 2m2AµAµ −1 4θ(τ)EµνρσFµνFρσ . (2.25) (2.25) A massive vector boson has three degrees of freedom, two transverse modes and a longitu- dinal mode, from which the time-like component is solved using the constraint ∇µAµ = 0. – 8 – Figure 1. The instantaneous eigenvalues of the Hamiltonian in dS. Here we have chosen m = 15H, κ = 60H, or, after translating to LZ model parameters, η = 15H, γ = 120H2. The gray band corresponds to the naive time scale 4η γ = 0.5H−1 while the red band corresponds to the actual production width ∆t∗= q 2π γ = 0.23H−1. As long as κ ≫m, the light pink band is narrow enough so that the expansion of spacetime becomes irrelevant, and the instantaneous energy eigenvalues approach to that of a LZ model. JHEP06(2021)129 Figure 1. The instantaneous eigenvalues of the Hamiltonian in dS. Here we have chosen m = 15H, κ = 60H, or, after translating to LZ model parameters, η = 15H, γ = 120H2. The gray band corresponds to the naive time scale 4η γ = 0.5H−1 while the red band corresponds to the actual production width ∆t∗= q 2π γ = 0.23H−1. As long as κ ≫m, the light pink band is narrow enough so that the expansion of spacetime becomes irrelevant, and the instantaneous energy eigenvalues approach to that of a LZ model. Decomposing the spatial components into helicity eigenstates, we have Ai(τ, x) = Z d3k (2π)3 eik·x X s=0,± ϵs i(ˆk) h fs(τ, k)as k + fs(τ, k)∗as† −k i . 2.3 Spin-1 (2.26) (2.26) The EoM then reads The EoM then reads f′′ ± + k2 ± 2kκa + m2a2 f± = 0 , f′′ 0 + k2 + m2a2 f0 = 0 (2.27) (2.27) As a result, the transverse modes are aﬀected by the chemical potential while the longitu- dinal mode is not. Focusing on the transverse modes, we can solve the EoMs analogously as (2.7), f±(τ, k) = e∓π˜κ/2 √ 2k W−i˜κ,iµ(2ikτ) , µ ≡ r ˜m2 −1 4 . (2.28) (2.28) The transverse particle production amount is then The transverse particle production amount is then \|β±(k)\|2 = e2π(µ∓˜κ) + 1 e4πµ −1 . (2.29) (2.29) In the large mass limit, the eﬀect of chemical potential again simpliﬁes to a linear bias over the eﬀective mass, In the large mass limit, the eﬀect of chemical potential again simpliﬁes to a linear bias over the eﬀective mass, \|β±(k)\|2 µ≫1 −−−→e−2π(µ±˜κ) . (2.30) (2.30) – 9 – – 9 – Here, without the protection of the exclusion principle, the production amount starts to become exponentially large if \|˜κ\| > µ. This is commonly recognized as a tachyonic insta- bility in the study of axions. In this parameter regime, the backreaction to the rolling θ(τ) must be taken into account. 3 The uniformly smoothed Stokes-line method In this section, we derive the uniformly smoothed Stokes-line method for both spin-1 vector bosons and spin-1/2 Majorana fermions, providing a framework to analyze the histories of particle production with chemical potentials. Mathematically, the evolution of spin-1 vector bosons (2.27) and spin-1/2 Majorana fermions (2.20) behave as second order diﬀerential equations and the Schrödinger equation with two quantum states respectively, and such systems can experience the emergence of negative-frequency part starting from an initial positive-frequency solution, known as the Stokes phenomenon: JHEP06(2021)129 Ψ+(τ) →α(τ)Ψ+(τ) + β(τ)Ψ−(τ) , (3.1) (3.1) where Ψ±(τ) are the positive/negative-frequency parts of either bosons or fermions, and the two time-dependent functions α(τ) and β(τ) can be regarded as the Bogoliubov coeﬃcients, associated with the particle production. The Stokes phenomena in second order diﬀerential equations [35, 40, 47, 48, 58, 59] and transitions between two quantum states [36, 37] have been studied in many works with the assumption that the magnitude of the emergent part \|β(τ)\| is exponentially small, and all of these works point out that the singulant F(τ) = −2i Z τ τc ω(τ1)dτ1 , (3.2) (3.2) deﬁned as the diﬀerence between the positive and negative phases accumulated from the complex turning point τc satisfying ω(τc) = 0, are important to describe the details of the Stokes phenomena. To be speciﬁc, when the systems evolve near the Stokes line, deﬁned as the line linking τc and τ ∗ c with ImF(τ) = 0, the negative-frequency part starts to produce, and the production histories including the amounts and widths can be calculated from the singulant F(τ), proved with the technique of optimally truncating the asymptotic series solution of (3.1). In the following subsections, we ﬁrst follow previous studies to apply the optimally truncated asymptotic series solution to calculate the particle production for vector bosons and fermions respectively, and this framework works properly when the particle produc- tion is exponentially small. We then analyze the situation when the exponential particle production is signiﬁcantly enhanced by the chemical potential, so that the optimal trun- cation technique is not applicable, and the Borel summation technique should be applied to obtain the particle production. 3In dS, it is more convenient to choose z = −kτ. Then the form of the singulant integral will diﬀer by a sign and the zc will be on the upper half complex plane. These convention diﬀerences do not change the physical results and one can choose the convention that best-suit the problem. We follow the guideline that Im F decreases as conformal time increases, and that the singulant integral always starts with the turning point on the lower half conformal time plane. 3.1 Bosonic case Consider the mode expansion for the transverse component of a massive vector boson, A⊥ i = Z d3k (2π)3 eik·x X s=± ϵs i(ˆk) h fs(τ, k)as k + fs(τ, k)∗as† −k i . (3.3) (3.3) – 10 – The mode function satisﬁes a second order EoM of the general form d2 dz2 f(z) + λ2w2(z)f(z) = 0 , w2 = 1 + 2sκa k + m2a2 k2 , (3.4) (3.4) where we use the dimensionless variable3 z = kτ, and we denote ′ as the derivative with respect to z starting from here. The helicity label s is also omitted for simplicity. The asymptotic parameter λ, deﬁned for later analysis of asymptotic series, is supposed to be large, and we can choose λ = m if the mass is the largest parameter in the problem. (3.4) has the same structure as the one-dimensional time-independent Schrödinger equation with a barrier, and therefore we adopt the method of analyzing waves near Stokes lines [35] to study the particle production. We can express the solution of (3.4) with the WKB form JHEP06(2021)129 f(z) = C p 2W(z)e−iλR z zc W(z1)dz1 , (3.5) (3.5) where C is a constant ﬁxed by the initial and normalization conditions, zc is the complex turning point deﬁned by w(zc) = 0 and located at the lower-half complex plane, and the function W(z) satisﬁes W 2(z) = w2(z) −1 λ2 " W ′′(z) 2W(z) −3 4 W ′(z) W(z) 2# . (3.6) (3.6) In general, we cannot obtain the exact solution of W(z), but we can apply the iterative adiabatic expansion to approximate it [40, 47] W (n+1)(z) = v u u u tw2(z) −1 λ2  W ′′(n)(z) 2W (n)(z) −3 4 W ′(n)(z) W (n)(z) !2  = v u u u tw2(z) − q W (n)(z) λ2 d2 dz2   1 q W (n)(z)  , (3.7) (3.7) with W (0)(z) = w(z). Such an iterative relation can be used to derive the asymptotic series solution of W(z) W(z) = w(z) ∞ X n=0 ϕ2n(z) λ2n , (3.8) (3.8) and the series truncation at the O(λ−2n) is W (n) deﬁned in (3.7). 3.1 Bosonic case Applying (3.8) can express the oscillating phase integral in (3.5) with the asymptotic series, and we will see later in section 4.1 that this is important for obtaining the 1/4 correction to the vector and the series truncation at the O(λ−2n) is W (n) deﬁned in (3.7). Applying (3.8) can express the oscillating phase integral in (3.5) with the asymptotic series, and we will see later in section 4.1 that this is important for obtaining the 1/4 correction to the vector 3In dS, it is more convenient to choose z = −kτ. Then the form of the singulant integral will diﬀer by a sign and the zc will be on the upper half complex plane. These convention diﬀerences do not change the physical results and one can choose the convention that best-suit the problem. We follow the guideline that Im F decreases as conformal time increases, and that the singulant integral always starts with the turning point on the lower half conformal time plane. – 11 – boson’s dS eﬀective mass. On the other hand, it is well-known that the terms in the asymptotic series (3.8) keep increasing when n is suﬃciently large [35, 40, 47, 58] boson’s dS eﬀective mass. On the other hand, it is well-known that the terms in the asymptotic series (3.8) keep increasing when n is suﬃciently large [35, 40, 47, 58] ϕ2n λ2n ≈−(2n −1)! πF 2n , (3.9) (3.9) where F(z) = −2iλ Z z zc w(z1)dz1 , (3.10) (3.10) is Dingle’s singulant variable [59]. So we can truncate the series sum at a suitable order n to approximate the solution (3.5). JHEP06(2021)129 The asymptotic series solution with an optimal truncation order n, f(n)(z) = e−iλR z zc W (n)(z1)z1/ q 2W (n)(z), cannot fully represent the WKB solution (3.5), but we can choose to expand the exact solution with the super-adiabatic basis formed by f(n) and f∗(n), f(z) = α(z)e −iλR zc zi W (n)(z1)dz1 √ λ f(n)(z) + β(z)e iλR zc zi W (n)(z1)dz1 √ λ f∗(n)(z) = α(z)e −iλR z zi W (n)(z1)dz1 + β(z)e iλR z zi W (n)(z1)dz1 q 2λW (n)(z) q ≡ ( ) ( ) + β( ) ∗( ) (3 11) ≡α(z)g(z) + β(z)g∗(z) , (3.11) ≡α(z)g(z) + β(z)g∗(z) , (3.11) (3.11) where zi in the value of z at initial time and g(z) can be viewed as the instantaneous positive-frequency solution. 3.1 Bosonic case Now the vector ﬁeld can be expanded in an alternative form using g(z): where zi in the value of z at initial time and g(z) can be viewed as the instantaneous positive-frequency solution. Now the vector ﬁeld can be expanded in an alternative form using g(z): A⊥ i = Z d3k (2π)3 eik·x X s=± ϵs i(ˆk) h gs(τ, k)bs k(τ) + gs(τ, k)∗bs† −k(τ) i , (3.12) (3.12) where the new annihilation operator acquires a time dependence through the Bogoliubov transformation bs k(z) ≡αs(z)as k + βs(z)∗as† k . (3.13) (3.13) And the original vacuum annihilated by as k now contains a spectrum of particles, ⟨ns k(τ)⟩ V = ⟨bs† k bs k(τ)⟩ V = \|βs(z)\|2 . (3.14) (3.14) Thus our aim is to solve the time dependence of the Bogoliubov coeﬃcients α(z), β(z). The solution satisﬁes the initial and normalization (Wronskian) conditions as Thus our aim is to solve the time dependence of the Bogoliubov coeﬃcients α(z), β(z). The solution satisﬁes the initial and normalization (Wronskian) conditions as f(zi) →e −iλR z zi W (n)(z1)dz1 q 2λW (n)(z) , ff∗′ −f∗f′ = i . (3.15) (3.15) – 12 – As pointed out in [40, 47], the constant Wronskian implies a degree of freedom of deﬁning the derivative of f: f′(z) = −iλW (n)(z)+V (z) α(z)e −iλR z zi W (n)(z1)dz1 q 2λW (n)(z) + iλW (n)(z)+V (z) β(z)e iλR z zi W (n)(z1)dz1 q 2λW (n)(z) , (3.16) (3.16) where V (z) is an arbitrary real function. Choosing the time-dependent function V (z) decides the evolution of the Bogoliubov coeﬃcients d dz α(z) β(z) ! = δ(z)    1 ∆(z) δ(z) + 1 e 2iλR z zi W (n)(z1)dz1 ∆(z) δ(z) −1 e −2iλR z zi W (n)(z1)dz1 −1    α(z) β(z) ! , (3.17) JHEP06(2021)129 (3.17) where δ(z) = λ 2iW (n)(z) w2(z)−(W (n)(z))2 + 1 λ2 V ′(z)+V 2(z) , ∆(z) = W ′(n)(z) 2W (n)(z) +V (z), (3.18) δ(z) = λ 2iW (n)(z) w2(z)−(W (n)(z))2 + 1 λ2 V ′(z)+V 2(z) , ∆(z) = W ′(n)(z) 2W (n)(z) +V (z), (3 18) (3.18) (3.18) and with the initial condition and with the initial condition and with the initial condition α(zi) = αi , β(zi) = βi . 3.1 Bosonic case (3.22)  (3.22) – 13 – – 13 – The term in the square bracket can be interpreted as the O(λ−2n−1) error when we ap- proximate the EoM (3.4) with the 2n-th order partial sum of the asymptotic series, and this can be calculated with the asymptotic series of the positive-frequency part f+ The term in the square bracket can be interpreted as the O(λ−2n−1) error when we ap- proximate the EoM (3.4) with the 2n-th order partial sum of the asymptotic series, and this can be calculated with the asymptotic series of the positive-frequency part f+ f+(z) = C e−iλR z zc w(z1)dz1 p 2w(z) ∞ X n=0 bn(z) λn . (3.23) (3.23) Substituting this series solution into the EoM (3.4) implies b′ n+1(z) = − i 2w(z)b′′ n(z) + iw′(z) 2w2(z)b′ n(z) + i w′′(z) 4w2(z) −3w′(z)2 8w3(z) ! bn(z) . (3.24) (3.24) JHEP06(2021)129 In all the scenarios that we study in section 4, w2(z) has a simple root at zc, so we approximate w(z) as w(z) = A(z −zc) 1 2 + O(\|z −zc\| 3 2 ) , (3.25) (3.25) and (3.24) is reduced to b′ n+1(z) ≈− i 2A(z −zc) 1 2 b′′ n(z) + iγ 4A(z −zc) 3 2 b′ n(z) − 5i 32A(z −zc) 5 2 bn(z) , (3.26) (3.26) and the solution of this recurrence relation with b0 = 1 is and the solution of this recurrence relation with b0 = 1 is bn λn ≈ 2−2n−13n i A n Γ n + 1 6 Γ n + 5 6 πΓ(n + 1)(z −zc) 3n 2 ≈ Γ n + 1 6 Γ n + 5 6 2πn!F n . (3.27) ≈ Γ n + 1 6 Γ n + 5 6 2πn!F n . 3.1 Bosonic case (3.19) (3.19) Therefore, the appropriate choice of V (z) should minimize the change of the Bogoliubov coeﬃcients as we intends to minimize the diﬀerence between the basis function f(n) and the exact solution, and such a choice is V (z) = −W ′(n)(z) 2W (n)(z), as suggested in [35, 40] with diﬀerent reasons. With this choice, the evolution of the Bogoliubov coeﬃcients satisﬁes d dz α(z) β(z) ! = λ(W (n+1)(z))2−(W (n)(z))2 2iW (n)(z)   1 e 2iλR z zi W (n)(z1)dz1 −e −2iλR z zi W (n)(z1)dz1 −1   α(z) β(z) ! = δ(z)   1 e 2iλR z zi W (n)(z1)dz1 −e −2iλR z zi W (n)(z1)dz1 −1   α(z) β(z) ! , (3.20) and the diagonal term can be removed by deﬁning variables d the diagonal term can be removed by deﬁning variables α(z) = e R z zi δ(z1)dz1S+(z) , β(z) = e −R z zi δ(z1)dz1e −2iλR zc zi W (n)(z1)dz1S−(z) , (3.21) α(z) = e R z zi δ(z1)dz1S+(z) , β(z) = e −R z zi δ(z1)dz1e −2iλR zc zi W (n)(z1)dz1S−(z) , (3.21) (3.21) implying that S± are the Stokes multipliers for the positive and negative modes respectively, so the evolution equation is simpliﬁed as implying that S± are the Stokes multipliers for the positive and negative modes respectively, so the evolution equation is simpliﬁed as dS± dF = ±iδ(z) 2λw exp ±2 Z z zc iλW (n)(z1)dz1 − Z z zi δ(z1)dz1 S∓ = ±e ∓2R z zi δ(z1)dz1 4wλ2 e ±iλR z zc W (n)(z1)dz1 √ W (n)    e ±iλR z zc W (n)(z1)dz1 √ W (n)   ′′ +λ2w2 e ±iλR z zc W (n)(z1)dz1 √ W (n)  S∓. 3.1 Bosonic case (3.27) (3.27) Applying the asymptotic series of f(z) (3.23), The term in the square bracket of (3.22) can be calculated explicitly Applying the asymptotic series of f(z) (3.23), The term in the square bracket of (3.22) can be calculated explicitly    e−iλR z zc W (n)(z1)dz1 √ W (n)   ′′ + λ2w2 e−iλR z zc W (n)(z1)dz1 √ W (n)    e−iλR z zc W (n)(z1)dz1 √ W (n)   −1 =    e−iλR z zc w(z1)dz1 √w 2n X j=0 bj λj   ′′ + λ2w2  e−iλR z zc w(z1)dz1 √w 2n X j=0 bj λj      e−iλR z zc w(z1)dz1 √w   −1 , (3 28)  √w X j=0 λj   √w X j=0 λj  √w  (3.28) implying that (3.28) implying that implying that  e−iλR z zc W (n)(z1)dz1 √ W (n)   ′′ + λ2w2 e−iλR z zc W (n)(z1)dz1 √ W (n) = e−iλR z zc W (n)(z1)dz1 λ2n√ W (n) " b′′ 2n −w′b′ 2n w + 3w′2 4w2 −w′′ 2w ! b2n # = 2iwe−iλR z zc W (n)(z1)dz1 λ2n√ W (n) b′ 2n+1 , (3.29) (3.29) R zc ( ) λ2n√ W (n) b′ 2n+1 , (3.29) – 14 – – 14 – where (3.24) is applied to obtain the last line. Therefore, the evolution of S± (3.22) is reduced to d dF S+ S− ! =    0 h d dF b2n+1 λ2n+1 i∗ w W (n) e2iλR z zc W (n)(z1)dz1 −d dF b2n+1 λ2n+1 w W (n) e−2iλR z zc W (n)(z1)dz1 0    S+ S− ! +O(δ2(z)) =   0 h d dF b2n+1 λ2n+1 i∗e−F −d dF b2n+1 λ2n+1 eF 0   S+ S− ! +O 1 λ2n+2 , (3.30) (3.30) where we keep only the term with O(λ−2n−1). 3.1 Bosonic case Assuming that the optimal truncation is applicable with ReF ≥2, we can thus approximate the integrand as a Gaussian function around the point with stationary phase, and thus the Stokes multiplier reduces to an error function. S(1) −(F) ≈−iRn 2 1 + Erf −ImF √ 2ReF . (3.35) (3.35) We are ready to calculate the ﬁrst-order term of S+ by solving dS(1) + dF = d dF b2n+1 λ2n+1 ∗ w W (n) e2iλR z zc W (n)(z1)dz1S(1) − = e4iλR z∗ zc W (n)(z1)dz1 d dF b2n+1 λ2n+1 w W (n) e−2iλR z zc W (n)(z1)dz1 ∗ S(1) − = e4iλR z∗ zc W (n)(z1)dz1 dS(1) − ∗ dF S(1) −, JHEP06(2021)129 = e4iλR z∗ zc W (n)(z1)dz1 dS(1) − ∗ dF S(1) −, (3.36) (3.36) where z∗is the intersection between the Stokes line and the real z axis, and the phase integral along the Stokes line e4iλR z∗ zc W (n)(z1)dz1 is real. Such a relation between S(1) + and S(1) − implies a much simpler form of the magnitude of S(0) + + S(1) + 2 ≈1+S(1) + +S(1) + ∗+O e8iλR z∗ zc W (n)(z1)dz1 = 1+e4iλR z∗ zc W (n)(z1)dz1 Z F ReF+i∞  dS(1) − ∗ dF S(1) −+ dS(1) − dF S(1) − ∗  dF +O e8iλR z∗ zc W (n)(z1)dz1 = 1+e4iλR z∗ zc W (n)(z1)dz1 S(1) −(F) 2+O e8iλR z∗ zc W (n)(z1)dz1 , (3.37) (3.37) and the deﬁnition of S± (3.21) implies that the normalization of the Bogoliubov coeﬃcients preserves under the ﬁrst-order perturbation. For the situations with ReF < 2, we expect that higher-order perturbations are re- quired to solve for (3.30), and the perturbation theory may break down when e−ReF →1, so we should analyze such situations carefully. In the cases with ReF < 2, we cannot choose an optimal truncation because of the failure of the stationary phase condition (3.34) and the magnitudes of the terms in the asymptotic series solution (3.23) bn increase, starting from the ﬁrst term. 3.1 Bosonic case It is clear that \|dS+/dF\| ≪\|dS−/dF\| as there is an exponential suppression e−F for the former, so we can solve (3.30) perturbatively starting from the initial values Si ± = S(0) ± , determined by αi and βi through (3.21), and the leading-order change S−is an integral along the straight line with constant positive ReF in the complex F plane: JHEP06(2021)129 S(0) −+ S(1) −(F) = Si −−Si + Z F ReF+i∞ d dF b2n+1 λ2n+1 eF dF = Si −+ Si + Rn 2π Z F ReF+i∞ (2n + 1)! F 2n+2 eF dF = Si −+ Si + Rn(2n + 1)! 2π ˜Γ(−1 −2n, −F) , (3.31) (3.31) where the value of prefactor where the value of prefactor Rn = Γ 2n + 7 6 Γ 2n + 11 6 (2n + 1)!Γ(2n + 1) , (3.32) (3.32) and the function ˜Γ(−1 −2n, −F) is the continuous version of the incomplete Gamma function, deﬁned as ˜Γ(−1−2n,−F) =    Γ(−1−2n,−F) ImF ≥0 Γ(−1−2n,−F)+limc→0+[Γ(−1−2n,−ReF −ic)−Γ(−1−2n,−ReF +ic)] ImF < 0. (3.33) (3.33) The prefactor Rn →1 when n ≫0, implying that limF→ReF−i∞S(0) −+S(1) −(F) = Si −−iSi +. The incomplete Gamma function of (3.31) can oscillate dramatically for general n, and we can choose an optimal truncation order n such that the phase is stationary at F = ReF, the moment when the integral receives dominant contribution d dF b2n+1 λ2n+1 −1 d2 dF 2 b2n+1 λ2n+1 F=ReF + 1 = 0 , (3.34) (3.34) and the solution is n = Int ReF 2 −1. It is noteworthy that the optimal truncation order cannot be determined by the stationary-phase condition when ReF < 2, and we will consider such a situation in the later part of this subsection. For simplifying the following perturbative calculation, we set the initial condition as (αi, βi) ≈(Si +, Si −) = (S(0) + , S(0) −) = (1, 0), and the cases with general initial conditions can be obtained easily based on (3.31). – 15 – Assuming that the optimal truncation is applicable with ReF ≥2, we can thus approximate the integrand as a Gaussian function around the point with stationary phase, and thus the Stokes multiplier reduces to an error function. 3.1 Bosonic case Such a divergent series deﬁned by (3.27) behaves like the general- ized hypergeometric function 2F0(a, b; ; F −1) which diverges everywhere from its original deﬁnition, but it can be deﬁned meaningfully by applying the Borel summation: B(F) = ∞ X n=0 bn λn = Z +∞ 0 e−t ∞ X n=0 Γ n + 1 6 Γ n + 5 6 2π(n!)2F n tndt = Z +∞ 0 e−t 2F1 1 6, 5 6; 1; t F dt = e−F/2√ −FK 1 3 −F 2 √π , (3.38) (3.38) – 16 – where Ka(z) is the modiﬁed Bessel function. The function B(F) is discontinuous when it crosses the Stokes line with ImF = 0, implying that the exact solution after crossing the Stokes line should depend on diﬀerent set of linear combination: where Ka(z) is the modiﬁed Bessel function. The function B(F) is discontinuous when it crosses the Stokes line with ImF = 0, implying that the exact solution after crossing the Stokes line should depend on diﬀerent set of linear combination: f(z) =          αi e −iλR z zi w(z1)dz1 √ 2λw(z) B(F(z)) + βi e iλR z zi w(z1)dz1 √ 2λw(z) B∗(F(z)) ImF > 0 C1 e −iλR z zi w(z1)dz1 √ 2λw(z) B(F(z)) + C2 e iλR z zi w(z1)dz1 √ 2λw(z) B∗(F(z)) ImF < 0 , (3.39) (3.39) where the C1 and C2 are constants. After knowing the expression of f(z), the Bogoliubov coeﬃcients can be solved by combining (3.11) and (3.16), and the results can be fully recorded by the singulant F JHEP06(2021)129 α(F) =    αi B(F) + dB(F) dF −βie−F+ReF dB(F) dF ∗ ImF > 0 C1 B(F) + dB(F) dF −C2e−F+ReF dB(F) dF ∗ ImF < 0 , (3.40) (3.40) and β(F) =    −αieF−ReF dB(F) dF + βi B(F) + dB(F) dF ∗ ImF > 0 −C1eF−ReF dB(F) dF + C2 B(F) + dB(F) dF ∗ ImF < 0 , (3.41) (3.41) and thus the constants C1 and C2 are chosen such that the Bogoliubov coeﬃcients and their derivatives are continuous at ImF = 0. To compare with the particle production in dS spacetime, we set the initial condition as (αi, βi) = (1, 0). 3.1 Bosonic case Since B(F) →1 when ImF →±∞, we can know that C1 and C2 are the ﬁnal values of α(F) and β(F) respectively, and numerical checking conﬁrms the normalization condition \|C1\|2 −\|C2\|2 = 1. As shown in ﬁgure 2, \|C2\| agrees with the tendency of e−ReF for the region with ReF ≳0.5, but large deviations appear when ReF →0. Such deviations may be partly attributed to the failure of the approximation of w(z) (3.25) when the two complex roots begin to merge when ReF →0, whereas part of the deviations are expected. For example, the \|β(k)\| in dS spacetime (2.29) cannot be fully described as an exponential factor in some parameter ranges, and thus the behavior of large \|β(k)\| depends on the details of scenarios. The universal property is that it approaches to the exponential form e−ReF when \|β(k)\| decreases, so we use the exponential form to describe the tendency of the production amount but not its exact value for the cases with small ReF. On the other hand, we also compare the Stokes multiplier Snum(F) obtain from the numerical result (3.41) with the approximations utilizing the incomplete gamma func- tion SΓ(F) with n = 0 from the ﬁrst-order perturbation (3.31) and the error function SErf(F) (3.35) respectively, as shown in ﬁgure 3. It is clear that only the imaginary part of β(k) remains non-zero after ﬁnishing the particle production, and diﬀerent approximations have signiﬁcant errors of describing the real part of β(k) which vanishes rapidly after cross- ing the Stokes line, but the error function can still describe the width of the production process. After knowing how the production amount and width depends on the singulant F(z), we can write down a simple form of β(z) which includes the tendencies of particle – 17 – Figure 2. The comparison between \|C2\| and e−ReF with 0 < ReF ≤2, and the vertical axis is in logarithmic scale. JHEP06(2021)129 Figure 2. The comparison between \|C2\| and e−ReF with 0 < ReF ≤2, and the vertical axis is in logarithmic scale. Figure 3. The comparison between Snum(F), SΓ(F) and SErf(F), where all of them are normalized such that the ﬁnal value approaches −i and ReF = 0.5. The solid lines represent the imaginary parts, whereas the dashed lines are the real parts. Figure 3. 3.1 Bosonic case The comparison between Snum(F), SΓ(F) and SErf(F), where all of them are normalized such that the ﬁnal value approaches −i and ReF = 0.5. The solid lines represent the imaginary parts, whereas the dashed lines are the real parts. production with vacuum initial condition when related parameters are changed: β(z) ≈−ie −2iR zc zi W (n)(z1)dz1 2 1 + Erf −ImF(z) √ 2ReF , (3.42) (3.42) where we replace λw →w in (3.4) for simplicity since λ and w(z) always appear together in the ﬁnal results, and the generalization to arbitrary initial conditions is straightforward based on the results (3.31) and (3.41). 3.2 Fermionic case The fermionic case is logically similar to the bosonic case, but with important diﬀerences in the mathematical details. To be more speciﬁc, let us consider the action (2.13) for a Majorana fermion with chemical potential in an FRW background. We will use the Van der Waerden notation for two-component spinors and the conventions follow from [60]. – 18 – We begin by rewriting the mode expansion of ψα in the Van der Waerden notation, ψα(τ, x) = Z d3k (2π)3 X s=± h us(τ, k)eik·xhs α(ˆk)bs k + vs(τ, k)∗e−ik·xϵαβhs† ˙γ (ˆk)¯σ0 ˙γβbs† k i . (3.43) Notice that it is sometimes customary to omit the zeroth Pauli matrix since it is an identity matrix in component form. However, for the sake of balancing the indices, we will keep them explicit here. The eigenvalue equation for the helicity basis can be written in a number of diﬀerent equivalent forms: −ˆkiσ0 α ˙β¯σi ˙βγhs γ(ˆk) = shs α(ˆk) (3.44a) hs† ˙γ (ˆk)¯σ0 ˙γβσi β ˙αˆki = shs† ˙α (ˆk) . (3.44b) (3.44a) JHEP06(2021)129 (3.44b) It is sometimes useful to choose an explicit component form of the helicity basis, h+ α (ˆk) = cos θ 2 eiφ sin θ 2 ! α , h− α (ˆk) = −e−iφ sin θ 2 cos θ 2 ! α , (3.45) (3.45) with ˆk pointing toward the (θ, φ) direction in spherical coordinates. Substituting the mode expansion into the equation of ψα obtained from varying the action (2.13), we obtain the EoM of the mode functions, Substituting the mode expansion into the equation of ψα obtained from varying th action (2.13), we obtain the EoM of the mode functions, i ∂ ∂τ us vs ! = −sk −aκ am am sk + aκ ! us vs ! . (3.46) (3.46) This EoM preserves the combination \|us\|2 + \|vs\|2, with the normalization constant deter- mined by the canonical quantization condition {ψα(τ, x), δS1/2 δ∂τψβ(τ,x′)} = iδβ αδ3(x−x′). After plugging in the mode expansion, this is reduced to a c-number equation δβ α = X s=± \|us\|2hs αhs† ˙β ¯σ0 ˙ββ + \|vs\|2σ0 α ˙βh−s† ˙βh−sβ . (3.47) (3.47) Taking the trace and the determinant of the above equation yields X s=± (\|us\|2 + \|vs\|2) = 2 , and Y s=± (\|us\|2 + \|vs\|2) = 1 , (3.48) (3.48) us ﬁxing the normalization condition \|us\|2 + \|vs\|2 = 1 separately for diﬀerent helicities. 3.2 Fermionic case A ti d i ti 2 2 th E M (3 46) b i t t d d ibi th thus ﬁxing the normalization condition \|us\|2 + \|vs\|2 = 1 separately for diﬀerent helicities. As mentioned in section 2.2, the EoM (3.46) can be interpreted as describing the transition of a two-level system with a Hamiltonian thus ﬁxing the normalization condition \|us\| + \|vs\| = 1 separately for diﬀerent helicities. As mentioned in section 2.2, the EoM (3.46) can be interpreted as describing the transition of a two-level system with a Hamiltonian H(τ) = Z(τ) X(τ) X(τ) −Z(τ) ! = E(τ) C(τ) S(τ) S(τ) −C(τ) ! , (3.49) (3.49) where E ≡ p Z2 + X2 = k2 + 2sκa + (m2 + κ2)a2 (3.50) E ≡ p Z2 + X2 = k2 + 2sκa + (m2 + κ2)a2 (3.50) and C ≡Z E , S ≡X E . The instantaneous eigenstates of H(τ) are given by and C ≡Z E , S ≡X E . The instantaneous eigenstates of H(τ) are given by H C S ! = E C S ! , H S −C ! = −E S −C ! . (3.51) (3.51) – 19 – Therefore, an ansatz of the solution of (3.46) can be constructed as us vs ! = αse−iR Esdτ ˜Cs ˜Ss ! + βseiR Esdτ ˜S∗ s −˜C∗ s ! , (3.52) (3.52) where ˜Cs, ˜Ss are slowly varying functions whose detailed form as a super-adiabatic basis will be computed later. If we choose \| ˜Cs\|2 + \| ˜Ss\|2 = 1, the coeﬃcient functions αs, βs will satisfy the normalization \|αs\|2 + \|βs\|2 = 1, as required by the normalization of us, vs and unitarity. Now we can insert the ansatz back into the mode expansion of ψα, JHEP06(2021)129 JHEP06(2021)129 ψα(τ,x) = Z d3k (2π)3 X s=± " αse−iR Esdτ ˜Cs+βseiR Esdτ ˜S∗ s eik·xhs α(ˆk)bs k + α∗ seiR Esdτ ˜S∗ s −β∗ se−iR Esdτ ˜Cs e−ik·xϵαβhs† ˙γ (ˆk)¯σ0 ˙γβbs† k # . (3.53) ψα(τ,x) = Z d3k (2π)3 X s=± " αse−iR Esdτ ˜Cs+βseiR Esdτ ˜S∗ s eik·xhs α(ˆk)bs k + α∗ seiR Esdτ ˜S∗ s −β∗ se−iR Esdτ ˜Cs e−ik·xϵαβhs† ˙γ (ˆk)¯σ0 ˙γβbs† k # . (3.53) (3.53) The time-dependent creation/annihilation operators are selected according to the instan- taneous negative/positive frequency parts of ψα. 3.2 Fermionic case Therefore, we can regroup the terms according to the dynamical phase e∓iR Esdτ. First, we note the relation ϵαβhs† ˙γ (−ˆk)¯σ0 ˙γβ ≡ηs(ˆk)hs α(ˆk) , (3.54) (3.54) where ηs(ˆk) is a phase factor satisfying ηs(−ˆk) = −ηs(ˆk) . (3.55) (3.55) This can be seen directly from left-multiplying (3.54) by hs† ˙α ¯σ0 ˙αα and solving out ηs(ˆk), or from directly inspecting the component form (3.45). After applying (3.54), we arrive at an alternative expansion, ψα(τ, x) = Z d3k (2π)3 X s=± h ˜Cse−iR Esdτeik·xhs α(ˆk)ds k + ˜S∗ seiR Esdτe−ik·xϵαβhs† ˙γ (ˆk)¯σ0 ˙γβds† k i , (3.56) (3.56) (3.56) where the new time-dependent annihilation operator is obtained as a Bogoliubov transfor- mation [22, 61], ds k(τ) ≡αs(τ, k)bs k −β∗ s(τ, k)ηs(ˆk)bs† k . (3.57) (3.57) The anti-commutation relation is preserved: {ds k, ds′† k′ } = \|αs\|2 + \|βs\|2 {bs k, bs′† k′ } = (2π)3δss′δ3(k −k′) . (3.58) (3.58) Therefore, the vacuum annihilated by the original operator bs k now contains a spectrum of particles with comoving number density ⟨ns k(τ)⟩ V = ⟨ds† k ds k(τ)⟩ V = \|βs(τ, k)\|2 , (3.59) (3.59) – 20 – – 20 – where V is the comoving volume. Here rotational symmetry demands the isotropy of particle production spectrum, as βs only depends on the magnitude of the momentum. The structure of the EoM of fermion (3.46) is similar to the two-state systems in quantum mechanics, and thus we adopt the framework of analyzing the quantum transition histories of such systems [36]. To calculate the particle production, it is convenient to use the bra-ket notation. Similar to the case of boson, we rewrite the equation of motion with the asymptotic parameter λ i d dz \|ψ⟩= λH\|ψ⟩, (3.60) (3.60) JHEP06(2021)129 where we use ψ⟩to denote the two-component mode function, and z = kτ. For the ansatz (3.52) in the bra-ket notation where we use ψ⟩to denote the two-component mode function, and z = kτ. For the ansatz (3.52) in the bra-ket notation \|ψ⟩= αe −iλR z zi E(z1)dz1\|ψα⟩+ βe iλR z zi E(z1)dz1\|ψβ⟩, (3.61) (3.61) the left hand side of (3.60) is i d dz \|ψ⟩= i (α′−iλEα)\|ψα⟩+α\|ψα⟩′e −iλR z zi E(z1)dz1+i (β′+iλEβ)\|ψβ⟩+β\|ψβ⟩′e iλR z zi E(z1)dz1 , (3.62) (3.62) whereas the right hand side is whereas the right hand side is λH\|ψ⟩= λ αe −iλR z zi E(z1)dz1H\|ψα⟩+ βe iλR z zi E(z1)dz1H\|ψβ⟩ . 3.2 Fermionic case (3.63) (3.63) If \|ψα⟩and \|ψβ⟩are the two exact solutions, the positive and negative modes evolve inde- pendently with constant α and β, implying that −iλ(H −E)\|ψα⟩−\|ψα⟩′ = 0 −iλ(H + E)\|ψβ⟩−\|ψβ⟩′ = 0 . (3.64) (3.64) Similar to the case of boson, we approximate the solutions with the asymptotic series Similar to the case of boson, we approximate the solutions with the asymptotic series \|ψα⟩= ∞ X j=0 cj(z)\|ψ(0) α ⟩+ dj(z)\|ψ(0) β ⟩ λj \|ψβ⟩= − ∞ X j=0 d∗ j(z)\|ψ(0) α ⟩−c∗ j(z)\|ψ(0) β ⟩ λj , (3.65) (3.65) where \|ψ(0) α ⟩and \|ψ(0) β ⟩are the instantaneous eigenstates found in (3.51), and they satisfy where \|ψ(0) α ⟩and \|ψ(0) β ⟩are the instantaneous eigenstates found in (3.51), and they satisfy \|ψ(0) α ⟩′ = −θ′ 2 \|ψ(0) β ⟩ \|ψ(0) β ⟩′ = θ′ 2 \|ψ(0) α ⟩, (3.66) (3.66) – 21 – – 21 – where θ(z) = arctan X(z) Z(z) , deﬁned in (3.49). Substitute (3.65) and (3.66) to (3.64), we have c′ j + dj 2 θ′ = 0 2iE(z)dj+1 −d′ j + cj 2 θ′ = 0 . (3.67) (3.67) Similar to the case of boson, we solve for cj(z) and dj(z) near the complex root zc of E2(z) = X2(z) + Z2(z) = 0. In all the scenarios that we study in section 4, zc is a ﬁrst-order root, implying that Similar to the case of boson, we solve for cj(z) and dj(z) near the complex root zc of E2(z) = X2(z) + Z2(z) = 0. In all the scenarios that we study in section 4, zc is a ﬁrst-order root, implying that JHEP06(2021)129 JHEP06(2021)129 X(z) ≈Xc + X′ c(z −zc) + O(\|z −zc\|2) Z(z) ≈iXc + Z′ c(z −zc) + O(\|z −zc\|2) , (3.68) (3.68) with X2 c + Z2 c = 0, and with X2 c + Z2 c = 0, and θ′(z) = Z2(z) X2(z) + Z2(z) X(z) Z(z) ′ i θ′(z) = Z2(z) X2(z) + Z2(z) X(z) Z(z) ′ ≈ i 2(z −zc) . (3.69) θ (z) X2(z) + Z2(z) Z(z) ≈ i 2(z −zc) . (3.69) (3.69) ≈ i 2(z −zc) . 3.2 Fermionic case (3.69) With this approximation of θ′(z), we obtain the recurrence relation of cj from (3.67) c′ j+1(z) = − i 2A(z −zc) 1 2 c′′ j (z) − i 2A(z −zc) 3 2 c′ j(z) + i 32A(z −zc) 5 2 cj(z) , (3.70) (3.70) where A is deﬁned by E ≈A(z −zc)1/2, and the solutions of cj(z) and dj(z) with c0 = 1 and d0 = 0 are cj(z) λj = − 3j−24−j−1 i A j 5 6 j−1 7 6 j−1 Γ(j + 1)(z −zc)3j/2 = − 5 6 j−1 7 6 j−1 36Γ(j + 1)F j , (3.71) (3.71) and dj(z) λj = iΓ j −1 6 Γ j + 1 6 2πΓ(j)F j , (3.72) and dj(z) λj = iΓ j −1 6 Γ j + 1 6 2πΓ(j)F j , (3.72) (3.72) respectively, where (a)k = Γ(a + k)/Γ(a) is the Pochhammer symbol, and the singulant is F(z) = −2i Z z zc E(z1)dz1 . (3.73) (3.73) Clearly cj(z) and dj(z) are divergent asymptotic series, and thus we truncate them at the order of n and denote the partial sums of (3.65) as \|ψ(n) α ⟩and \|ψ(n) β ⟩. Equating (3.62) – 22 – and (3.63) with the truncated asymptotic series can derive the coupled diﬀerential equations of α and β and (3.63) with the truncated asymptotic series can derive the coupled diﬀerential equations of α and β d dz α(z) β(z) ! ( ) ! =    iλE(z)−⟨ψ(n) α \|ψ(n) α ⟩′−iλ⟨ψ(n) α \|H\|ψ(n) α ⟩ − ⟨ψ(n) α \|ψ(n) β ⟩′+iλ⟨ψ(n) α \|H\|ψ(n) β ⟩ e 2iλR z zi E(z1)dz1 − ⟨ψ(n) β \|ψ(n) α ⟩′+iλ⟨ψ(n) β \|H\|ψ(n) α ⟩ e −2iλR z zi E(z1)dz1 −iλE(z)−⟨ψ(n) β \|ψ(n) β ⟩′−iλ⟨ψ(n) β \|H\|ψ(n) β ⟩    α(z) β(z) ! . 3.2 Fermionic case Similar to the bosonic case (3.30), (3.79) can be solved perturbatively starting with the initial condition (Si +, Si −) = (S(0) + , S(0) −): which converges to 1 for n →∞. Similar to the bosonic case (3.30), (3.79) can be solved perturbatively starting with the initial condition (Si +, Si −) = (S(0) + , S(0) −): S(0) −+ S(1) −(F) = Si −+ iSi +Tn Z F ReF+i∞ n! 2πF n+1 eF dF = Si −+ Si +Tn i(−1)n+1n! 2π ˜Γ(−n, −F) . (3.81) (3.81) By setting the vacuum initial condition (Si +, Si −) = (S(0) + , S(0) −) = (1, 0), the ﬁrst-order perturbation of S+ implies that JHEP06(2021)129 S(0) + + S(1) + 2 = 1 −e−2ReF Z F ReF+i∞ dS(1) − ∗ dF S(1) −+ dS(1) − dF S(1) − ∗dF + O e−4ReF = 1 −e−2ReF S(1) −(F) 2 + O e−4ReF , (3.82) (3.82) which agrees with the normalization of the Bogoliubov coeﬃcients. By choosing the optimal truncation order as n = ReF −1, the integrand is stationary at F = ReF, and thus S(1) − can be approximated as S(1) −(z) ≈Tn 2 1 + Erf −ImF(z) √ 2ReF . (3.83) (3.83) Similar to the bosonic case, the situations with ReF < 1 implies the failure of choosing an optimal truncation for the asymptotic series (3.65), and we may apply the Borel sum to evaluate such divergent series: Similar to the bosonic case, the situations with ReF < 1 implies the failure of choosing an optimal truncation for the asymptotic series (3.65), and we may apply the Borel sum to evaluate such divergent series: I(F) = ∞ X j=0 cj λj = U −1 6, 2 3, −F 6√ −F , J(F) = ∞ X j=0 dj λj = 6iF dI(F) dF , (3.84) = 6iF dI(F) dF , (3.84) (3.84) where U(a, b, z) is the conﬂuent hypergeometric function, and the last line is obtained from the relation between cj and dj (3.67). 3.2 Fermionic case (3.74) (3.74) The values of matrix elements can be obtained by evaluating (3.64) with the truncated series The values of matrix elements can be obtained by evaluating (3.64) with the truncated series JHEP06(2021)129 −iλ(H −E)\|ψ(n) α ⟩−\|ψ(n) α ⟩′ = 1 λn cn(z) 2 θ′(z) −d′ n(z) \|ψ(0) β ⟩ −iλ(H −E)\|ψ(n) α ⟩−\|ψ(n) α ⟩′ = 1 λn cn(z) 2 θ′(z) −d′ n(z) \|ψ(0) β ⟩ −iλ(H + E)\|ψ(n) β ⟩−\|ψ(n) β ⟩′ = 1 λ cn(z)θ′(z) −d′ n(z) ∗ \|ψ(0) α ⟩, (3.75 −iλ(H −E)\|ψ(n) α ⟩−\|ψ(n) α ⟩′ = 1 λn cn(z) 2 θ′(z) −d′ n(z) \|ψ(0) β ⟩ −iλ(H + E)\|ψ(n) β ⟩−\|ψ(n) β ⟩′ = 1 λn cn(z) 2 θ′(z) −d′ n(z) ∗ \|ψ(0) α ⟩, (3.75) −iλ(H + E)\|ψ(n) β ⟩−\|ψ(n) β ⟩′ = 1 λn cn(z) 2 θ′(z) −d′ n(z) ∗ \|ψ(0) α ⟩, (3.75) (3.75) and thus d dz α(z) β(z) ! =   δ2(z) −∆∗ 2(z)e 2iλR z zi E(z1)dz1 ∆2(z)e −2iλR z zi E(z1)dz1 δ∗ 2(z)   α(z) β(z) ! , (3.76) (3.76) where δ2(z) = 1 λn cn(z) 2 θ′(z) −d′ n(z) n X j=1 d∗ j(z) λj ∆2(z) = 1 λn cn(z) 2 θ′(z) −d′ n(z) n X j=0 cj(z) λj (3.77) (3.77) which has the same structure as vector boson (3.20). We remove the diagonal term by deﬁning the Stokes multipliers α(z) = e R z zi δ2(z1)dz1S+(z) , β(z) = e R z zi δ∗ 2(z1)dz1e −2iλR zc zi E(z1)dz1S−(z) , (3.78) (3.78) and d dF S+ S− ! =   0 ∆∗ 2 2iλE e −F+R z zi δ∗(z1)−δ(z1)dz1 −∆2 2iλE e F+R z zi δ(z1)−δ∗(z1)dz1 0   S+ S− ! =   0 −iTn n! 2πF n+1 ∗e−F iTn n! 2πF n+1 eF   S+ S− ! + O 1 λn+1 , (3.79) (3.79) where we only keep the O(λ−n) term, and the prefactor where we only keep the O(λ−n) term, and the prefactor Tn = Γ n + 5 6 Γ n + 7 6 Γ(n + 1)2 , (3.80) (3.80) – 23 – – 23 – which converges to 1 for n →∞. 3.2 Fermionic case Since I(F) and J(F) are discontinuous at ImF = 0, we rewrite the exact solution (3.61) into two parts: \|ψ⟩=                    αie −iλR z zi E(z1)dz1 I(F)\|ψ(0) α ⟩+ J(F)\|ψ(0) β ⟩ −βie iλR z zi E(z1)dz1 J∗(F)\|ψ(0) α ⟩−I∗(F)\|ψ(0) β ⟩ , ImF > 0 C1e −iλR z zi E(z1)dz1 I(F)\|ψ(0) α ⟩+ J(F)\|ψ(0) β ⟩ + C2e iλR z zi E(z1)dz1 J∗(F)\|ψ(0) α ⟩−I∗(F)\|ψ(0) β ⟩ , ImF < 0 , (3.85) (3.85) – 24 – – 24 – Figure 4. The comparison between \|C2\| and e−ReF with 0 < ReF ≤1, and the vertical axis is in logarithmic scale. JHEP06(2021)129 Figure 4. The comparison between \|C2\| and e−ReF with 0 < ReF ≤1, and the vertical axis is in logarithmic scale. where αi and βi are the initial Bogoliubov coeﬃcients, C1 and C2 are constants to let \|ψ⟩ and its derivative continuous at ImF = 0. The time-dependent Bogoliubov coeﬃcients deﬁned with respect to the eigenstates \|ψ(0) α ⟩and \|ψ(0) β ⟩are thus: where αi and βi are the initial Bogoliubov coeﬃcients, C1 and C2 are constants to let \|ψ⟩ and its derivative continuous at ImF = 0. The time-dependent Bogoliubov coeﬃcients deﬁned with respect to the eigenstates \|ψ(0) α ⟩and \|ψ(0) β ⟩are thus: α(F) =    αiI(F) −βie−F+ReF J∗(F) , ImF > 0 C1I(F) + C2e−F+ReF J∗(F) , ImF < 0 β(F) =    αieF−ReF J(F) + βiI∗(F) , ImF > 0 C1eF−ReF J(F) −C2I∗(F) , ImF < 0 . (3.86) (3.86) With the vacuum initial condition (αi, βi) = (1, 0), the two unknown constants are deter- mined by matching the two parts. Since I(F) →1 and J(F) →0 for Im F →±∞, it is clear that C1 and −C2 are the ﬁnal values of α(k) and β(k) respectively, and numerical checking conﬁrms the normalization \|C1\|2 + \|C2\|2 = 1. As shown in ﬁgure 4, the amount of particle production ﬁts the tendency of e−ReF for ReF ≳0.2, but large deviations ap- pear when ReF →0, similar to the case of boson. 3.2 Fermionic case We also compare the Stokes multiplier Snum(F) obtain from the numerical result (3.86) with the approximations utilizing the in- complete gamma function SΓ(F) with n = 0 from the ﬁrst-order perturbation (3.81) and the error function SErf(F) (3.83) respectively, as shown in ﬁgure 5. With the vacuum initial condition (αi, βi) = (1, 0), the two unknown constants are deter- mined by matching the two parts. Since I(F) →1 and J(F) →0 for Im F →±∞, it is clear that C1 and −C2 are the ﬁnal values of α(k) and β(k) respectively, and numerical checking conﬁrms the normalization \|C1\|2 + \|C2\|2 = 1. As shown in ﬁgure 4, the amount of particle production ﬁts the tendency of e−ReF for ReF ≳0.2, but large deviations ap- pear when ReF →0, similar to the case of boson. We also compare the Stokes multiplier Snum(F) obtain from the numerical result (3.86) with the approximations utilizing the in- complete gamma function SΓ(F) with n = 0 from the ﬁrst-order perturbation (3.81) and the error function SErf(F) (3.83) respectively, as shown in ﬁgure 5. Similar to the bosonic case, the asymptotic parameter λ always appear with the in- stantaneous eigenvalue E(z), so we may set λH →H in (3.60). We summarize a simple form of β(z) which reﬂects the tendency of the particle production starting from vacuum initial condition β(z) ≈e−F(zi) 2 1 + Erf −ImF(z) √ 2ReF , (3.87) (3.87) where it is noteworthy that there is no additional prefactor −i compared to the bosonic case (3.42), and the generalization to arbitrary initial conditions can be easily done base on (3.81) and (3.86). – 25 – – 25 – Figure 5. The comparison between Snum(F), SΓ(F) and SErf(F), where all of them are normalized such that the ﬁnal value approaches 1 and ReF = 0.5. The solid lines represent the real parts, whereas the dashed lines are the imaginary parts. JHEP06(2021)129 Figure 5. The comparison between Snum(F), SΓ(F) and SErf(F), where all of them are normalized such that the ﬁnal value approaches 1 and ReF = 0.5. The solid lines represent the real parts, whereas the dashed lines are the imaginary parts. 3.2 Fermionic case Note that the production histories solely depend on the singulant F, and the deﬁnition of the singulant F in the fermion case is similar to the boson case, with E(z) = s 1 + 2sκa k + (m2 + κ2)a2 k2 (3.88) (3.88) in replacement of in replacement of W(z) ≈w(z) = s 1 + 2sκa k + m2a2 k2 . (3.89) (3.89) Therefore, combining this observation with (3.42) and (3.87), we arrive at a simple replace- ment rule for the production histories (i.e., \|β(z)\|2) of vector bosons and fermions: m2 ↔m2 + κ2 m2 ↔m2 + κ2 (bosons) (fermions) . (3.90) (bosons) (fermions) . (3.90) (3.90) 4 Analysis of particle production in various spacetimes Armed with these powerful mathematical tools, we are now in a position to compute the ﬁne-grained particle production histories of both massive vector bosons and Majorana fermions in various setups. In this section, we have in mind that the chemical potential is provided as a external source by, for instance, a rolling scalar ﬁeld. The backreaction to the external ﬁeld that generates the chemical potential is also assumed to be negligible. In particular, we will assume the chemical potential κ is a constant in spacetime. The reason for such a choice is three-fold. First of all, this is indeed true in some cases. For example, the Hubble friction during inﬂation drives a rolling scalar to an attractor phase with constant speed ˙φ, which corresponds to a constant κ when coupled to vectors or fermions. In a radiation/matter- dominated universe, speciﬁcally chosen scalar potentials also give rise to constant rolling – 26 – speeds. Second, physically speaking, for any slowly-varying κ(τ) ̸= 0, a constant chemical potential is always a leading order approximation. As long as the typical time scale of κ(τ) is longer than the particle production time scale, this approximation will be valid. Third, mathematically speaking, a constant κ leads to simple and analytical results that already contain lots of information in the general cases, which can always be dealt with using numerical methods. We will focus on ﬁve familiar types of FRW spacetimes whose singulant integrals are exactly computable. The resulting production amount, time and width are given explicitly as analytical expressions. Some of these results are exact while others are approximate or empirical with percent-level error in most parameter regimes. To distinguish them from each other, we will use = when the result is exact. We use ≃for results which are easily computable to any desired precision but which are shown with ﬁnite accuracy. And ≈will be used for empirical results whose relative error is at percent-level. JHEP06(2021)129 Due to the replacement rule mentioned in section 3, we will only work out the spin- 1 case with \|κ\| < m, and obtain the spin-1/2 results for all parameter regions by simple substitutions. Also because diﬀerent helicities are related by a sign ﬂip of κ, we will focus on the negative helicity state, whose production is enhanced if κ is positive. 4This choice has interesting implications for dS. See section 4.1 and appendix B for more details. 4 Analysis of particle production in various spacetimes Throughout this section, we will be working in comoving coordinates and using conformal time rather than cosmic time.4 Tilde variables will be used to deﬁne dimensionless parameters measured in units of a certain Hubble scale, e.g., ˜m ≡m H , ˜κ ≡ κ H , etc. And we will typically expand quantities in powers of ˜κ ˜m = κ m. 4.1 dS In an exact dS spacetime, the scale factor has a time dependence a(τ) = −1 Hτ = eHt. The EoM for spin-1 particles with chemical potential can be written in terms of a dimensionless variable z = −kτ = k aH , In an exact dS spacetime, the scale factor has a time dependence a(τ) = −1 Hτ = eHt. The EoM for spin-1 particles with chemical potential can be written in terms of a dimensionless variable z = −kτ = k aH , d2f(z) dz2 + w2(z)f(z) = 0 , w2(z) = 1 −2˜κ z + ˜m2 z2 . (4.1) (4.1) Notice that the variable z now runs from the right to the left, and the physical region is the positive real axis z > 0. After analytical continuation, one can deﬁne w(z) on the whole complex plane. The two roots of w2(z) = 0 lie at zc and z∗ c, with Notice that the variable z now runs from the right to the left, and the physical region is the positive real axis z > 0. After analytical continuation, one can deﬁne w(z) on the whole complex plane. The two roots of w2(z) = 0 lie at zc and z∗ c, with zc = ˜κ + i p ˜m2 −˜κ2 . (4.2) (4.2) Notice that zc lies on the upper half complex plane since the original τc is on the lower half complex plane and they diﬀer by a sign. The absence of tachyonic instability requires ˜m > \|˜κ\|. Therefore, the two complex turning points lie in a symmetric fashion across the real axis. Starting from zc and z∗ c are two branch cuts that meet their ends at the pole at z = 0. In the z-domain the singulant is evaluated as In the z-domain, the singulant is evaluated as F(z) = 2i Z z zc w(z)dz . (4.3) (4.3) 4This choice has interesting implications for dS. See section 4.1 and appendix B for more details. – 27 – Figure 6. The super-adiabatic singulant F(z) in dS for ˜κ = 5 (upper left panel) and ˜κ = 9.9 (upper right panel), with the dS-corrected mass µ = 10. The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. 4.1 dS Plug in the ansatz ∞b ˜n z∗( ˜m, ˜κ) = ∞ X n=0 bn˜κn ˜mn−1 (4.7) (4.7) JHEP06(2021)129 into the equation (4.6) and collect the terms order-by-order in ˜κ ˜m, we are able to solve the coeﬃcients iteratively, 1 = 1 q 1 + b2 0 tanh−1   1 q 1 + b2 0   (4.8a) b1 = b0 q 1 + b2 0 tanh−1   b0 q 1 + b2 0   (4.8b) b2 = b1 + 2b1b2 0 −b2 0 2b0 + 2b3 0 (4.8c) b3 = b1 + b2 0 −2b1 + 6b0b2 + b2 0 (2 −3b1) b1 + 6b3 0b2 + b4 0 6b2 0 1 + b2 0 2 (4.8d) 1 = 1 q 1 + b2 0 tanh−1   1 q 1 + b2 0   (4.8a) b1 = b0 q 1 + b2 0 tanh−1   b0 q 1 + b2 0   (4.8b) b2 = b1 + 2b1b2 0 −b2 0 3 (4.8c) (4.8a) (4.8b) b2 = b1 + 2b1b2 0 −b2 0 2b0 + 2b3 0 (4.8c) (4.8c) 0 b3 = b1 + b2 0 −2b1 + 6b0b2 + b2 0 (2 −3b1) b1 + 6b3 0b2 + b4 0 6b2 0 1 + b2 0 2 (4.8d) (4.8d) · · · . In this way, we obtain the particle production time as In this way, we obtain the particle production time as z∗( ˜m, ˜κ) ≃0.6627 ˜m + 0.3435˜κ −0.0102˜κ2 ˜m + 0.0064 ˜κ3 ˜m2 + · · · . (4.9) (4.9) Clearly, with a larger eﬀective mass and a larger positive chemical potential, the pro- duction time becomes earlier. Clearly, with a larger eﬀective mass and a larger positive chemical potential, the pro- duction time becomes earlier. • Production width. The derivative of the singulant is none other than the frequency itself: s 2 • Production width. The derivative of the singulant is none other than the frequency itself: s 2 • Production width. The derivative of the singulant is none other than the frequenc itself: s 2 Im F ′(z∗) = 2Re w(z∗) = 2 s 1 −2˜κ z∗ + ˜m2 z2∗ . 4.1 dS The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. JHEP06(2021)129 Figure 6. The super-adiabatic singulant F(z) in dS for ˜κ = 5 (upper left panel) and ˜κ = 9.9 (upper right panel), with the dS-corrected mass µ = 10. The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. Notice the sign change compared to (3.10). The phase integral is exactly solvable: F(z) = 2i "p ˜m2−2z˜κ+z2+ ˜mtanh−1 z˜κ−˜m2 ˜m √ ˜m2−2z˜κ+z2 ! +˜κtanh−1 ˜κ−z p ˜m2+z (z−2˜κ) !# +π( ˜m−˜κ), (4.4) (4.4) where the ﬁrst line is purely imaginary for z lying on the positive real axis. Its behavior on the right-half z-plane as well as the Stokes lines are shown in ﬁgure 6. Production amount. The production amount is straightforwardly given by the formula \|β\|2 = e−2Re F(zi) = e−2π( ˜m−˜κ) , (4.5) (4.5) with zi ≫˜m taken on the real axis. with zi ≫˜m taken on the real axis. – 28 – • Production time. The crossing time z∗lies on the real axis, and therefore satisﬁes a real-numbered equation: • Production time. The crossing time z∗lies on the real axis, and therefore satisﬁes a real-numbered equation: • Production time. The crossing time z∗lies on the real axis, and therefore satisﬁes a real-numbered equation: 0 = q ˜m2−2z∗˜κ+z2∗+ ˜mtanh−1 z∗˜κ−˜m2 ˜m p ˜m2−2z∗˜κ+z2∗ ! +˜κtanh−1 ˜κ−z∗ p ˜m2+z∗(z∗−2˜κ) ! . (4.6) ( ) This equation can be solved using a perturbative expansion in powers of ˜κ. 4.1 dS (4.10) (4.10) Thus the production width in the z-domain is Thus the production width in the z-domain is ∆z∗= 2 p 2\|Re F(z∗)\| \|Im F ′(z∗)\| = v u u t 2π ( ˜m −˜κ) 1 − 2˜κ z∗( ˜m,˜κ) + ˜m2 z2∗( ˜m,˜κ) . (4.11) (4.11) We can also translate the production width into the t-domain by ∆t∗= ∆z∗ Hz∗. Another useful measure of production width is the e-folding numbers during which the production is complete. In terms of a power series in ˜κ/ ˜m, we have ∆N∗= H∆t∗= ∆z∗ z∗ ≃2.0895 √ ˜m −0.4131˜κ ˜m3/2 + 0.1323˜κ2 ˜m5/2 −0.0523˜κ3 ˜m7/2 +O ˜κ4 ˜m9/2 ! . (4.12) (4.12) – 29 – – 29 – Now the alert readers may ﬁnd an inconsistency here. If one compares the produc- tion amount (4.5) computed from the Stokes-line method with that of the exact result, namely (2.29) or (2.30), one ﬁnds that there is a mismatch of mass: the true result should contain the dS-corrected mass µ = q ˜m2 −1 4 instead of ˜m as given by the Stokes-line method. This mismatch implies that the results obtained by the naive application of Stokes- line method are subjected to a relative error O( ˜m−2), which can be important if ˜m is small. ( ) To trace the origin of this 1/4 puzzle, let us go back to the super-adiabatic basis for the EoM (4.1), f(z) = 1 p 2W(z)e iR z zi W(z′)dz′ −iS(z)e−F(zi) p 2W(z) e −iR z zi W(z′)dz′ , (4.13) (4.13) JHEP06(2021)129 where S(z) is the Stokes multiplier. W(z) is solved order-by-order as speciﬁed in section 3.1. The leading order reads W (0)(z) = w(z) and f(0)(z) = 1 p 2w(z)e iR z zi w(z′)dz′ −iS(z)e−F(zi) p 2w(z) e −iR z zi w(z′)dz′ . (4.14) (4.14) This leading order solution is sometimes called WKB approximation. The late-time be- havior of the frequency function is W (0)(z) = w(z) z→0 −−−→˜m z + O(z0) . (4.15) (4.15) Upon integration over z, the phase of the positive-frequency mode becomes linearly in- creasing with cosmic time t: Z z zi w(z′)dz′ = Z z zi ˜m z′ + O(z′0) dz′ = ˜m ln z + const = −mt + const , (4.16) (4.16) where we have used z = −kτ = −k H e−Ht. Hence the late time behavior of the WKB basis is f(0) ∼e∓imt. 4.1 dS This, however, corresponds to a wrong oscillation frequency for spin-1 vector particles. The correct frequency can be easily obtained by inspecting the late-time behavior of the EoM itself. Namely we can plug in the ansatz f ∼z∆−1 and expand (4.1) to leading order in z, ((∆−1)(∆−2) + ˜m)z∆−3 + O(z∆−2) = 0 . (4.17) (4.17) This gives ∆± = 3 2 ± iµ and the correct IR behavior f ∼z(1/2±iµ)t ∝e∓iµHt ̸= e∓imt. As a result, the leading order super-adiabatic solution does not capture the correct IR oscillation frequency, which is dictated by dS symmetries. Particles in dS are classiﬁed according to the unitary irreducible representation of the dS group [62, 63] and a massive spin-S particle in the principal series has a conformal weight [10] This gives ∆± = 3 2 ± iµ and the correct IR behavior f ∼z(1/2±iµ)t ∝e∓iµHt ̸= e∓imt. As a result, the leading order super-adiabatic solution does not capture the correct IR oscillation frequency, which is dictated by dS symmetries. Particles in dS are classiﬁed according to the unitary irreducible representation of the dS group [62, 63] and a massive spin-S particle in the principal series has a conformal weight [10] ∆(S) ± = 3 2 ± iµS , µS = s m2 H2 − S −1 2 2 , S ⩾1 . (4.18) (4.18) In other words, the geometry of dS modiﬁes the eﬀective mass of spinning particles in the IR (small z), and this fact is not taken into account by the naive WKB approximation (4.14). – 30 – Fortunately, the advantage of the smoothed Stokes-line method is that the higher-order terms in the super-adiabatic basis can, and actually do, give essential corrections to the leading order solution. The ﬁrst order correction to W is δW (1)(z) = 3w′(z)2 −2w(z)w′′(z) 8w(z)3 = 6z ˜m2 (˜κ −z) −˜m4 + z2˜κ (4z −3˜κ) 8 (z2 −2˜κz + ˜m2)3 s 1 −2˜κ z + ˜m2 z2 . (4.19) This function has two third order poles at z and z∗The branch cuts brought by w(z) are δW (1)(z) = 3w′(z)2 −2w(z)w′′(z) 8w(z)3 = 6z ˜m2 (˜κ −z) −˜m4 + z2˜κ (4z −3˜κ) 8 (z2 −2˜κz + ˜m2)3 s 1 −2˜κ z + ˜m2 z2 . (4 19) ( ) This function has two third-order poles at zc and z∗ c. • Production amount 4.1 dS The branch cuts brought by w(z) are still present and they connect the third-order poles to a simple pole at z = 0, JHEP06(2021)129 δW (1)(z) = − 1 8 ˜mz + O(z0) , Re z > 0 . (4.20) (4.20) The eﬀect of this pole at the origin is exactly to give an O( ˜m−2) correction to the IR oscillation frequency: W (1)(z) = W (0)(z) + δW (1)(z) = ˜m −1 8 ˜m 1 z + O(z0) . (4.21) (4.21) Including the higher order corrections in the super-adiabatic series, we recover the correct IR oscillation frequency, W(z) = W (0)(z) + δW (1)(z) + δW (2)(z) + · · · = ˜m −1 8 ˜m − 1 128 ˜m3 + · · · 1 z + O(z0) = r ˜m2 −1 4 × 1 z + O(z0) . (4.22) (4.22) Therefore, the 1/4 puzzle can be resolved by taking into account the full super-adiabatic basis and resumming the higher-order corrections to the mass. We can recover the correct production amount by redeﬁning the singulant as an integral of W instead of W (0) = w. Technically, since this integral is ill-deﬁned around the complex turning points where δW (n)(z) diverges, we need to manually impose a principal value prescription. We deform the integration contour to lie along the branch cut and tour along a semi-circle around the pole at origin (see ﬁgure 7 for illustration). Then by some arguments of complex analysis, the only non-zero contribution comes from the semi-circle at the origin, where the IR frequency correction is at work. Thus the corrected production amount can be written as half of the residue of W(z) at z = 0+, subtracting half of the residue at z = ∞, ln \|β2\| = −2i Z z∗ c zc W(z)dz = −2π (Res z→0+W(z) −Res z→∞W(z)) = −2π(µ −κ) . (4.2 For a detailed mathematical discussion of these arguments, we refer the readers to appendix B. In summary, the super-adiabatic corrections completely ﬁx the mass mismatch and we only need to replace ˜m →µ in (4.5), (4.9) and (4.12) to recover the true results. We list them below for the sake of clarity. For a detailed mathematical discussion of these arguments, we refer the readers to appendix B. 4.1 dS In both panels, the chemical potential ranges over ˜κ = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin-1/2 fermion mass µ = ˜m = 10. The gray region with Re z < 0 cannot be reached physically in dS. JHEP06(2021)129 Figure 8. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel). In both panels, the chemical potential ranges over ˜κ = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin-1/2 fermion mass µ = ˜m = 10. The gray region with Re z < 0 cannot be reached physically in dS. Figure 9. The production histories in exact dS. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: production width mea- sured in e-folding numbers. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. Figure 9. The production histories in exact dS. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: production width mea- sured in e-folding numbers. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. production histories when the chemical potential is small. 4.1 dS In summary, the super-adiabatic corrections completely ﬁx the mass mismatch and we only need to replace ˜m →µ in (4.5), (4.9) and (4.12) to recover the true results. We list them below for the sake of clarity. \|β\|2 = e−2π(µ−˜κ) , (4.24) (4.24) – 31 – – 31 – Figure 7. The integration contour can be deformed from along the Stokes line (C0) to along the branch cut (C1 ∪C2), collecting half of the residue at z = 0+. The black dashed line indicates the branch cut. JHEP06(2021)129 Figure 7. The integration contour can be deformed from along the Stokes line (C0) to along the branch cut (C1 ∪C2), collecting half of the residue at z = 0+. The black dashed line indicates the branch cut. Figure 7. The integration contour can be deformed from along the Stokes line (C0) to along the branch cut (C1 ∪C2), collecting half of the residue at z = 0+. The black dashed line indicates the branch cut. • Production time z∗(µ, ˜κ) ≃0.6627µ + 0.3435˜κ −0.0102˜κ2 µ + 0.0064 ˜κ3 µ2 + · · · . (4.25) (4.25) • Production width • Production width ∆N∗= H∆t∗= ∆z∗ z∗ ≃2.0895 √µ −0.4131˜κ µ3/2 + 0.1323˜κ2 µ5/2 −0.0523˜κ3 µ7/2 +O ˜κ4 µ9/2 ! . (4.26) (4.26) For spin-1/2 fermions, however, there is no such a problem, since the leading order WKB result already capture the correct IR behavior. The instantaneous eigenvalue of the Hamiltonian for its EoM (2.20) written in z-domain is E±(z) = s 1 ± 2˜κ z + ˜m2 + ˜κ2 z2 = √ ˜m2 + ˜κ2 z + O(z0) . (4.27) (4.27) Thus the mode functions behave as u, v ∼e±iR E±(z)dz ∼e∓i √ ˜m2+˜κ2t. This oscillation frequency indeed agrees with the late-time behavior of (2.20). Therefore, higher orders in the super-adiabatic basis do not oﬀer any O( ˜m−n) corrections to the oscillation phase, hence to the production history. The production amount, time and width can simply be obtained by replacing µ → √ ˜m2 + ˜κ2 in (4.24–4.26). Its Stokes lines are shown together with spin-1 particles in ﬁgure 8. Finally, we plot the parameter dependence of production histories for both bosons and fermions in ﬁgure 9. From the plots, one can see that bosons and fermions share similar – 32 – Figure 8. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel). 4.2 Deviation from dS of the ϵ-type The exact and rigid dS spacetime is maximally symmetric with a simple time dependence in the scale factor. The mode functions for free ﬁelds are exactly solvable, even though knowing the production time and width still requires the technique of smoothed Stokes phenomenon. However, dS only works as a leading order approximation to certain stages of cosmic evolution such as inﬂation or dark-energy dominated era. The actual evolution can deviate from that of dS in diﬀerent ways, with distinctive impacts on the particle production history that we are after. In this subsection and the next, we will focus on two simplest ways to deform the dS geometry, namely with a constant ϵ parameter and with a constant η parameter. We will call these deviations the ϵ-type and the η-type. JHEP06(2021)129 The simplest kind of deformation is to introduce a (small) constant ϵ parameter, kind of deformation is to introduce a (small) constant ϵ parameter, ϵ(τ) = − H′(τ) a(τ)H2(τ) = ϵ = const , H(τ) = a′(τ) a2(τ) . (4.28) (4.28) Integrating over τ yields a scale factor a(τ) = (−Hpτ)− 1 1−ϵ , (4.29) (4.29) where Hp ≡H(τp) is the Hubble parameter evaluated at the time τp when the scale factor is a(τp) = (1 −ϵ)−1/ϵ. Because the Hubble parameter is decreasing with conformal time as a power-law, modes that exit the horizon experience a slightly diﬀerent gravitational background. This soft breaking of scale invariance, as we will see, manifests itself in the scale dependence of production history. Deﬁning the dimensionless variable z = −kτ, the EoM of a massive spin-1 particle reads 2 Deﬁning the dimensionless variable z = −kτ, the EoM of a massive spin-1 particle d d2f(z) dz2 + w2(z, k)f(z) = 0 , w2(z, k) = 1 −2˜κp(k) z 1 1−ϵ + ˜m2 p(k) z 2 1−ϵ , (4.30) (4.30) here we have denoted the scale-dependent dimensionless mass and chemical potential as ˜κp(k) = κ Hp k Hp ! ϵ 1−ϵ , ˜mp(k) = m Hp k Hp ! ϵ 1−ϵ . (4.31) (4.31) The complex turning points lie at zc, z∗ c, with zc = ˜κp + i q ˜m2p −˜κ2p 1−ϵ . (4.32) zc = ˜κp + i q ˜m2p −˜κ2p 1−ϵ . 4.1 dS However, they begin to depart from each other as \|˜κ\| becomes large, and in the end their large chemical-potential limits are drastically diﬀerent: bosons enter the tachyonic regime, while fermions saturate and approach an asymptotic limit where z∗∝\|κ\| and ∆N∗∝\|κ\|−1/2. Note that for bosons, both z∗and ∆N∗are monotonic functions of the chemical potential, whereas neither are monotonic for fermions. – 33 – 4.2 Deviation from dS of the ϵ-type (4.32) The phase integral is Z w(z,k)dz = (1−ϵ) Z dz′ z′ϵ s 1−2˜κp z′ + ˜m2p z′2 = −z(1−ϵ) ϵ s 1−2˜κp z 1 1−ϵ + ˜m2p z 2 1−ϵ (4.33) + z ϵ F1 −1+ϵ; 1 2, 1 2;ϵ;z 1 −1+ϵ ˜κp +i q ˜m2p −˜κ2p ,z 1 −1+ϵ ˜κp −i q ˜m2p −˜κ2p + z ϵ −1+ϵ (1−ϵ)˜κp ϵ2 F1 ϵ; 1 2, 1 2;1+ϵ;z 1 −1+ϵ ˜κp +i q ˜m2p −˜κ2p ,z 1 −1+ϵ ˜κp −i q ˜m2p −˜κ2p . Z w(z,k)dz (4.33) – 34 – Figure 10. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in dS with ϵ-type deviation and ϵ = 0.1. In both panels, the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κp = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin- 1/2 fermion mass µp = ˜mp = 10. The gray region with Re z < 0 cannot be reached physically. JHEP06(2021)129 Figure 10. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in dS with ϵ-type deviation and ϵ = 0.1. In both panels, the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κp = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin- 1/2 fermion mass µp = ˜mp = 10. The gray region with Re z < 0 cannot be reached physically. Here F1(a; b1, b2; c; x, y) is the Appell hypergeometric function deﬁned by the double series F1(a; b1, b2; c; x, y) = ∞ X m=0 ∞ X n=0 (a)m+n(b1)m(b2)n (c)m+n xmyn m!n! , (4.34) (4.34) where (a)n = Γ(a+n) Γ(a) is the Pochhammer symbol. The Stokes lines are plotted in ﬁgure 10. Comparing to ﬁgure 8, one ﬁnds that with a non-zero ϵ, the distribution of turning points becomes asymmetric around the imaginary axis and the Stokes lines become more squeezed with a larger positive chemical potential. where (a)n = Γ(a+n) Γ(a) is the Pochhammer symbol. The Stokes lines are plotted in ﬁgure 10. 4.2 Deviation from dS of the ϵ-type Comparing to ﬁgure 8, one ﬁnds that with a non-zero ϵ, the distribution of turning points becomes asymmetric around the imaginary axis and the Stokes lines become more squeezed with a larger positive chemical potential. • Production amount. For zi lying on the real axis, the above integral is real, suggesting no contribution to the particle production amount e−2Re F(zi), zi ∈R. Therefore, the particle production amount only receives contribution at the lower end zc. Converting this to an integral along the whole Stokes line, we obtain • Production amount. For zi lying on the real axis, the above integral is real, suggesting no contribution to the particle production amount e−2Re F(zi), zi ∈R. Therefore, the particle production amount only receives contribution at the lower end zc. Converting this to an integral along the whole Stokes line, we obtain 2ReF(zi) = 2i Z z∗ c zc w(z,k)dz = −2√π(1−ϵ)Γ(−ϵ) ˜m1−2ϵ p ×Im   ˜κp+i q ˜m2p−˜κ2p ϵ 2 ˜F1  −1 2,−ϵ; 3 2 −ϵ;−1+ 2˜κp ˜κp−i q ˜m2p−˜κ2p ˜m2p    . (4.35)  (4.35) For k Hp not far away from unity, a small dS-deformation parameter ϵ ≪1 can be used as an expansion parameter. After a further expansion in powers of ˜κp and resummation, For k Hp not far away from unity, a small dS-deformation parameter ϵ ≪1 can be used as an expansion parameter. After a further expansion in powers of ˜κp and resummation, – 35 – we ﬁnd we ﬁnd 2Re F(zi) = 2π ˜mp −˜κp + ϵ ( ˜mp + ˜κp) ln ˜mp + ˜κp 2 −2 ˜mp ln ˜mp + O(ϵ2) . Therefore, the particle production amount is now scale dependent: \|β(k)\|2 = e−2π µp−˜κp+ϵ( ˜mp+˜κp)ln ˜ mp+˜κp 2 −2 ˜mp ln ˜mp+O( ˜m−1 p ) +O(ϵ2) = exp  −2π Hp   s m2−H2p 4 −κ   1+ϵln k m +ϵm+κ m−κ ln m+κ 2m +O ϵH2 p m2 ,ϵ2 !! , (4.36) )\| = exp  −2π Hp   s m2−H2p 4 −κ   1+ϵln k m +ϵm+κ m−κ ln m+κ 2m +O ϵH2 p m2 ,ϵ2 !! , (4.36) JHEP06(2021)129  (4.36) where we have resummed the O(ϵ0) super-adiabatic corrections to the mass and replaced ˜mp(k) →µp(k) = s m2 H2p −1 4 k Hp ! ϵ 1−ϵ . 4.2 Deviation from dS of the ϵ-type (4.37) (4.37) The O(ϵ) order, however, cannot be treated in the same way as in (4.23), because of the presence of a branch cut extending from the origin all the way to inﬁnity. Therefore, considering the fact that the adiabatic parameter is w′(z) w(z)2 ∼˜m−2 p , we expect that the relative error of (4.36) is of order ϵH2 p m2 , which is negligible if the mass m is large. For a positive ϵ, the production amount necessarily drops with scale k, because the eﬀective mass in the Boltzmann factor is measured in units of the time-dependent Hubble parameter, which is decreasing during inﬂation. • Production time. This can only be solved numerically from Im F(z∗) = 0. However, by an educated guess, we found an empirical formula that describes the O(ϵ) contribution to z∗very well, with an error of 2% on average (see ﬁgure 11 for example). Namely, z∗( ˜mp,˜κp,ϵ) ≡z(0) ∗+z(1) ∗+··· ≈0.6627µp+0.3435˜κp−0.0102˜κ2 p µp + 0.0064˜κ3 p µ2p +··· +ϵ h 0.82 ˜mp−0.42˜κp−0.32 ˜mp ln ˜mp−(0.33 ˜mp+0.30˜κp)ln( ˜mp+˜κp)+O ˜m−1 p i +O(ϵ2). (4.38) (4.38) Here the ﬁrst line (z(0) ∗) has the same form as the exact solution found in dS spacetime and the second line (z(1) ∗) represents the leading-order slow-roll correction as an empirical formula, with an uncertainty due to the unresummed super-adiabatic corrections. Here the ﬁrst line (z(0) ∗) has the same form as the exact solution found in dS spacetime and the second line (z(1) ∗) represents the leading-order slow-roll correction as an empirical formula, with an uncertainty due to the unresummed super-adiabatic corrections. • Production width. The production width can also be expressed partially analyti- cally. On the real axis, the singulant function accumulates no real parts and therefore – 36 – Figure 11. The relative error of the O(ϵ) empirical formula (4.38) as compared to the numeric result. The bright band cutting across the plot is where the O(ϵ) correction crosses zero, hence the large relative error. JHEP06(2021)129 Figure 11. The relative error of the O(ϵ) empirical formula (4.38) as compared to the numeric result. The bright band cutting across the plot is where the O(ϵ) correction crosses zero, hence the large relative error. Re F(z∗) = Re F(zi). Plugging in the expression for Im F ′(z∗), we obtain Re F(z∗) = Re F(zi). 4.2 Deviation from dS of the ϵ-type Plugging in the expression for Im F ′(z∗), we obtain ∆z∗= 2 p 2\|Re F(z∗)\| \|Im F ′(z∗)\| = v u u u u t 2π h µp −˜κp + ϵ ( ˜mp + ˜κp) ln ˜mp+˜κp 2 −2 ˜mp ln ˜mp i 1 − 2˜κp z 1 1−ϵ ∗ + ˜m2p z 2 1−ϵ ∗ . (4 39 Re F(z∗) = Re F(zi). Plugging in the expression for Im F ′(z∗), we obtain z∗= 2 p 2\|Re F(z∗)\| \|Im F ′(z∗)\| = v u u u u t 2π h µp −˜κp + ϵ ( ˜mp + ˜κp) ln ˜mp+˜κp 2 −2 ˜mp ln ˜mp i 1 − 2˜κp z 1 1−ϵ ∗ + ˜m2p z 2 1−ϵ ∗ . (4 39 (4.39) ( ) When expressed in units of e-folding numbers and expanded into powers of κ m, we have the empirical expression ( ) When expressed in units of e-folding numbers and expanded into powers of κ m, we have the empirical expression = 1 1−ϵ ∆z∗ z∗ ≡∆N(0) ∗ +∆N(1) ∗ +··· ≈ 1 (m2/H2p −1/4)1/4 ( 2.0895− 0.4131κ (m2−H2p/4)1/2 + 0.1323κ2 m2−H2p/4 − 0.0523κ3 (m2−H2p/4)3/2 +··· +ϵ " −1 2 (2.0895−···)ln k m + 1.4κ m −0.70κ2 m2 + 0.32κ3 m3 +O H2 p m2 !# +O(ϵ2) ) . (4.40) (4.40) Again, the ﬁrst line (= ∆N(0) ∗) is exact while the second line (= ∆N(1) ∗) is approximate, with the exception of the coeﬃcient before the running term ln k m, which inherits its exactness from the ﬁrst line. The fermionic case is then easily obtained by applying the replacement µp, ˜mp → q ˜m2p + ˜κ2p to the above results. Since the zeroth order in ϵ is the same as the exact dS result with a Hubble constant Hp, we focus on the O(ϵ) corrections due to the deformation and plot them in ﬁgure 12. As shown in the plots, the behavior of bosons and fermions are again diﬀerent for large chemical potentials. – 37 – Figure 12. The O(ϵ) corrections to production histories in deformed dS with ϵ = 0.1. Left panel: production amount excluding zeroth-order contribution as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the correction to z-domain production time dependence on chemical potential and mass. Right panel: the correction to production width measured in e-folding numbers. 4.2 Deviation from dS of the ϵ-type In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. JHEP06(2021)129 Figure 12. The O(ϵ) corrections to production histories in deformed dS with ϵ = 0.1. Left panel: production amount excluding zeroth-order contribution as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the correction to z-domain production time dependence on chemical potential and mass. Right panel: the correction to production width measured in e-folding numbers. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. Before ending this subsection, we note that the validity of our perturbative expansion is actually controlled by ϵ ln max{ ˜mp, ˜κp} instead of just ϵ. Thus, for particles with extremely large mass or chemical potential, i.e., ˜mp, ˜κp ≳e1/ϵ, perturbation theory fails and one would have to rely on the full result (4.35) for production amount and numerically solve production time and width. For parameters chosen in ﬁgure 12, the corrections are small compared to the zeroth order dS results, suggesting the perturbative expansion is valid. 4.3 Deviation from dS of the η-type The second type of deformation of dS is obtained from introducing a weak time dependence in ϵ(τ) described by a non-zero η parameter, η(τ) = ϵ′(τ) a(τ)H(τ)ϵ(τ) . (4.41) (4.41) A further simpliﬁcation appears if ϵ(τ) ≪η(τ) for τ lying in the range of interest. We will call this type of deviation the η-type. It is reasonable to analyze particle production in such a scenario since for inﬂation, the Planck 2018 data [64] favors a smaller ﬁrst slow-roll parameter compared to the second slow-roll parameter. To study η-type deviation from dS, we suppose the scale factor can be expressed as a series with variable −Hiτ, where Hi has the dimension of Hubble. So the leading-order deviation can be approximated as a(τ) = −1 Hiτ + 1 (−Hiτ)1+ηi , ‘ (4.42) (4.42) – 38 – – 38 – Figure 13. The Hubble parameter and slow-roll parameters of the hybrid model (4.42). Here we have chosen ηi = 0.1 for illustration. Clearly, the hybrid model is separated into two stages, with early time dominated by η-type and late time dominated by ϵ-type. JHEP06(2021)129 JHEP06(2021)129 Figure 13. The Hubble parameter and slow-roll parameters of the hybrid model (4.42). Here we have chosen ηi = 0.1 for illustration. Clearly, the hybrid model is separated into two stages, with early time dominated by η-type and late time dominated by ϵ-type. with ηi > 0 and τ < 0. With this scale factor, the Hubble parameter is H(τ) = Hi(−Hiτ)ηi [1 + ηi + (−Hiτ)ηi] [1 + (−Hiτ)ηi]2 , (4.43) (4.43) and therefore the physical meaning of Hi is the Hubble parameter at τ →−∞. On the other hand, the ﬁrst and second slow-roll parameters are and therefore the physical meaning of Hi is the Hubble parameter at τ →−∞. On the other hand, the ﬁrst and second slow-roll parameters are ϵ(τ) = ηi 1 + ηi + (1 −ηi)(−Hiτ)ηi [1 + ηi + (−Hiτ)ηi]2 η(τ) = ηi (−Hiτ)ηi [1 + (−Hiτ)ηi] (1 + ηi)2 + (1 −ηi)(−Hiτ)ηi [1 + ηi + (−Hiτ)ηi]2 [1 + ηi + (1 −ηi)(−Hiτ)ηi] . (4.44) [1 + ηi + (−Hiτ)ηi] η(τ) = ηi (−Hiτ)ηi [1 + (−Hiτ)ηi] (1 + ηi)2 + (1 −ηi)(−Hiτ)ηi [1 + ηi + (−Hiτ)ηi]2 [1 + ηi + (1 −ηi)(−Hiτ)ηi] . 4.3 Deviation from dS of the η-type (4.44) (4.44) The asymptotic behaviors of the slow-roll parameters in the early-/late- time limit are (ϵ(τ), η(τ)) →    (0, ηi) , (−Hiτ)ηi ≫1 ηi 1+ηi , 0 , (−Hiτ)ηi ≪1 . (4.45) (4.45) Thus (4.42) describes a hybrid spacetime in which the deviation starts as η-type and then becomes ϵ-type (see ﬁgure 13 for illustration). In this section, we will be interested in the early stage where η-type deviation dominates. We ﬁrst evaluate the phase integrals in the limit with (−Hiτ)ηi ≫1. With the variable z ≡−kτ, the time-dependent frequency is w2(z, k) = 1 −2˜κ z 1 + c(k) zηi + ˜m2 z2 1 + c(k) zηi 2 , (4.46) (4.46) where ˜κ ≡ κ Hi , ˜m ≡m Hi and c(k) ≡( k Hi )ηi. The phase integral with this frequency cannot be evaluated analytically. However, for the modes which cross the horizon at z = −kτ ∼1 – 39 – early, when i.e., (−Hiτ)−ηi = c(k)z−ηi ∼c(k) ≪1, we can solve perturbatively. This suggests expanding the frequency in the series of c(k), early, when i.e., (−Hiτ)−ηi = c(k)z−ηi ∼c(k) ≪1, we can solve perturbatively. This suggests expanding the frequency in the series of c(k), w(z, k) = s 1 −2˜κ z + ˜m2 z2 + c(k)z−ηi−1 ˜m2 −˜κz p ˜m2 + z(z −2˜κ) + O (c(k)z−ηi)2 . (4.47) (4.47) Thus the phase integral reads Thus the phase integral reads JHEP06(2021)129 Z w(z, k)dz = p ˜m2 + z2 −2˜κz + m tanh−1 ˜κz −˜m2 ˜m √ ˜m2 + z2 −2˜κz ! + ˜κ tanh−1 ˜κ −z p ˜m2 + z(z −2˜κ) ! + c(k)z−ηi ˜m   z2F1 2 −ηi; 1 2, 1 2; 3 −ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi −2 − z˜κF1 1 −ηi; 1 2, 1 2; 2 −ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi −1 − ˜m2F1 −ηi; −1 2, −1 2; 1 −ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi   Z w(z, k)dz = p ˜m2 + z2 −2˜κz + m tanh−1 ˜κz −˜m2 ˜m √ ˜m2 + z2 −2˜κz ! + ˜κ tanh−1 ˜κ −z p ˜m2 + z(z −2˜κ) ! 4.3 Deviation from dS of the η-type + c(k)z−ηi ˜m   z2F1 2 −ηi; 1 2, 1 2; 3 −ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi −2 + O c(k)2 . + O c(k)2 . (4.4 (4.48) To calculate zc, we expand it in the order of c(k): To calculate zc, we expand it in the order of c(k): zc = z(0) c + c(k)z(1) c + O(c(k)2) , (4.49) (4.49) and the roots of (4.46) satisfy and the roots of (4.46) satisfy and the roots of (4.46) satisfy and the roots of (4.46) satisfy 1 zc 1 + c(k) zηi c = ˜κ2 −i √ ˜m2 −κ2 ˜m2 , (4.50) (4.50) and thus zc = ˜κ + i p ˜m2 −κ2 + c(k) ˜κ + i p ˜m2 −κ2 1−ηi + O(c(k)2) . (4.51) (4.51) – 40 – The singulant can now be solved up to the ﬁrst order in c(k): F(z) = 2i "Z z z(0) c +c(k)z(1) c w(0)(z,k)+c(k)w(1)(z,k) dz # +O(c(k)2) = 2i Z z z(0) c w(0)(z,k)+c(k)w(1)(z,k) dz−w(0)(z(0) c ,k)c(k)z(1) c +O(c(k)2) = 2i        p ˜m2+z2−2˜κz+mtanh−1 ˜κz−˜m2 ˜m √ ˜m2+z2−2˜κz ! +˜κtanh−1 ˜κ−z p ˜m2+z(z−2˜κ) ! + c(k)z−ηi ˜m   z2F1 2−ηi; 1 2, 1 2;3−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi−2 z) = 2i "Z z z(0) c +c(k)z(1) c w(0)(z,k)+c(k)w(1)(z,k) dz # +O(c(k)2) = 2i Z z z(0) c w(0)(z,k)+c(k)w(1)(z,k) dz−w(0)(z(0) c ,k)c(k)z(1) c +O(c(k)2) = 2i        p ˜m2+z2−2˜κz+mtanh−1 ˜κz−˜m2 ˜m √ ˜m2+z2−2˜κz ! +˜κtanh−1 ˜κ−z p ˜m2+z(z−2˜κ) ! + c(k)z−ηi ˜m   z2F1 2−ηi; 1 2, 1 2;3−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi−2 − z˜κF1 1−ηi; 1 2, 1 2;2−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi−1 − ˜m2F1 −ηi;−1 2,−1 2;1−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi   + √πc(k) ˜mΓ(2−ηi) ˜κ+i √ ˜m2−˜κ2 −ηi 2 ˜F1 −1 2,−ηi; 3 2 −ηi; (˜κ+i √ ˜m2−˜κ2) 2 ˜m2 2ηi        +π( ˜m−˜κ) +O(c(k)2). 4.3 Deviation from dS of the η-type (4.52) JHEP06(2021)129 − z˜κF1 1−ηi; 1 2, 1 2;2−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi−1 − ˜m2F1 −ηi;−1 2,−1 2;1−ηi; z(˜κ+i √ ˜m2−˜κ2) ˜m2 , z(˜κ−i √ ˜m2−˜κ2) ˜m2 ηi +O(c(k)2). (4.52) The Stokes lines are shown in ﬁgure 14. The distribution of Stokes lines on the complex plane is very similar to that of dS, except for a slightly decreased Hubble parameter. The Stokes lines are shown in ﬁgure 14. The distribution of Stokes lines on the complex plane is very similar to that of dS, except for a slightly decreased Hubble parameter. • Production amount. The singulant evaluated at zi gives • Production amount. The singulant evaluated at zi gives • Production amount. The singulant evaluated at zi gives 2ReF(zi) = 2i   Z z∗(0) c +c(k)z∗(1) c z(0) c +c(k)z(1) c w(0)(z,k)+c(k)w(1)(z,k) dz  +O(c(k)2) = 2i   Z z∗(0) c z(0) c w(0)(z,k)+c(k)w(1)(z,k) dz+w(0)(z∗(0) c ,k)c(k)z∗(1) c −w(0)(z(0) c ,k)c(k)z(1) c   +O(c(k)2) = 2π( ˜m−˜κ) −2√πc(k) ˜mΓ(2−ηi) ηi Im " ˜κ+i p ˜m2−˜κ2 −ηi 2 ˜F1 −1 2,−ηi; 3 2 −ηi; ˜κ+i √ ˜m2−˜κ2 2 ˜m2 !# +O(c(k)2). (4.53) (4.53) Thus, the production amount of particles that exit the horizon during the early η-phase Thus, the production amount of particles that exit the horizon during the early η-phase – 41 – Figure 14. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in dS with η-type deviation and η = 0.1, c(k) = 0.05. In both panels, the dimensionless chemical potential ranging over ˜κ = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin-1/2 fermion mass µ = ˜m = 10. The gray region with Re z < 0 cannot be reached physically. JHEP06(2021)129 Figure 14. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in dS with η-type deviation and η = 0.1, c(k) = 0.05. In both panels, the dimensionless chemical potential ranging over ˜κ = −8, −6, · · · , 6, 8 (from left to right, and the green lines correspond to the case without chemical potential), with spin-1 boson mass and spin-1/2 fermion mass µ = ˜m = 10. 4.3 Deviation from dS of the η-type The gray region with Re z < 0 cannot be reached physically. is computable to the linear order in c(k) ≪1: \|β(k)\|2 = exp        −2π  µ−˜κ−c(k) ˜mΓ(2−ηi) √πηi Im 2 ˜F1 −1 2,−ηi; 3 2 −ηi; (˜κ+i √ ˜m2−˜κ2)2 ˜m2 ˜κ+i √ ˜m2−˜κ2 ηi +O c(k) ˜m ,c(k)2          , (4.54) where we have resummed the O(c(k)0) super-adiabatic corrections to the mass and replaced ˜m →µ = s m2 H2 i −1 4 . (4.55) (4.55) The O(c(k)1) super-adiabatic corrections are complicated by the branch cut from z = 0 to z = −∞and we are not able to analyze them. Therefore, as in the ϵ-deviation case, we have indicated in (4.54) the presence of an O( ˜m−2) relative error. • Production time. The equation Im F(z∗) = 0 can only be solved numerically. However, as in the ϵ-type deviation case, we have found a useful empirical formula to the ﬁrst order in ηi, whose relative error is less than 2% within most parameter regions (see • Production time. The equation Im F(z∗) = 0 can only be solved numerically. However, as in the ϵ-type deviation case, we have found a useful empirical formula to the ﬁrst order in ηi, whose relative error is less than 2% within most parameter regions (see – 42 – Figure 15. The relative error of the empirical formula (4.56) as compared to the numeric result. The bright band cutting across the plot is where the z(1,1) ∗ crosses zero, hence the large relative error. JHEP06(2021)129 Figure 15. The relative error of the empirical formula (4.56) as compared to the numeric result. The bright band cutting across the plot is where the z(1,1) ∗ crosses zero, hence the large relative error. ﬁgure 15 for illustration). Up to O(c(k), ηi), we found ﬁgure 15 for illustration). Up to O(c(k), ηi), we found ﬁgure 15 for illustration). Up to O(c(k), ηi), we found z∗( ˜m, ˜κ,ηi;k) z∗( ˜m, ˜κ,ηi;k) ≡z(0,0) ∗ +z(1,0) ∗ +z(1,1) ∗ +··· ≈(1+c(k)) 0.6627µ+0.3435˜κ−0.0102˜κ2 µ + 0.0064˜κ3 µ2 +··· ! +c(k)ηi 0.84 ˜m−0.37˜κ−(0.35 ˜m+0.33˜κ)ln(˜κ+ ˜m)−0.30 ˜mln ˜m+O( ˜m−1) +O(c(k)η2 i ,c(k)2). (4.56) (4.56) (4.56) Here the ﬁrst line (= z(0,0) ∗ + z(1,0) ∗ ) is exact while the second line (= z(1,1) ∗ ) is approxi- mate.5 Here the ﬁrst line (= z(0,0) ∗ + z(1,0) ∗ ) is exact while the second line (= z(1,1) ∗ ) is approxi- mate.5 • Production width. Using the empirical formula (4.56), this can also be obtained ap- proximately as a truncated power series, ∆z∗( ˜m, ˜κ, ηi; k) √ ˜m ≈1.385 + 0.444˜κ ˜m −0.07554˜κ2 ˜m2 + 0.02839˜κ3 ˜m3 + · · · + c(k) ( 1 2 (1.385 + · · · ) + ηi " 0.34 −1.38˜κ ˜m + 0.19˜κ2 ˜m2 −0.05˜κ3 ˜m3 + −0.67 −0.21˜κ ˜m + 0.02˜κ2 ˜m2 ! ln ˜m # + O(η2 i ) ) 2 (4.57) 5Note that in this subsection, we will denote X(m,n) as the O(c(k)mηn i ) correction to a certain quantity X. Since X(1,0) follows from an exact dS spacetime with a rescaled Hubble constant, it can always be trivially obtained by rescaling Hi in X(0,0). In contrast, X(1,1) represents the leading order nontrivial η-corrections. – 43 – – 43 – In terms of e-folding numbers, the width is In terms of e-folding numbers, the width is In terms of e-folding numbers, the width is ∆N∗(m,κ,ηi;k) ≡∆N(0,0) ∗ +∆N(1,0) ∗ +∆N(1,1) ∗ +··· ≈∆z∗× d dz ln 1+c(k)z−ηi z z=z∗ ≈ 1 ( m2 H2 i −1 4)1/4 2.089− 0.4131κ (m2−H2 i /4)1/2 + 0.1323κ2 m2−H2 i /4 − 0.05226κ3 (m2−H2 i /4)3/2 ··· +c(k) ( −1 2 (2.089−···) +ηi " −0.04+ 1.4κ m −0.68κ2 m2 + 0.30κ3 m3 + 1.0−0.21κ m + 0.08κ2 m2 −0.04κ3 m3 ! ln m Hi # +O ηiH2 i m2 ,η2 i !) +O(c(k)2) ! . (4.58) JHEP06(2021)129 +O(c(k)2) ! . (4.58) (4.58) Again, the ﬁrst line (= ∆N(0,0) ∗ ) and the second line (= ∆N(1,0) ∗ ) are exact, whereas the third line (= ∆N(1,1) ∗ ) is approximate. Again, the ﬁrst line (= ∆N(0,0) ∗ ) and the second line (= ∆N(1,0) ∗ ) are exact, whereas the third line (= ∆N(1,1) ∗ ) is approximate. The results for fermions are again obtained via the corresponding replacement. ﬁgure 15 for illustration). Up to O(c(k), ηi), we found The O(c(k)ηi) corrections to production histories in η-type deformed dS are plotted in ﬁgure 16. Comparing to ﬁgure 12, one can see that the overall behavior is similar to that of ϵ-type deformed dS. Indeed, if we choose ϵ ∼c(k)ηi, the corrections roughly match in size. This interesting fact will be discussed below. Nevertheless, we notice that there are important diﬀerences in the scale dependence of various production history parameters. Comparison between ϵ-type and η-type. Gravitational particle production histories are crucially inﬂuenced by the Hubble parameter H(τ), as it directly enters the expression of dimensionless mass and chemical potential. Two spacetimes with diﬀerent H(τ) are intrinsically diﬀerent for any process that is non-local in time, including the smoothed Stokes phenomenon. Thus to test this intrinsic diﬀerence of particle production for the two types of dS deviations, we carefully select their Hubble parameters to be tangent to each other at time τ0, so that both the Hubble and the ﬁrst slow-roll parameter are equal, H(ϵ)(τ0, Hp, ϵ) = H(η)(τ0, Hi, ηi) (4.59) ϵ(ϵ)(τ0, Hp, ϵ) = ϵ(η)(τ0, Hi, ηi) (4.60) (4.59) (4.60) (4.59) (4.60) For the purpose of demonstration, we will choose the following solution, ϵ = 0.005 , Hp ≊0.8204Hi , ηi = 0.1 , τ0 = −2.042 × 1012H−1 i . (4.61) (4.61) – 44 – – 44 – Figure 16. The O(c(k)ηi) corrections to production histories in deformed dS with c(k) = 0.05 and ηi = 0.1. Left panel: production amount excluding O(η0 i ) contribution as a function of the dimen- sionless chemical potential for diﬀerent particle masses. Middle panel: the correction to z-domain production time dependence on chemical potential and mass. Right panel: the correction to pro- duction width measured in e-folding numbers. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. JHEP06(2021)129 Figure 16. The O(c(k)ηi) corrections to production histories in deformed dS with c(k) = 0.05 and ηi = 0.1. ﬁgure 15 for illustration). Up to O(c(k), ηi), we found Left panel: production amount excluding O(η0 i ) contribution as a function of the dimen- sionless chemical potential for diﬀerent particle masses. Middle panel: the correction to z-domain production time dependence on chemical potential and mass. Right panel: the correction to pro- duction width measured in e-folding numbers. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. For bosons, we limit the range of chemical potential to be smaller than the mass, so that no tachyonic instability is induced. For fermions, the chemical potential is not restricted and we allow it to take arbitrarily large values. Figure 17. The Hubble parameter for two spacetimes are chosen to be tangent to each other at the time τ0, which is indicated by the vertical black line. Figure 17. The Hubble parameter for two spacetimes are chosen to be tangent to each other at the time τ0, which is indicated by the vertical black line. The Hubble parameter for the choice (4.61) is shown in ﬁgure 17. The production amount and width can be computed using formula given above. Their scale dependence is shown in ﬁgure 18. As shown in this ﬁgure, particles created near the tangent time τ0 approximately have the same production amount and width, with mismatches of the same order as the higher-order errors in (4.54) and (4.58). The Hubble parameter for the choice (4.61) is shown in ﬁgure 17. The production amount and width can be computed using formula given above. Their scale dependence is shown in ﬁgure 18. As shown in this ﬁgure, particles created near the tangent time τ0 approximately have the same production amount and width, with mismatches of the same order as the higher-order errors in (4.54) and (4.58). As a result, one can approximate the η-deformed dS by a tangential ϵ-deformed dS locally in time, and obtain the details of particles produced then, up to some higher-order errors. However, this can only be done mode-by-mode, since the scale dependence for these two types of deviations is diﬀerent. Conversely, if we wish to observationally distinguish – 45 – Figure 18. The scale dependence of production amount and width for the two spacetimes in ﬁgure 17. 6If the initial condition is non-trivial with a non-zero particle number density, one can model this initial particle population by a non-zero βi. The phase of βi may be ﬁxed if the initial particles are prepared coherently, whereas it is random for thermally prepared particles. In such cases, one can return to (3.31), (3.41), (3.81), (3.86), and add to β(τ) a term proportional to βi and then compute the change of particle number by taking the square and performing an additional ensemble average over the phase of βi if it is thermally prepared. ﬁgure 15 for illustration). Up to O(c(k), ηi), we found The opaque bands stand for the higher-order errors present in (4.36), (4.40), (4.54) and (4.58). Notice that the error band for ϵ-type is too thin to be observed. The vertical black line indicates the mode k0 = −τ −1 0 . The other parameters are chosen to be m = 10Hi, κ = 9Hi. JHEP06(2021)129 Figure 18. The scale dependence of production amount and width for the two spacetimes in ﬁgure 17. The opaque bands stand for the higher-order errors present in (4.36), (4.40), (4.54) and (4.58). Notice that the error band for ϵ-type is too thin to be observed. The vertical black line indicates the mode k0 = −τ −1 0 . The other parameters are chosen to be m = 10Hi, κ = 9Hi. the two scenarios, we can either accurately measure the production amount/width for a single mode up to some higher orders (e.g., O(c(k)2)), or probe the scale dependence by looking at diﬀerent modes. 4.4 Radiation-domination era Now we turn to other completely diﬀerent FRW backgrounds, namely those describing the post-inﬂationary evolution of the universe. Roughly speaking, the scale factor evolves as a(τ) ∝τ 2 3w+1 , where w is the equation of state of the dominating component. In this section and the next, we will consider radiation domination era with w = 1 3 and matter domination era with w = 0, respectively. The radiation-dominated universe has a scale factor linearly dependent on the confor- mal time, a(τ) = crτ, where cr > 0 and τ runs from 0 to +∞. The Hubble parameter is H(τ) = 1 crτ 2 . At the origin lies the Big Bang Singularity, H(0+) →∞. This singularity can be removed by continuously deforming the spacetime to other geometries such as that of inﬂationary [1–4], ekpyrotic [65, 66], bouncing [67, 68] and string gas cosmology [69, 70]. These physical continuations provide a cutoﬀtime τi > 0 with H(τi) ≪Mp. As we are interested in particle production during the later stage of the radiation-dominated era, we will limit ourselves to the modes with kτi ≪1 and assume that they have a vacuum initial condition at τ = τi ≈0+ prepared by the earlier evolution history.6 Thus for τ ≫τi, – 46 – any particle mode with comoving momentum k has a time-dependent physical momentum k a(τ) = k crτ . This scale is to be compared with H(τ), m and κ. Deﬁning z = kτ, the EoM now reads any particle mode with comoving momentum k has a time-dependent physical momentum k a(τ) = k crτ . This scale is to be compared with H(τ), m and κ. r eﬁning z = kτ, the EoM now reads g z = kτ, the EoM now reads d2f(z) dz2 + w2(z, k)f(z) = 0 , w2(z, k) = 1 −2˜κ(k)z + ˜m2(k)z2 , (4.62) (4.62) with scale-dependent eﬀective chemical potential and mass ˜κ(k) = κcr k2 , ˜m(k) = mcr k2 . (4.63) (4.63) Their physical meaning is the mass and chemical potential measured in units of Hubble parameter at horizon re-entry. The lower zero of w(z; k) lies at JHEP06(2021)129 Their physical meaning is the mass and chemical potential measured in units of Hubble parameter at horizon re-entry. The lower zero of w(z; k) lies at zc = ˜κ −i √ ˜m2 −˜κ2 ˜m2 . 4.4 Radiation-domination era (4.64) (4.64) The singulant integral reads F(z) = −2i Z z zc dz′w(z′;k) = i  ˜κ−˜m2z ˜m2 p 1−2˜κz+ ˜m2z2+ ˜m2−˜κ2 ˜m3 ln √ ˜m2−˜κ2 ˜m √ 1−2˜κz+ ˜m2z2+z ˜m −˜κ  + π ˜m2−˜κ2 2 ˜m3 . (4.65) (4.65) The super-adiabatic corrections to W(z) yield no simple pole at the origin or at inﬁnity: δW (n)(z) z→0 −−−→O(z0) , δW (n)(z) z→∞ −−−→O(z1−4n), n ⩾1 . (4.66) (4.66) Therefore, unlike the case of dS in conformal time coordinates, there is no super-adiabatic mass correction and hence no need for resummation. The behavior of the singulant as well as the Stokes multiplier are shown in ﬁgure 19. The details of particle production can be easily obtained as follows. • Production amount. Taking the imaginary part of F on the real axis, we have uction amount. Taking the imaginary part of F on the real axis, we have \|β(k)\|2 = e−π ˜ m3 ( ˜m2−˜κ2) = exp " −πk2 crm3 m2 −κ2# . (4.67) (4.67) There are two interesting aspects in this formula. First, \|β(k)\|2 is symmetric under the ﬂip κ ↔−κ (equivalent to a parity transformation that ﬂips helicities), seemingly suggesting an equal enhancement for both helicities. However, this turns out to be superﬂuous as we will see from inspecting the production history below. Second, for a given mode k, the production amount does not seem to increase with mass monotoni- cally. This phenomenon may be somewhat counterintuitive, as heavier particles naively should be more diﬃcult to produce. This puzzle is resolved when one recalls that the radiation-dominated universe does not have a constant background temperature like dS. It eﬀective “temperature” is likened to the decreasing Hubble parameter H(τ) = 1 crτ 2 . – 47 – Figure 19. The super-adiabatic singulant F(z) in radiation-dominated era for ˜κ = 0.05 (upper left panel) and ˜κ = 0.099 (upper right panel), with the dimensionless mass ˜m = 0.1. The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. JHEP06(2021)129 Figure 19. The super-adiabatic singulant F(z) in radiation-dominated era for ˜κ = 0.05 (upper left panel) and ˜κ = 0.099 (upper right panel), with the dimensionless mass ˜m = 0.1. 4.4 Radiation-domination era The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in radiation-dominated universe. In both panels, the mass is set to ˜m = 0.1, and the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κ = −0.08, −0.06, · · · , 0.06, 0.08 (from left to right, and the green lines correspond to the case without chemical potential). The gray region with Re z < 0 cannot be reached physically. Solid lines represent scenarios where the Ginzburg criterion (4.70) is fulﬁlled and vacuum initial condition is satisﬁed, while dotted lines can only be understood in the analytically continued sense. JHEP06(2021)129 Figure 20. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in radiation-dominated universe. In both panels, the mass is set to ˜m = 0.1, and the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κ = −0.08, −0.06, · · · , 0.06, 0.08 (from left to right, and the green lines correspond to the case without chemical potential). The gray region with Re z < 0 cannot be reached physically. Solid lines represent scenarios where the Ginzburg criterion (4.70) is fulﬁlled and vacuum initial condition is satisﬁed, while dotted lines can only be understood in the analytically continued sense. and admits a straightforward continuation to the unphysical region −∞< z < 0. This mathematically continued EoM enjoys a Z2 symmetry that is well-deﬁned at the origin:7 κ ↔−κ, z ↔−z. This is the cause of the apparent parity symmetry in \|β(k)\|2. In fact, the production time z∗is also in the unphysical region for κ < 0. Thus for negative κ, both (4.67) and (4.68) must be understood in the analytically-continued sense. Namely, only if the initial condition at z = 0+ is prepared so as to match the solution of the analytically-continued EoM with Bunch-Davies initial condition at z →−∞, the Stokes- line method results are valid. For κ > 0 and 0 < ∆z∗ 2 ≲z∗, these results agree with that of the usual vacuum initial condition at z = 0+ since the Stokes line is far right to the origin and particles do not get produced until a late time. Otherwise, a direct application of the Stokes-line method may be inaccurate. 4.4 Radiation-domination era The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. Raising the mass may lead to two competing eﬀects, one being increasing the diﬃculty of producing a real particle, the other being pushing the production time earlier, when the eﬀective “temperature” is higher. We will see later that for the radiation-dominated universe, the ﬁrst eﬀect dominates the applicable range of our method and heavier particles come with a smaller production amount. However, in the matter-dominated universe, this is not the case. Finally, the production amount sharply drops to zero for k ≳ q 2crm3 m2−κ2 , for which the “temperature” is too low to support any real particles. Raising the mass may lead to two competing eﬀects, one being increasing the diﬃculty of producing a real particle, the other being pushing the production time earlier, when the eﬀective “temperature” is higher. We will see later that for the radiation-dominated universe, the ﬁrst eﬀect dominates the applicable range of our method and heavier particles come with a smaller production amount. However, in the matter-dominated universe, this is not the case. Finally, the production amount sharply drops to zero for k ≳ q 2crm3 m2−κ2 , for which the “temperature” is too low to support any real particles. • Production time. Since the Stokes lines are vertical lines on the complex plane (see ﬁgure 20), the crossing time solved from Im F(z∗) = 0 is simply • Production time. Since the Stokes lines are vertical lines on the complex plane (see ﬁgure 20), the crossing time solved from Im F(z∗) = 0 is simply z∗= ˜κ ˜m2 , or τ∗= kκ crm2 . (4.68) (4.68) Now let us recall that the physical time domain is 0 < z < +∞, with z = 0 being the Big Bang singularity. The EoM (4.62) by itself, however, is regular at z = 0 Now let us recall that the physical time domain is 0 < z < +∞, with z = 0 being the Big Bang singularity. The EoM (4.62) by itself, however, is regular at z = 0 – 48 – Figure 20. 7In the dS case, this Z2 symmetry is also present in the EoM, but then it is not well-deﬁned at the dS boundary z = 0, which is a singularity for the EoM (not a singularity for the spacetime). The lack of a smooth continuation to z < 0 breaks the Z2 symmetry spontaneously, and therefore leads to the parity asymmetry in particle production. 4.4 Radiation-domination era In that case, the particle production history depends on the actual initial condition set at zi = kτi ≪1 by an earlier cosmic evolution, which is a physical continuation to the region z < 0. • Production width. The production width in z-domain is simply calculated as Production width. The production width in z-domain is simply calculated as ∆z∗= 2 p 2\|Re F(z∗)\| \|Im F ′(z∗)\| = r π ˜m , or ∆τ∗= r π crm . (4.69) (4.69) 7In the dS case, this Z2 symmetry is also present in the EoM, but then it is not well-deﬁned at the dS boundary z = 0, which is a singularity for the EoM (not a singularity for the spacetime). The lack of a smooth continuation to z < 0 breaks the Z2 symmetry spontaneously, and therefore leads to the parity asymmetry in particle production. – 49 – – 49 – Figure 21. The parameter region where the Ginzburg criterion (4.70) is satisﬁed. The blue region stands for spin-1 bosons while the yellow region stands for spin-1/2 fermions. JHEP06(2021)129 Figure 21. The parameter region where the Ginzburg criterion (4.70) is satisﬁed. The blue region stands for spin-1 bosons while the yellow region stands for spin-1/2 fermions. Interestingly, the production width in the z-domain for spin-1 boson does not depend on the chemical potential κ. The parameter region where our continuation interpretation matches that of the vacuum initial condition at z = 0+ is where the Ginzburg criterion is satisﬁed, Interestingly, the production width in the z-domain for spin-1 boson does not depend on the chemical potential κ. The parameter region where our continuation interpretation matches that of the vacuum initial condition at z = 0+ is where the Ginzburg criterion is satisﬁed, 0 < ∆z∗ 2 ≲z∗⇒ √π 2 ˜m3/2 ≲˜κ < ˜m , or √πcrm3/2 2k ≲κ < m . (4.70) (4.70) These conditions actually limits ˜m < 4 π. The parameter region satisfying the Ginzburg criterion is shown in ﬁgure 21, and one can see that both ˜κ and ˜m are bounded from above, i.e., ˜κ, ˜m ≲O(1). This suggests that the modes are still relativistic at horizon re-entry, and becomes non-relativistic only after the production time z∗. The fermion case is again obtained by a simple substitution ˜m → √ ˜m2 + ˜κ2. Here one useful check is to go to the large-chemical-potential limit with κ ≫m. 4.4 Radiation-domination era There the production amount of fermions reduces to \|β(k)\|2 = e − π ˜ m2 ( ˜ m2+˜κ2)3/2 κ≫m −−−→e−πk2m2 crκ3 = e− πm2 κH(τ∗(k)) , (4.71) (4.71) which is exactly what we expect from the LZ model (2.22). which is exactly what we expect from the LZ model (2.22). The Ginzburg criterion must also be applied to fermions. Hence, unlike the previous scenarios in dS and its deviations, the chemical potential of fermions is bounded from above (as well as below) by the applicability of our method. We plot the production histories in ﬁgure 22. In the valid parameter region, we found that the production amount is monotonically decreasing with mass. Bosons are produced later with larger chemical potential, while the production time of fermions ﬁrst increase and then decrease with chemical potential. The boson production width is independent of the chemical potential whereas that of fermions decreases with chemical potential, since ˜κ enters the expression for eﬀective mass. – 50 – Figure 22. The production histories in radiation-dominated era. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: the z- domain production width. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. The parameter range is chosen according to ﬁgure 21. Due to the imposed constraint, the purple dashed line for fermions with ˜m = 0.5 is absent in all three plots, since it would correspond to artiﬁcial initial conditions, as discussed in the main text above. JHEP06(2021)129 Figure 22. The production histories in radiation-dominated era. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: the z- domain production width. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. The parameter range is chosen according to ﬁgure 21. 4.4 Radiation-domination era Due to the imposed constraint, the purple dashed line for fermions with ˜m = 0.5 is absent in all three plots, since it would correspond to artiﬁcial initial conditions, as discussed in the main text above. 4.5 Matter-dominated era Now we turn to the matter-dominated era with w = 0 and a scale factor quadratically dependent on the conformal time, a(τ) = cmτ 2. The Hubble parameter is H(τ) = 2 cmτ 3 . Here cm > 0 and τ runs from 0 to +∞. The initial Big Bang singularity is understood to be removed by attaching a period of radiation domination era and some former primordial eras. A vacuum initial condition is still assumed, therefore we will still impose a Ginzburg criterion so that the Stokes-line method gives physical results. Deﬁning z = kτ, the EoM of a massive vector boson reads Deﬁning z = kτ, the EoM of a massive vector boson reads d2f(z) dz2 + w2(z, k)f(z) = 0 , w2(z, k) = 1 −2˜κ(k)z2 + ˜m2(k)z4 , (4.72) (4.72) with scale-dependent chemical potential and mass ˜κ(k) = κcm k3 , ˜m(k) = mcm k3 . (4.73) (4.73) The w(z, k) in the matter-dominated universe has four simple zeros on the complex z-plane, namely, ±zc and ±z∗ c with √ √ zc = √˜m + ˜κ −i√˜m −˜κ √ 2 ˜m . (4.74) (4.74) The zc-z∗ c pair lies in the right-half plane with positive real parts. Therefore, they are joined by a Stokes line that cross the real axis in the physical region (see the Stokes lines in ﬁgure 23 and ﬁgure 24). On the other hand, the Stokes line joining −zc and −z∗ c crosses the real axis in the unphysical region, which can only be understood in the aforementioned continuation sense. The zc-z∗ c pair lies in the right-half plane with positive real parts. Therefore, they are joined by a Stokes line that cross the real axis in the physical region (see the Stokes lines in ﬁgure 23 and ﬁgure 24). On the other hand, the Stokes line joining −zc and −z∗ c crosses the real axis in the unphysical region, which can only be understood in the aforementioned continuation sense. – 51 – We are interested in the zc-z∗ c pair since these determines the physical particle produc- tion details. As κ approaches −m, the two pairs tend to merge into one, and an analysis of the “physical” Stokes line alone seems insuﬃcient. In our setup, however, this is not a problem. 4.5 Matter-dominated era The Ginzburg criterion keeps the production time late enough so that the turning points at −zc, −z∗ c do not have signiﬁcant inﬂuence on the production history at leading order. The singulant integral takes a relatively complicated form, F(z) = −2 3iz p 1−2˜κz2+ ˜m2z4 + 4 3 ˜m s ˜m2−˜κ2 ˜κ−i √ ˜m2−˜κ2 " K −˜m2−2˜κ2+2i˜κ √ ˜m2−˜κ2 ˜m2 −F arcsin z ˜m p ˜κ−i √ ˜m2−˜κ2 ! −˜m2−2˜κ2+2i˜κ √ ˜m2−˜κ2 ˜m2 !# + 4i˜κ p ˜κ+i √ ˜m2−˜κ2 3 ˜m2 " E −˜m2−2˜κ2+2i˜κ √ ˜m2−˜κ2 ˜m2 −E arcsin z ˜m p ˜κ−i √ ˜m2−˜κ2 ! −˜m2−2˜κ2+2i˜κ √ ˜m2−˜κ2 ˜m2 !# , (4.75) JHEP06(2021)129 where K(M), E(M) and E(x\|M), F(x\|M) are the complete and incomplete elliptic integrals. • Production amount. Taking the real part gives \|β(k)\|2 = e−2Re F(zi) with • Production amount. Taking the real part gives \|β(k)\|2 = e−2Re F(zi) with Re F(zi) = 4 3 ˜m2 "p ˜m2 −˜κ2Re q ˜κ + i p ˜m2 −˜κ2K −˜m2 −2˜κ2 + 2i˜κ √ ˜m2 −˜κ2 ˜m2 !! −˜κIm q ˜κ + i p ˜m2 −˜κ2E −˜m2 −2˜κ2 + 2i˜κ √ ˜m2 −˜κ2 ˜m2 !! # = Γ 5 4 Γ 7 4 r π 2 ˜m  1 − 2Γ 3 4 Γ 7 4 ˜κ Γ 1 4 Γ 5 4 ˜m −3˜κ2 8 ˜m2 + O ˜κ3 ˜m3 ! . (4.76) In terms of the original parameters, we have In terms of the original parameters, we have \|β(k)\|2 = exp   − Γ 5 4 Γ 7 4 s 2πk3 cmm  1 − 2Γ 3 4 Γ 7 4 κ Γ 1 4 Γ 5 4 m −3κ2 8m2 + O κ3 m3 !    . (4.77) (4.77) Thus in the matter-dominated era, the superﬁcial degeneracy between κ > 0 and κ < 0 in the radiation-dominated era is explicitly broken, since there is no Z2 symmetry in the EoM (4.72) now. This shows again that the degeneracy is unphysical. Not only the z →−∞phase before radiation domination (inﬂation, etc.) can break the degeneracy, the z →+∞stage after it (matter-domination) can do so, too. The production amount drops to zero quickly for k ≫(cmm)1/3. – 52 – Figure 23. 4.5 Matter-dominated era The super-adiabatic singulant F(z) in matter-dominated era for ˜κ = 0.05 (upper left panel) and ˜κ = 0.099 (upper right panel), with the dimensionless mass ˜m = 0.1. The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. JHEP06(2021)129 Figure 23. The super-adiabatic singulant F(z) in matter-dominated era for ˜κ = 0.05 (upper left panel) and ˜κ = 0.099 (upper right panel), with the dimensionless mass ˜m = 0.1. The hue represents the phase arg F(z) while the brightness represents the modulus \|F(z)\|. In both panels, the dark red line is the Stokes line joining the complex turning points. The lower two panels show the Stokes multiplier S(z) corresponding to the parameters chosen above. • Production time. The production time is solved from Im F(z∗) = 0. The result can be expressed as a power series • Production time. The production time is solved from Im F(z∗) = 0. The result can be expressed as a power series z∗( ˜m, ˜κ) ≃ 1 √ ˜m " 0.6098 + 0.4659˜κ ˜m −0.1011˜κ2 ˜m2 + 0.03446˜κ3 ˜m3 + O ˜κ4 ˜m4 !# . (4.78) (4.78) Although z∗is always positive for \|˜κ\| < ˜m, the Ginzburg criterion must still be imposed to match the vacuum initial condition. Although z∗is always positive for \|˜κ\| < ˜m, the Ginzburg criterion must still be imposed to match the vacuum initial condition. • Production width. This can also be obtained as a power series, • Production width. This can also be obtained as a power series, ∆z∗( ˜m, ˜κ) ≃ 1 ˜m1/4 " 1.474 −0.2971˜κ ˜m + 0.09144˜κ2 ˜m2 −0.06400˜κ3 ˜m3 + O ˜κ4 ˜m4 !# . (4.79) (4.79) – 53 – – 53 – Figure 24. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in matter-dominated universe. In both panels, the mass is set to ˜m = 0.1, and the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κ = −0.08, −0.06, · · · , 0.06, 0.08 (from left to right, and the green lines correspond to the case without chemical potential). The gray region with Re z < 0 cannot be reached physically. 4.5 Matter-dominated era Solid lines represent scenarios where the Ginzburg criterion (4.80) is fulﬁlled and vacuum initial condition is satisﬁed, while dotted lines can only be understood in the analytically continued sense. JHEP06(2021)129 Figure 24. The Stokes lines for spin-1 bosons (left panel) and spin-1/2 fermions (right panel) in matter-dominated universe. In both panels, the mass is set to ˜m = 0.1, and the momentum k is ﬁxed to have a dimensionless chemical potential ranging over ˜κ = −0.08, −0.06, · · · , 0.06, 0.08 (from left to right, and the green lines correspond to the case without chemical potential). The gray region with Re z < 0 cannot be reached physically. Solid lines represent scenarios where the Ginzburg criterion (4.80) is fulﬁlled and vacuum initial condition is satisﬁed, while dotted lines can only be understood in the analytically continued sense. The Ginzburg criterion is now 0 < ∆z∗ 2 ≲z∗, or k3 cmm !1/4 ≳1.208× 1−0.2017 κ m +0.06205 κ2 m2 −0.04343 κ3 m3 +··· 1+0.7641 κ m −0.1660 κ2 m2 +0.05652 κ3 m3 +··· . (4.80) (4.80) ( ) We plot this region in ﬁgure 25 for both bosons and fermions. ( We plot this region in ﬁgure 25 for both bosons and fermions. The generalization to fermions is the same as above. To check the large-chemical- potential behavior, we set ˜κ ≫˜m, and obtain The generalization to fermions is the same as above. To check the large-chemical- potential behavior, we set ˜κ ≫˜m, and obtain \|β(k)\|2 ˜κ≫˜m −−−→e −π ˜ m2 2˜κ5/2 = e −πm2 2κ q k3 cmκ3 = e− πm2 κH(τ∗(k)) . (4.81) (4.81) This again agrees with the LZ model result (2.22). In fact, we can take advantage of the LZ model and obtain some interesting sum rules for the numeric coeﬃcients in (4.78) and (4.79). According to the exact solution of LZ model, the ˜κ ≫˜m limit production time becomes z∗ ˜κ≫˜m −−−→˜κ−1/2. Matching this with (4.78) in the fermion case, we have 1 ≃0.6098 + 0.4659 −0.1011 + 0.03446 + · · · . (4.82) (4.82) The production width predicted by the LZ model is ∆z∗= q π 2 ˜κ−1/4. Matching this with (4.79), we have rπ 2 ≃1.474 −0.2971 + 0.09144 −0.06400 + · · · . (4.83) (4.83) – 54 – – 54 – Figure 25. The parameter region where the Ginzburg criterion (4.80) is satisﬁed. 4.5 Matter-dominated era The blue region stands for spin-1 bosons while the yellow region stands for spin-1/2 fermions. JHEP06(2021)129 Figure 25. The parameter region where the Ginzburg criterion (4.80) is satisﬁed. The blue region stands for spin-1 bosons while the yellow region stands for spin-1/2 fermions. Figure 26. The production histories in matter-dominated era. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: the z- domain production width. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. The parameter range is chosen according to ﬁgure 25. Figure 26. The production histories in matter-dominated era. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: the z- domain production width. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. The parameter range is chosen according to ﬁgure 25. Figure 26. The production histories in matter-dominated era. Left panel: production amount as a function of the dimensionless chemical potential for diﬀerent particle masses. Middle panel: the z-domain production time dependence on chemical potential and mass. Right panel: the z- domain production width. In all three plots, solid lines represent spin-1 bosons while dashed lines represent spin-1/2 fermions. Particles with diﬀerent masses are distinguished by the colors of the lines according to the legend in the left panel. The parameter range is chosen according to ﬁgure 25. Another piece of information attainable from the LZ model is an upper bound on ˜κ due to the Ginzburg criterion: ∆ ∆z∗ 2 < z∗ ˜κ≫˜m −−−→˜κ < 64 π2 , (4.84) (4.84) in agreement with ﬁgure 25. in agreement with ﬁgure 25. As before, we plot the production histories for both bosons and fermions in ﬁgure 26. The left panel clearly demonstrates the seemingly counter-intuitive mass dependence men- tioned previously. Here, the valid region covers the part where the production amount increases with the mass. This shows the fact that production amount decreases with mass in the radiation-domination era is just a coincidence, and that with a time-dependent Hub- – 55 – ble parameter, there are two opposite eﬀects competing against each other, in which case the resulting mass dependence can be subtle. Another notable aspect is that the production time can be either earlier or later than horizon re-entry. 5 Summary and outlook Ranging from cosmological collider physics in the primordial era to baryogenesis in the late universe, chemical potential plays an important role in the process of spontaneous creation of particles. In this paper, we focused on the impact of chemical potential on gravitational massive particle production. We ﬁrst introduced the general form of chemical potential term and gave a necessary condition for its physical eﬀects. After reviewing the chemical potential for particles with diﬀerent spins, we extracted their essential features and likened the corresponding Bogoliubov coeﬃcients to the coeﬃcients of instantaneous positive/negative frequency solutions. Then the mathematical tools such as asymptotic series, Berry’s smoothing techniques of Stokes-lines and Borel resummation were introduced to solve the coeﬃcients. Having checked the applicability of this method at \|β\|2 ≲1, we obtain the recipe of particle production histories for both spin-1 bosons and spin- 1/2 fermions, which are related by a simple replacement formula. At last, applying this recipe to cosmology, we gave a ﬁne-grained analysis of chemical-potential-assisted particle production in ﬁve common FRW spacetimes. The production amount, time and width are obtained as analytic/semi-analytical expressions, each with characteristic dependences on chemical potential and mass. JHEP06(2021)129 In summary, our method demonstrates the application of uniformly smoothed Stokes- line method to ﬁne-grained particle production. In addition, our results serve as valu- able theoretical data for future studies of chemical potential as well as general particle production. Despite the heavy mathematical machinery and the detailed analysis in this current work, there are still many questions left unanswered which we hope to address in the future. We list a few of them as outlooks below. • Starting with vacuum initial condition, the introduction of chemical potential invites the interesting possibility of signiﬁcant particle production with \|β\|2 ∼1, even with large masses. This mathematically corresponds to the failure of choosing an optimal truncation order n (ReF < 2 for vector bosons and ReF < 1 for fermions) determined by either the stationary phase condition of the ﬁrst-order perturbation or the minimum term in the asymptotic series solution, and we proposed to use the Borel summation to evaluate the whole divergent asymptotic series, extending the workable parameter regions to ReF ≳0.5 for bosons and ReF ≳0.2 for fermions respectively. 5 Summary and outlook However, large errors in evaluating the particle production amount were still found when the particle production is too large and runs outside the mentioned workable regions, and part of the reason may be attributed to the failure of approximating w(z) or E(z) around its complex root zc. Although we are currently unable to fully resolve the problem when \|β\|2 ∼1, it is interesting to note that the result with \|β\|2 ≲1, when naively extrapolated • Starting with vacuum initial condition, the introduction of chemical potential invites the interesting possibility of signiﬁcant particle production with \|β\|2 ∼1, even with large masses. This mathematically corresponds to the failure of choosing an optimal truncation order n (ReF < 2 for vector bosons and ReF < 1 for fermions) determined by either the stationary phase condition of the ﬁrst-order perturbation or the minimum term in the asymptotic series solution, and we proposed to use the Borel summation to evaluate the whole divergent asymptotic series, extending the workable parameter regions to ReF ≳0.5 for bosons and ReF ≳0.2 for fermions respectively. However, large errors in evaluating the particle production amount were still found when the particle production is too large and runs outside the mentioned workable regions, and part of the reason may be attributed to the failure of approximating w(z) or E(z) around its complex root zc. Although we are currently unable to fully resolve the problem when \|β\|2 ∼1, it is interesting to note that the result with \|β\|2 ≲1, when naively extrapolated – 56 – to the \|β\|2 ∼1 case, actually gives very accurate answers for the production amount (e.g., in dS and radiation domination era). A systematic method of calculating particle production which can link to the limit with \|β\|2 ∼1, the tachynonic instability for bosons and the exact Landau-Zener model for fermions (2.21), may require new techniques, and we leave it for future works. • Throughout the analysis of particle production in section 4, we have assumed a chemical potential constant in space and time. Although this can be justiﬁed as leading order approximations, the full understanding can only be acquired by introducing appropriate spacetime dependences according to diﬀerent contexts. JHEP06(2021)129 • In dS and its two types of deviations, it is natural to assume a vacuum initial condition. However, in radiation domination era and matter domination era, quantum ﬁelds do not necessarily evolve from the vacuum. 5 Summary and outlook In fact, it is expected to have some initial particle population produced in earlier stages of the universe such as inﬂation or (p)reheating. Yet a non-vacuum initial condition is highly model-dependent. In this work, we choose vacuum initial condition because we focus more on a model-independent analysis of par- ticle production due to a later eﬀect of chemical potential. The treatment of other initial conditions is brieﬂy described in section 4.4. However, it is worthwhile to note some in- teresting behaviors. If the initial particles are thermally prepared, the interference term in \|β(z)\|2 is averaged out by taking the ensemble average over the phase of βi. Thus particle number generally increases due to chemical potential, as expected. In contrast, if the initial particles are coherently prepared with a common phase of βi, chemical potential can serve to produce or destroy particles, depending on the sign of the inter- ference term. If the particle number decreases, one can understand it as the “decay” of particles with energy injection into the background chemical potential sector. It would be interesting to investigate these possibilities with concrete models in the future. • In dS and its two types of deviations, it is natural to assume a vacuum initial condition. However, in radiation domination era and matter domination era, quantum ﬁelds do not necessarily evolve from the vacuum. In fact, it is expected to have some initial particle population produced in earlier stages of the universe such as inﬂation or (p)reheating. Yet a non-vacuum initial condition is highly model-dependent. In this work, we choose vacuum initial condition because we focus more on a model-independent analysis of par- ticle production due to a later eﬀect of chemical potential. The treatment of other initial conditions is brieﬂy described in section 4.4. However, it is worthwhile to note some in- teresting behaviors. If the initial particles are thermally prepared, the interference term in \|β(z)\|2 is averaged out by taking the ensemble average over the phase of βi. Thus particle number generally increases due to chemical potential, as expected. In contrast, if the initial particles are coherently prepared with a common phase of βi, chemical potential can serve to produce or destroy particles, depending on the sign of the inter- ference term. If the particle number decreases, one can understand it as the “decay” of particles with energy injection into the background chemical potential sector. 5 Summary and outlook It would be interesting to investigate these possibilities with concrete models in the future. • The knowledge of the production time and width can be helpful in the estimation of signal strength in cosmological collider physics. As mentioned before, the O(\|β\|) oscilla- tory signatures on the cosmological collider originate from the interference between the positive frequency part and the negative frequency part, whose presence is controlled by the Stokes multiplier S(z). This fact can be useful when estimating the loop diagrams. At loop level, the momentum integral receives contribution from the UV region with z ≫1. Usually, this UV divergent part can be regularized by a momentum cutoﬀat Hubble scale, i.e., k a < H. Then the signal strength follows from dimensional analy- sis. This is convenient if the mass of the particle running the loop is close to Hubble scale [23]. However, if the particle is much heavier, the dimensionless parameter µ > 1 can enter in complicated ways. Adding chemical potential introduces yet another dimen- sionless parameter ˜κ, thus invalidating the naive dimensional analysis [25]. However, with the knowledge of particle production history, the momentum cutoﬀcan be posed more precisely at z∗. This is because physically speaking, the particles that generate the signals do not get produced until their momentum drops below the production scale, k a < H(z∗± ∆z∗). This potentially oﬀers a better way to estimation signal strength, which deserves further explorations. • The knowledge of the production time and width can be helpful in the estimation of signal strength in cosmological collider physics. As mentioned before, the O(\|β\|) oscilla- tory signatures on the cosmological collider originate from the interference between the positive frequency part and the negative frequency part, whose presence is controlled by the Stokes multiplier S(z). This fact can be useful when estimating the loop diagrams. At loop level, the momentum integral receives contribution from the UV region with z ≫1. Usually, this UV divergent part can be regularized by a momentum cutoﬀat Hubble scale, i.e., k a < H. Then the signal strength follows from dimensional analy- sis. This is convenient if the mass of the particle running the loop is close to Hubble scale [23]. However, if the particle is much heavier, the dimensionless parameter µ > 1 can enter in complicated ways. Adding chemical potential introduces yet another dimen- sionless parameter ˜κ, thus invalidating the naive dimensional analysis [25]. 5 Summary and outlook However, with the knowledge of particle production history, the momentum cutoﬀcan be posed more precisely at z∗. This is because physically speaking, the particles that generate the signals do not get produced until their momentum drops below the production scale, k a < H(z∗± ∆z∗). This potentially oﬀers a better way to estimation signal strength, which deserves further explorations. – 57 – • Aside from chemical potential, there are many other sophisticated mechanisms of cos- mological particle production, to which the smoothed version of Stokes-line method can be applied. For example, parametric resonance is widely used in models of preheat- ing [71–73], generation of primordial black holes [74, 75] and primordial gravitational waves [76, 77]. For a periodic eﬀective frequency w(z), the turning points form periodic pairs on the complex z-plane, joined by periodic Stokes lines. Then the resonance con- dition can be viewed as the constructive interference of particle production amplitudes when crossing each Stokes line. In the literature, there are already preliminary attempts in this direction [50, 52], but using the traditional Stokes-line method without uniform smoothing. This will be accurate if the Stokes lines are well-separated so that one can apply the “dilute gas” approximation, treating each crossing separately as sudden jumps in particle number. However, if the production widths are as wide as the separation be- tween two neighboring Stokes lines, one may need to go to the ﬁne-grained picture and perform the analysis using the smoothed version of Stokes-line method. It is interesting to compare this method with traditional ones such as the Floquet theory. JHEP06(2021)129 Acknowledgments We would like to thank Kaifeng Zheng and Kun-Feng Lyu for helpful discussions. This work is supported in part by GRF Grants 16301917, 16304418 and 16303819 from the Research Grants Council of Hong Kong, and the NSFC Excellent Young Scientist (EYS) Scheme (Hong Kong and Macau) Grant No. 12022516. A A checklist of results In this appendix, we assemble our main results into a checklist in table 1. We have explicitly spell out the schematic form of various quantities and given the full expressions/plots as references jumping into the text. B The 1/4 puzzle As mentioned in section 4.1, the mismatch of the 1/4 term in the dS eﬀective mass is due to the non-vanishing correction to the IR frequency near z = 0+. These corrections are hidden in the asymptotic series. In order to take them into account, we must use W instead of W (0) = w and compute Dingle’s singulant F order-by-order. We ﬁrst deform the integration contour from C0 : Im F(z) = 0 to lie along the branch cut joining zc, 0 and z∗ c, which is deﬁned by arg w2(z) = ±π . (B.1) (B.1) This is illustrated by the path C1 ∪C2 in ﬁgure 27. All the super-adiabatic corrections δW (n) are proportional to odd powers of w and therefore possess the same branch cut. It is straightforward to check that 0, zc, z∗ c and the branch cut are the only singularities for δW (n) with n ⩾1. – 58 – Amount Time Width Plot Valid region dS =(4.24) ≃(4.25) ≃(4.26) e−2π(m−κ) 0.6m + 0.3κ + · · · 2.1 √m −0.4κ m3/2 + · · · Fig. 9 ϵ-dS =(4.36) ≈(4.38) ≈(4.40) e−2π(m−κ)[1+ϵ(ln k+··· )] 0.6m + 0.3κ + · · · 2.1 √m −0.4κ m3/2 + · · · Fig. 12 ϵ ≪1 + ϵ(0.8m + · · · ) + ϵ √m(−ln k + · · · ) η-dS =(4.54) ≈(4.56) ≈(4.58) e−2π(m−κ)[1+c(k)(··· )] (1 + c(k)) (0.6m + 0.3κ + · · · ) 1 −c(k) 2 2.1 √m −0.4κ m3/2 + · · · Fig. 16 c(k), ηi ≪1 +c(k)ηi (0.8m + · · · ) +c(k)ηi √m ln m Hi + · · · RD =(4.67) =(4.68) =(4.69) Ginzburg criterion: e−πk2 crm3 (m2−κ2) k2κ crm2 q πk2 crm Fig. 22 (4.70) MD =(4.77) ≃(4.78) ≃(4.79) Ginzburg criterion: e −1.0 q 2πk3 cmm(1−0.7κ m +···) q k3 cmm 0.6 + 0.5κ m + · · · k3 cmm 1/4 1.5 −0.3κ m + · · · Fig. 26 (4.80) JHEP06(2021)129 Table 1. A checklist of results for spin-1 vector bosons. Here RD and MD stand for radiation domination era and matter domination era, respectively. m is the mass while κ is the chemical potential. The =, ≃, ≈symbols are used to indicate whether the result is exact, numerically exact, or empirical with 2% error. B The 1/4 puzzle In the schematic expressions, to display the most salient features, we have omitted the Hubble parameter and blurred the diﬀerence between m H and µ = q m2 H2 −1 4 in dS. The widths in dS, ϵ-type dS and η-type dS are measured in e-folds, whereas the widths in RD and MD are measured in z-domain. The results for spin-1/2 fermion is obtained via a simple replacement rule m2 →m2+κ2. Setting κ = 0 also gives the purely gravitational production results. Table 1. A checklist of results for spin-1 vector bosons. Here RD and MD stand for radiation domination era and matter domination era, respectively. m is the mass while κ is the chemical potential. The =, ≃, ≈symbols are used to indicate whether the result is exact, numerically exact, or empirical with 2% error. In the schematic expressions, to display the most salient features, we have omitted the Hubble parameter and blurred the diﬀerence between m H and µ = q m2 H2 −1 4 in dS. The widths in dS, ϵ-type dS and η-type dS are measured in e-folds, whereas the widths in RD and MD are measured in z-domain. The results for spin-1/2 fermion is obtained via a simple replacement rule m2 →m2+κ2. Setting κ = 0 also gives the purely gravitational production results. Now consider a closed integration path C1 ∪C2 ∪C3 ∪C4 that tours around the branch cut. The integral can be converted to a residue at z →∞, I C1,2,3,4 W(z)dz = I C′ 1,2,3,4 W(1/z′)dz′ z′2 = −2πiRes z′→0 W(1/z′) z′2 . (B.2) (B.2) Separating the full frequency into the zeroth order and higher orders, we have Res z′→0 W(1/z′) z′2 = Res z′→0 " w(1/z′) z′2 + ∞ X n=1 δW (n)(1/z′) z′2 # = −˜κ , (B.3) (B.3) where we have used the fact that δW (n)(z) drops as where we have used the fact that δW (n)(z) drops as δW (n)(z) z→∞ −−−→O 1 z2n+1 (B.4) (B.4) at inﬁnity and therefore does not contribute to the residue. Then we separate the closed contour into two parts, Then we separate the closed contour into two parts, 2πi˜κ = I C1,2,3,4 W(z)dz = Z C1∪C3 W(z)dz + Z C2∪C4 W(z)dz . (B.5) (B.5) – 59 – – 59 – Figure 27. The original integration contour C0 and its deformations. The black dashed line is the branch cut. B The 1/4 puzzle The angle θ at which C1 and C3 meet is chosen carefully according to a principal value prescription. JHEP06(2021)129 Figure 27. The original integration contour C0 and its deformations. The black dashed line is the branch cut. The angle θ at which C1 and C3 meet is chosen carefully according to a principal value prescription. Across the branch cut, the phase of W(z) jumps by π and its modulus remains continuous. Therefore, along the branch cut, the integral on C1 and C3 gives the same result. The arcs around zc and z∗ c on C1 and C3, however, must be taken with a grain of salt. As these two singularities are of high orders, the integral there is ill-deﬁned when the radius of the arc goes to zero. As a result, we need to manually impose a principal value prescription. By symmetry, the function H(θ) ≡i Z C1(θ) W(z)dz (B.6) (B.6) is always real. Hence when we adjust the meeting points of C1 and C3 by changing θ (see ﬁgure 27), the integral varies from H(0+) to H(2π−). Then by the intermediate value theorem of a continuous real function, there always exists an angle 0 < θm < 2π such that H(θm) = H(0+) + H(2π−) 2 . (B.7) (B.7) If we choose θm to be the angle of the meeting point of C1 and C3, then the integrals along them give the same result: If we choose θm to be the angle of the meeting point of C1 and C3, then the integrals along them give the same result: Z C1(θm) W(z)dz = Z C3(θm) W(z)dz . (B.8) (B.8) – 60 – – 60 – In addition, due to the sign ﬂip across the branch cut, the integral along C2 and C4 are related by In addition, due to the sign ﬂip across the branch cut, the integral along C2 and C4 are related by Z C2 W(z)dz = − Z C4 W(z)dz = −πiRes z→0+W(z) = −iπµ . (B.9) (B.9) Thus (B.5) simpliﬁes into Thus (B.5) simpliﬁes into Z C1(θm) W(z)dz = iπ˜κ . (B.10) (B.10) Finally, we obtain the original integral as Z C0 W(z)dz = Z C1(θm)∪C2 W(z)dz = −iπ(µ −κ) , (B.11) ntially gives the resummed production amount (4.23) in dS. Z C0 W(z)dz = Z C1(θm)∪C2 W(z)dz = −iπ(µ −κ) , (B.11) (B.11) JHEP06(2021)129 which essentially gives the resummed production amount (4.23) in dS. B The 1/4 puzzle Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. References [1] A.H. Guth, The Inﬂationary Universe: A Possible Solution to the Horizon and Flatness Problems, Phys. Rev. D 23 (1981) 347 [INSPIRE]. [1] A.H. Guth, The Inﬂationary Universe: A Possible Solution to the Horizon and Flatness Problems, Phys. Rev. D 23 (1981) 347 [INSPIRE]. [2] A.D. Linde, A New Inﬂationary Universe Scenario: A Possible Solution of the Horizon, Flatness, Homogeneity, Isotropy and Primordial Monopole Problems, Phys. Lett. B 108 (1982) 389 [INSPIRE]. [3] A. Albrecht and P.J. Steinhardt, Cosmology for Grand Uniﬁed Theories with Radiatively Induced Symmetry Breaking, Phys. Rev. Lett. 48 (1982) 1220 [INSPIRE]. [4] A.A. Starobinsky, A New Type of Isotropic Cosmological Models Without Singularity, Phys. Lett. B 91 (1980) 99 [INSPIRE]. [5] X. Chen and Y. Wang, Large non-Gaussianities with Intermediate Shapes from Quasi-Single Field Inﬂation, Phys. Rev. D 81 (2010) 063511 [arXiv:0909.0496] [INSPIRE]. [6] X. Chen and Y. Wang, Quasi-Single Field Inﬂation and Non-Gaussianities, JCAP 04 (2010) 027 [arXiv:0911.3380] [INSPIRE]. [7] D. Baumann and D. Green, Signatures of Supersymmetry from the Early Universe, Phys. Rev. D 85 (2012) 103520 [arXiv:1109.0292] [INSPIRE]. [8] T. Noumi, M. Yamaguchi and D. Yokoyama, Eﬀective ﬁeld theory approach to quasi-single ﬁeld inﬂation and eﬀects of heavy ﬁelds, JHEP 06 (2013) 051 [arXiv:1211.1624] [INSPIRE]. [9] N. Arkani-Hamed and J. Maldacena, Cosmological Collider Physics, arXiv:1503.08043 [INSPIRE]. [10] H. Lee, D. Baumann and G.L. Pimentel, Non-Gaussianity as a Particle Detector, JHEP 12 (2016) 040 [arXiv:1607.03735] [INSPIRE]. [11] E. Silverstein and A. Westphal, Monodromy in the CMB: Gravity Waves and String Inﬂation, Phys. Rev. D 78 (2008) 106003 [arXiv:0803.3085] [INSPIRE]. – 61 – [12] L. McAllister, E. Silverstein and A. Westphal, Gravity Waves and Linear Inﬂation from Axion Monodromy, Phys. Rev. D 82 (2010) 046003 [arXiv:0808.0706] [INSPIRE]. [13] R. Flauger, M. Mirbabayi, L. Senatore and E. Silverstein, Productive Interactions: heavy particles and non-Gaussianity, JCAP 10 (2017) 058 [arXiv:1606.00513] [INSPIRE]. [14] A. Berera and L.-Z. Fang, Thermally induced density perturbations in the inﬂation era, Phys. Rev. Lett. 74 (1995) 1912 [astro-ph/9501024] [INSPIRE]. [15] A. Berera, Warm inﬂation, Phys. Rev. Lett. 75 (1995) 3218 [astro-ph/9509049] [INSPIRE] [16] X. Tong, Y. Wang and S. Zhou, Unsuppressed primordial standard clocks in warm quasi-single ﬁeld inﬂation, JCAP 06 (2018) 013 [arXiv:1801.05688] [INSPIRE]. JHEP06(2021)129 [17] M.S. Turner and L.M. Widrow, Inﬂation Produced, Large Scale Magnetic Fields, Phys. Rev. D 37 (1988) 2743 [INSPIRE]. [18] W.D. Garretson, G.B. Field and S.M. Carroll, Primordial magnetic ﬁelds from pseudoGoldstone bosons, Phys. Rev. References D 46 (1992) 5346 [hep-ph/9209238] [INSPIRE]. [19] N. Barnaby and M. Peloso, Large NonGaussianity in Axion Inﬂation, Phys. Rev. Lett. 106 (2011) 181301 [arXiv:1011.1500] [INSPIRE]. [20] T. Liu, X. Tong, Y. Wang and Z.-Z. Xianyu, Probing P and CP-violations on the Cosmological Collider, JHEP 04 (2020) 189 [arXiv:1909.01819] [INSPIRE]. [21] L.-T. Wang and Z.-Z. Xianyu, Gauge Boson Signals at the Cosmological Collider, JHEP 11 (2020) 082 [arXiv:2004.02887] [INSPIRE]. [22] P. Adshead and E.I. Sfakianakis, Fermion production during and after axion inﬂation, JCAP 11 (2015) 021 [arXiv:1508.00891] [INSPIRE]. [23] X. Chen, Y. Wang and Z.-Z. Xianyu, Neutrino Signatures in Primordial Non-Gaussianities, JHEP 09 (2018) 022 [arXiv:1805.02656] [INSPIRE]. [24] P. Adshead, L. Pearce, M. Peloso, M.A. Roberts and L. Sorbo, Phenomenology of fermion production during axion inﬂation, JCAP 06 (2018) 020 [arXiv:1803.04501] [INSPIRE]. [25] A. Hook, J. Huang and D. Racco, Searches for other vacua. Part II. A new Higgstory at the cosmological collider, JHEP 01 (2020) 105 [arXiv:1907.10624] [INSPIRE]. [26] T. Kobayashi and N. Afshordi, Schwinger Eﬀect in 4D de Sitter Space and Constraints on Magnetogenesis in the Early Universe, JHEP 10 (2014) 166 [arXiv:1408.4141] [INSPIRE]. [27] J.-J. Geng, B.-F. Li, J. Soda, A. Wang, Q. Wu and T. Zhu, Schwinger Pair Production by Electric Field Coupled to Inﬂaton, JCAP 02 (2018) 018 [arXiv:1706.02833] [INSPIRE]. [28] W.Z. Chua, Q. Ding, Y. Wang and S. Zhou, Imprints of Schwinger Eﬀect on Primordial Spectra, JHEP 04 (2019) 066 [arXiv:1810.09815] [INSPIRE]. [29] R.T. Co, A. Pierce, Z. Zhang and Y. Zhao, Dark Photon Dark Matter Produced by Axion Oscillations, Phys. Rev. D 99 (2019) 075002 [arXiv:1810.07196] [INSPIRE]. [30] C.S. Machado, W. Ratzinger, P. Schwaller and B.A. Stefanek, Audible Axions, JHEP 01 (2019) 053 [arXiv:1811.01950] [INSPIRE]. [31] C.S. Machado, W. Ratzinger, P. Schwaller and B.A. Stefanek, Gravitational wave probes of axionlike particles, Phys. Rev. D 102 (2020) 075033 [arXiv:1912.01007] [INSPIRE]. – 62 – [32] B. Salehian, M.A. Gorji, S. Mukohyama and H. Firouzjahi, Analytic study of dark photon and gravitational wave production from axion, JHEP 05 (2021) 043 [arXiv:2007.08148] [INSPIRE]. [33] P. Adshead and E.I. Sfakianakis, Leptogenesis from left-handed neutrino production during axion inﬂation, Phys. Rev. Lett. 116 (2016) 091301 [arXiv:1508.00881] [INSPIRE]. [34] M.V. Barry, Uniform Asymptotic Smoothing of Stokes’s Discontinuities, Proc. Roy. Soc. Lond. A 422 (1989) 7 [INSPIRE]. [35] M.V. Berry, Waves near stokes lines, Proc. Roy. Soc. Lond. A 427 (1990) 265. [35] M.V. Berry, Waves near stokes lines, Proc. Roy. Soc. References Lond. A 427 (1990) 265. [36] M.V. Berry, Histories of adiabatic quantum transitions, Proc. Roy. Soc. Lond. A 429 (1990) 61 [36] M.V. Berry, Histories of adiabatic quantum transitions, Proc. Roy. Soc. Lond. A 429 (1990 61. JHEP06(2021)129 [37] M.V. Berry and R. Lim, Universal transition prefactors derived by superadiabatic renormalization, J. Phys. A 26 (1993) 4737. [38] C.K. Dumlu and G.V. Dunne, The Stokes Phenomenon and Schwinger Vacuum Pair Production in Time-Dependent Laser Pulses, Phys. Rev. Lett. 104 (2010) 250402 [arXiv:1004.2509] [INSPIRE]. [39] C.K. Dumlu, Schwinger Vacuum Pair Production in Chirped Laser Pulses, Phys. Rev. D 82 (2010) 045007 [arXiv:1006.3882] [INSPIRE]. [40] R. Dabrowski and G.V. Dunne, Time dependence of adiabatic particle number, Phys. Rev. D 94 (2016) 065005 [arXiv:1606.00902] [INSPIRE]. [41] S.P. Kim, Schwinger Pair Production via Polons and the Origin of Stokes Phenomena, New Phys. Sae Mulli 68 (2018) 1225. [42] R.-G. Cai, S.P. Kim and W. Kim, Spin Eﬀect and Stokes Phenomenon for Fermion Production in Electric Fields, New Phys. Sae Mulli 69 (2019) 1235. [43] H. Taya, T. Fujimori, T. Misumi, M. Nitta and N. Sakai, Exact WKB analysis of the vacuum pair production by time-dependent electric ﬁelds, JHEP 03 (2021) 082 [arXiv:2010.16080] [INSPIRE]. [44] C.K. Dumlu, Stokes phenomenon and Hawking radiation, Phys. Rev. D 102 (2020) 125006 [arXiv:2009.09851] [INSPIRE]. [45] S.P. Kim, The Stokes Phenomenon and Quantum Tunneling for de Sitter Radiation in Nonstationary Coordinates, JHEP 09 (2010) 054 [arXiv:1006.4004] [INSPIRE]. [46] S.P. Kim, Geometric Origin of Stokes Phenomenon for de Sitter Radiation, Phys. Rev. D 88 (2013) 044027 [arXiv:1307.0590] [INSPIRE]. [47] R. Dabrowski and G.V. Dunne, Superadiabatic particle number in Schwinger and de Sitter particle production, Phys. Rev. D 90 (2014) 025021 [arXiv:1405.0302] [INSPIRE]. [48] L. Li, T. Nakama, C.M. Sou, Y. Wang and S. Zhou, Gravitational Production of Superheavy Dark Matter and Associated Cosmological Signatures, JHEP 07 (2019) 067 [arXiv:1903.08842] [INSPIRE]. [49] S. Enomoto and T. Matsuda, The exact WKB for cosmological particle production, JHEP 03 (2021) 090 [arXiv:2010.14835] [INSPIRE]. [50] S. Enomoto and T. Matsuda, The Exact WKB and the Landau-Zener transition for asymmetry in cosmological particle production, arXiv:2104.02312 [INSPIRE]. – 63 – [51] S. Hashiba, Y. Yamada and J. Yokoyama, Particle production induced by vacuum decay in real time dynamics, Phys. Rev. D 103 (2021) 045006 [arXiv:2006.10986] [INSPIRE]. [52] S. Hashiba and Y. References Yamada, Stokes phenomenon and gravitational particle production — How to evaluate it in practice, JCAP 05 (2021) 022 [arXiv:2101.07634] [INSPIRE]. [53] L.-T. Wang and Z.-Z. Xianyu, In Search of Large Signals at the Cosmological Collider, JHEP 02 (2020) 044 [arXiv:1910.12876] [INSPIRE]. [54] A. Bodas, S. Kumar and R. Sundrum, The Scalar Chemical Potential in Cosmological Collider Physics, JHEP 02 (2021) 079 [arXiv:2010.04727] [INSPIRE]. [55] L.D. Landau, Zur Theorie der Energieubertragung II, Z. Sowjetunion 2 (1932) 46 [https://ci.nii.ac.jp/naid/10011873546/en/]. JHEP06(2021)129 JHEP06(2021)129 [56] C. Zener, Nonadiabatic crossing of energy levels, Proc. Roy. Soc. Lond. A 137 (1932) 696 [INSPIRE]. [57] F. Wilczek, Two Applications of Axion Electrodynamics, Phys. Rev. Lett. 58 (1987) 1799 [INSPIRE]. [58] S. Winitzki, Cosmological particle production and the precision of the WKB approximation, Phys. Rev. D 72 (2005) 104011 [gr-qc/0510001] [INSPIRE]. [59] R. Dingle, Asymptotic Expansions: Their Derivation and Interpretation, Academic Press (1973). [60] H.K. Dreiner, H.E. Haber and S.P. Martin, Two-component spinor techniques and Feynman rules for quantum ﬁeld theory and supersymmetry, Phys. Rept. 494 (2010) 1 [arXiv:0812.1594] [INSPIRE]. [61] A. Landete, J. Navarro-Salas and F. Torrenti, Adiabatic regularization for spin-1/2 ﬁelds, Phys. Rev. D 88 (2013) 061501 [arXiv:1305.7374] [INSPIRE]. [62] L.H. Thomas, On unitary representations of the group of de sitter space, Annals Math. 42 (1941) 113. [63] T.D. Newton, A note on the representations of the de sitter group, Annals Math. 51 (1950) 730. [64] Planck collaboration, Planck 2018 results. X. Constraints on inﬂation, Astron. Astrophys. 641 (2020) A10 [arXiv:1807.06211] [INSPIRE]. [65] J. Khoury, B.A. Ovrut, P.J. Steinhardt and N. Turok, The Ekpyrotic universe: Colliding branes and the origin of the hot big bang, Phys. Rev. D 64 (2001) 123522 [hep-th/0103239] [INSPIRE]. [66] E.I. Buchbinder, J. Khoury and B.A. Ovrut, New Ekpyrotic cosmology, Phys. Rev. D 76 (2007) 123503 [hep-th/0702154] [INSPIRE]. [67] D. Wands, Duality invariance of cosmological perturbation spectra, Phys. Rev. D 60 (1999) 023507 [gr-qc/9809062] [INSPIRE]. [68] F. Finelli and R. Brandenberger, On the generation of a scale invariant spectrum of adiabatic ﬂuctuations in cosmological models with a contracting phase, Phys. Rev. D 65 (2002) 103522 [hep-th/0112249] [INSPIRE]. [69] R.H. Brandenberger and C. Vafa, Superstrings in the Early Universe, Nucl. Phys. B 316 (1989) 391 [INSPIRE]. – 64 – [70] A. Nayeri, R.H. Brandenberger and C. Vafa, Producing a scale-invariant spectrum of perturbations in a Hagedorn phase of string cosmology, Phys. Rev. Lett. 97 (2006) 021302 [hep-th/0511140] [INSPIRE]. [71] J.H. Traschen and R.H. References Brandenberger, Particle Production During Out-of-equilibrium Phase Transitions, Phys. Rev. D 42 (1990) 2491 [INSPIRE]. [72] A.D. Dolgov and D.P. Kirilova, On particle creation by a time dependent scalar ﬁeld, Sov. J. Nucl. Phys. 51 (1990) 172 [INSPIRE]. [73] L. Kofman, A.D. Linde and A.A. Starobinsky, Towards the theory of reheating after inﬂation, Phys. Rev. D 56 (1997) 3258 [hep-ph/9704452] [INSPIRE]. JHEP06(2021)129 [74] Y.-F. Cai, X. Tong, D.-G. Wang and S.-F. Yan, Primordial Black Holes from Sound Speed Resonance during Inﬂation, Phys. Rev. Lett. 121 (2018) 081306 [arXiv:1805.03639] [INSPIRE]. [75] Z. Zhou, J. Jiang, Y.-F. Cai, M. Sasaki and S. Pi, Primordial black holes and gravitational waves from resonant ampliﬁcation during inﬂation, Phys. Rev. D 102 (2020) 103527 [arXiv:2010.03537] [INSPIRE]. [76] C. Lin and M. Sasaki, Resonant Primordial Gravitational Waves Ampliﬁcation, Phys. Lett. B 752 (2016) 84 [arXiv:1504.01373] [INSPIRE]. [77] Y.-F. Cai, C. Lin, B. Wang and S.-F. Yan, Sound speed resonance of the stochastic gravitational wave background, Phys. Rev. Lett. 126 (2021) 071303 [arXiv:2009.09833] [INSPIRE]. – 65 –
https://openalex.org/W1520762665	https://thejournalofheadacheandpain.biomedcentral.com/counter/pdf/10.1186/1129-2377-14-S1-P137	English	null	Chronic migraine is not associated with cerebellar infarct-like lesions	The Journal of headache and pain	2,013	cc-by	637	Introduction Following the same MRI methodology of the previous studies, we demonstrate that, at least for migrainous women, there is no relationship between migraine frequency and the presence of cerebellar infarct-like lesions. Therefore, at least for the cerebellum, frequency of migraine attacks itself does not seem to be a factor increasing the risk of vascular brain lesions. These findings call for caution when extrapolating findings from the gen- eral population into current clinical practice. Two general population studies [1,2] have found that migraine in general, and especially women with aura, has an increased risk of cerebellar infarct-like lesions by MRI. In addition, there was a trend for higher risk of cerebellar lesions in those migraine subjects with a higher migraine attack frequency, which would have obvious clinical and management implications. Objective d Supported by the PI11/00889 FISSS grant (ISCIII) To determine whether chronic migraine patients are at increased risk of cerebellar infarct-like lesions on MRI. Published: 21 February 2013 Published: 21 February 2013 Published: 21 February 2013 Methods After signed informed consent, brain MRIs were obtained in 50 women from our headache clinic meeting chronic migraine according to 2006 IHC-II revised criteria. Six had a history of migraine with aura attacks and 19 meet overuse criteria. Their ages ranged from 16 to 63 years (mean 40.9 years) and the length of chronic migraine range from 6 months to 27 years (mean 7.5 years). At least 11 patients had a minimum of one vascular risk factor and the prevalence of right to left shunt with transcranial echo was 58%. Brain MRIs were acquired on a 1.5T unit Signa LX 9.1 (General Electric Systems, USA). Protocol includes whole brain weighted images in saggital T1 (5 mm slices), axial FLAIR T2 (3 mm) and combined proton density and T2 fast spin echo (3 mm). Two independent neuroradiolo- gists carefully analysed all the cerebellar images. References doi:10.1186/1129-2377-14-S1-P137 Cite this article as: Meilan et al.: Chronic migraine is not associated with cerebellar infarct-like lesions. The Journal of Headache and Pain 2013 14(Suppl 1):P137. doi:10.1186/1129-2377-14-S1-P137 Cite this article as: Meilan et al.: Chronic migraine is not associated with cerebellar infarct-like lesions. The Journal of Headache and Pain 2013 14(Suppl 1):P137. Submit your manuscript to a journal and beneﬁ t from: 7 Convenient online submission 7 Rigorous peer review 7 Immediate publication on acceptance 7 Open access: articles freely available online 7 High visibility within the ﬁ eld 7 Retaining the copyright to your article Submit your next manuscript at 7 springeropen.com Meilan et al. The Journal of Headache and Pain 2013, 14(Suppl 1):P137 http://www.thejournalofheadacheandpain.com/content/14/S1/P137 Meilan et al. The Journal of Headache and Pain 2013, 14(Suppl 1):P137 http://www.thejournalofheadacheandpain.com/content/14/S1/P137 Open Access Open Access Chronic migraine is not associated with cerebellar infarct-like lesions A Meilan, E Santamarta, A Saiz, D Larrosa, E Cernuda, J Pascual* From The European Headache and Migraine Trust International Congress London, UK. 20-23 September 2012 © 2013 Meilan et al; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Results After an in depth review of all posterior fossa slices, we were unable to find even one cerebellar infarct-like lesion in any of these chronic migraine patients. University Hospital Central de Asturias, Spain © 2013 Meilan et al; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Meilan et al; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
https://openalex.org/W3030000693	https://link.springer.com/content/pdf/10.1007/s11661-020-05847-5.pdf	English	null	Using SXRR to Probe the Nature of Discontinuities in SLM Additive Manufactured Inconel 718 Specimens	Metallurgical and materials transactions. A, Physical metallurgy and materials science	2,020	cc-by	8,190	RENE´ LAQUAI, BERND R. MU¨ LLER , JUDITH A. SCHNEIDER, ANDREAS KUPSCH, and GIOVANNI BRUNO The utilization of additive manufacturing (AM) to fabricate robust structural components relies on understanding the nature of internal anomalies or discontinuities, which can compromise the structural integrity. While some discontinuities in AM microstructures stem from similar mechanisms as observed in more traditional processes such as casting, others are unique to the AM process. Discontinuities in AM are challenging to detect, due to their submicron size and orientation dependency. Toward the goal of improving structural integrity, minimizing discontinuities in an AM build requires an understanding of the mechanisms of formation to mitigate their occurrence. This study utilizes various techniques to evaluate the shape, size, nature and distribution of discontinuities in AM Inconel 718, in a non-hot isostatic pressed (HIPed) as-built, non-HIPed and direct age, and HIPed with two step age samples. Non-destructive synchrotron radiation refraction and transmission radiography (SXRR) provides additional information beyond that obtained with destructive optical microscopy. SXRR was able to distinguish between voids, cracks and lack of melt in, due to its sensitivity to the orientation of the discontinuity. https://doi.org/10.1007/s11661-020-05847-5 The Author(s) 2020 Using SXRR to Probe the Nature of Discontinuities in SLM Additive Manufactured Inconel 718 Specimens RENE´ LAQUAI, BERND R. MU¨ LLER , JUDITH A. SCHNEIDER, ANDREAS KUPSCH, and GIOVANNI BRUNO https://doi.org/10.1007/s11661-020-05847-5 The Author(s) 2020 I. INTRODUCTION That means that the smallest detectable object size is within the nanometer range.[9] This is not to be confused with the spatial resolution or the size of objects that can be imaged. With the typical set-up at a synchrotron source, the spatial resolution is limited by the pixel size of the detector system. It must be also emphasized that because of the inevitable background noise, it is impossible to conclusively detect one single discontinu- ity. Therefore, a certain population of objects is neces- sary to yield an integrated signal above the background noise. XRR and SXRR yield a 2.5D signal that is integrated over the sample thickness. This results in the detection and imaging of populations rather than single discontinuities. While SXCT would yield truly 3D information, it would not be able to capture disconti- nuities below its spatial resolution; in this sense SXRR yields information (again, not images) on nano-sized objects but at a meso to macro length scale. This length scale matches that of many analytical models, so that the microstructural changes detected by SXRR have been quantitatively correlated with micromechanical models.[20] Furthermore, X-ray techniques combining refraction and transmission are sensitive to the orienta- tion of the discontinuity, thereby allowing the source of the discontinuity to be identiﬁed.[17,21] For example, a spherical discontinuity would yield a refraction signal in any orientation of the sample. In contrast, cracks or elongated voids would yield a signal that strongly decreases when their surface normal is not parallel to the scattering vector of the instrumental set-up as identiﬁed by the bisectrix of the angle between the incident and refracted beam directions. Finally, it must be remarked that the X-ray refraction signal is sensitive to all kind of discontinuities (cracks, pores, grain boundaries) in a confounded manner. In practice, however, the largest contrast (diﬀerence in refraction index) is given by cavities and cracks, which are the most important discontinuities inﬂuencing the mechanical properties of materials. A complementary route to detect microstructural features and discontinuities is represented by X-ray refraction techniques. In particular, using a synchrotron source, refraction radiography (SXRR)[11–13] can be implemented. This technique has been successfully shown[12] to detect submicron inhomogeneities such as voids, cracks, and entrapped particles in bulk light materials such as ceramics and epoxy composites.[14] Its main advantage over SXCT is the detectability of much ﬁner discontinuities (even nano-sized) over much larger ﬁelds of view (centimeter size). I. INTRODUCTION non-optimized process parameters[3] or irregularities in laser power. VARIOUS discontinuities can result from fabrication methods in which molten metal is solidiﬁed. During the AM process, voids can appear as they do in the case of cast components or fusion welded joints due to trapped gasses or gas formation during the process. Other discontinuities can arise from trapped particles or oxides. In AM processing, discontinuities such as porosity, voids, and inclusions occur but at a smaller, submicron scale than in traditional processes.[1–3] In addition, other discontinuities unique to AM are found, including layer defects, lack of melt in, and entrapment of unconsolidated powders often resulting result from p Discontinuities can be characterized in a spatially resolved manner using classic two-dimensional (2D) image analysis techniques (e.g., based on optical or electron microscopy). These techniques can be expanded to three dimensions (3D) using destructive serial sec- tioning, often with focused ion beam (FIB) preparation. However, these destructive techniques are time-consum- ing and suﬀer from a limited ﬁeld-of-view. To obtain (nearly) non-destructive 3D imaging, computed tomog- raphy (CT) has been successfully used.[4–7] Laboratory X-ray or Neutron CT can detect and image discontinu- ities down to 5 to 10 lm in samples with dimensions about 5 to 20 mm.[4–6] Recently, advanced parallel beam synchrotron computed X-ray tomography (SXCT) has been used.[8–10] However, its spatial resolution, of approximately 1 lm3 or slightly below, can only reliably identify volumetric discontinuities, and not other dis- continuities such as cracks or inclusions. SXCT also requires miniature samples[10] (cylinders with 1 to 2 mm diameter) to achieve the highest spatial resolution. In spite of the many advantages of SXCT, identifying micron size discontinuities in meso- or macro-sized samples becomes impracticable. RENE´ LAQUAI, BERND R. MU¨ LLER, and ANDREAS KUPSCH are with the Bundesanstalt fu¨ r Materialforschung und - pru¨ fung (BAM), Unter den Eichen 87, 12205 Berlin, Germany. Contact e-mail: bernd.mueller@bam.de JUDITH A. SCHNEIDER is with the Department of Mechanical and Aerospace Engineering, University of Alabama in Huntsville, Huntsville, AL 35899. GIOVANNI BRUNO is with Bundesanstalt fu¨ r Materialforschung und -pru¨ fung (BAM) and also with the University of Potsdam, Institute of Physics and Astronomy, Karl-Liebknecht-Str. 24-25, 14476 Potsdam, Germany. Contact e-mail: Giovanni.bruno@bam.de Manuscript submitted December 14 2019 METALLURGICAL AND MATERIALS TRANSACTIONS A 4146—VOLUME 51A, AUGUST 2020 opening) of a few X-ray wavelengths in order for the wave to ‘‘notice’’ a density diﬀerence at the interface. I. INTRODUCTION Recently, the use of SXRR has been extended to metal matrix composites,[15] metallic alloys[16] and to determine the void shape and size in AM Titanium 6V/4Al.[17] Previously,[10] SXCT was compared with SXRR highlighting their comple- mentary nature with regards to the detectability of submicrometer objects in macroscopic samples. It must be noted that a direct comparison is not possible, since the SXRR signal mostly comes from objects that are not detected in the CT scans. In this study, the application of SXRR is used to characterize AM samples of Inconel 718, which has a higher density than other alloys investigated to date. In this study, SXRR was able to distinguish between discontinuities that resulted from lack of melt during the AM process, which were not sealed in subsequent HIP processing, from those of voids whose shape changed during HIP processing. II. FUNDAMENTALS OF X-RAY REFRACTION X-ray refraction techniques were introduced a couple of decades ago,[12] and have been successfully used for both materials characterization and non-destructive testing.[18] X-ray refraction techniques such as X-ray refraction radiography (XRR) are used to obtain the amount of the relative internal speciﬁc surface, i.e., surface per unit volume, relative to a reference state of a sample. They are therefore a proven non-destructive characterization method to investigate and identify discontinuities such as inhomogeneities, inclusions, cracks and voids. Obviously, for practical reasons specimens are always reduced in size to ﬁt the equipment, but the potential for fully non-destructive use of XRR has been repeatedly shown in the literature.[11,14,19] III. EXPERIMENTAL TECHNIQUES Additional imaging using a Keyence VK-X was correlated with the size and shape analysis obtained with the digital imaging process. The starting powder density was measured using He-Pycnometry. g g y Round cylinders nominally 2 mm diameter for XCT and ﬂat slices nominally 9.5 9 5.5 mm for SXRR studies were removed from samples #508 and #523 using wire EDM. To minimize the surface roughness, the ﬂat samples were ground using P4000 SiC paper to a thickness of 0.160 to 0.180 mm for sample #508 and 0.145 to 0.160 mm for sample #523. Round sub-sized samples were used for tensile testing. A total of 6 samples were machined for each heat treatment outlined in Table I. Specimen 508 was given a full heat treatment common to post processing of Inconel 718.[22] Due to the expected residual stresses, a stress relief and quench (SR/Q) is conducted prior to removal of the AM part from the build plate to minimize distortion. Hot isostatic pressing (HIP) is used to minimize volumetric voids. Due to the typical slow cooling in a HIP furnace, this is followed by a homogenization and quench (homo/Q) step to drive solidiﬁcation phases back into solution with the matrix. A solutionizing and quench (ST/Q) is applied to precipitate some d phase out along grain boundaries to minimize grain growth. After redistribution of the d phase, an aging (Age) process is used to precipitate out the strengthening phases of c¢ and c¢¢. In this study, a comparison was made between direct aging[23,24] and the recommended two step aging per the SAE/ASM stan- dard for wrought Inconel 718.[22] SXRR measurements were carried out at the BAM synchrotron laboratory BAMline at Helmholtz-Zen- trum Berlin, Germany.[25,26] The two prepared samples were mounted in a slide frame as shown in Figure 2. A highly collimated monochromatic X-ray beam with an energy bandwidth of 0.2 pct was used.[26,27] The beam energy was set to 30 keV to achieve a sample X-ray transmission of about 25 pct (on the 0.2 mm thick plates). A pco.1600-Camera (1600 9 1200 pixel) in combination with a lens system and a 50 lm thick CWO scintillator screen provided a pixel size of 4 lm 9 4 lm. III. EXPERIMENTAL TECHNIQUES The incident beam was narrowed to the ﬁeld-of-view of about 6.3 mm 9 4.7 mm[27] by a slit system to avoid detector backlighting.[28] In contrast to transmission-based radiographic mea- surements, a Si (111) analyzer crystal was placed in the beam path between the sample and the camera system to perform refraction radiographs as shown in Figure 2. The analyzer crystal reﬂects the beam transmitted by the sample into the detector system, if the incidence angle is set to the Bragg angle (hB = 3.779 deg at 30 keV). By tilting the analyzer crystal around an axis perpendicular to the scattering plane, a rocking curve is recorded. This describes the scattered beam intensity as a function of the deviation from the Bragg angle, Dh = h–hB. The rocking curve was recorded for each sample by taking 41 radiographs between h = 3.775 deg and h = 3.783 deg with a step size of Dh = 0.0002 deg and exposure time of 5 seconds. Both samples were measured in two orthog- onal orientations, with their build direction (Z) parallel and perpendicular to the scattering vector. In addition, the following images were acquired: dark ﬁeld (beam oﬀ) and ﬂat ﬁeld (beam on, but without sample). The dark ﬁeld image is used to subtract the dark current and detector readout noise from the sample and ﬂat ﬁeld acquisitions. The ﬂat ﬁeld images quantify the instru- mental artifacts and noise and are used to correct the X-ray refraction radiography images of the sample. g The samples had a nominal 6.4 mm diameter and a 76 mm nominal length and the tensile axis aligned with the build direction (Z). All tensile tests were conducted on a screw-driven mechanical actuator, which used a linear variable diﬀerential transformer for displacement feed- back. Stress measurements were based on loads obtained from an 89 kN load cell and sample dimension measurements. Strain measurements were obtained from a 25 mm extensometer calibrated to 50 pct strain. All tests were run in displacement control at a constant crosshead velocity of 1.3 mm/min. After testing, sections were removed from the grip ends of the tensile samples oriented as shown in Figure 1(a). The optical microscopy (OM) samples were mounted in two orientations, parallel (Z) and perpen- dicular to the build direction (X–Y) and prepared using standard metallography practices. All samples were imaged using a Zeiss XioVert.A1m Inverted Microscope for Reﬂected Light Techniques. III. EXPERIMENTAL TECHNIQUES Micromelt Inconel 718 powders with a mesh size of -325M, or 44 lm, were used to fabricate SLM AM samples, using a Concept Laser M2 Cusing with a 1064 nm wavelength neodymium-doped yttrium aluminum garnet (Nd-YAG) laser. Based on prior optimization studies, which considered density and surface ﬁnish, the build parameters utilized in this study were: 180 Watt laser power, 600 mm/s laser speed, 0.105 mm hatch spacing, and 0.035 mm layer thickness. Each layer was formed by the laser locally melting the powder using a pattern of alternately scanned x and y (i.e., in-plane perpendicular) directions. Cylindrical samples were printed in an 8 9 8 grid with a nominal 15 mm diameter, and 15 mm spacing, to a height of 200 mm. Each 200-mm-long sample was cut into half to generate 128 cylinders, which underwent varying heat treatments prior to machining into sub-sized tensile samples. Each X-ray refraction occurs whenever X-rays interact with interfaces between materials of diﬀerent densities, as in the case of cracks, voids, or particles in a matrix. This is analogous to the behavior of visible light in transparent materials, such as lenses or prisms. Diﬀerences in the refraction indices between the two interfacing materials determine the refraction angle at the interface, which is also dependent on the wavelength of the radiation. Since the diﬀerence between the refraction indices of two neighboring materials for X-ray radiation is typically 10-5, X-ray optical eﬀects can only be observed at very small scattering angles, which lie between several sec- onds and a few arc minutes. Since typical X-ray wavelengths are approximately 0.1 nm, voids and cracks can be detected as soon as they exceed a size (or METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4147 heat treatment lot was randomly chosen from the print grid to avoid biasing. magniﬁcation to obtain 100 voids rejecting indications of less than 6 pixels (or 2.7 lm) as noise. To reveal the grain structure, images were recorded for the samples etched with waterless Kalling’s reagent. After etching the grain size was determined based on the Feret diameter using Image J analysis software. The size and shape analysis of the starting powders was carried out with a Horiba LA-9, laser diﬀraction particle size analyzer. The powders were measured dry with an air-jet dispersion method at 0.40 MPa for 5000 seconds. Approximately 100 particles were used in the analysis. 4148—VOLUME 51A, AUGUST 2020 III. EXPERIMENTAL TECHNIQUES 2—Experimental set-up of the X-ray refraction station at the BAMline. The blue arrows indicate the beam path. The samples are mounted in a slide frame shown on the right. A sketch of the set-up is also indicated at the bottom of the photo (Color ﬁgure onine). Fig. 3—Rocking curves measured at one (arbitrary) detector pixel. Open circles: without sample. Filled circles: with sample. The hatched area represents the calculated rocking curve assuming that the sample would not have any internal interfaces and thus would not display any refraction eﬀect. Using the image analysis software Fiji ImageJ[29,30] the attenuation properties (l d) of a sample of thick- ness d, its refraction value ðCm dÞ, and its relative speciﬁc surface (Cm=lÞ are evaluated for each pixel according to Eqs. [1] through [3]), respectively. l d ¼ ln I I0 ½1 Cm d ¼ 1 IR IR0 I0 I ½2 Cm=l ¼ 1 IR IR0 I0 I = ln I I0 ½3 l d ¼ ln I I0 ½1 Cm d ¼ 1 IR IR0 I0 I ½2 m=l ¼ 1 IR IR0 I0 I = ln I I0 ½3 l d ¼ ln I I0 ½1 Cm d ¼ 1 IR IR0 I0 I ½2 ½1 ½2 Cm=l ¼ 1 IR IR0 I0 I = ln I I0 ½3 ½3 Fig. 3—Rocking curves measured at one (arbitrary) detector pixel. Open circles: without sample. Filled circles: with sample. The hatched area represents the calculated rocking curve assuming that the sample would not have any internal interfaces and thus would not display any refraction eﬀect. A detailed description of the data conditioning and evaluation can be found.[31,32] Note that the local relative speciﬁc internal surface of the sample is independent of its thickness d and more importantly of its local variations. We notice that by analyzing the rocking curve, we separate the linear absorption coeﬃcient by applying Eq. [1], and the refraction/scatter fraction by applying Eq. [3]. While The background-corrected rocking curve images were analyzed using an in-house software code based on LabView. III. EXPERIMENTAL TECHNIQUES Bright ﬁeld images were taken of un-etched samples to record the void size and morphology. Suﬃcient images were recorded at 200X Table I. Heat Treatment Parameters (Temperature C/time h)[23,24] Sample ID SR/Q HIP/SC Homo/Q ST/Q Age 1 Age 2 554 — — — — — — 508 1066/1 1163/3 1163/1 1066/1 760/10 649/10 523 — — — 1010/1 718/2 — 4148—VOLUME 51A, AUGUST 2020 METALLURGICAL AND MATERIALS TRANSACTIONS A Table I. Heat Treatment Parameters (Temperature C/time h)[23,24] Table I. Heat Treatment Parameters (Temperature C/time h)[23,24] 4148—VOLUME 51A, AUGUST 2020 Fig. 1—Build direction of samples in this study (a); location of SXRR and OM samples (b). Fig. 1—Build direction of samples in this study (a); location of SXRR and OM samples (b). extracted from one arbitrary detector pixel, with and without the sample in the beam. The intrinsic beam divergence was extracted from the rocking curve without the sample in place. The beam pos- sesses a full width half maximum (FWHM) of about 0.00064 deg. The rocking curve with sample has an increased FWHM of about 0.001 deg, due to refrac- tion at interfaces inside the sample. Without inter- faces in the sample, the shaded curve in Figure 3 would have been measured. This means that the reduced peak height of the measured rocking curve (ﬁlled circles) is caused by both X-ray attenuation caused by the sample and X-ray refraction resulting from inner surfaces inside the sample. The reduced peak integral intensity (area beneath ﬁlled circles) is related to the attenuation properties of the sample. The analysis software calculates the values of the rocking curve integral, the peak height, the peak position, and the FWHM. The material’s linear attenuation coeﬃcient l [mm-1] and the relative speciﬁc surface content of the samples are calculated from the quantities listed in Table II. h f j [29 30] Fig. 2—Experimental set-up of the X-ray refraction station at the BAMline. The blue arrows indicate the beam path. The samples are mounted in a slide frame shown on the right. A sketch of the set-up is also indicated at the bottom of the photo (Color ﬁgure onine). Fig. 2—Experimental set-up of the X-ray refraction station at the BAMline. The blue arrows indicate the beam path. The samples are mounted in a slide frame shown on the right. A sketch of the set-up is also indicated at the bottom of the photo (Color ﬁgure onine). Fig. IV. RESULTS AND DISCUSSION other works have dealt with the detailed character of the refracted signal (reﬂection, refraction, small angle scat- tering), see References 33 and 34, the correlation between X-ray refraction signal and speciﬁc surface has been previously quantitatively established.[13,14] This is based on the hypothesis that the main scattering mechanism is refraction. A comparison was made of the non-HIPed direct age vs the HIP two step age as various studies have shown that similar mechanical properties can be obtained with shorter heat treatments of Inconel 718.[23,24] The particle size analysis of the starting powders showed a Gaussian distribution from 10 to 70 lm. The median size was about 27 lm with a mean of 28 lm, and an aspect ratio of 0.98. The density obtained from the Pycnometry measurements was 8.14 g/cm3. This is slightly less than the reported 8.19 g/cm3 reported for wrought Inconel 718, and corresponds to a void volume of 0.5 pct. Figure 4 shows an example of the above-mentioned quantities, calculated for sample #523 using Eqs. [1] through [3], as two-dimensional gray value images. The local relative speciﬁc surface combined with the local attenuation properties (also obtained from SXRR) was used to identify the discontinuities observed. Irreg- ular discontinuities aligned perpendicular to the build direction are indicative of lack of melt in-between successive layers. This type of discontinuity can result from non-optimized build parameters such as layer thickness or hatch spacing, or irregularities in the laser beam. If a large refraction signal is displayed, it indicates the region possesses a large relative speciﬁc surface within its volume. This would result from a complex shape or some type of internal structure, expected to be inclusions from entrapped splatter or unmelted powder in an AM sample. Crescent-shaped structures in the relative speciﬁc surface image correspond to circu- lar-shaped voids with no internal structure, such as the case of empty voids (which can be caused by trapped gas). The binarized optical microscopy images of represen- tative un-etched samples are shown in Figure 5. Table III summarizes the void sizes and area fractions measured. In the as-built and direct aged non-HIPed samples, almost round voids are observed that are similar in size and area fraction. After the heat treat- ment with HIP, the voids are reduced in both size and area fraction but are slightly elongated. Optical microscopy images of corresponding repre- sentative etched samples are shown in Figure 6. III. EXPERIMENTAL TECHNIQUES Figure 3 shows typical rocking curves METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4149 Table II. Rocking Curve Parameter Symbol Quantity Physical Meaning IR peak height with sample in the beam (curve with filled circles) reduced maximum peak intensity caused by the attenuation properties and internal interfaces of the sample IR0 peak height without sample in the beam (curve with open circles) maximum peak intensity of the monochromatic synchrotron beam I peak integral with sample in the beam (curve with filled circles) reduced peak intensity caused by the attenuation properties of the sample I0 peak integral without sample in the beam (curve with open circles) overall peak intensity of the monochromatic synchrotron beam Fig. 4—Example of data treatment steps for sample #523 (non-HIPed with direct age). (a) Image of the attenuation property (l d); (b) refraction value (Cm dÞ; (c) relative speciﬁc internal surface ðCm=lÞ. The build direction Z was oriented parallel to the scattering vector. Fig. 4—Example of data treatment steps for sample #523 (non-HIPed with direct age). (a) Image of the attenuation property (l d); (b) refraction value (Cm dÞ; (c) relative speciﬁc internal surface ðCm=lÞ. The build direction Z was oriented parallel to the scattering vector. Fig. 4—Example of data treatment steps for sample #523 (non-HIPed with direct age). (a) Image of the attenuation property (l d); (b) refraction value (Cm dÞ; (c) relative speciﬁc internal surface ðCm=lÞ. The build direction Z was oriented parallel to the scattering vector. IV. RESULTS AND DISCUSSION Evi- dence of columnar grains is apparent in samples 554 and 523, neither of which were HIPed. Strong textures, evident in the almost uniform response to etching, have been correlated with this columnar grain structure in the METALLURGICAL AND MATERIALS TRANSACTIONS A 4150—VOLUME 51A, AUGUST 2020 presentative optical microscopy images of the un-etched samples. (a) Non-HIPed as-built (#554), (b) non-HIPed and direct a (c) HIP and heat treated per SAE/AMS 5663 (#508). Fig. 5—Representative optical microscopy images of the un-etched samples. (a) Non-HIPed as-built (#554), (b) non-HIPed and direct aged (#523), and (c) HIP and heat treated per SAE/AMS 5663 (#508). Table III. Summary of Void Size Based on Optical Microscopy Table III. Summary of Void Size Based on Optical Microscopy Sample ID Area Fraction (Pct) Maximum Void Size (lm) Average Void Size (lm) 554 (XY) 0.29 64.5 12.3 ± 13.9 554 (Z) 0.17 50.0 11.7 ± 12.5 523 (XY) 0.30 77.7 14.0 ± 16.1 523 (Z) 0.21 67.8 14.6 ± 14.7 508 (XY) 0.15 39.0 5.3 ± 4.5 508 (Z) 0.15 43.2 5.4 ± 5.4 Z build direction.[35–37] Due to the retention of the scan pattern, grain sizes cannot be determined from the etched optical images. After a full heat treatment including HIP, equiaxed grains are observed in both the build direction (Z) and build plane (XY). The grains are on average 20 lm, with METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4151 l microscopy images of the etched samples. (a) Non-HIPed as-built (554), (b) non-HIPed and direct age (523), an r SAE/AMS 5663. Fig. 6—Optical microscopy images of the etched samples. (a) Non-HIPed as-built (554), (b) non-HIPed and direct age (523), and (c) HIP and heat treated per SAE/AMS 5663 several large grains on the order of 430 lm in both orientations. Equiaxed grains imply that the texture has been randomized, as evidenced by the random variation in response to the etching. non-HIPed and heat-treated sample #523 reveals dark circular spots of apparent low density. The average size of these dark spots is approximately 40 lm. Such spots are visible at the same location in both orientations of the sample as shown by rectangles no. 1 and 2 in Figure 7 (left, top and bottom), indicating identical positions within the sample. The dark spots correspond to the bright spots in the of Cm/l images as shown in Figure 7, middle, top and bottom images. 4152—VOLUME 51A, AUGUST 2020 IV. RESULTS AND DISCUSSION The crescents seem to be slightly elongated in the parallel orientation than in the perpendicular orientation as shown by comparison with Figure 7 (right top and bottom). This is also suggested by the values reported in Table IV. This means that the discontinuities or cavities are slightly elongated, and oriented along the build plane X–Y. This slight elonga- tion in the build plane could result from non-optimized hatch spacing used in the build or variations in laser power. indicates, that the dark spots in the lÆd images are empty voids or cavities (i.e., not ﬁlled with unmelted powder, or representing high-density grains). There is pro- nounced orientation dependence of the various features observed. While the lÆd image is independent of the sample orientation, the relative speciﬁc surface image has a slight dependency. The crescents seem to be slightly elongated in the parallel orientation than in the perpendicular orientation as shown by comparison with Figure 7 (right top and bottom). This is also suggested by the values reported in Table IV. This means that the discontinuities or cavities are slightly elongated, and oriented along the build plane X–Y. This slight elonga- tion in the build plane could result from non-optimized hatch spacing used in the build or variations in laser power. The comparison between the lÆd and Cm/l image for the HIPed and heat-treated sample #508 does not show the single voids or cavities that were visible in Figure 7 for the non-HIPed and direct aged sample #523. Rather some bright features are visible in the lÆd image which possess a certain amount of relative speciﬁc surface that relates to grain or sub-grain boundaries and/or to forests of dislocations (similar structures have been observed in Al alloys).[32,38] The very weak and blurred refraction intensity in the relative speciﬁc surface image, shown by the magniﬁed image in Figure 8 right, is thus probably related to grain boundaries oriented perpen- dicular to the scattering vector, or alternatively to pores that have not closed after HIPing. In contrast, these small dark circular spots are not observed in the HIPed and heat-treated sample #508 shown in the attenuation images in Figure 8. Rather, bright spots of larger size, with apparent high density, are observed. Such bright spots have irregular bound- aries and are not stationary with respect to the sample as noted by comparison with the rectangle no. IV. RESULTS AND DISCUSSION The enlarge- ment of rectangle no. 1 in Figure 7 right, shows crescents in both orientations at the upper and lower edges of the voids, while their center is dark. This The local values of l d are visualized as 2D gray scale images in Figure 7 (left) for sample non-HIPed and direct age sample #523 and in Figure 8 for the HIPed and heat-treated sample #508 oriented with its build direction Z parallel (top) and perpendicular (bottom) to the scattering plane, respectively. A homogenous l d value is observed across both samples (the edges show exceptions due to the polishing artifacts). The METALLURGICAL AND MATERIALS TRANSACTIONS A 4152—VOLUME 51A, AUGUST 2020 Fig. 7—Comparison of two orientations of the build direction Z for the non-HIPed and direct aged sample #523 with respect to the scattering plane of the detector system (Top—parallel, bottom—perpendicular). Left: lÆd (attenuation), middle: speciﬁc surface (refraction), right: shows the enlargement of the rectangle no. 1. Numerical values are given in Table IV. Fig. 7—Comparison of two orientations of the build direction Z for the non-HIPed and direct aged sample #523 with respect to the scattering plane of the detector system (Top—parallel, bottom—perpendicular). Left: lÆd (attenuation), middle: speciﬁc surface (refraction), right: shows the enlargement of the rectangle no. 1. Numerical values are given in Table IV. grains do not fulﬁll the Bragg condition if the sample is rotated 90 deg. They can, however, be used to estimate the size of crystallites or coherent domains. The size of these bright spots is approximately 200 lm. This corresponds to the largest grain sizes observed in the equiaxed sample that was HIPed and heat treated (see Figure 6(c) left). We also notice that, since the sample thickness of the SXRR measurement is 160 lm, the spots in Figure 8 correspond to single crystallites. Such bright spots are not observed in the attenuation maps of sample #523 simply because the grain orientation is not favorable. indicates, that the dark spots in the lÆd images are empty voids or cavities (i.e., not ﬁlled with unmelted powder, or representing high-density grains). There is pro- nounced orientation dependence of the various features observed. While the lÆd image is independent of the sample orientation, the relative speciﬁc surface image has a slight dependency. IV. RESULTS AND DISCUSSION 1 shown in (left, top and bottom images). Thus, they do not correspond to dense regions, but rather to grains that are oriented to fulﬁll the Bragg condition for the monochromatic radiation used. This phenomenon removes intensity from the transmitted beam so that these regions appear as absorbing or bright. In fact these In the optical microscopy images, the void area and size are similar in the non-HIPed samples for both the as-built and the direct age. Due to their round shape and METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4153 Fig. 8—Comparison of two orientations of the HIPed and heat-treated sample #508 with respect to the scattering vector (Top—parallel, bottom—perpendicular). Left: lÆd (attenuation), middle: speciﬁc surface (refraction), right: enlargement of the rectangle no. 1, which marks the same area of the sample. Fig. 8—Comparison of two orientations of the HIPed and heat-treated sample #508 with respect to the scattering vector (Top—parallel, bottom—perpendicular). Left: lÆd (attenuation), middle: speciﬁc surface (refraction), right: enlargement of the rectangle no. 1, which marks the same area of the sample. of strongly diﬀracting grains in the equiaxed sample #508. smooth interior surface, these may be retained from initial porosity in the argon gas atomized powders used in the build.[1,10,39–41] Analysis of the powders indicates an ‘initial’ void volume of around 0.5 pct. This is slightly higher than the maximum measured porosity by optical microscopy (0.3 pct). Considering the maximum voids measured with optical microscopy and SXRR, a 14 lm diameter void could easily reside within the 70 lm maximum size of the starting powder. This would indicate voids present in the gas atomized powder could be retained in the AM build and do not necessarily result from the AM process. The slight increase in size of the void in the non-HIPed sample after heat treatment would result from the expansion of trapped gas. Obviously, a detailed analysis of the origin of the pores would need in situ monitoring (e.g., by means of active thermography[42,43]), but this is outside the scope of this work. The values of the relative speciﬁc surface are small for both samples in both orientations. IV. RESULTS AND DISCUSSION Mechanical Properties of Samples Sample ID Ultimate Tensile Stress (MPa) Yield Stress (MPa) Elongation at Fracture (Pct) 554 995 ± 13 698 ± 15 33.2 ± 1.10 523 1362 ± 12 1233 ± 12 19.1 ± 0.74 508 1406 ± 7 1170 ± 6 24.9 ± 0.39 Table IV. Linear Attenuation Coeﬃcient and Relative Internal Speciﬁc Surface Calculated from X-ray Transmission and Refraction Measurements Table V. Mechanical Properties of Samples Sample ID Ultimate Tensile Stress (MPa) Yield Stress (MPa) Elongation at Fracture (Pct) 554 995 ± 13 698 ± 15 33.2 ± 1.10 523 1362 ± 12 1233 ± 12 19.1 ± 0.74 508 1406 ± 7 1170 ± 6 24.9 ± 0.39 Table V. Mechanical Properties of Samples HIPed and aged sample is similar to the non-HIPed and direct aged sample, the size, area fraction and morphol- ogy do not greatly aﬀect the quasi-static strength. However, it should be noted that since the morphology of the voids changed as a result of the HIP process, the eﬀect of the elongated voids maybe more pronounced in dynamic fatigue tests. The slightly lower ductility measured in quasi-static tension tests may also corre- spond to decreased fatigue life. Interestingly, a primary diﬀerence between the non-HIPed and HIPed processes is the higher temperature of the HIP process of 1163 C. It is this higher temperature that is attributed to forming the reﬁned equiaxed grain morphology observed in the HIPed and aged sample #508. In fact, in spite of the high temperature, no grain growth occurs during HIPing. This is attributed to grain boundary pinning by carbides or precipitates which limit grain boundary mobility. sample decreased to a 5 lm average size with a maximum average of 40 lm in the HIPed samples after the two step aging heat treatment. X-ray attenuation images of the non-HIPed and direct age sample show 40 lm dark spots that correspond to very low-density objects with smooth circular boundaries that can be identiﬁed as voids. In the relative speciﬁc surface image these discontinuities correspond to juxtaposed crescents surrounding a dark center indicating the objects are empty voids, which correlate with the 2D optical microscopy images. From the SXRR, the speciﬁc surface value of the non-HIPed sample was about 2 times higher than the HIPed sample. IV. RESULTS AND DISCUSSION This eﬀect is ascribed to the diﬀerences in the porosity ranging from 0.20 to 0.30 pct for the non-HIPed samples, with respect to the 0.15 pct in the HIPed sample. The SXRR also detected a change in morphology from a circular shape to a slightly elon- gated shape in the HIPed sample. The change in grain morphology, size and area fraction did not appreciably change the quasi-static tensile strength although is expected to be more pronounced in dynamic fatigue tests. IV. RESULTS AND DISCUSSION In spite of the fact that such values lie just above the detection limit of the experimental set-up, it can be stated that the non-HIPed direct aged sample #523 has a two times larger value of the relative speciﬁc surface compared to the HIPed and aged sample #508. In both samples, the perpendicular orientation shows smaller values of Cm/l, thereby indicating some preferential discontinuity alignment with the powder layer plane. This could result from non-optimized hatch spacing or variations in laser power during the build. The results of the mechanical tests are summarized in Table V. The non-HIPed as-built sample has the lowest strength, as expected for the heat treatable Inconel 718 alloy with a corresponding larger elongation at fracture. Although the area fraction of voids and their size and morphology are similar between the non-HIPed as-built #554 and the heat-treated #523, the elongation at fracture is notably aﬀected. This is expected as increases in strength are usually accompanied by a reduction in ductility. This implies that porosity is not the most important factor determining the mechanical properties in uniaxial static tests. Table IV summarizes the mean values (integrated across the ﬁeld-of-view) of l d, l, d and Cm/l of the two samples in the two orientations. Diﬀerences in l d are not solely due to the variation of d between the two samples. Using their average measured thickness, a mean value of l for each sample could be calculated as shown in Table IV. As expected, the measured value of l d or l is invariant against sample orientation. Remarkably, the attenuation coeﬃcient is diﬀerent in the two materials; this can be due not only to the diﬀerent amount of porosity but also to the presence The heat-treated and HIPed sample #508 has both the highest strength and elongation at fracture with the lowest void area and size. Since the strength of the METALLURGICAL AND MATERIALS TRANSACTIONS A 4154—VOLUME 51A, AUGUST 2020 Table IV. Linear Attenuation Coeﬃcient and Relative Internal Speciﬁc Surface Calculated from X-ray Transmission and Refraction Measurements Sample Orientation of Build Direction to the Scattering Vector l d d (mm) l (mm) Cm/l #508 parallel 1.629 0.17 9.850 0.00183 #508 perpendicular 1.632 9.602 0.00074 #523 parallel 1.320 0.15 8.803 0.00485 #523 perpendicular 1.323 8.820 0.00324 Table V. VI. CONCLUSIONS 1. S. Tammas-Williams, P.J. Withers, and P.B. Prangnell: Scripta Mater., 2016, vol. 122, pp. 72–76. For optimization of AM processing parameters, an understanding of the underlying mechanisms of discon- tinuity formation is required. The use of non-destructive synchrotron X-ray refraction and transmission radiog- raphy has been shown to be an extremely valuable tool in identifying the nature of discontinuities in a build, especially when anchored with optical microscopy. As this study showed, quasi-static mechanical properties are not suﬃcient to quantify the quality of a printed AM part. Although optimized parameters were used in the build of the specimens in this study, lack of melt in indications were observed which did not heal during the HIP process. Once identiﬁed, further parameter devel- opment can alleviate this discontinuity. Other disconti- nuities such as voids could be easily distinguished due to their lack of internal surface features. Additional infor- mation is also obtained regarding grain morphology, size and texture. Since the ﬁeld-of-view investigated by X-ray refraction techniques is much larger than that obtainable by optical microscopy, this technique yields better statistics and can be readily correlated with the mechanisms of discontinuity formation. Expanding this study to the rationalization of dynamic mechanical properties in fatigue tests due to diﬀerences in discon- tinuity formation is left for future work. 2. C. Schaak, W. Tillmann, M. Schaper, and M.E. Sydinoez: Forum Rapid Technol., 2016, vol. 13, p. 4403. 3. B. Zhang, Y. Li, and Q. Bai: J. Chin. J. Mech. Eng., 2017, vol. 30, pp. 515–27. 4. S. Siddique, M. Imran, M. Rauer, M. Kaloudis, E. Wycisk, C. Emmelmann, and F. Walther: Mater. Des., 2015, vol. 83, pp. 661–69. 5. A. Yadollahi and N. Shamsaei: Int. J. Fatigue, 2017, vol. 98, pp. 14–31. 6. S. Beretta and S. Romano: Int. J. Fatigue, 2017, vol. 94, pp. 178–91. 7. R.C. Cooper, G. Bruno, Y. Onel, A. Lange, T.R. Watkins, and A. Shyam: J. Mater. Sci., 2016, vol. 51, pp. 9749–60. 8. T. Thiede, T. Mishurova, S. Evsevleev, I. Serrano-Munoz, C. Gollwitzer, and G. Bruno: Quantum Beam Sci., 2019, vol. 3, pp. 1–12. 9. B.R. Mu¨ ller, R.C. Cooper, A. Lange, A. Kupsch, M. Wheeler, M.P. Hentschel, A. Staude, A. Pandey, A. Shyam, and G. Bruno: Acta Mater., 2018, vol. 144, pp. 627–41. 10. G. Kasperovich, J. Haubrich, J. Gussone, and G. Requena: Mater. Des., 2016, vol. 105, pp. 160–70. 11. A. Kupsch, B.R. Mu¨ ller, A. VI. CONCLUSIONS Lange, and G. Bruno: J. Eur. Ceram. Soc., 2017, vol. 37, pp. 1879–89. 12. M.P. Hentschel, R. Hosemann, A. Lange, B. Uther, and R. Bruckner: Acta Crystallogr. A, 1987, vol. 43, pp. 506–13. 13. F.E. Fensch-Kleemann, K.-W. Harbich, and M.P. Hentschel: Ceram. Forum Int., 2002, vol. 79, pp. E35–38. 14. K.W. Harbich, M.P. Hentschel, and J. Schors: NDT & E Int., 2001, vol. 34, pp. 297–302. pp 15. B.R. Mu¨ ller and M.P. Hentschel: X-Ray Spectrom., 2004, vol. 33, pp. 402–06. 16. R. Laquai, T. Schaupp, B.R. Mu¨ ller, A. Griesche, G. Bruno, and T. Kannengießer: in NDT-CE 2015 - International Symposium Non-destructive Testing in Civil Engineering (Proceedings), Tech- nische Universita¨ t Berlin / Bundesanstalt fu¨ r Materialforschung und -pru¨ fung: Berlin, Germany, 2015, pp. 1217–24. V. SUMMARY The use of 2.5D synchrotron radiation refraction radiography (SXRR), yielding both attenuation and refraction information, allows a great deal of insight into the nature of the discontinuities observed in the microstructure of AM SLM Inconel 718. Samples were investigated in the non-HIPed as-built, non-HIPed direct age and HIPed with two step aging treatments. Although the diﬀerences in the quasi-static tension tests were attributed to the aging heat treatments to promote precipitate formation and subsequent grain reﬁnement in age hardenable alloys rather than a diﬀerence in the nature of the discontinuities within the samples, the nature of the discontinuities is expected to aﬀect the dynamic fatigue life. Noticeable diﬀerences in the microstructure were observed in the optical images of the non-HIPed and HIPed samples. After the HIP and aging heat treatment, the microstructure consisted of reﬁned, equiaxed grains with an average grain size of 20 lm and maximum of 430 lm. Although similar strengths were noted in the non-HIPed direct age and HIPed two step age, the ductilities diﬀer. The higher ductility in the HIPed sample is attributed to the reﬁned equiaxed grains. A noted diﬀerence between the heat treatments was the HIP temperature of 1163 C as compared to the homogenization/solutionizing temperature of 1066 C. Whether this higher temperature triggers the grain reﬁnement vs the pressure of the HIP process is unknown as mechanisms driving grain reﬁnement in AM processing remain under investigation. y g Optical micrographs show a slight increase in void size and area fraction in the non-HIPed samples after heat treatment which could be due to expansion of trapped gases. The average 14 lm void size with a maximum of 70 lm in the non-HIPed direct aged p g g X-ray attenuation images of the HIP process and two step age sample shows bright spots about 200 lm in size METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4155 OPEN ACCESS with irregular boundaries. The bright spots in the attenuation images also display a large relative speciﬁc surface, related to grain or sub-grain boundaries and/or forests of dislocations and are thus correlated with speciﬁcally oriented grains. The non-homogenous dis- tribution of these bright spots correlates with the random orientation of the equiaxed grains. Similar characteristics are not observed in the non-HIPed direct aged sample which retains columnar grains. with irregular boundaries. The bright spots in the attenuation images also display a large relative speciﬁc surface, related to grain or sub-grain boundaries and/or forests of dislocations and are thus correlated with speciﬁcally oriented grains. The non-homogenous dis- tribution of these bright spots correlates with the random orientation of the equiaxed grains. Similar characteristics are not observed in the non-HIPed direct aged sample which retains columnar grains. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appro- priate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat ivecommons.org/licenses/by/4.0/. SXRR also identiﬁed discontinuities with preferential alignment along the powder layers. While this charac- teristic was observed in both the SXRR analysis of the non-HIPed and HIPed sample, it was not observed in the optical microscopy images. These types of discon- tinuities could result from non-optimized hatch spacing or variations in laser power during the build. Under- standing and mitigating the cause of this discontinuity, which persisted after the HIP process, is critical to the structural integrity of AM components. ACKNOWLEDGMENTS 17. R. Laquai, B.R. Mu¨ ller, G. Kasperovich, J. Haubrich, G. Re- quena, and G. Bruno: Mater. Res. Lett., 2018, vol. 6, pp. 130–35. Open Access funding provided by Projekt DEAL. We thank HZB colleagues for their support as well as Ralf Britzke and Thomas Wolk (BAM) for their assistance during beam time at BAMline. One author (JAS) acknowledges the opportunity as a guest scientist at the Federal Institute for Materials Research and Testing (BAM) to participate in this collaborative research. 18. B.R. Mu¨ ller and M.P. Hentschel: in Handbook of Technical Diagnostics—Fundamentals and Application to Structures and Systems, H. Czichos, ed., Springer, 2013, pp. 287–300. y p g pp 19. S. Gu¨ nzel, S. Hickmann, C. Wittemeyer, and V. Trappe: Adv. Eng. Mater., 2012, vol. 14, pp. 867–72. 20. R.C. Cooper, G. Bruno, M.R. Wheeler, A. Pandey, T.R. Watkins, and A. Shyam: Acta Mater., 2017, vol. 135, pp. 361–71. 4156—VOLUME 51A, AUGUST 2020 METALLURGICAL AND MATERIALS TRANSACTIONS A 21. A. Kupsch, A. Lange, M.P. Hentschel, Y. Onel, T. Wolk, A. Staude, K. Ehrig, B.R. Mu¨ ller, and G. Bruno: J. Ceram. Sci. Technol., 2013, vol. 4, pp. 169–76. 33. F. Zhang, G.G. Long, L.E. Levine, J. Ilavsky, and P.R. Jemian: J. Appl. Crystallogr. 2008, vol. 41. pp y g , 34. O. Oltulu, Z. Zhong, M. Hasnah, M. Wernick, and D. Chapman: J. Phys. D, 2003, vol. 36, p. 2152. , , , pp 22. SAE/AMS 5664, 2009. , , , p 22. SAE/AMS 5664, 2009. y p 35. L.N. Carter, C. Martin, P.J. Withers, and M.M. Attallah: J. Alloys Compd., 2014, vol. 615, pp. 338–47. 23. J.A. Schneider, B. Lund, and M. Fullen: AM J., 2018, vol. 21, pp. 248–54. 24. I. Kelbassa, E.W. Kreutz, P. Albus, and L. Zhu: in Proc. 24th Intl. Congr. Appl. Lasers & Electro-optics (ICALEO), Laser Institute of America, Orlando, pp. 660–65. 36. N. Nadammal, S. Cabeza, T. Mishurova, T. Thiede, A. Kromm, C. Seyfert, L. Farahbod, C. Haberland, J.A. Schneider, P. Portella, and G. Bruno: Mater. Des., 2017, vol. 134, pp. 139–50. pp 37. L. Yanjin, W. Songquan, Y. Gan, T. Huang, C. Yang, J. Lin, and J. Lin: Opt. Laser Technol., 2015, vol. 75, pp. 197–206. 25. B.R. Mu¨ ller, A. Lange, M. Harwardt, and M.P. Hentschel: Adv. Eng. Mater., 2009, vol. 11, pp. 435–40. 26. W. Go¨ rner, M.P. Hentschel, B.R. Mu¨ ller, H. Riesemeier, M. Krumrey, G. Ulm, W. VOLUME 51A, AUGUST 2020—4157 METALLURGICAL AND MATERIALS TRANSACTIONS A ACKNOWLEDGMENTS Diete, U. Klein, and R. Frahm: Nucl. Instrum. Methods Phys. Res. Sect. A, 2001, vol. 467, pp. 703–06. 38. R. Ferna´ ndez, G. Bruno, and G. Gonza´ lez-Doncel, 2018, vol. 123, p. 145108. 39. W.J. Sames, F.A. List, S. Pannala, R.R. Dehoﬀ, and S.S. Babu: Int. Mater. Rev., 2016, vol. 61, pp. 315–60. 27. A. Rack, S. Zabler, B.R. Mu¨ ller, H. Riesemeier, G. Weidemann, A. Lange, J. Goebbels, M. Hentschel, and W. Go¨ rner: Nucl. Instrum. Methods A, 2008, vol. 586, pp. 327–44. 40. W.J. Sames, F. Medina, W.H. Peter, S.S. Babu, and R.R. Dehoﬀ: in Superalloy 718 & Derivatives, Pittsburgh, PA, 2014. in Superalloy 718 & Derivatives, Pittsburgh, PA, 2014 28. A. Lange, M.P. Hentschel, A. Kupsch, and B.R. Mu¨ ller: Int. J. Mater. Res., 2012, vol. 103, pp. 174–78. 41. H. Qi, M. Azer, and A. Ritter: Metall. Mater. Trans. A, 2009, vol. 40A, pp. 2410–22. 29. Wayne Rasband, 1997. 42. N. Scheuschner, S. Altenburg, A. Straße, A. Gumenyuk, and C. Maierhofer: in II International Conference on Simulation for Additive Manufacturing—Sim-AM 2019, Pavia, Italy, 2019, pp. 246–55. 30. C.A. Schneider, W.S. Rasband, and K.W. Eliceiri: Nat. Methods, 2012, vol. 9, p. 671. 31. J. Nellesen, R. Laquai, B.R. Mu¨ ller, A. Kupsch, M.P. Hentschel, N.B. Anar, E. Soppa, W. Tillmann, and G. Bruno: J. Mater. Sci., 2018, vol. 53, pp. 6021–61. 43. U. Hassler, D. Gruber, O. Hentschel, F. Sukowski, T. Grulich, and L. Seifert: Phys. Procedia, 2016, vol. 83, pp. 1244–52. 32. Sandra Cabeza, Bernd Randolf Mu¨ ller, Rocı´o Pereyra, Ricardo Fernandez, G. Gonzalez-Doncel and Giovanni Bruno, Journal of Applied Crystallography 2018, vol. 51, p. https://doi.org/10.1107/ S1600576718001449. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional aﬃliations. METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 51A, AUGUST 2020—4157
https://openalex.org/W1845673467	https://europepmc.org/articles/pmc1450293?pdf=render	English	null	Gender effect on clinical features of achalasia: a prospective study	BMC gastroenterology	2,006	cc-by	4,535	BioMed Central BioMed Central BioMed Central Open Acc Research article Gender effect on clinical features of achalasia: a prospective study Javad Mikaeli, Farnoosh Farrokhi, Faraz Bishehsari, Mahboobeh Mahdavinia and Reza Malekzadeh* Address: Achalasia Research Unit, Digestive Disease Research Center, Tehran University of Medical Sciences, Shariati Hospital, Tehran 14114, Iran Email: Javad Mikaeli - mikaeli@ams.ac.ir; Farnoosh Farrokhi - farrokhi@ddrcir.org; Faraz Bishehsari - bishehsari@ddrcir.org; Mahboobeh Mahdavinia - mahdavi@ddrcir.org; Reza Malekzadeh* - malek@ams.ac.ir * Corresponding author * Corresponding author Published: 01 April 2006 BMC Gastroenterology2006, 6:12 doi:10.1186/1471-230X-6-12 Received: 15 November 2005 Accepted: 01 April 2006 This article is available from: http://www.biomedcentral.com/1471-230X/6/12 © 2006Mikaeli et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Received: 15 November 2005 Accepted: 01 April 2006 © 2006Mikaeli et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BMC Gastroenterology Open Access Abstract Background: Achalasia is a well-characterized esophageal motor disorder but the rarity of the disease limits performing large studies on its demographic and clinical features. Methods: Prospectively, 213 achalasia patients (110 men and 103 women) were enrolled in the study. The diagnosis established by clinical, radiographic, and endoscopic as well as manometry criteria. All patients underwent a pre-designed clinical evaluation before and within 6 months after the treatment. Results: Solid dysphagia was the most common clinical symptom in men and women. Chest pain was the only symptom which was significantly different between two groups and was more complained by women than men (70.9% vs. 54.5% P value= 0.03). Although the occurrence of chest pain significantly reduced after treatment in both groups (P < 0.001), it was still higher among women (32% vs. 20.9% P value= 0.04). In both sexes, chest pain did not relate to the symptom duration, LES pressure and type of treatment patients received. Also no significant relation was found between chest pain and other symptoms expressed by men and women before and after treatment. Chest pain was less frequently reported by patients over 56 yrs of age in comparison to those less than 56 yrs (p < 0.05). Conclusion: It seems that chest pain is the distinct symptom of achalasia which is affected by sex as well as age and does not relate to the duration of illness, LESP and the type of treatment achalasia patients receive. result in several clinical symptoms including dysphagia as the most common symptom, postprandial and/or noctur- nal regurgitation, weight loss, chest pain and cough [4-7]. Background g Achalasia is a primary esophageal motor disorder of unknown etiology characterized by absent esophageal peristalsis and abnormal lower esophageal sphincter (LES) relaxation [1-3]. The disease can occur at any age but it is usually diagnosed in patients who are between 25 and 60 years. The motor abnormalities found in this dis- ease compromise the normal esophageal emptying and Men and women are affected with equal frequencies [8], but data regarding the characteristics differences between males and females in achalasia is spars. Page 1 of 6 (page number not for citation purposes) BMC Gastroenterology 2006, 6:12 http://www.biomedcentral.com/1471-230X/6/12 Table 1: Scoring system for evaluation of clinical symptoms Score 0 1 2 3 Symptoms Dysphagia to solids none weekly Daily each meal Dysphagia to liquids none weekly Daily each meal Active regurgitation none weekly Daily each meal Passive regurgitation none monthly Weekly daily Chest pain none monthly Weekly daily Table 1: Scoring system for evaluation of clinical symptoms standard questionnaire was filled for each of the patients. Evaluation of their clinical features was done on the bases of five clinical symptoms; solid and liquid dysphagia, active and passive regurgitation and chest pain. The fre- quency of each of these symptoms was scored on a scale of 0–3 to obtain the total symptomatic score, as described elsewhere [10,11] (table 1). The highest obtainable score was 15. standard questionnaire was filled for each of the patients. l f h l l f d h b standard questionnaire was filled for each of the patients. Evaluation of their clinical features was done on the bases of five clinical symptoms; solid and liquid dysphagia, active and passive regurgitation and chest pain. The fre- quency of each of these symptoms was scored on a scale of 0–3 to obtain the total symptomatic score, as described elsewhere [10,11] (table 1). The highest obtainable score was 15. Although studies on esophageal motility in normal indi- viduals have shown some differences between males and females [9], this was not widely studied among achalasia patients. Demographic studies in achalasia have been hampered by low incidence and rarity of the disease, which limits the number of cases to be studied and there- fore weakens the probability to find significant differences in subgroups of achalasia patients. Additionally the history of other symptoms such as noc- turnal cough, weight loss as well as any evidence of the previous treatment was taken. Background Our center as the major referral center for achalasia in Iran provides us with the opportunity to include relatively large number of patients during these years. In this study, we aimed to evaluate the differences of the clinical fea- tures between men and women before and after treat- ment, and assess whether the dissimilarities are related to the clinical course, demographic features, or manometric characteristics, and determine how treatment affects the discrepancies. The patients underwent botulinum toxin injection or pneumatic dilation as described bellow. Subsequently the total symptomatic score was compared among men and women patients before and within 6 months after the treatment. Botulinm toxin injection Botulinum toxin injection (BT) was offered as the thera- peutic choice in achalasia patients with tortuous meg- aesophagus, epiphrenic diverticula, previous cardiomyotomy, age > 60 yrs with cardiopulmonary or comorbid diseases. Four-hundred units of Dysport (Ipsen, UK) equal to 160 units of Botox [12] was injected follow- ing an overnight fast, using 100 units in each quadrant at the level of LES under videoendoscopic guide. The study protocol was reviewed and approved by the ethic committee of digestive disease research center of Tehran University of Medical Sciences and informed con- sent was obtained from all patients. Esophageal manometry Over a period of ten years from July 1994 to December 2004 all symptomatic patients with idiopathic achalasia referred to the Shariati Hospital of Tehran University of Medical Sciences were consecutively enrolled in the study. The diagnosis of achalasia was established by clinical (dysphagia, regurgitation, chest pain), radiographic (bird's beak appearance of the lower esophageal sphinc- ter, decreased esophageal peristalsis, delayed esophageal emptying) and endoscopic as well as manometric criteria (aperistalsis of the esophageal body, increased lower esophageal sphincter pressure and incomplete relaxation (<50%) on swallowing). Manometric assessment was added to achalasia patient's evaluation from 1997 in our center. After an overnight fast, manometry was carried out by a standard eight chan- nel low compliance catheter (Synecticw, Synectics, Stock- holm, Sweden) perfused with hydraulic pump. Lower esophageal sphincter pressure (LESP) was studied by rapid pull through (RPT) technique in deep expiration. The average of five RPT analyses was recorded as resting LESP. An average of 10 water swallows was used to assess body contraction waves and LES relaxation. NS: Not Significant cant differences were found in the characteristics of men and women before their initial treatment (table 2). Before treatment Solid dysphagia was noted as the first clinical symptom in the disease course in 81% of men and 76.4% of women. No significant difference was found in the clinical features of achalasia between two groups except for the chest pain which was more common in women than in men (70.9% vs. 54.5% P = 0.03) (table 3). LESP was measured in 103 patients (50 women and 53 men), Mean LESP was 53.7 ± 17.4 and 58.6 ± 19.5 mmHg in men and women respectively (P = 0.1). The LESP did not differ significantly between patients with and without chest pain, neither in men nor in women (P = 0.6 & 0.8 respectively). Pneumatic dilatation After a clear liquid diet for 12 h and an overnight fast, pneumatic dilatations (PD) were performed using 35 or 30 mm Rigiflex balloon (The technique changed from using 35 mm balloon diameter to 30 mm diameter Since 1997) under a conscious sedation with diazepam and meperidin (5–10 and 25–50 mg I.V respectively). Follow- ing a complete upper gastrointestinal endoscopy, balloon dilators were passed over a guide wire and were posi- tioned such that LES places at the midpoint of the balloon under videoendoscopic guide. Rigiflex balloons were gradually inflated up to 10 psi in 30s and maintained for another 60 s. After emptying and pulling out the balloons, patients were endoscoped again to assess the LES opening (relaxation) and any evidence of bleeding or perforation and were discharged after a 6 h observation if remained asymptomatic. The frequencies of the clinical symptoms were compared between men and women before and within 6 months of treatment. Statistical analysis Continuous variables were summarized as mean ± stand- ard deviation. Qualitative variables were summarized as a percentage of the group total. Comparing numerical vari- ables with assumed equal variances and different vari- ances tested by t student and mann-withney respectively. Out of 110 men and 103 women, 47.3% vs. 54.4% had received no previous treatment, 3.6% vs. 10.6% had undergone surgical myotomy, 47.3% vs. 30.2% had been recommended to use medications such as nitrates or cal- cium channel blockers, 0.9% vs. 1.9% and 0.9% vs. 2.9% had undergone previous PD and BT respectively. A p-value of less than 0.05 was considered significant. Clinical evaluations A structured interview evaluating demographic as well as clinical features was conducted at the time of their first visit and within 6 months following their initial treatment (before the second treatment following relapse), and the Page 2 of 6 (page number not for citation purposes) http://www.biomedcentral.com/1471-230X/6/12 http://www.biomedcentral.com/1471-230X/6/12 BMC Gastroenterology 2006, 6:12 Table 2: Comparing the characteristics between men and women before the initial treatment Variable Women (n = 103, 48.4%) Men (n = 110, 51.6%) P Value Age (SD) (min-max) 39.3 (14.5) (16–77 y) 37.7 (14.1) (14–75 y) NS* (0.4) Mean symptom duration (min-max) (Years) 4.8 (0.5–28) 4.6 (0.4–30) NS (0.9) Mean weight loss (Kg) (SD) 3.2 (4) 3.2 (4.9) NS (0.7) NS: Not Significant Age (SD) (min-max) Mean symptom duration (min-max) (Years) Mean weight loss (Kg) (SD) Results 213 patients including 110 (51.6%) men and 103 (48.4%) women were enrolled in the study. No signifi- Page 3 of 6 (page number not for citation purposes) Table 3: Comparing the presenting symptoms between men and women before therapy Variable Women (n = 103, 48.4%) Men (n = 110, 51.6%) P Value Mean pre treatment total score (SD) 9.9 (2.7) 9 (3.1) 0.01 Pre treatment Solid dysphagia 100% 96.4% NS (0.2) Pre treatment Liquid dysphagia 98.1% 90.9% NS (0.5) Pre treatment active regurgitation 71.8% 73.6% NS (0.2) Pre treatment passive regurgitation 77.7% 63.6% NS (0.2) Pre treatment chest pain Daily 12.6% 70.9% 8.2% 54.5% 0.03 Weekly 44.7% 38.2% Monthly 13.6% 8.1% Never 29.1% 45.5% Nocturnal cough 51.5% 40% NS (0.09) NS: Not Significant Table 3: Comparing the presenting symptoms between men and women before therapy http://www.biomedcentral.com/1471-230X/6/12 http://www.biomedcentral.com/1471-230X/6/12 BMC Gastroenterology 2006, 6:12 The frequency of chest pain experienced by men and women had no significant relationship with the type of pretreatment therapy they had received (P values > 0.05). tionnaire could be more informative in term of finding differences between men and women. tionnaire could be more informative in term of finding differences between men and women. It has been suggested that gender and age affect the clini- cal presentation of achalasia [13-15]. A study by d' Altero- che et al. showed that chest pain was 1.7 times more frequent in women than in men, but no explanation was suggested by the authors [14]. Occurrence of chest pain did not relate to the sympto- matic duration before treatment in both sexes; mean symptom duration in women with (n = 73) and without chest pain (n = 30) was 4.5 ± 3.6 and 5.5 ± 6.2 yrs respec- tively (P = 0.3), while 4.5 ± 4.7 and 4.7 ± 6.4 yrs in 60 men with and 50 men without chest pain (P = 0.8). In the present study, we found chest pain is the only symptom that differs between men and women; it is more common in women before or after treatment. The occur- rence of chest pain did not relate to the LESP in both sexes. No significant relation was found between chest pain and other symptoms such as solid dysphagia, liquid dys- phagia, active and passive regurgitation (P > 0.05 in all cases). Results Some previous studies have tried to figure out the under- lying causes of chest pain in achalasia patients. Eckardt et al evaluated achalasia patients with and without com- plaining chest pain and found that the occurrence of this symptom was unrelated to LESP. In addition, They showed that repetitive esophageal contractions were more common in patients experiencing chest pain, although this finding barely reached the specific cut off [16]. Another study found that 38% of patients with reappear- ance of peristalsis waves in manometry performed 1 year after dilation experienced the symptom of chest pain, while only 6% of those with esophageal aperistalsis com- plained about this symptom. In addition, post dilation peristalsis was strongly correlated to higher contraction waves amplitude before treatment [6]. Chest pain was less frequently reported by patients over 56 yrs of age in comparison to those less than 56 yrs (42.8% vs. 65.5% P = 0.02). After treatment In total, 171 (80.5%) patients underwent pneumatic dila- tation and 42 (19.5%) underwent botulinum toxin injec- tion; 62.4% and18.8 % of patients underwent one and two sessions of PD, and 12.2% and 5.6% of the cases received one and two BT injections respectively. Symptoms of the patients were evaluated again at 6 months after treatment, as depicted in table 4. Frequency of chest pain after treatment did not relate to the type of therapy that achalasia patients received (P = 0.4). There was no significant relation between chest pain and other symptoms such as solid dysphagia, liquid dysphagia, active and passive regurgitation (all P values > 0.05). Results of Dantas et al. study on normal subjects could help to explain the observed difference in frequency of chest pain between men and women in our series. They performed manometry on 20 men and 20 women with the same age distribution. There was no difference in LESP between men and women, but a higher duration of con- tractions 5 cm above the LES was observed in women (women: 4.5 ± 0.3s men: 3.7 ± 0.2s p < 0.05). On the other hand no gender difference in esophageal functional anatomy or innervations has been described [9]. The occurrence of chest pain significantly reduced after treatment in both sexes (P = 0.0001) but it was still higher among women (32% vs. 20.9% P = 0.04). During 6 months after treatment, the occurrence of chest pain in patients over 56 was significantly lower in com- parison to younger group; 7.1% vs. 29.2% P = 0.01 All these data suggest that higher occurrence of chest pain among women may be interrelated to higher amounts or amplitude of contractions in this group. Page 4 of 6 (page number not for citation purposes) Discussion is more frequent in women but also responds less favora- ble to the current treatments. It has been previously sug- gested that these patients could also benefit from accompanying behavioral and emotional treatments to reduce their pain [22,23]. Although around 70% of patients in our study reported some extent of solid dys- phagia after therapy, the total dysphagia score has signifi- cantly decreased after treatment and only in 18% of patients the dysphagia score was greater than 1. Previous studies also pointed that chest pain decreases sig- nificantly with ageing [14,16]. Clouse and Simmons reported that not only chest pain is less frequent in elderly subjects but also the pain is less sever in this age group[13,19]. It has been shown that there is no difference in LESP and other manometric fea- tures in older comparing younger achalasia patients except the lower LES residual pressure in elderly [13,20]. The lower LES residual pressure along with lesser visceral perception in elderly were considered to explain lesser expression of chest pain in the older achalasia patients [16]. Many efforts have been done in treatment of achalasia during the past decade, but the affected patients still suffer from chest pain even after otherwise successful treatment. Timed barium swallow as an objective means of assessing esophageal emptying would show whether the continued chest pain in achalasia patients after treatment is associ- ated with unfavorable esophageal emptying. Studies to clarify the underlying mechanism of this symptom would help us to perform more efficient managements for acha- lasia. In agreement with previous studies, our results showed that the occurrence and intensity of chest pain was unre- lated to the intensity of other esophageal symptoms [16]. This was the case in both sexes, either before or after treat- ment. In agreement with previous studies, our results showed that the occurrence and intensity of chest pain was unre- lated to the intensity of other esophageal symptoms [16]. This was the case in both sexes, either before or after treat- ment. Conclusion It seems that chest pain is the only symptom of achalasia which is affected by sex as well as age. The duration of ill- ness, lower esophageal sphincter pressure and the type of therapeutic procedure do not relate to the occurrence of chest pain in achalasia and can not explain its higher occurrence in women. The relation between chest pain and sex, observed in this study needs to be explored in detail to understand the physiologic- based content of this symptom in achalasia. There was no significant relationship between symptom duration and chest pain in our study. Rakita et al. and Alteroche et al. also found that duration of symptoms had little impact on premyotomy symptom frequency [14,15]. However, results of the other study showed that patients with chest pain had a shorter duration of symptoms before the diagnosis of achalasia. The authors suggested that more pronounced disease in patients with chest pain could call for sooner investigation to receive a diagnosis at an earlier stage [16]. Competing interests The author(s) declare that they have no competing inter- ests. Some studies have indicated that a successful treatment improves chest pain but rarely makes this symptom abol- ish [1,16,21,22]. We found that occurrence of chest pain decreased after therapeutic session in both sexes, but remained still higher among women. The occurrence of other symptoms was not different between men and women after therapy. This shows that not only chest pain Discussion We evaluated the clinical features of achalasia in relation to the demographic aspects of the affected individuals, studying the consecutively enrolled achalasia patients in our center. Potential epidemiological limitations of stud- ies on rare diseases as achalasia include inadequate sam- ple size and the consequent selective presentation of results due to multiple comparisons. Regarding to these concerns, we avoided multilayer comparisons and tried to limit the number of comparisons in order to be supported by our sample size. We have tried to evaluate the clinical profile of our patients in detail, but more extensive ques- In addition, women are more likely to experience recur- rent pains or report more intense and longer duration pain comparing to men [17,18]. This might be due to some biological and psychological differences between these two groups which can lead to lower pain tolerance and threshold in women [18]. Thus the different pain expression in women could also play a role in the differ- ence seen in the chest pain between two genders in acha- lasia. Page 4 of 6 (page number not for citation purposes) Page 4 of 6 (page number not for citation purposes) BMC Gastroenterology 2006, 6:12 http://www.biomedcentral.com/1471-230X/6/12 http://www.biomedcentral.com/1471-230X/6/12 BMC Gastroenterology 2006, 6:12 Table 4: Comparing the presenting symptoms between men and women after treatment Variable Women (n= 103, 48.4%) Men (n = 110, 51.6%) P Value Mean post treatment total score (SD) 2.4 (2.2) 2.7 (2.7) NS (0.3) Post solid dysphagia 68.9% 76.4% NS (0.1) Post liquid dysphagia 25.2% 33.6% NS (0.09) Post active regurgitation 12.6% 15.5% NS (0.5) Post passive regurgitation 7.8% 8.2% NS (0.9) Post chest pain Daily 1% 32% 0% 20.9% 0.04 Weekly 11.7% 5.4% Monthly 19.3% 15.5% Never 68% 79.1% NS: Not Significant Table 4: Comparing the presenting symptoms between men and women after treatment We found that chest pain is conversely related to the age and is less frequent in elderly. is more frequent in women but also responds less favora- ble to the current treatments. It has been previously sug- gested that these patients could also benefit from accompanying behavioral and emotional treatments to reduce their pain [22,23]. Although around 70% of patients in our study reported some extent of solid dys- phagia after therapy, the total dysphagia score has signifi- cantly decreased after treatment and only in 18% of patients the dysphagia score was greater than 1. All authors read and approved the final manuscript. 23. Clouse RE, Lustman PJ: Psychiatric illness and contraction abnormalities of the esophagus. N Engl J Med 1983, 309:1337-1342. http://www.biomedcentral.com/1471-230X/6/12 BMC Gastroenterology 2006, 6:12 FF conducted the study, interviewed the patients, ana- lyzed the data, and drafted the manuscript. 17. Philpott S, Boynton PM, Feder G, Hemingway H: Gender differ- ences in descriptions of angina symptoms and health prob- lems immediately prior to angiography: the ACRE study. Appropriateness of Coronary Revascularisation study. Soc Sci Med 2001, 52:1565-1575. FB assisted in conducting the study and interviewing the patients and revised the manuscript. 18. Unruh AM: Gender variations in clinical pain experience. Pain 1996, 65:123-167. 19. Simmons DB, Schuman BM, Griffin JWJ: Achalasia in patients over 65. J Fla Med Assoc 1997, 84:101-103. MM assisted in interviewing the patients and helped in revising the manuscript. MM assisted in interviewing the patients and helped in revising the manuscript. 20. Ribeiro AC, Klingler PJ, Hinder RA, DeVault K: Esophageal man- ometry: a comparison of findings in younger and older patients. Am J Gastroenterol 1998, 93:706-710. p J 21. Lai IR, Lee WJ, Huang MT: Laparoscopic Heller myotomy with fundoplication for achalasia. J Formos Med Assoc 2002, 101:332-336. RM supervised the study scientifically, has been involved in preparing the manuscript and has given final approval of the version to be published. 22. Shabsin HS, Katz PO, Schuster MM: Behavioral treatment of intractable chest pain in a patient with vigorous achalasia. Am J Gastroenterol 1988, 83:970-973. Authors' contributions JM visited the patients and performed endoscopies and therapeutic procedures. Page 5 of 6 (page number not for citation purposes) Page 5 of 6 (page number not for citation purposes) Page 5 of 6 (page number not for citation purposes) Pre-publication history This study was supported by local funds from DDRC with no external financial support. The pre-publication history for this paper can be accessed here: The pre-publication history for this paper can be accessed here: References 1. Clouse R DN: Esophageal motor and sensory function and motor disorders of the esophagus. In Sleisenger & fordtran's Gas- trointestinal and liver disease 7th ed, edition. Edited by: Feldman M SBFSMHFJS. Philadelphia, Saunders; 2002:577-578. http://www.biomedcentral.com/1471-230X/6/12/pre pub http://www.biomedcentral.com/1471-230X/6/12/pre pub J p 2. Cohen S: Motor disorders of the esophagus. N Engl J Med 1979, 301:184-192. 3. Vantrappen G, Hellemans J: Treatment of achalasia and related motor disorders. Gastroenterology 1980, 79:144-154. 3. Vantrappen G, Hellemans J: Treatment of achalasia and related motor disorders. Gastroenterology 1980, 79:144-154. 4. Birgisson S, Richter JE: Achalasia: what's new in diagnosis and treatment? Dig Dis 1997, 15 Suppl 1:1-27.:1-27. gy 4. Birgisson S, Richter JE: Achalasia: what's new in diagnosis and treatment? Dig Dis 1997, 15 Suppl 1:1-27.:1-27. g , pp 5. Howard PJ, Maher L, Pryde A, Cameron EW, Heading RC: Five year prospective study of the incidence, clinical features, and diagnosis of achalasia in Edinburgh. Gut 1992, 33:1011-1015. g g 6. Papo M, Mearin F, Castro A, Armengol JR, Malagelada JR: Chest pain and reappearance of esophageal peristalsis in treated acha- lasia. Scand J Gastroenterol 1997, 32:1190-1194. J 7. Vaezi MF, Richter JE: Current therapies for achalasia: compari- son and efficacy. J Clin Gastroenterol 1998, 27:21-35. y J 8. Spechler SJ: Clinical manifestations of achalasia. In Up To Date Edited by: Rose BDE. Wellesley, MA, Up To Date; 2005. y y p 9. Dantas RO, Ferriolli E, Souza MA: Gender effects on esophageal motility. Braz J Med Biol Res 1998, 31:539-544. y J 10. Mikaeli J, Bishehsari F, Montazeri G, Yaghoobi M, Malekzadeh R: Pneumatic balloon dilatation in achalasia: a prospective comparison of safety and efficacy with different balloon diameters. Aliment Pharmacol Ther 2004, 20:431-6. 11. Yaghoobi M, Mikaeli J, Montazeri G, Nouri N, Sohrabi MR, Malekza- deh R: Correlation between clinical severity score and the lower esophageal sphincter relaxation pressure in idiopathic achalasia. Am J Gastroenterol 2003, 98:278-283. J 12. Annese V, Bassotti G, Coccia G, D'onofrio V, Gatto G, Repici A, Andriulli A: Comparison of two different formulations of bot- ulinum toxin A for the treatment of oesophageal achalasia. The Gismad Achalasia Study Group. Aliment Pharmacol Ther 1999, 13:1347-1350. References Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Page 6 of 6 (page number not for citation purposes) Publish with BioMed Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp BioMedcentral Publish with BioMed Central and every scientist can read your work free of charge 13. Clouse RE, Abramson BK, Todorczuk JR: Achalasia in the elderly. Effects of aging on clinical presentation and outcome. Dig Dis Sci 1991, 36:225-228. 14. d'Alteroche L, Oung C, Fourquet F, Picon L, Lagasse JP, Metman EH: Evolution of clinical and radiological features at diagnosis of achalasia during a 19-year period in central France. Eur J Gas- troenterol Hepatol 2001, 13:121-126. p 15. Rakita S, Bloomston M, Villadolid D, Thometz D, Boe B, Rosemurgy A: Age affects presenting symptoms of achalasia and out- comes after myotomy. Am Surg 2005, 71:424-429. y y g 16. Eckardt VF, Stauf B, Bernhard G: Chest pain in achalasia: patient characteristics and clinical course. Gastroenterology 1999, 116:1300-1304.
https://openalex.org/W4361246762	https://aacr.figshare.com/ndownloader/files/39815646	English	null	Supplementary Legends for Figures 1-3, Tables 1-5 from Protein Phosphatase 2A and Rapamycin Regulate the Nuclear Localization and Activity of the Transcription Factor GLI3	null	2,023	cc-by	406	Suppl. Fig. 1: Expression of SHH pathway members in tumor cell lines. Semi-quantitative RT-PCR analysis of cyclin D1 (upper panel), PTC1 (lower panel) and GLI3 (middle panel) expression in different tumor cell lines. Two melanoma cell lines (MeWo and SKMel29), three basal cell carcinoma (BCC-1, BCC-3 and BCC-5), small cell carcinoma derived from the adrenal gland/cortex (SW13), neuroblastoma (SH-SY5Y), glioblastoma-astrocytoma (U373MG) and HeLa with or without SHH stimulation / inhibition (cyclopamine) were used. RT-PCR results of the respective genes are shown in relation to levels of 18S RNA. Suppl. Fig. 1: Expression of SHH pathway members in tumor cell lines. Semi-quantitative RT-PCR analysis of cyclin D1 (upper panel), PTC1 (lower panel) and GLI3 (middle panel) expression in different tumor cell lines. Two melanoma cell lines (MeWo and SKMel29), three basal cell carcinoma (BCC-1, BCC-3 and BCC-5), small cell carcinoma derived from the adrenal gland/cortex (SW13), neuroblastoma (SH-SY5Y), glioblastoma-astrocytoma (U373MG) and HeLa with or without SHH stimulation / inhibition (cyclopamine) were used. RT-PCR results of the respective genes are shown in relation to levels of 18S RNA. Suppl. Fig. 3: Several SHH targets genes do not respond on GLI3 knock down Relative mRNA amounts of different SHH targets after transfection with nonsilencing control siRNA (white columns) or with GLI3-specific siRNA (gray columns). Columns represent mean values of 4 samples measured in parallel ± s.d. GAPDH was used for normalization. Efficiency of the GLI3 knock down is shown in Fig. 5. Suppl. Fig. 2: Subcellular localization after wortmannin treatment Nuclear localization of overexpressed GFP-GLI3 in HeLa cells without (white or with treatment with 0.1 µM wortmannin (gray column) as determined by fluorescence microscopy. Visualization and scoring were performed as described in the legend to Fig. 1. Suppl. Table 1b Suppl. Table 2 List of SHH target genes that were not detectable in HeLa cells by means of real-time PCR. Suppl. Table 4 Suppl. Table 1a pp Summary of the subcellular localization of GLI3 with different tags in HeLa cells after diverse treatments. Visualization and scoring were performed as described in the legend to Fig. 1. Suppl. Table 1b Summary of the subcellular localization of GLI3 with different tags after diverse treatments in U373MG, MeWo, and SKMel29 cells. Visualization and scoring were performed as described in the legend to Fig. 1. Suppl. Table 4 List of primers used for real-time PCR Suppl. Table 5 List of primers used for RT-PCR Suppl. Table 5 List of primers used for RT-PCR
https://openalex.org/W3115979224	https://pure.rug.nl/ws/files/154957452/microorganisms_09_00007_v3.pdf	English	null	Growth Inhibition by Amino Acids in Saccharomyces cerevisiae	Microorganisms	2,020	cc-by	12,935	University of Groningen University of Groningen Growth Inhibition by Amino Acids in Saccharomyces cerevisiae Ruiz, Stephanie J; van 't Klooster, Joury S; Bianchi, Frans; Poolman, Bert IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publication date: 2020 Publication date: 2020 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Ruiz, S. J., van 't Klooster, J. S., Bianchi, F., & Poolman, B. (2020). Growth Inhibition by Amino Acids in Saccharomyces cerevisiae. Microorganisms, 9(1), 1-17. Article 9010007. https://doi.org/10.3390/microorganisms9010007 Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Stephanie J. Ruiz , Joury S. van ’t Klooster, Frans Bianchi and Bert Poolman * Department of Biochemistry, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands; s.j.ruiz@outlook.com (S.J.R.); j.s.van.t.klooster@rug.nl (J.S.v.t.K.); f.bianchi@rug.nl (F.B.) * Correspondence: b.poolman@rug.nl; Tel.: +31-50-363-4190; Fax: +31-50-363-4165 Abstract: Amino acids are essential metabolites but can also be toxic when present at high levels intracellularly. Substrate-induced downregulation of amino acid transporters in Saccharomyces cerevisiae is thought to be a mechanism to avoid this toxicity. It has been shown that unregulated uptake by the general amino acid permease Gap1 causes cells to become sensitive to amino acids. Here, we show that overexpression of eight other amino acid transporters (Agp1, Bap2, Can1, Dip5, Gnp1, Lyp1, Put4, or Tat2) also induces a growth defect when speciﬁc single amino acids are present at concentrations of 0.5–5 mM. We can now state that all proteinogenic amino acids, as well as the important metabolite ornithine, are growth inhibitory to S. cerevisiae when transported into the cell at high enough levels. Measurements of initial transport rates and cytosolic pH show that toxicity is due to amino acid accumulation and not to the inﬂux of co-transported protons. The amino acid sensitivity phenotype is a useful tool that reports on the in vivo activity of transporters and has allowed us to identify new transporter-speciﬁc substrates. Keywords: amino acid transport; amino acid toxicity; growth inhibition; Saccharomyces cerevisiae As well as being the building blocks for proteins, amino acids provide raw materials for energy generation, nitrogen metabolism, and the biosynthesis of structural, signaling, or defensive compounds [1]. Although they are essential metabolites, it has been known for over half a century that the addition of excess amino acids to both prokaryotic and eukary- otic cell cultures can cause growth inhibition and/or cell death [2–9]. The dependence of toxicity on transport activity indicates that the effect is exerted intracellularly [10,11]. In hu- mans, several inherited metabolic diseases are caused by elevated levels of amino acids and/or closely related metabolites, with perhaps the most well-known being phenylke- tonuria [12]. Citation: Ruiz, S.J.; van ’t Klooster, J.S.; Bianchi, F.; Poolman, B. Growth Inhibition by Amino Acids in Saccharomyces cerevisiae. Microorganisms 2021, 9, 7. https://dx.doi.org/10.3390/ microorganisms9010007 Received: 3 December 2020 Accepted: 18 December 2020 Published: 22 December 2020 Received: 3 December 2020 Accepted: 18 December 2020 Published: 22 December 2020 Some amino acids are known to cause toxicity via speciﬁc mechanisms. Valine and phenylalanine inhibit the growth of bacterial cells by repressing enzymes involved in synthesis of isoleucine and tyrosine, respectively [13,14]. Histidine toxicity in yeast cells is caused by a reduction in copper availability in vivo [11]. There is also evidence for a more general mechanism involving the target of rapamycin complex 1 (TORC1), a master controller of cell metabolism that is conserved among eukaryotes and responds to various nutrient stimuli, including amino acids [15–17]. Deregulation of TORC1 signaling is linked to many human diseases [18], and it has been recently shown that in some cancer cell lines, this deregulation is caused by abnormal amino acid transport [19–21]. Phenylalanine sensitivity in a mammalian cell line, and rescue by valine, has been shown to involve mTORC1 [22]. All amino acids except glutamine are growth inhibitory to the plant Nico- tiana silvestris, and glutamine itself can rescue some (but not all) of the inhibition caused by other amino acids [23–25]. Glutamine is now known to be an important activator of TORC1 [26–28]. Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institutional afﬁliations. Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institutional afﬁliations. Copyright: © 2020 by the authors. Li- censee MDPI, Basel, Switzerland. Copyright strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of th or copyright holder(s), unless the work is under an open content license (like Creative Commons). The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date: 24-10-2024 microorganisms Name Transport Substrates Reference(s) * Agp1 His, Asp, Glu, Ser, Thr, Asn (0.29 mM), Gln (0.79 mM), Cys, Gly, Pro, Ala, Val, Ile (0.6 mM), Leu (0.16 mM), Met, Phe (0.6 mM), Tyr, Trp [42–46] Bap2 Cys, Ala, Val, Ile, Leu (37 µM), Met, Phe, Tyr, Trp [46–49] Can1 His, Arg (10–20 µM), Lys (150–250 µM), Orn [50,51] Dip5 Glu (48 µM), Asp (56 µM), Ser, Asn, Gln, Gly, Ala [52] Gnp1 Ser, Thr, Asn, Gln (0.59 mM), Cys, Pro, Leu, Met [44,45,53] Lyp1 Lys (10–25 µM), Met [50,51,54,55] Put4 Gly, Pro, Ala [45,56–58] Tat2 Gly, Ala, Phe, Tyr, Trp, Cys [44,59] * All proteins in this table were studied in Regenberg et al. [60]. * All proteins in this table were studied in Regenberg et al. [60]. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/ licenses/by/4.0/). https://www.mdpi.com/journal/microorganisms Microorganisms 2021, 9, 7. https://dx.doi.org/10.3390/microorganisms9010007 2 of 17 Microorganisms 2021, 9, 7 Wild-type yeast cells are able to synthesize all amino acids de novo but can also scavenge them from the environment using a host of broad- and narrow-range transporters located in the plasma membrane [29,30]. Saccharomyces cerevisiae amino acid permeases have been studied extensively not only in terms of their transport activity but also their regu- lation. Many are transcriptionally up- or downregulated by three interconnected pathways that respond to the availability of amino acids and other nitrogen sources: nitrogen catabo- lite repression (NCR), general amino acid control (GAAC), and SPS-signaling [29,31–33]. Some transporters are also post-translationally regulated in response to the external con- centration of their substrates e.g., Can1 (Arg), Dip5 (Glu), Gap1 (various), Lyp1 (Lys), Mup1 (Met), and Tat2 (Trp) [34–38]. In this pathway, substrate binding triggers transporter ubiquitination and subsequent removal of the proteins from the plasma membrane via en- docytosis [39]. This causes a decrease in transport activity and thus limits the accumulation of certain substrates. Similar regulation has been observed for Ptr2, which imports di- and tri-peptides that can be broken down into free amino acids inside the cell [40]. It has been suggested that this is a mechanism to avoid amino acid-induced toxicity [40,41]. Risinger et al. [41] showed that S. cerevisiae cells expressing Gap1K9R,K16R, a ubiquitination- and endocytosis-deﬁcient mutant that is constitutively localized to the plasma membrane, experience severe growth defects when individual amino acids are added to the medium at high concentrations (3 mM). The same effect, with varying levels of severity, was triggered by the metabolite citrulline as well as all proteinogenic amino acids except alanine. We decided to investigate the amino acid sensitivity of strains constitutively overexpressing eight different narrow- and broad-range amino acid transporters: Agp1, Bap2, Can1, Dip5, Gnp1, Lyp1, Put4, or Tat2 (Table 1). Table 1. Transporters used in this study and their reported substrates. Values in brackets indicate published Michaelis constants (Km). Table 1. Transporters used in this study and their reported substrates. Values in brackets indicate published Michaelis constants (Km). published Michaelis constants (Km). 2.1. Strains and Growth Conditions 2.1. Strains and Growth Conditions Escherichia coli strain MC1061 (E. coli Genetic Stock Center) was used for cloning and plasmid storage. Experiments were performed with S. cerevisiae strains BY4709 (MATα ura3∆0) and BY4741 (MATa his∆1 leu2∆0 met15∆0 ura3∆0) [61]. BY4709 was used for growth experiments and transport assays, while BY4741 was used for measurements of the internal pH. p All experiments were done in YNBD, a minimal medium containing 6.9 g/L Yeast Nitrogen Base without amino acids (FormediumTM, Norfolk, UK), 20 g/L D-glucose (Sigma-Aldrich, St. Louis, MO, USA), and 100 mM of potassium phosphate (pH 6.0). Single amino acids or a standard synthetic complete (SC) mixture (Kaiser Drop-out minus uracil, FormediumTM) were added for growth assays. For pHluorin experiments a low ﬂuorescence version of YNBD was made using Yeast Nitrogen Base without amino acids and without folic acid and riboﬂavin (FormediumTM), and methionine and histidine were Microorganisms 2021, 9, 7 3 of 17 added at 76 mg/L each (509 and 490 µM, respectively). For growth assays using Arg and Lys at concentration ≤500 µM, the YNBD did not contain potassium phosphate buffer, but the pH was set to 6.0 using HCl/NaOH before sterilization by ﬁltration. All media contain ammonium sulfate (5 g/L) as nitrogen source. g g Agar was added at 20 g/L for solid cultures. Liquid cultures (5–15 mL) were inoculated from single colonies on agar plates and incubated in 50 mL CELLreactorTM ﬁlter top tubes (Greiner Bio-One, Kremsmünster, Austria) at 30 ◦C with shaking (~200 rpm). 2.3. Microscopy Live cell imaging was performed on a LSM 710 commercial laser scanning confocal microscope (Zeiss, Oberkochen, Germany) equipped with a C-Apochromat 40×/1.2 NA objective. Cells were immobilized between a glass slide and coverslip. Images were obtained with the focal plane positioned at the mid-section of the cells. For ﬂuorescence images, samples were excited with a blue argon ion laser (488 nm) and emission collected at 493–797 nm. 2.2. Plasmids Name Sequence (5′ to 3′) Description 1 5273 GGAGGGGAAAATTTATATTTTCAAGGTTC pFB022 (F) 5954 CATTTTGGGATCCACTAGTTCTAG pFB022 (R) 5480 TCTAGAACTAGTGGATCCCAAAATGTCGTCGTCGAAGTCTC AGP1 (F) 5481 TCTAGAACTAGTGGATCCCAAAATGCTATCTTCAGAAGATTTTGGATC BAP2 (F) 5561 TCTAGAACTAGTGGATCCCAAAATGGGAACAAATTCAAAAGAAG CAN1 (F) 5482 TCTAGAACTAGTGGATCCCAAAATGAAGATGCCTCTAAAGAAGATG DIP5 (F) 5483 TCTAGAACTAGTGGATCCCAAAATGACGCTTGGTAATAGACGC GNP1 (F) 5484 TCTAGAACTAGTGGATCCCAAAATGGTAAATATACTGCCCTTCC PUT4 (F) 5485 TCTAGAACTAGTGGATCCCAAAATGACCGAAGACTTTATTTCTTCTG TAT2 (F) 5486 ACCTTGAAAATATAAATTTTCCCCTCCACACCAGAAGGCAACGAC AGP1 (R) 5487 ACCTTGAAAATATAAATTTTCCCCTCCACACCAGAAATGATAAGCTTTTCTC BAP2 (R) 5562 ACCTTGAAAATATAAATTTTCCCCTCCTGCTACAACATTCCAAAATTTG CAN1 (R) 5488 ACCTTGAAAATATAAATTTTCCCCTCCGAAGATATTACCCAAAAATTTTTCATAG DIP5 (R) 5489 ACCTTGAAAATATAAATTTTCCCCTCCACACCAGAAATCAAGAACTCTTTTC GNP1 (R) 5490 ACCTTGAAAATATAAATTTTCCCCTCCCAACAAGGCGTCCAAGAAC PUT4 (R) 5491 ACCTTGAAAATATAAATTTTCCCCTCCACACCAGAAATGGAACTGTCTC TAT2 (R) 4258 ACCACCACCAUCATCATCATCATTAACTGCAGGAATTC pFB023-A (F) 3631 AGCACTACCCUTTAGCTGTTCTATATGCTGCC pFB023-A (R) 3630 AGGGTAGTGCUGAAGGAAGCATACGATACCC pFB023-B (F) 5159 ATTTTGGGAUCCACTAGTTCTAGAGCGGCCAGCTTGGAGTTGATTG pFB023-B (R) 5087 ATCCCAAAAUGCATGGGTCATTGCAAGGTGG LYP1(62-590) (F) 5089 ATGGTGGTGGUGTCCCCCTCCTTCGATTTCTCTTCTGTCGGAATC LYP1(62-590) (R) 1 F = forward primer, R = reverse primer. Ratiometric pHluorin [67] was expressed from pYES2-PACT1-pHluorin [66]. For these experiments a truncated Lyp1 with no ﬂuorescent tag was used [67]. First, the C-terminal YPet tag was removed from pFB023 using uracil excision-based cloning [68]. Primer pairs 4158/3631, 3630/5159 and 5087/5089 (Table 3) and the PfuX7 polymerase [69] were used to amplify pFB023 in three fragments excluding the YPet coding region. The crude PCR products were treated with USERTM enzyme (New England Biolabs, Ipswich, MA, USA) as per the manufacturer’s instructions and transformed into E. coli MC1061 for in vivo assembly. After the resulting plasmid (pSR053) was conﬁrmed by sequencing of the entire coding region, the area containing LYP1(62-590) and the ﬂanking promoter and terminator sequences were sub cloned into pRSII425 using SacI/KpnI digestion to yield plasmid pSR057. The expressed protein (from N- to C-terminus) includes a starting Met, residues 62-590 of Lyp1, a short three-residue glycine linker, and an eight-residue His epitope. 2.2. Plasmids The plasmids used in this study are listed in Table 2. Vectors for the constitutive expression of transporters are based on pFB022 and pFB023 [62]. These are pRS426 deriva- tives (URA3, 2µ) that allow for expression of ﬂuorescently-tagged Lyp1, either full length or truncated, from the constitutive ADH1 promoter [63]. The C-terminal tag contains a TEV protease recognition site, followed by the ﬂuorescent protein YPet [64] and an eight-residue His epitope (for the full sequence, see the Supplementary). Plasmids pSR045–051 are iden- tical to pFB022 except for replacement of LYP1 with other transporter genes as indicated in Table 1. They were constructed by in vivo homologous recombination using crude PCR products. The plasmid backbone was ampliﬁed from pFB022 [62], using primers binding immediately upstream and downstream of the LYP1 coding region. The reaction was then treated with DpnI (as per the manufacturer’s instructions) to remove the template DNA. Each target gene was PCR ampliﬁed from S. cerevisiae BY4742 chromosomal DNA, using for- ward and reverse primers that added approximately 25 bp of sequence homologous to the plasmid backbone (Table 3). The backbone and insert were simultaneously transformed into BY4709 and positive transformants screened by growth on uracil dropout media, colony PCR and ﬂuorescence microscopy. Fusion genes were conﬁrmed by sequencing of the entire open reading frame. A single base pair mutation (T to C) was observed at the end of the ADH1 promoter, but this did not appear to affect expression. Table 2. Plasmids used in this study. Table 2. Plasmids used in this study. Name Description Source pRSII425 LEU2, 2µ (multicopy) [65] pRSII426 URA3, 2µ (multicopy) [65] pFB022 pRS426 (URA3, 2µ) derivative with Lyp1YPet under ADH1 promoter [62] pFB023 pFB022 derivative containing Lyp1(62-590)YPet [62] pSR053 pFB023 derivative containing Lyp1(62-590) This study pSR045 pFB022 derivative containing Agp1YPet This study pSR046 pFB022 derivative containing Bap2YPet This study pSR047 pFB022 derivative containing Dip5YPet This study pSR048 pFB022 derivative containing Gnp1YPet This study pSR049 pFB022 derivative containing Put4YPet This study pSR050 pFB022 derivative containing Tat2YPet This study pSR051 pFB022 derivative containing Can1YPet This study pSR057 pRSII425 derivative containing Lyp1(62-590) under ADH1 promoter This study pYES2-PACT1-pHluorin pYES2 (URA3, 2µ) derivative with pHluorin under ACT1 promoter [66] 4 of 17 Microorganisms 2021, 9, 7 Table 3. Oligonucleotide primers used in PCR ampliﬁcation. 2.4. Growth Assays Growth assays were performed using 120 µL liquid cultures in CELLSTAR® 96-well ﬂat-bottom microplates (Greiner Bio-One). Each plate was covered with a Breathe-Easy® sealing membrane (Sigma-Aldrich), as well as the provided lid, and incubated in a 30 ◦C room at 400 rpm on an Excella™E1 benchtop open-air shaker (Eppendorf, Hamburg, Ger- many). OD600 measurements were made in a PowerWave 340 spectrophotometer (BioTek, Winooski, VT, USA) without the microplate lid but with the (optically clear) membrane. p p y To prepare inocula for microplate experiments, strains were cultured in YNBD for approximately 24 h, with one round of dilution, to an OD600 of 0.3–0.7. Each culture was Microorganisms 2021, 9, 7 5 of 17 5 of 17 diluted in fresh media to an OD600 of 0.1 and 60 µL aliquots were added to microplate wells containing 60 µL of media with or without either single amino acids or the SC mixture. The ﬁnal concentrations are given in Table 4 and the relevant ﬁgure legends. All measurements were done in biological triplicate (three independent inoculations made on different days from different pre-cultures). Table 4. Concentration (mM) of components in synthetic complete (SC) mixture, when used at 1x. Amino Acids (3-letter, 1-letter code) Alanine (Ala, A) 0.853 Arginine (Arg, R) 0.361 Asparagine (Asn, N) 0.575 Aspartic acid (Asp, D) 0.571 Cysteine (Cys, C) 0.627 Glutamine (Gln, Q) 0.520 Glutamate (Glu, E) 0.517 Glycine (Gly, G) 1.012 Histidine (His, H) 0.490 Isoleucine (Ile, I) 0.579 Leucine (Leu, L) 2.897 Lysine (Lys, K) 0.520 Methionine (Met, M) 0.509 Phenylalanine (Phe, F) 0.460 Proline (Pro, P) 0.460 Serine (Ser, S) 0.723 Threonine (Thr, T) 0.638 Tryptophan (Trp, W) 0.372 Tyrosine (Tyr, Y) 0.419 Valine (Val, V) 0.649 Other Adenine 0.133 myo-Inositol 0.422 4-Aminobenzoic acid 0.058 Table 4. Concentration (mM) of components in synthetic complete (SC) mixture, when used at 1x. Raw OD600 values were background corrected by subtracting the average value for all blank (media only) wells from the same plate (n = 12–24). This observed value (ODobs) was then corrected for non-linearity at higher cell densities using the polynomial equation: ODcor = 0.319 × ODobs3 + 0.089 × ODobs2 + 0.959 × ODobs (Figure S1); this correction method was previously discussed in Warringer and Blomberg [70]. ODcor values were then normalized within each strain and replicate. Three independent experiments were performed but, due to technical error, some results had to be discarded. 2.5. Transport Assays In vivo transport assays were performed as previously described [62]. The radioac- tive substrates used were L-(14C(U))-phenylalanine, L-(14C(U))-lysine and L-(methyl-3H)- methionine (PerkinElmer, USA) and L-(14C(U))-alanine (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK). Transport was assayed at the following concentrations: 50 µM lysine with and without 100 mM of histidine or ornithine, 500 µM alanine or methionine, and 2.5 mM phenylalanine. 2.4. Growth Assays For this reason, n = 2 for Bap2 1 mM Gly/His/Ile/Leu/Lys/Met/Orn/Phe, and also for all Tat2 1 mM and 0.5 mM conditions. 2.6. Measurement of Cytosolic pH Strains were cultured in liquid media to an OD600 of 0.3–0.7, diluted to an OD600 of 0.2 in pre-warmed media, and then immediately transferred to the spectrophotometer sample holder (also pre-warmed). After 15 min, 10 µL of either distilled water or 100 mM lysine was added (ﬁnal concentration 500 µM) and the measurement continued for another 3 h. Microorganisms 2021, 9, 7 6 of 17 Fluorescence measurements were made using a JASCO FP-8300 ﬂuorescence spec- trophotometer with the following settings: sensitivity, high; response, 0.1 s; excitation/ emission bandwidths, 5 nm; emission, 508 nm; excitation, 355–495 nm in 5 nm steps. All samples were 2 mL in 4.5 mL plastic cuvettes (catalogue number 1961, Kartell Labware, Noviglio, Italy) with a magnetic bar for stirring. The temperature of the sample holder was maintained at 20 ◦C for calibration measurements and 30 ◦C for experiments with growing cells. To prepare calibration samples, strains were cultured as above. Cells were then washed twice and resuspended in 5 volumes of ice-cold PBS (10 mM Na2HPO4, 137 mM NaCl, 2.7 mM KCl, 1.8 mM KH2PO4, pH 7.35). Digitonin was added to a ﬁnal concentration of 0.02% w/v (from a 2% w/v stock in PBS) and the samples incubated for 1 h at 30 ◦C with mixing. Cells were then washed twice with 1 mL of ice-cold PBS, resuspended to 0.25 mg/mL with the same (1 µL added per mg wet weight of cell pellet) and kept on ice. 10 µL of cells was added to 2 mL of room-temperature buffer (either 100 mM potassium phosphate or PBS) and a ﬂuorescence measurement made once per minute for ten minutes. p p p We found that background subtraction was very important, especially for long- term measurements with growing cells. Strains carrying pHluorin/pRSII425 or pHlu- orin/ Lyp1(62-590) were compared to strains carrying pRSII426/pRSII425 and pRSII426/ Lyp1(62-590), respectively. To generate the calibration curve (Figure S2), the median emission intensities at 395 and 475 nm excitation (I395 and I475) were background subtracted, and the ratio (R395/475) plotted against the pH. For measurements with growing cells, the I395 and I475 values from each strain and timepoint were background subtracted using individual values from the corresponding strain and timepoint. 3. Results and Discussion 3.1. All Standard Proteinogenic Amino Acids, as Well as Ornithine and Citrulline, Inhibit Growth of S. Cerevisiae 3.1. All Standard Proteinogenic Amino Acids, as Well as Ornithine and Citrulline, Inhibit Growth of S. Cerevisiae Eight different S. cerevisiae amino acid transporters were overexpressed by introduc- ing their genes on multicopy plasmids under control of the constitutive ADH1 promoter (Figure 1). A C-terminal tag containing the ﬂuorescent protein YPet (https://www.fpbase. org/protein/ypet/) was added to allow visualization of the transporters in the cell. Confo- cal ﬂuorescence microscopy conﬁrmed that all eight transporters localize primarily to the cell periphery (Figure 1A), indicating that our tagged proteins are delivered to the plasma membrane. Some ﬂuorescent signal is seen in internal membranes and the vacuole, which is not uncommon for plasma membrane transporters, given that their normal life cycle involves being trafﬁcked between these different locations [39,71]. We observed substantial variation in protein expression between cells from the same culture, with a signiﬁcant proportion showing little or no ﬂuorescence (Figure S3). We believe that this reﬂects a sub-population of cells that have lost the expression vector. For the high copy pRS plasmids it is known that plasmid-free cells make up 20–30% of the population, even in selective media [72]. Expression of particular proteins can increase this population signiﬁcantly, with reported values of up to 50% [73]. Using ﬂow cytometry, we found that the fraction of cells in the ﬂuorescent population (30–65%) and the median ﬂuorescence signal from individual cells in this population (1.5–5.6 arbitrary units) was reasonably stable between independent cultures of each strain (Figure 1B,C). In minimal media without amino acids, growth of the overexpression strains over 24 h was up to 40% lower than that of BY4709 containing the empty plasmid pRSII426 (Figure 1D). This is likely due to both the plasmid instability discussed earlier (different populations of plasmid-free cells translates to a difference in the starting cell density of each culture), as well as general effects caused by overexpres- sion [74,75]. These variations within and between strains do not affect the conclusions of this paper as we only make comparisons of the same strains grown in different conditions. 7 of 17 we only Microorganisms 2021, 9, 7 Figure 1. Overexpression of amino acid permeases in BY4709. All proteins co YPet tag and were expressed from a multicopy plasmid under control of the promoter. 3. Results and Discussion (A) Fluorescence confocal microscopy images showing that the tra mainly to the cell periphery. Scale bars are 2 µm. (B,C) Data from flow cytom ing the estimated fraction of fluorescent cells in each culture (B), and the med nal of cells in this sub-population (C). (D) Growth of overexpression strains Figure 1. Overexpression of amino acid permeases in BY4709. All proteins contained a C-terminal YPet tag and were expressed from a multicopy plasmid under control of the constitutive ADH1 promoter. (A) Fluorescence confocal microscopy images showing that the transporters localize mainly to the cell periphery. Scale bars are 2 µm. (B,C) Data from ﬂow cytometry analysis showing the estimated fraction of ﬂuorescent cells in each culture (B), and the median ﬂuorescence signal of cells in this sub-population (C). (D) Growth of overexpression strains in minimal media. Normalized growth refers to cell density after 24 h, normalized to BY4709 carrying the empty plasmid pRSII426. Values shown are the mean ± standard deviation (n = 5 for (B,C), n = 7 for (D)). Figure 1. Overexpression of amino acid permeases in BY4709. All proteins Figure 1. Overexpression of amino acid permeases in BY4709. All proteins contained a C-terminal Figure 1. Overexpression of amino acid permeases in BY4709. All proteins co YPet tag and were expressed from a multicopy plasmid under control of the promoter. (A) Fluorescence confocal microscopy images showing that the tra mainly to the cell periphery. Scale bars are 2 µm. (B,C) Data from flow cytom ing the estimated fraction of fluorescent cells in each culture (B), and the me nal of cells in this sub-population (C). (D) Growth of overexpression strains Figure 1. Overexpression of amino acid permeases in BY4709. All proteins contained a C-terminal YPet tag and were expressed from a multicopy plasmid under control of the constitutive ADH1 promoter. (A) Fluorescence confocal microscopy images showing that the transporters localize mainly to the cell periphery. Scale bars are 2 µm. (B,C) Data from ﬂow cytometry analysis showing the estimated fraction of ﬂuorescent cells in each culture (B), and the median ﬂuorescence signal of cells in this sub-population (C). (D) Growth of overexpression strains in minimal media. Normalized growth refers to cell density after 24 h, normalized to BY4709 carrying the empty plasmid pRSII426. Values shown are the mean ± standard deviation (n = 5 for (B,C), n = 7 for (D)). 3. Results and Discussion We compared the growth of each strain in minimal medium with or without the addi- tion of either single amino acids or a synthetic complete (SC) mixture (Figure 2). It should be noted that BY4709 contains all the endogenous S. cerevisiae amino acid transporters and is in the yeast S288C background. We therefore expect that at the start of the experiment all strains, which were precultured in minimal medium containing ammonium as the sole nitrogen source, would contain signiﬁcant levels of active Gap1 [76,77]. For the control strain, the SC mixture increased growth and only cysteine or histidine at 5 mM was in- hibitory. For all overexpression strains except Put4, at least two different individual amino acids (other than cysteine or histidine) inhibited growth by more than 50%. We observed no citrulline-mediated growth inhibition, consistent with the fact that Gap1 is the only citrulline transporter [78]. Risinger et al. [41] observed no alanine-mediated toxicity and only a 20% reduction in cell growth caused by phenylalanine (all other amino acids caused a ≥65% reduction in growth). Here, we show that 5 mM of phenylalanine reduces the growth of Agp1 and Gnp1 strains by 70%, while 5 mM of alanine reduces the growth of Agp1 strains by 30%. Ornithine, a basic amino acid and important metabolite, inhibited the growth of strains overexpressing Can1 and Lyp1. Microorganisms 2021, 9, 7 8 of 17 mM of p f alanin e effect of amino acids on the growth of strains overexpressing amino acid transporters. BY4709 mid pRSII426 was used as a control. All proteins were expressed from the constitutive ADH1 prom YPet tag. Heat maps show cell density (OD600) after 24 h, normalized such that the same strain in Y without amino acids = 1. Raw values are given in Figure S4. The proteinogenic amino acids are rep etter code (see Table 4). Orn = ornithine, Cit = citrulline, SC = synthetic complete media minus ur dded at final concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates for eferences, see Table 1). Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test). Figure 2. The effect of amino acids on the growth of strains overexpressing amino acid transporters. BY4709 carrying the empty plasmid pRSII426 was used as a control. All pro- teins were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. 3. Results and Discussion Heat maps show cell density (OD600) after 24 h, normalized such that the same strain in YNBD minimal media without amino acids = 1. Raw values are given in Figure S4. The proteinogenic amino acids are represented by their single letter code (see Table 4). Orn = ornithine, Cit = citrulline, SC = synthetic complete media minus uracil. Amino acids were added at ﬁnal concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates for each transporter (for references, see Table 1). Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test). ffect of amino acids on the growth of strains overexpressing amino acid transporters. BY470 pRSII426 was used as a control. All proteins were expressed from the constitutive ADH1 pr et tag. Heat maps show cell density (OD600) after 24 h, normalized such that the same strain in hout amino acids = 1. Raw values are given in Figure S4. The proteinogenic amino acids are r er code (see Table 4). Orn = ornithine, Cit = citrulline, SC = synthetic complete media minus ed at final concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates f rences, see Table 1). Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test Figure 2. The effect of amino acids on the growth of strains overexpressing amino acid transporters. BY4709 carrying the empty plasmid pRSII426 was used as a control. All pro- teins were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. Heat maps show cell density (OD600) after 24 h, normalized such that the same strain in YNBD minimal media without amino acids = 1. Raw values are given in Figure S4. The proteinogenic amino acids are represented by their single letter code (see Table 4). Orn = ornithine, Cit = citrulline, SC = synthetic complete media minus uracil. Amino acids were added at ﬁnal concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates for each transporter (for references, see Table 1). Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test). ded at final concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates rences, see Table 1). 3. Results and Discussion A series of elegant experiments using Can1 and Gap1 mutants showed that this process requires ligand binding but not transport, supporting the hypothesis that substrate binding induces conformational changes that promote en- docytosis [38]. The same study demonstrated that different substrates are more or less effective at triggering endocytosis of the same transporter, and it follows that the same amino acid could be more or less effective at triggering endocytosis of different transport- ers mutants showed that this process requires ligand binding but not transport, supporting the hypothesis that substrate binding induces conformational changes that promote endocyto- sis [38]. The same study demonstrated that different substrates are more or less effective at triggering endocytosis of the same transporter, and it follows that the same amino acid could be more or less effective at triggering endocytosis of different transporters. intracellular trafficking [34–36,38]. A series of elegant experiments using Can1 and Gap1 mutants showed that this process requires ligand binding but not transport, supporting the hypothesis that substrate binding induces conformational changes that promote en- docytosis [38]. The same study demonstrated that different substrates are more or less effective at triggering endocytosis of the same transporter, and it follows that the same amino acid could be more or less effective at triggering endocytosis of different transport- ers Figure 3. Amino acid uptake by S. cerevisiae cells. Transporters were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. BY4709 carrying the empty plasmid pRSII426 was used as a vector control strain. (A) Uptake of phenylalanine (2.5 mM). (B) Uptake of lysine (50 µM) in the absence (black symbols) or presence of histidine (100 mM, grey symbols) or ornithine (100 mM, white symbols); circles, Lyp1 overexpression; squares, vector control strain. (C) Up- take of either alanine (500 µM, black symbols) or methionine (500 µM, white symbols); diamonds, Agp1 overexpression; squares, vector control strain. The straight black lines in (B,C) represent the calculated initial uptake rates, which are shown in (D,E), respectively. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Rates in (D) were compared using a t-test, and asterisks () represent p < 0.001. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Specificity Figure 3. Amino acid uptake by S. cerevisiae cells. Transporters were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. 3. Results and Discussion BY4709 carrying the empty plasmid pRSII426 was used as a vector control strain. (A) Uptake of phenylalanine (2.5 mM). (B) Uptake of lysine (50 µM) in the absence (black symbols) or presence of histidine (100 mM, grey symbols) or ornithine (100 mM, white symbols); circles, Lyp1 overexpression; squares, vector control strain. (C) Uptake of either alanine (500 µM, black symbols) or methionine (500 µM, white symbols); diamonds, Agp1 overexpression; squares, vector control strain. The straight black lines in (B,C) represent the calculated initial uptake rates, which are shown in (D,E), respectively. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Rates in (D) were compared using a t-test, and asterisks () represent p < 0.001. Figure 3. Amino acid uptake by S. cerevisiae cells. Transporters were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. BY4709 carrying the empty plasmid pRSII426 was used as a vector control strain. (A) Uptake of phenylalanine (2.5 mM). (B) Uptake of lysine (50 µM) in the absence (black symbols) or presence of histidine (100 mM, grey symbols) or ornithine (100 mM, white symbols); circles, Lyp1 overexpression; squares, vector control strain. (C) Up- take of either alanine (500 µM, black symbols) or methionine (500 µM, white symbols); diamonds, Agp1 overexpression; squares, vector control strain. The straight black lines in (B,C) represent the calculated initial uptake rates, which are shown in (D,E), respectively. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Rates in (D) were compared using a t-test, and asterisks () represent p < 0.001. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Specificity Figure 3. Amino acid uptake by S. cerevisiae cells. Transporters were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. BY4709 carrying the empty plasmid pRSII426 was used as a vector control strain. (A) Uptake of phenylalanine (2.5 mM). (B) Uptake of lysine (50 µM) in the absence (black symbols) or presence of histidine (100 mM, grey symbols) or ornithine (100 mM, white symbols); circles, Lyp1 overexpression; squares, vector control strain. (C) Uptake of either alanine (500 µM, black symbols) or methionine (500 µM, white symbols); diamonds, Agp1 overexpression; squares, vector control strain. 3. Results and Discussion The straight black lines in (B,C) represent the calculated initial uptake rates, which are shown in (D,E), respectively. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Rates in (D) were compared using a t-test, and asterisks () represent p < 0.001. 3. Results and Discussion Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test g g The proteinogenic amino acids are represented by their single letter code (see Table 4). Orn = ornithine, Cit = citrulline, SC = synthetic complete media minus uracil. Amino acids were added at ﬁnal concentrations of 0.5, 1, or 5 mM. Thick black boxes indicate reported substrates for each transporter (for references, see Table 1). Asterisks () indicate p < 0.05 when compared to growth in YNBD (t-test). We were surprised that proline or valine did not cause higher levels of growth in- hibition. Both of these amino acids caused a 90% reduction in growth when supplied at 3 mM to an S. cerevisiae strain expressing a endocytosis-resistant Gap1 mutant [41]. Agp1, Gnp1, and Put4 are all proline transporters, with Put4 estimated to have an approximately 500-fold lower Km for proline than Gap1 [45,56]. Valine is known to be a substrate for both Agp1 and Bap2, with transport in sub-mM concentrations occurring at rates equal or higher to that of other substrates for which we did observe growth inhibition (e.g., Ile, Leu) [43,47,60]. Sensitivity to other amino acids and increases in whole cell uptake (Fig- ures 2 and 3A) indicate that Agp1, Bap2, and Gnp1 are all present in these strains. The largest effect caused by proline or valine, however, was an approximately 10% reduction in the growth of Agp1 strains. It is possible that this difference is due to variations in substrate-regulated trafﬁcking. Although the use of a constitutive promoter minimizes transcriptional regulation, the levels of transporter at the plasma membrane and thus the rate of substrate transport and accumulation is subject to post-translational regulation via intracellular trafﬁcking [34–36,38]. A series of elegant experiments using Can1 and Gap1 Microorganisms 2021, 9, 7 9 of 17 s 9 of 17 s e mutants showed that this process requires ligand binding but not transport, supporting the hypothesis that substrate binding induces conformational changes that promote endocyto- sis [38]. The same study demonstrated that different substrates are more or less effective at triggering endocytosis of the same transporter, and it follows that the same amino acid could be more or less effective at triggering endocytosis of different transporters. intracellular trafficking [34–36,38]. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity They also suggest that Lyp1 has a lower afﬁnity for these substrates than Can1, which has a reported inhibition constant (Ki) of 3 mM for both His and Orn [80], or Gap1. It has previously been reported that Lyp1 does not transport Orn or His, but this could be explained by a lower afﬁnity and the fact that competition for Lys transport by Lyp1 was assayed with only a ten-fold excess in comparison to the 1000-fold excess used in the competition for Arg transport by Can1 [50,80]. Subsequent studies, using Lyp1 overexpression strains, tested for His transport using concentrations of 50–100 µM, which is likely to be far below the Km [54,60]. y In further experiments using lower concentrations, Lyp1 strains were inhibited (≥50% reduction in growth) by 16 µM lysine and Can1 strains by 16 µM of arginine or 125 µM of lysine (Figure 4). This is in the range of the measured Km values (Table 1) [51,62]. The growth advantage seen for Lyp1 overexpression strains in the presence of external arginine is due to import of this amino acid, which can be used as a carbon or nitrogen source by other endogenous transporters. The same effect is observed for the control strain, although it is only apparent during the exponential growth phase (Figure S5) and thus not seen at the 24 h time-point used in Figure 4. ER REVIEW 11 of 17 Figure 4. The effect of external arginine (A) and lysine (B) on the growth of strains overexpressing Can1 or Lyp1. BY4709 carrying the empty plasmid pRSII426 was used as a control. Can1 and Lyp1 were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. Normalized growth refers to cell density after 24 h, normalized to minimal media without amino acids. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. 3.3. Toxicity Is Caused by Amino Acid Accumulation, Not Proton Transport Figure 4. The effect of external arginine (A) and lysine (B) on the growth of strains overexpressing Can1 or Lyp1. BY4709 carrying the empty plasmid pRSII426 was used as a control. Can1 and Lyp1 were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. Normalized growth refers to cell density after 24 h, normalized to minimal media without amino acids. Values are the mean of biological triplicates. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity We expected that the pattern of amino acid sensitivity for each strain would match the known substrate speciﬁcity (Table 1) of the overexpressed transporter and, for the most part, this was true (Figure 2). As discussed in the previous section, some known substrates did not cause substantial growth inhibition. Even more surprising was that many strains were sensitive to amino acids not predicted to be transport substrates. While it is possible that these results are affected by altered expression levels of endogenous systems, we tentatively conclude that these transporters are able to transport a much broader range of substrates than previously described, but with a high Michaelis constant (Km > 1 mM). This can be rationalized by structural and mutational studies, which indicate that the residues interacting with the α-amino and α-carboxyl groups of the substrate are conserved [49,51,79]. The reason why this broader speciﬁcity has not been seen before could be simply because previous studies have tested a limited range of substrates and/or concentrations. The screening study published by Regenberg et al. [60], for example, only assayed transport at 100 or 250 µM of substrate. These lower-afﬁnity transport activities would be important to consider in a laboratory setting where synthetic media often contains amino acids at (sub-)mM concentrations. Some of these novel substrate speciﬁcities (the transport of Phe by Gnp1 and of His and Orn by Lyp1) were further investigated by monitoring the transport of radioactive Microorganisms 2021, 9, 7 10 of 17 10 of 17 substrates into whole cells (Figure 3). Again, it should be noted that some endogenous transporters, including Gap1, are expected to be present under these conditions but rapidly endocytosed in response to amino acid addition. For this reason, we only measured initial rates of transport. Overexpression of Gnp1 indeed increased the uptake of phenylalanine into whole cells (Figure 3A), supporting its identiﬁcation as a phenylalanine transporter. Lysine uptake was 4-fold faster in cells overexpressing Lyp1 (Figure 3B,D), indicating that the YPet-tagged Lyp1 is active. When His or Orn was present at 100 mM (2000-fold excess) Lys transport by Lyp1 cells was reduced by 56% and 43%, respectively. At the same concentrations, Lys transport by the control strain was reduced to background levels (Figure 3B). These results, in conjunction with the growth inhibition, are consistent with Lyp1 transporting both His and Orn. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity Error bars represent standard deviation and, in some cases, are obscured by the data point. Figure 4. The effect of external arginine (A) and lysine (B) on the growth of strains overexpressing Can1 or Lyp1. BY4709 carrying the empty plasmid pRSII426 was used as a control. Can1 and Lyp1 were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. Normalized growth refers to cell density after 24 h, normalized to minimal media without amino acids. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Figure 4. The effect of external arginine (A) and lysine (B) on the growth of strains overexpressing Can1 or Lyp1. BY4709 carrying the empty plasmid pRSII426 was used as a control. Can1 and Lyp1 were expressed from the constitutive ADH1 promoter with a C-terminal YPet tag. Normalized growth refers to cell density after 24 h, normalized to minimal media without amino acids. Values are the mean of biological triplicates. Error bars represent standard deviation and, in some cases, are obscured by the data point. Microorganisms 2021, 9, 7 11 of 17 11 of 17 3.3. Toxicity Is Caused by Amino Acid Accumulation, Not Proton Transport 3.3. Toxicity Is Caused by Amino Acid Accumulation, Not Proton Transport The amino acid transporters studied here, as well as Gap1, are thought to be amino acid: proton symporters. One possible explanation for the toxicity of amino acid transport is that the corresponding inﬂux of protons interferes with cellular homeostasis. This has been previously demonstrated in E. coli where excessive, uncontrolled transport of galactosides by the lactose:proton symporter LacY decreases the electrochemical proton gradient across the cell membrane and lowers the intracellular ATP concentration [81,82]. In the case of S. cerevisiae amino acid transport, it is possible that this effect would be ampliﬁed by the transport of excess amino acids into the vacuole, which is mediated by amino acid:proton antiporters would thus cause further movement of protons into the cytoplasm [83,84]. Risinger et al. [41] argued that proton inﬂux was not the mechanism for Gap1- mediated amino acid toxicity by showing that growth was not inhibited by amino acid mixtures, a condition where the overall transport rate should remain high but individual amino acids would accumulate to lower levels. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity We did observe toxicity when cells were grown in a mixture of amino acids; in SC media, the growth of Can1 and Lyp1 strains was reduced 70% and 85%, respectively, while Dip5 did not have the growth advantage seen for the control and other overexpression strains (Figure 2). However, this likely reﬂects the narrow speciﬁcities of these transporters and their relatively low Km. The other amino acids are presumably not present in high enough concentrations to effectively compete for transport, and thus one or two substrates are still accumulated to high levels. Whole cell transport assays did not show any direct correlation between growth inhibition and initial transport rate (Figure 3), and this suggests that toxicity is not mediated by proton inﬂux. Overexpression of Agp1 increased the transport of both Ala and Met 5- to 6-fold when each amino acid was supplied at 500 µM (Figure 3C,E), with the initial transport of Ala 3-fold higher than that of Met, yet at the same external concentration only Met inhibits growth. 50 µM of Lys is inhibitory to Lyp1 overexpression strains, yet the initial rate of transport is less than half that of the transport of Ala by Agp1 strains (Figure 3D,E). These transport assays also suggest differences in the intracellular concentration at which various amino acids become toxic. Assuming an internal volume of 70 fL per cell [85], we calculated that the amino acid pool of the Agp1 overexpression strain increased by 69 mM of Ala and 24 mM of Met in 24 min, and by 84 mM of Phe in 15 min. Under the same conditions, Met and Phe, but not Ala, cause growth inhibition of this strain. g To investigate the effect of increased proton ﬂux more directly, we used a pH-sensitive GFP variant called ratiometric pHluorin [66,67] to monitor the cytoplasmic pH (pHc) in growing cells with or without the addition of amino acid to the media (Figure 5A). For these experiments, a truncation mutant of Lyp1 was used which is more stably maintained at the plasma membrane after lysine addition [62]. Overexpression of this variant also causes lysine sensitivity (data not shown). For control cells or Lyp1 cells without the addition of lysine, the pHc was 7.0–7.2 and stayed fairly constant over three hours. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity (B) Normalized bulk intensity at 425 nm excita- tion/508 nm emission, at which pHluorin is largely insensitive to pH. Values shown are the mean of two independent experiments (individual results are shown in Figure S6). Dotted lines indicate the time at which either distilled water or lysine (final concentration 500 µM) was added. Figure 5. The effect of external lysine on the cytosolic pH of cells overexpressing Lyp1. BY4741 expressing the pH-sensitive ratiometric pHluorin and carrying the empty plasmid pRSII425 was used as a control. The Lyp1 strain constitutively expresses both pHluorin and a truncated, non- ﬂuorescently tagged Lyp1. (A) Cytosolic pH. (B) Normalized bulk intensity at 425 nm excitation/508 nm emission, at which pHluorin is largely insensitive to pH. Values shown are the mean of two independent experiments (individual results are shown in Figure S6). Dotted lines indicate the time at which either distilled water or lysine (ﬁnal concentration 500 µM) was added. We were able to obtain more information from the pHluorin experiments by analyzing the ﬂuorescence signal at 425 nm excitation/508 nm emission. Under these conditions, pHluorin is insensitive to pH [67]. This means that we can monitor the amount of pHluorin independent of the pH changes. For the samples where pHc remained constant, the bulk ﬂuorescence ﬁts well to an exponential curve (Figure 5B). We believe that this value is reporting the combined rates of protein production plus degradation in dividing cells, with the amount of pHluorin in individual cells staying stable over time. In contrast, the bulk ﬂuorescence of Lyp1 cells after lysine addition begins to plateau at approximately the same time that the cytosolic pH begins to decrease. This suggests that the mechanisms leading to growth inhibition are occurring in the hour after lysine addition, and that the decrease in cytosolic pH is a consequence, rather than a cause. One hypothesis that would ﬁt our data is that lysine (and also cysteine) toxicity is due to interference with ubiquitination pathways. Protein modiﬁcation by the attachment of ubiquitin (Ub) is a regulatory mechanism involved in a wide range of essential processes in eukaryotic cells [90]. A key step in these pathways is the transfer of Ub from a Ub- conjugating (E2) enzyme to a target protein. Some human E2~Ubs are able to react with free lysine and cysteine molecules in such a way that Ub is irreversibly transferred to the amino acid [91,92]. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity In contrast, the pHc of Lyp1 cells given 500 µM of lysine began to decrease after one hour, and continued to drop steadily over the course of the experiment with a total change of 0.4 pH units. Neither the magnitude of the pH change or the time scale on which it happens is consistent with the hypothesis of rapid proton inﬂux as a causative mechanism. When yeast is exposed to glucose after a period of starvation, the cytoplasmic pH transiently drops to as low as pH 6 and cells are able to recover to normal levels within minutes [86,87]. Several groups have demonstrated a direct correlation between pHc and growth rate in yeast, but S. cerevisiae is still able to grow reliably at pHc between 6.5 and 7 [88,89]. 12 of 17 ansiently minutes Microorganisms 2021, 9, 7 y g y p Figure 5. The effect of external lysine on the cytosolic pH of cells overexpressing Lyp1. BY4741 expressing the pH-sensitive ratiometric pHluorin and carrying the empty plasmid pRSII425 was used as a control. The Lyp1 strain constitutively expresses both pHluorin and a truncated, non- fluorescently tagged Lyp1. (A) Cytosolic pH. (B) Normalized bulk intensity at 425 nm excita- tion/508 nm emission, at which pHluorin is largely insensitive to pH. Values shown are the mean of two independent experiments (individual results are shown in Figure S6). Dotted lines indicate the time at which either distilled water or lysine (final concentration 500 µM) was added. Figure 5. The effect of external lysine on the cytosolic pH of cells overexpressing Lyp1. BY4741 expressing the pH-sensitive ratiometric pHluorin and carrying the empty plasmid pRSII425 was used as a control. The Lyp1 strain constitutively expresses both pHluorin and a truncated, non- ﬂuorescently tagged Lyp1. (A) Cytosolic pH. (B) Normalized bulk intensity at 425 nm excitation/508 nm emission, at which pHluorin is largely insensitive to pH. Values shown are the mean of two independent experiments (individual results are shown in Figure S6). Dotted lines indicate the time at which either distilled water or lysine (ﬁnal concentration 500 µM) was added. Figure 5. The effect of external lysine on the cytosolic pH of cells overexpressing Lyp1. BY4741 expressing the pH-sensitive ratiometric pHluorin and carrying the empty plasmid pRSII425 was used as a control. The Lyp1 strain constitutively expresses both pHluorin and a truncated, non- fluorescently tagged Lyp1. (A) Cytosolic pH. 4. Conclusions Our study expands on the work of Risinger et al. [41] who showed that all 20 proteino- genic amino acids, as well as the non-proteinogenic amino acids citrulline and ornithine, can be growth inhibitory to S. cerevisiae. We have shown that this effect can be mediated by various amino acid transporters and is not speciﬁc to Gap1. We have also demonstrated that amino acid-mediated growth inhibition is not dependent on the initial rate of trans- port, or triggered by the rapid inﬂux of protons, but is instead caused by the longer-term accumulation of single amino acids. For at least some amino acids, S. cerevisiae has mecha- nisms in place to prevent their over-accumulation by substrate-dependent removal of the corresponding transporters from the plasma membrane and/or partitioning in speciﬁc membrane domains (as shown for example for Can1 and Lyp1) [94–96]. Amino acid sensitivity is an important phenomenon that should be considered in the design and analysis of studies of amino acid and peptide transport. It is also a useful tool for assessing the in vivo activity of transporters as it reports on their levels at the plasma membrane and their transport kinetics for speciﬁc substrates. We have used it to develop growth-based screens to conﬁrm the activity of overexpressed wild-type and mutant transporters, including when expressed in Pichia pastoris [62]. Our results from screening eight different amino acid transporters did vary from what we predicted based on the current literature suggesting (i) a much broader speciﬁcity than previously thought, and (ii) transporter/substrate-speciﬁc variations that may reﬂect differences in substrate-induced downregulation. Supplementary Materials: The following are available online at https://www.mdpi.com/2076-260 7/9/1/7/s1, Figure S1: OD600 correction for linearity at high cell density. Figure S2: Calibration of pHluorin ﬂuorescence. Figure S3: Confocal microscopy images showing the variation in ﬂuorescent signal between cells in the same culture. Figure S4: Raw data from experiments presented in Figure 2, showing the effect of amino acids on the growth of strains overexpressing amino acid transporters. Figure S5: Growth curves from the same experiment shown in Figure 4. Figure S6: Data from the two experiments used to generate Figure 5. Author Contributions: S.J.R., F.B. and B.P. designed the research; S.J.R. performed the majority of the experiments and data analysis; J.S.v.t.K. performed the transport assays, and F.B. designed and constructed some strains. B.P. supervised the work, and S.J.R. and B.P. wrote the manuscript. 3.2. Amino Acid Sensitivity Reports on Transporter Activity and Substrate Speciﬁcity Intrinsic reactivity with free lysine has also been observed for the S. cerevisiae E2 enzymes Ubc4 and Pex4 (Chris Williams, personal communication). This reaction Microorganisms 2021, 9, 7 13 of 17 occurs on the minute timescale, in the cytoplasmic pH range and at concentrations as low as 50 mM. Extrapolation of our transport assay data suggests that our overexpression strains import this amount of lysine within 30–60 min, which is in agreement with previous work [62]. Overlapping activities mean that individual E2 enzymes are not essential in S. cerevisiae but multiple knockouts, for example ∆ubc4∆ubc5, are lethal [93]. 1. Bender, D.A. Amino Acid Metabolism, 3rd ed.; John Wiley & Sons: Chichester, West Sussex, UK, 2012. 2. Rowley, D. Inhibition of E. coli strains by amino-acids. Nature 1953, 171, 80–81. [CrossRef] [PubMed] e de , mino cid Metabolism, 3 d ed ; Jo W ey & So s C c este , West Susse , U , 0 2. Rowley, D. Inhibition of E. coli strains by amino-acids. Nature 1953, 171, 80–81. [CrossRef] [PubMed] 4. Conclusions All authors have read and agreed to the published version of the manuscript. Funding: This work was carried out within the BE-Basic R&D Program, which was granted a FES subsidy from the Dutch Ministry of Economic Affairs, Agriculture and Innovation (EL&I). The research was also funded by NWO TOP-GO program (project number 700.10.53). Institutional Review Board Statement: Not applicable. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Please refer to suggested Data Availability Statements in section “MDPI Research Data Policies” at https://www.mdpi.com/ethics. Acknowledgments: We thank Gertien J. Smits for providing the plasmid pYES2-PACT1-pHluorin, and Chris Williams for general discussion of our work. Acknowledgments: We thank Gertien J. Smits for providing the plasmid pYES2-PACT1-pHluorin, and Chris Williams for general discussion of our work. Conﬂicts of Interest: The authors declare that there is no conﬂict of interest. References Valine inhibition of the growth of Es- h i hi li t i K 12 J B t i l 1962 83 624 630 [C R f] [P bM d] p y y 13. Leavitt, R.I.; Umbarger, H.E. Isoleucine and valine metabolism in Escherichia coli. XI. Valine inhibition of the growth of Es- cherichia coli strain K-12. J. Bacteriol. 1962, 83, 624–630. [CrossRef] [PubMed] 14. Bhatnagar, R.K.; Berry, A.; Hendry, A.T.; Jensen, R.A. The biochemical basis for growth inhibition by L-phenylalanine in Neisseria gonorrhoeae. Mol. Microbiol. 1989, 3, 429–435. [CrossRef] [PubMed] nagar, R.K.; Berry, A.; Hendry, A.T.; Jensen, R.A. The biochemical basis for growth inhibition by L-phe 15. Duan, Y.; Li, F.; Tan, K.; Liu, H.; Li, Y.; Liu, Y.; Kong, X.; Tang, Y.; Wu, G.; Yin, Y. Key mediators of intracellular amino acids signaling to mTORC1 activation. Amino Acids 2015, 47, 857–867. [CrossRef] [PubMed] 16. Powis, K.; De Virgilio, C. Conserved regulators of Rag GTPases orchestrate amino acid-dependent TOR 2016, 2, 15049. [CrossRef] [PubMed] 16. Powis, K.; De Virgilio, C. Conserved regulators of Rag GTPases orchestrate amino acid-dependent TORC1 signaling. Cell Discov. 2016, 2, 15049. [CrossRef] [PubMed] 17 Zh L Zh W Zh Y Li F W i H P J R d i d di i id i h i h 16. Powis, K.; De Virgilio, C. Conserved regulators of Rag GTPases orchestrate amino acid dependent TORC1 signaling. Cell Discov. 2016, 2, 15049. [CrossRef] [PubMed] 17. Zheng, L.; Zhang, W.; Zhou, Y.; Li, F.; Wei, H.; Peng, J. Recent advances in understanding amino acid sensing mechanisms that 2016, 2, 15049. [CrossRef] [PubMed] 17. Zheng, L.; Zhang, W.; Zhou, Y.; Li, F.; Wei, H.; Peng, J. Recent advances in understanding amino acid sensing mechanisms that regulate mTORC1. Int. J. Mol. Sci. 2016, 17, 1636. [CrossRef] 17. Zheng, L.; Zhang, W.; Zhou, Y.; Li, F.; Wei, H.; Peng, J. Recent advances in understanding amino acid regulate mTORC1. Int. J. Mol. Sci. 2016, 17, 1636. [CrossRef] 18. Laplante, M.; Sabatini, D.M. MTOR signaling in growth control and disease. Cell 2012, 149, 274–293. [CrossRef] 19. Park, Y.-Y.; Sohn, B.H.; Johnson, R.L.; Kang, M.-H.; Kim, S.B.; Shim, J.-J.; Mangala, L.S.; Kim, J.H.; Yoo, J.E.; Rodriguez-Aguayo, C.; et al. Yes-associated protein 1 and transcriptional coactivator with PDZ-binding motif activate the mammalian target of rapamycin complex 1 pathway by regulating amino acid transporters in hepatocellular carcinoma. Hepatology 2016, 63, 159–172. [CrossRef] 20. References Cormerais, Y.; Giuliano, S.; LeFloch, R.; Front, B.; Durivault, J.; Tambutté, E.; Massard, P.-A.; de la Ballina, L.R.; Endou, H.; Wempe, M.F.; et al. Genetic disruption of the multifunctional CD98/LAT1 complex demonstrates the key role of essential amino acid transport in the control of mTORC1 and tumor growth. Cancer Res. 2016, 76, 4481–4492. [CrossRef] 21. Krall, A.S.; Xu, S.; Graeber, T.G.; Braas, D.; Christofk, H.R. Asparagine promotes cancer cell proliferation acid exchange factor. Nat. Commun. 2016, 7, 11457. [CrossRef] 22. Sanayama, Y.; Matsumoto, A.; Shimojo, N.; Kohno, Y.; Nakaya, H. Phenylalanine sensitive K562-D cells for the analysis of the biochemical impact of excess amino acid. Sci. Rep. 2014, 4, 6941. [CrossRef] [PubMed] p p 23. Bonner, C.A.; Rodrigues, A.M.; Miller, J.A.; Jensen, R.A. Amino acids are general growth inhibitors of Nicotiana silvestris in tissue culture. Physiol. Plant. 1992, 84, 319–328. [CrossRef] y 24. Bonner, C.A.; Williams, D.S.; Aldrich, H.C.; Jensen, R.A. Antagonism by L-glutamine of toxicity and growth inhibition caused by other amino acids in suspension cultures of Nicotiana silvestris. Plant Sci. 1996, 113, 43–58. [CrossRef] p 25. Bonner, C.A.; Jensen, R.A. Recognition of speciﬁc patterns of amino acid inhibition of growth in hig glutamine-reversible ‘general amino acid inhibition’. Plant Sci. 1997, 130, 133–143. [CrossRef] 25. Bonner, C.A.; Jensen, R.A. Recognition of speciﬁc patterns of amino acid inhibition of growth in higher plants, uncomplicated by glutamine-reversible ‘general amino acid inhibition’. Plant Sci. 1997, 130, 133–143. [CrossRef] 26. Durán, R.V.; Oppliger, W.; Robitaille, A.M.; Heiserich, L.; Skendaj, R.; Gottlieb, E.; Hall, M.N. Glutaminolysis activates Rag- mTORC1 signaling. Mol. Cell 2012, 47, 349–358. [CrossRef] g g 27. Stracka, D.; Jozefczuk, S.; Rudroff, F.; Sauer, U.; Hall, M.N. Nitrogen source activates TOR (target of r glutamine and independently of Gtr/Rag proteins. J. Biol. Chem. 2014, 289, 25010–25020. [CrossRef] , ; J , ; , ; , ; , g ( g p y ) p glutamine and independently of Gtr/Rag proteins. J. Biol. Chem. 2014, 289, 25010–25020. [CrossRef] 28. Jewell, J.L.; Kim, Y.C.; Russell, R.C.; Yu, F.-X.; Park, H.W.; Plouffe, S.W.; Tagliabracci, V.S.; Guan, K.-L. Differential regulation of mTORC1 by leucine and glutamine. Science 2015, 347, 194–198. [CrossRef] glutamine and independently of Gtr/Rag proteins. J. Biol. Chem. 2014, 289, 25010 25020. [CrossRef] 28. Jewell, J.L.; Kim, Y.C.; Russell, R.C.; Yu, F.-X.; Park, H.W.; Plouffe, S.W.; Tagliabracci, V.S.; Guan, K.-L. Differential regulation of mTORC1 by leucine and glutamine. Science 2015, 347, 194–198. [CrossRef] 29. References 14 of 17 14 of 17 Microorganisms 2021, 9, 7 3. Rowley, D. Interrelationships between amino-acids in the growth of coliform organisms. J. Gen. Microbiol. 1953, 9, 37–43. [CrossRef] [PubMed] [ ] [ ] 4. Johnson, C.L.; Vishniac, W. Growth inhibition in Thiobacillus neapolitanus by histidine, methionine, phenylalanine, and threonine. J. Bacteriol. 1970, 104, 1145–1150. [CrossRef] [PubMed] [ ] [ ] 4. Johnson, C.L.; Vishniac, W. Growth inhibition in Thiobacillus neapolitanus by histidine, methionine, phenylalanine, and threonine. 5. Jensen, R.A.; Stenmark-Cox, S.; Ingram, L.O. Mis-regulation of 3-deoxy-D-arabino-heptulosonate 7-phosphate synthetase does not account for growth inhibition by phenylalanine in Agmenellum quadruplicatum. J. Bacteriol. 1974, 120, 1124–1132. [CrossRef] [PubMed] [ ] 6. Englesberg, E.; Bass, R.; Heiser, W. Inhibition of the growth of mammalian cells in culture by amino acids and the isolation and characterization of L-phenylalanine transport. Somatic Cell Genet. 1976, 2, 411–428. [CrossRef] [PubMed] 6. Englesberg, E.; Bass, R.; Heiser, W. Inhibition of the growth of mammalian cells in culture by amino acids and the isolation and h i i f L h l l i S i C ll G 1 2 411 428 [C R f] [P bM d] 7. Sumrada, R.; Cooper, T. Basic amino acid inhibition of growth in Saccharomyces cerevisiae. Biochem. Biophys. Res. Commun. 1976, 68, 598–602. [CrossRef] 8. Nishiuch, Y.; Sasaki, M.; Nakayasu, M.; Oikawa, A. Cytotoxicity of cysteine in culture media. In Vitro 197 8. Nishiuch, Y.; Sasaki, M.; Nakayasu, M.; Oikawa, A. Cytotoxicity of cysteine in culture media. In Vitro 1976, 12, 635–638. [CrossRef] 9. Miles, D.O.; Dyer, J.K.; Wong, J.C. Inﬂuence of amino acids on the growth of Bacteroides melaninogenicus. J. Bacteriol. 1976, 127, 899–903. [CrossRef] y y y y 9. Miles, D.O.; Dyer, J.K.; Wong, J.C. Inﬂuence of amino acids on the growth of Bacteroides melaninogenicus. J. Bacteriol. 1976, 127, 899–903. [CrossRef] 10. Kaur, J.; Bachhawat, A.K. Yct1p, a novel, high-afﬁnity, cysteine-speciﬁc transporter from the yeast Saccharomyces cerevisiae. Genetics 2007, 176, 877–890. [CrossRef] 11. Watanabe, D.; Kikushima, R.; Aitoku, M.; Ohtsu, I.; Nasuno, R.; Takagi, H. Exogenous addition of histidine reduces copper availability in the yeast Saccharomyces cerevisiae. Microbial Cell 2014, 1, 241–246. [CrossRef] 12. Blau, N.; van Spronsen, F.J.; Levy, H.L. Phenylketonuria. Lancet 2010, 376, 1417–142 , ; p , J ; y, y , , [ ] 13. Leavitt, R.I.; Umbarger, H.E. Isoleucine and valine metabolism in Escherichia coli. XI. References Ljungdahl, P.O.; Daignan-Fornier, B. Regulation of amino acid, nucleotide, and phosphate metabolism in Saccharomyces cerevisiae. Genetics 2012, 190, 885–929. [CrossRef] 30. Gournas, C.; Prévost, M.; Krammer, E.-M.; André, B. Function and regulation of fungal amino acid transporters: Insights from predicted structure. In Yeast Membrane Transport; Ramos, J., Sychrová, H., Kschischo, M., Eds.; Springer International Publishing: Cham, Switzerland, 2016; Volume 892, pp. 69–106. pp 31. Magasanik, B.; Kaiser, C.A. Nitrogen regulation in Saccharomyces cerevisiae. Gene 2002, 290, 1–18. [CrossRef] 15 of 17 15 of 17 Microorganisms 2021, 9, 7 32. Hinnebusch, A.G. Translational regulation of GCN4 and the general amino acid control of yeast. Annu. Rev. Microbiol. 2005, 59, 407–450. [CrossRef] 33. Ljungdahl, P.O. Amino-acid-induced signalling via the SPS-sensing pathway in yeast. Biochem. Soc. Trans. 2009, 37, 242–247. [CrossRef] [PubMed] 34. Lin, C.H.; MacGurn, J.A.; Chu, T.; Stefan, C.J.; Emr, S.D. Arrestin-related ubiquitin-ligase adaptors regulate endocytosis and protein turnover at the cell surface. Cell 2008, 135, 714–725. [CrossRef] [PubMed] 35. Nikko, E.; Pelham, H.R.B. Arrestin-mediated endocytosis of yeast plasma membrane transporters. Trafﬁc 2009, 10, 1856–1867. [CrossRef] [PubMed] 36. Hatakeyama, R.; Kamiya, M.; Takahara, T.; Maeda, T. Endocytosis of the aspartic acid/glutamic acid transporter Dip5 is triggered by substrate-dependent recruitment of the Rsp5 ubiquitin ligase via the arrestin-like protein Aly2. Mol. Cell. Biol. 2010, 30, 5598–5607. [CrossRef] 37. Keener, J.M.; Babst, M. Quality control and substrate-dependent downregulation of the nutrient transporter Fur4. Trafﬁc 2013, 14, 412–427. [CrossRef] [PubMed] 38. Ghaddar, K.; Merhi, A.; Saliba, E.; Krammer, E.M.; Prevost, M.; André, B. Substrate-induced ubiquitylation and endocytosis of yeast amino acid permeases. Mol. Cell. Biol. 2014, 34, 4447–4463. [CrossRef] 39. MacGurn, J.A.; Hsu, P.-C.; Emr, S.D. Ubiquitin and membrane protein turnover: From cradle to grave. 81, 231–259. [CrossRef] 40. Melnykov, A.V. New mechanisms that regulate Saccharomyces cerevisiae short peptide transporter achieve balanced intracellular amino acid concentrations. Yeast 2016, 33, 21–31. [CrossRef] 41. Risinger, A.L.; Cain, N.E.; Chen, E.J.; Kaiser, C.A. Activity-dependent reversible inactivation of the general amino acid permease. Mol. Biol. Cell 2006, 17, 4411–4419. [CrossRef] 42. Schreve, J.L.; Sin, J.K.; Garrett, J.M. The Saccharomyces cerevisiae YCC5 (YCL025c) gene encodes an amino acid permease, Agp1, which transports asparagine and glutamine. J. Bacteriol. 1998, 180, 2556–2559. [CrossRef] p p g g 43. Iraqui, I.; Vissers, S.; Bernard, F.; de Craene, J.O.; Boles, E.; Urrestarazu, A.; André, B. References Amino acid signaling in Saccharomyces cere- visiae: A permease-like sensor of external amino acids and F-Box protein Grr1p are required for transcriptional induction of the AGP1 gene, which encodes a broad-speciﬁcity amino acid permease. Mol. Cell. Biol. 1999, 19, 989–1001. [CrossRef] [PubMed] g p y p [ ] [ ] 44. Düring-Olsen, L.; Regenberg, B.; Gjermansen, C.; Kielland-Brandt, M.C.; Hansen, J. Cysteine uptake by Saccharomyces cerevisiae is accomplished by multiple permeases. Curr. Genet. 1999, 35, 609–617. [CrossRef] [PubMed] 45. Andréasson, C.; Neve, E.P.A.; Ljungdahl, P.O. Four permeases import proline and the toxic proline analogue azetidine-2- carboxylate into yeast. Yeast 2004, 21, 193–199. [CrossRef] [PubMed] 46. Sáenz, D.A.; Chianelli, M.S.; Stella, C.A. L-Phenylalanine transport in Saccharomyces cerevisiae: Participation of GAP1, BAP2, and AGP1. J. Amino Acids 2014, 2014, 283962. [CrossRef] 47. Grauslund, M.; Didion, T.; Kielland-Brandt, M.C.; Andersen, H.A. BAP2, a gene encoding a permea acids in Saccharomyces cerevisiae. Biochim. Biophys. Acta 1995, 1269, 275–280. [CrossRef] 48. Schreve, J.; Garrett, J.M. The branched-chain amino acid permease gene of Saccharomyces cerevisiae, BAP2, encodes the high-afﬁnity leucine permease (S1). Yeast 1997, 13, 435–439. [CrossRef] p ( ) [ ] 49. Usami, Y.; Uemura, S.; Mochizuki, T.; Morita, A.; Shishido, F.; Inokuchi, J.-I.; Abe, F. Functional mapping and implications of substrate speciﬁcity of the yeast high-afﬁnity leucine permease Bap2. Biochim. Biophys. Acta 2014, 1838, 1719–1729. [CrossRef] p 49. Usami, Y.; Uemura, S.; Mochizuki, T.; Morita, A.; Shishido, F.; Inokuchi, J.-I.; Abe, F. Functional mapping and implications of substrate speciﬁcity of the yeast high-afﬁnity leucine permease Bap2. Biochim. Biophys. Acta 2014, 1838, 1719–1729. [CrossRef] 50. Grenson, M. Multiplicity of the amino acid permeases in Saccharomyces cerevisiae. II. Evidence for a speciﬁc lysine-transporting Bi hi Bi h A t 1966 127 339 346 [C R f] substrate speciﬁcity of the yeast high-afﬁnity leucine permease Bap2. Biochim. Biophys. Acta 2014, 1838, 1719–1729. [CrossRef] 50. Grenson, M. Multiplicity of the amino acid permeases in Saccharomyces cerevisiae. II. Evidence for a speciﬁc lysine-transporting system. Biochim. Biophys. Acta 1966, 127, 339–346. [CrossRef] 50. Grenson, M. Multiplicity of the amino acid permeases in Saccharomyces cerevisiae. II. Evidence for a system. Biochim. Biophys. Acta 1966, 127, 339–346. [CrossRef] y p y 51. Ghaddar, K.; Krammer, E.-M.; Mihajlovic, N.; Brohée, S.; André, B.; Prévost, M. Converting the yeast arginine Can1 permease to a lysine permease. J. Biol. Chem. 2014, 289, 7232–7246. [CrossRef] mberg, S.; Olsen, L.D.; Kielland-Brandt, M.C. References Brachmann, C.B.; Davies, A.; Cost, G.J.; Caputo, E.; Li, J.; Hieter, P.; Boeke, J.D. Designer deletion strains derived from Saccharomyces cerevisiae S288C: A useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 1998, 14, 115–132. [CrossRef] [ ] 62. Bianchi, F.; Klooster, J.S.; Ruiz, S.J.; Luck, K.; Pols, T.; Urbatsch, I.L.; Poolman, B. Asymmetry in inward- and outward-afﬁnity constant of transport explain unidirectional lysine ﬂux in Saccharomyces cerevisiae. Sci. Rep. 2016, 6, 31443. [CrossRef] 63. Mumberg, D.; Müller, R.; Funk, M. Yeast vectors for the controlled expression of heterologous proteins in different genetic backgrounds. Gene 1995, 156, 119–122. [CrossRef] g 64. Nguyen, A.W.; Daugherty, P.S. Evolutionary optimization of ﬂuorescent proteins for intracellular FRET. Nat. Biotechnol. 2005, 23, 355–360. [CrossRef] [PubMed] 65. Chee, M.K.; Haase, S.B. New and redesigned pRS plasmids shuttle vectors for genetic manipulation of Saccharomyces cerevisiae. G3: Genes, Genomes, Genetics 2012, 2, 515–526. [CrossRef] [PubMed] 66. Orij, R.; Postmus, J.; Beek Ter, A.; Brul, S.; Smits, G.J. In vivo measurement of cytosolic and mitochondrial pH using a pH-sensitive GFP derivative in Saccharomyces cerevisiae reveals a relation between intracellular pH and growth. Microbiology 2009, 155, 268–278. [CrossRef] 67. Miesenböck, G.; De Angelis, D.A.; Rothman, J.E. Visualizing secretion and synaptic transmission with pH-sensitive green ﬂuorescent proteins. Nature 1998, 394, 192–195. [CrossRef] 68. Bitinaite, J.; Nichols, N.M. DNA cloning and engineering by uracil excision. Curr. Protoc. Mol. Biol. 2009, 86, 3–21. [CrossRef] 69. Nørholm, M.H.H. A mutant Pfu DNA polymerase designed for advanced uracil-excision DNA engineering. BMC Biotechnol. 2010, 10, 21–27. [CrossRef] g g g y 69. Nørholm, M.H.H. A mutant Pfu DNA polymerase designed for advanced uracil-excision DNA engineering. BMC Biotechnol. 2010, 10, 21–27. [CrossRef] , , [ ] 70. Warringer, J.; Blomberg, A. Automated screening in environmental arrays allows analysis of quantitative phenotypic proﬁles in Saccharomyces cerevisiae. Yeast 2003, 20, 53–67. [CrossRef] 71. O’Donnell, A.F.; Apffel, A.; Gardner, R.G.; Cyert, M.S. α-Arrestins Aly1 and Aly2 regulate intracellular trafﬁcking in response to nutrient signaling. Mol. Biol. Cell 2010, 21, 3552–3566. [CrossRef] g g 72. Christianson, T.W.; Sikorski, R.S.; Dante, M.; Shero, J.H.; Hieter, P. Multifunctional yeast high-copy-num 1992, 110, 119–122. [CrossRef] 72. Christianson, T.W.; Sikorski, R.S.; Dante, M.; Shero, J.H.; Hieter, P. Multifunctional yeast high-copy-number shuttle vectors. Gene 1992, 110, 119–122. [CrossRef] 73. Ro, D.-K.; Ouellet, M.; Paradise, E.M.; Burd, H.; Eng, D.; Paddon, C.J.; Newman, J.D.; Keasling, J.D. References Dip5p mediates high-afﬁnity and high-capacity transport of aspartate in Saccharomyces cerevisiae. Curr. Genet. 1998, 33, 171–177. [CrossRef] 52. Regenberg, B.; Holmberg, S.; Olsen, L.D.; Kielland-Brandt, M.C. Dip5p mediates high-afﬁnity and h L-glutamate and L-aspartate in Saccharomyces cerevisiae. Curr. Genet. 1998, 33, 171–177. [CrossRef] L-glutamate and L-aspartate in Saccharomyces cerevisiae. Curr. Genet. 1998, 33, 171–177. [CrossRef] 53. Zhu, X.; Garrett, J.; Schreve, J.; Michaeli, T. GNP1, the high-afﬁnity glutamine permease of S. cerevisiae. Curr. Genet. 1996, 30, 107 114 [C R f] [P bM d] L glutamate and L aspartate in Saccharomyces cerevisiae. Curr. Genet. 1998, 33, 171 177. [CrossRef] 53. Zhu, X.; Garrett, J.; Schreve, J.; Michaeli, T. GNP1, the high-afﬁnity glutamine permease of S. cerevisiae. Curr. Genet. 1996, 30, 107–114. [CrossRef] [PubMed] Schreve, J.; Michaeli, T. GNP1, the high-afﬁnity glutamine permease of S. cerevisiae. Curr. Genet. 1996, 30 [PubMed] 53. Zhu, X.; Garrett, J.; Schreve, J.; Michaeli, T. GNP1, the high-afﬁnity glutamine permease of S. cerevis 107–114. [CrossRef] [PubMed] 54. Sychrová, H.; Chevallier, M.R. Cloning and sequencing of the Saccharomyces cerevisiae gene LYP1 coding for a lysine-speciﬁc permease. Yeast 1993, 9, 771–782. [CrossRef] [PubMed] 55. Sychrová, H.; Mat˘ejˇcková, A.; Kotyk, A. Kinetic properties of yeast lysine permeases coded by genes on multi-copy vectors. FEMS Microbiol. Lett. 1993, 113, 57–61. [CrossRef] [PubMed] 56. Lasko, P.F.; Brandriss, M.C. Proline transport in Saccharomyces cerevisiae. J. Bacteriol. 1981, 148, 241–2 56. Lasko, P.F.; Brandriss, M.C. Proline transport in Saccharomyces cerevisiae. J. Bacteriol. 1981, 148, 241–247. [CrossRef] [PubMed] 57. Jauniaux, J.C.; Vandenbol, M.; Vissers, S.; Broman, K.; Grenson, M. Nitrogen catabolite regulation of proline permease in Saccharomyces cerevisiae. Cloning of the PUT4 gene and study of PUT4 RNA levels in wild-type and mutant strains. Eur. J. Biochem. 1987, 164, 601–606. [CrossRef] [PubMed] 58. Vandenbol, M.; Jauniaux, J.C.; Grenson, M. Nucleotide sequence of the Saccharomyces cerevisiae PUT4 pr gene: Similarities between CAN1, HIP1 and PUT4 permeases. Gene 1989, 83, 153–159. [CrossRef] 59. Schmidt, A.; Hall, M.N.; Koller, A. Two FK506 resistance-conferring genes in Saccharomyces cerevisiae, TAT1 and TAT2, encode amino acid permeases mediating tyrosine and tryptophan uptake. Mol. Cell. Biol. 1994, 14, 6597–6606. [CrossRef] 16 of 17 16 of 17 Microorganisms 2021, 9, 7 60. Regenberg, B.; Düring-Olsen, L.; Kielland-Brandt, M.C.; Holmberg, S. Substrate speciﬁcity and gene expression of the amino-acid permeases in Saccharomyces cerevisiae. Curr. Genet. 1999, 36, 317–328. [CrossRef] p y 61. References Induction of multiple pleiotropic drug resistance genes in yeast engineered to produce an increased level of anti-malarial drug precursor, artemisinic acid. BMC Biotechnol. 2008, 8, 83. [CrossRef] [PubMed] 74. Österberg, M.; Kim, H.; Warringer, J.; Melén, K.; Blomberg, A.; von Heijne, G. Phenotypic effects of membrane protein overexpression in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 2006, 103, 11148–11153. [CrossRef] [PubMed] 75. White, M.A.; Clark, K.M.; Grayhack, E.J.; Dumont, M.E. Characteristics affecting expression and solubiliz proteins. J. Mol. Biol. 2007, 365, 621–636. [CrossRef] [PubMed] p 76. Courchesne, W.E.; Magasanik, B. Ammonia regulation of amino acid permeases in Saccharomyces cerevisiae. Mol. Cell. Biol. 1983, 3, 672–683. [CrossRef] [PubMed] 77. Chen, E.J.; Kaiser, C.A. Amino acids regulate the intracellular trafﬁcking of the general amino acid permease of Saccharomyces cere- visiae. Proc. Natl. Acad. Sci. USA 2002, 99, 14837–14842. [CrossRef] 78. Jauniaux, J.C.; Grenson, M. GAP1, the general amino acid permease gene of Saccharomyces cerevisiae. Nucleotide sequence, protein similarity with the other bakers yeast amino acid permeases, and nitrogen catabolite repression. Eur. J. Biochem. 1990, 190, 39–44. [CrossRef] [ ] 79. Kanda, N.; Abe, F. Structural and functional implications of the yeast high-afﬁnity tryptophan permease Tat2. Biochemistry 2013, 52, 4296–4307. [CrossRef] 80. Grenson, M.; Mousset, M.; Wiame, J.M.; Bechet, J. Multiplicity of the amino acid permeases in Saccharomyces cerevisiae. I. Evidence for a speciﬁc arginine-transporting system. Biochim. Biophys. Acta 1966, 127, 325–338. [CrossRef] 81. Wilson, D.M.; Kusch, M.; Flagg-Newton, J.L.; Wilson, T.H. Control of lactose transport in Escherichia coli. FEBS Lett. 1980, 117, K37–K44. [CrossRef] 82. Wilson, D.M.; Putzrath, R.M.; Wilson, T.H. Inhibition of growth of Escherichia coli by lactose and other galactosides. Biochim. Bio- phys. Acta 1981, 649, 377–384. [CrossRef] p y 83. Ohsumi, Y.; Anraku, Y. Active transport of basic amino acids driven by a proton motive force in vacuo Saccharomyces cerevisiae. J. Biol. Chem. 1981, 256, 2079–2082. [PubMed] 84. Sato, T.; Ohsumi, Y.; Anraku, Y. Substrate speciﬁcities of active transport systems for amino acids in vacuolar-membrane vesicles of Saccharomyces cerevisiae. Evidence of seven independent proton/amino acid antiport systems. J. Biol. Chem. 1984, 259, 11505–11508. [PubMed] 85. Sherman, F. Getting started with yeast. Meth. Enzymol. 2002, 350, 3–41. 86. Dechant, R.; Binda, M.; Lee, S.S.; Pelet, S.; Winderickx, J.; Peter, M. Cytosolic pH is a second messenger for glucose and regulates the PKA pathway through V-ATPase. EMBO J. 2010, 29, 2515–2526. [CrossRef] [PubMed] 87. References Tarsio, M.; Zheng, H.; Smardon, A.M.; Martínez-Muñoz, G.A.; Kane, P.M. Consequences of loss of Vph1 protein-containing vacuolar ATPases (V-ATPases) for overall cellular pH homeostasis. J. Biol. Chem. 2011, 286, 28089–28096. [CrossRef] [PubMed] 17 of 17 17 of 17 Microorganisms 2021, 9, 7 88. Orij, R.; Urbanus, M.L.; Vizeacoumar, F.J.; Giaever, G.; Boone, C.; Nislow, C.; Brul, S.; Smits, G.J. Genome-wide analysis of intracellular pH reveals quantitative control of cell division rate by pHc in Saccharomyces cerevisiae. Genome Biol. 2012, 13, R80. [CrossRef] [PubMed] [ ] [ ] 89. Dechant, R.; Saad, S.; Ibáñez, A.J.; Peter, M. Cytosolic pH regulates cell growth through distin to promote Ras/PKA and TORC1 activity. Mol. Cell 2014, 55, 409–421. [CrossRef] 89. Dechant, R.; Saad, S.; Ibáñez, A.J.; Peter, M. Cytosolic pH regulates cell growth through distinct GTPases, Arf1 and Gtr1, to promote Ras/PKA and TORC1 activity. Mol. Cell 2014, 55, 409–421. [CrossRef] 90. Finley, D.; Ulrich, H.D.; Sommer, T.; Kaiser, P. The ubiquitin-proteasome system of Saccharomyces cerevisiae. Genetics 2012, 192, 319 360 [CrossRef] p / y , , [ ] 90. Finley, D.; Ulrich, H.D.; Sommer, T.; Kaiser, P. The ubiquitin-proteasome system of Saccharomyces cerevisiae. Genetics 2012, 192, 319–360. [CrossRef] 91. Pickart, C.M.; Rose, I.A. Functional heterogeneity of ubiquitin carrier proteins. J. Biol. Chem. 1985, 92 W l D M Li A B i PS Kl i R E UBCH7 i i ﬁl l ki d 91. Pickart, C.M.; Rose, I.A. Functional heterogeneity of ubiquitin carrier proteins. J. Biol. Chem. 1985, 260, 1573–1581. 92. Wenzel, D.M.; Lissounov, A.; Brzovic, P.S.; Klevit, R.E. UBCH7 reactivity proﬁle reveals parkin and HHARI to be RING/HECT hybrids. Nature 2011, 474, 105–108. [CrossRef] 93. Stoll, K.E.; Brzovic, P.S.; Davis, T.N.; Klevit, R.E. The essential Ubc4/Ubc5 function in yeast is HECT E E3-dependent pathways require only monoubiquitin transfer by Ubc4. J. Biol. Chem. 2011, 286, 15165–15 93. Stoll, K.E.; Brzovic, P.S.; Davis, T.N.; Klevit, R.E. The essential Ubc4/Ubc5 function in yeast is HECT E3-dependent, and RING E3-dependent pathways require only monoubiquitin transfer by Ubc4. J. Biol. Chem. 2011, 286, 15165–15170. [CrossRef] [PubMed] 94. Gournas, C.; Gkionis, S.; Carquin, M.; Twyffels, L.; Tyteca, D.; André, B. Conformation-dependent partitioning of yeast nutrient transporters into starvation-protective membrane domains. Proc. Natl. Acad. Sci. USA 2018, 115, E3145–E3154. [CrossRef] [PubMed] E3 dependent pathways require only monoubiquitin transfer by Ubc4. J. Biol. Chem. 2011, 286, 15165 15170. [CrossRef] [PubMed] 94. References Gournas, C.; Gkionis, S.; Carquin, M.; Twyffels, L.; Tyteca, D.; André, B. Conformation-dependent partitioning of yeast nutrient transporters into starvation-protective membrane domains. Proc. Natl. Acad. Sci. USA 2018, 115, E3145–E3154. [CrossRef] [PubMed] oster, J.S.; Ruiz, S.J.; Poolman, B. Regulation of amino acid transport in Saccharomyces cerevisiae. Microbiol. Mol 20. [CrossRef] [PubMed] [ ] 95. Bianchi, F.; Van't Klooster, J.S.; Ruiz, S.J.; Poolman, B. Regulation of amino acid transport in Saccharomyce Biol. Rev. 2019, 83, 720. [CrossRef] [PubMed] 96. Bianchi, F.; Syga, Ł.; Moiset, G.; Spakman, D.; Schavemaker, P.E.; Punter, C.M.; Seinen, A.-B.; van Oijen, A.M.; Robinson, A.; Poolman, B. Steric exclusion and protein conformation determine the localization of plasma membrane transporters. Nat. Commun. 2018, 9, 501. [CrossRef] [PubMed]
https://openalex.org/W4384341029	https://www.qeios.com/read/5HVILT/pdf	English	null	Review of: "Deep Learning Modeling for Prediction of Cognitive Task Related Features from Resting-state fMRI Data"	null	2,023	cc-by	279	Qeios, CC-BY 4.0 · Review, July 14, 2023 Qeios ID: 5HVILT · https://doi.org/10.32388/5HVILT Review of: "Deep Learning Modeling for Prediction of Cognitive Task Related Features from Resting-state fMRI Data" Yunjiang Ge Yunjiang Ge Yunjiang Ge Potential competing interests: No potential competing interests to declare. The authors proposed a modified connectivity-based graph convolutional network (cGCN) and tested the performance of the model through an data example. The structure looks fine, but the writing can be improved. I have the following comments for the authors. In the first paragraph in the introduction, the authors may need to reorganize the structure and provide reference for the assertions. The connection between the gcn and and brain connectivity is unclear with current statement. In section 2.2, the title is ‘MRI data’. However, the real data example is an application on the fMRI scan. The authors may want to clarify or unify the term. Also, it’ll be helpful if the authors can specify the image type (i.e. T1, T2 image). The authors may add a brief introduction for the image data, including the study purpose. It’ll be better if the authors can provide the reference of the atlas. The authors may add a brief introduction for the image data, including the study purpose. It’ll be better if the authors can provide the reference of the atlas. When reporting the real data example results, it may not be necessary to list out that many regions. Also to interpret the findings, the author may provide some reference with the similar setting and study findings to support their claim. The authors may consider adding the competing methods result and compare the performance. Qeios ID: 5HVILT · https://doi.org/10.32388/5HVILT 1/1
https://openalex.org/W3026994471	https://bibliotecadigital.ipb.pt/bitstream/10198/18516/1/30.pdf	English	null	Sex Differences in Swimming Disciplines—Can Women Outperform Men in Swimming?	International journal of environmental research and public health/International journal of environmental research and public health	2,020	cc-by	12,343	International Journal of Environmental Research and Public Health Received: 25 March 2020; Accepted: 19 May 2020; Published: 22 May 2020 Abstract: In recent years, the interest of female dominance in long-distance swimming has grown where several newspaper articles have been published speculating about female performance and dominance—especially in open-water ultra-distance swimming. The aim of this narrative review is to review the scientiﬁc literature regarding the diﬀerence between the sexes for all swimming strokes (i.e., butterﬂy, backstroke, breaststroke, freestyle and individual medley), diﬀerent distances (i.e., from sprint to ultra-distances), extreme conditions (i.e., cold water), diﬀerent ages and swimming integrated in multi-sports disciplines, such as triathlon, in various age groups and over calendar years. The inﬂuence of various physiological, psychological, anthropometrical and biomechanical aspects to potentially explain the female dominance was also discussed. The data bases Scopus and PUBMED were searched by April 2020 for the terms ’sex–diﬀerence–swimming’. Long-distance open-water swimmers and pool swimmers of diﬀerent ages and performance levels were mainly investigated. In open-water long-distance swimming events of the ’Triple Crown of Open Water Swimming’ with the ’Catalina Channel Swim’, the ’English Channel Swim’ and the ’Manhattan Island Marathon Swim’, women were about 0.06 km/h faster than men. In master swimmers (i.e., age groups 25–29 to 90–94 years) competing in the FINA (Fédération Internationale de Natation) World Championships in pool swimming in freestyle, backstroke, butterﬂy, breaststroke, individual medley and in 3000-m open-water swimming, women master swimmers appeared able to achieve similar performances as men in the oldest age groups (i.e., older than 75–80 years). In boys and girls aged 5–18 years—and listed in the all-time top 100 U.S. freestyle swimming performances from 50 m to 1500 m—the ﬁve fastest girls were faster than the ﬁve fastest boys until the age of ~10 years. After the age of 10 years, and until the age of 17 years, however, boys were increasingly faster than girls. Therefore, women tended to decrease the existing sex diﬀerences in speciﬁc age groups (i.e., younger than 10 years and older than 75–80 years) and swimming strokes in pool-swimming or even to overperform men in long-distance open-water swimming (distance of ~30 km), especially under extreme weather conditions (water colder than ~20 ◦C). Two main variables may explain why women can swim faster than men in open-water swimming events: (i) the long distance of around 30 km, (ii) and water colder than ~20 ◦C. Future studies may investigate more detailed (e.g., anthropometry) the very young (<10 years) and very old (>75–80 years) age groups in swimming Int. J. Environ. Res. Int. J. Environ. Res. Public Health 2020, 17, 3651; doi:10.3390/ijerph17103651 Sex Diﬀerences in Swimming Disciplines—Can Women Outperform Men Gallen, Switzerland 2 Institute of Primary Care, University of Zurich, 8091 Zurich, Switzerland; thomas.rosemann@usz.ch 3 Faculty of Physical Education and Sport Sciences, School of Physical Education and Sports Science, Aristotle University of Thessaloniki, 55535 Thessaloniki, Greece; dalammi9@hotmail.com 4 Physical Education and Sport Science Academic Group, National Institute of Education, Nanyang Technological University, Singapore 637616, Singapore; tiago.barbosa@nie.edu.sg 5 Research Centre in Sports, Health and Human Development, 5001-801 Vila Real, Portugal 6 Department of Sport Sciences, Polytechnic Institute of Bragança, 5300-253 Bragança, Portugal 7 Bouve College of Health Sciences, Northeastern University, Boston, MA 02115, USA; cvsousa89@gmail.com 8 Exercise Physiology Laboratory, 18450 Nikaia, Greece; pademil@hotmail.com 9 School of Health and Caring Sciences, University of West Attica, 12243 Athens, Greece * Correspondence: beat.knechtle@hispeed.ch; Tel.: +41-(0)-71-226-93-00 Beat Knechtle 1,2,* , Athanasios A. Dalamitros 3 , Tiago M. Barbosa 4,5,6, Caio Victor Sousa 7 , Thomas Rosemann 2 and Pantelis Theo Nikolaidis 8,9 1 Medbase St. Gallen Am Vadianplatz, 9001 St. Gallen, Switzerland 2 Institute of Primary Care, University of Zurich, 8091 Zurich, Switzerland; thomas.rosemann@usz.ch 3 Faculty of Physical Education and Sport Sciences, School of Physical Education and Sports Science, Aristotle University of Thessaloniki, 55535 Thessaloniki, Greece; dalammi9@hotmail.com 4 Physical Education and Sport Science Academic Group, National Institute of Education, Nanyang Technological University, Singapore 637616, Singapore; tiago.barbosa@nie.edu.sg 5 Research Centre in Sports, Health and Human Development, 5001-801 Vila Real, Portugal 6 Department of Sport Sciences, Polytechnic Institute of Bragança, 5300-253 Bragança, Portugal 7 Bouve College of Health Sciences, Northeastern University, Boston, MA 02115, USA; cvsousa89@gmail.com 8 Exercise Physiology Laboratory, 18450 Nikaia, Greece; pademil@hotmail.com 9 School of Health and Caring Sciences, University of West Attica, 12243 Athens, Greece * Correspondence: beat.knechtle@hispeed.ch; Tel.: +41-(0)-71-226-93-00 Sex Diﬀerences in Swimming Disciplines—Can Women Outperform Men Beat Knechtle 1,2,* , Athanasios A. Dalamitros 3 , Tiago M. Barbosa 4,5,6, Caio Victor Sousa 7 , Thomas Rosemann 2 and Pantelis Theo Nikolaidis 8,9 1 Medbase St. Gallen Am Vadianplatz, 9001 St. Gallen, Switzerland 2 Institute of Primary Care, University of Zurich, 8091 Zurich, Switzerland; thomas.rosemann@usz.ch 3 Faculty of Physical Education and Sport Sciences, School of Physical Education and Sports Science, Aristotle University of Thessaloniki, 55535 Thessaloniki, Greece; dalammi9@hotmail.com 4 Physical Education and Sport Science Academic Group, National Institute of Education, Nanyang Technological University, Singapore 637616, Singapore; tiago.barbosa@nie.edu.sg 5 Research Centre in Sports, Health and Human Development, 5001-801 Vila Real, Portugal 6 Department of Sport Sciences, Polytechnic Institute of Bragança, 5300-253 Bragança, Portugal 7 Bouve College of Health Sciences, Northeastern University, Boston, MA 02115, USA; cvsousa89@gmail.com 8 Exercise Physiology Laboratory, 18450 Nikaia, Greece; pademil@hotmail.com 9 School of Health and Caring Sciences, University of West Attica, 12243 Athens, Greece * Correspondence: beat.knechtle@hispeed.ch; Tel.: +41-(0)-71-226-93-00 Beat Knechtle 1,2,* , Athanasios A. Dalamitros 3 , Tiago M. Barbosa 4,5,6, Caio Victor Sousa 7 , Thomas Rosemann 2 and Pantelis Theo Nikolaidis 8,9 1 Medbase St. Gallen Am Vadianplatz, 9001 St. Gallen, Switzerland 2 Institute of Primary Care, University of Zurich, 8091 Zurich, Switzerland; thomas.rosemann@usz.ch 3 Faculty of Physical Education and Sport Sciences, School of Physical Education and Sports Science, Aristotle University of Thessaloniki, 55535 Thessaloniki, Greece; dalammi9@hotmail.com 4 Physical Education and Sport Science Academic Group, National Institute of Education, Nanyang Technological University, Singapore 637616, Singapore; tiago.barbosa@nie.edu.sg 5 Research Centre in Sports, Health and Human Development, 5001-801 Vila Real, Portugal 6 Department of Sport Sciences, Polytechnic Institute of Bragança, 5300-253 Bragança, Portugal 7 Bouve College of Health Sciences, Northeastern University, Boston, MA 02115, USA; cvsousa89@gmail.com 8 Exercise Physiology Laboratory, 18450 Nikaia, Greece; pademil@hotmail.com 9 School of Health and Caring Sciences, University of West Attica, 12243 Athens, Greece * Correspondence: beat.knechtle@hispeed.ch; Tel.: +41-(0)-71-226-93-00 Beat Knechtle 1,2,* , Athanasios A. Dalamitros 3 , Tiago M. Barbosa 4,5,6, Caio Victor Sousa 7 , Thomas Rosemann 2 and Pantelis Theo Nikolaidis 8,9 1 Medbase St. Gallen Am Vadianplatz, 9001 St. 1. Introduction Swimming is a speciﬁc sports discipline which can be performed in a range of styles, usually referred to as ’strokes’ [1–4], over diﬀerent lengths [5,6] and in both pools (i.e., indoor, outdoor) of diﬀerent lengths (mainly 25 m and 50 m) and in open water (i.e., sea, lake, river) [7–10]. The diﬀerent swimming strokes are butterﬂy [1], backstroke [2], breaststroke [3], freestyle [4] and individual medley referred to the combination of the four diﬀerent strokes [11]. In addition to these individual events, four swimmers can take part in either a freestyle or medley relay. In pool-swimming competitions, the distances for butterﬂy, backstroke, breaststroke and freestyle usually include 50 m to 200 m, whereas individual medley is held over 200 m and 400 m [11]. In freestyle, the 800 m and the 1500 m are further race distances in pool-swimming [4,12]. Indoor-swimming events with a deﬁned time limit (i.e., 12 h) are also held [13]. In open-water swimming, master swimmers most often compete in 3000 m [14], whereas 5 km [6,10], 10 km [10,15] and 25 km [10,15] races were held for elite swimmers. Open-water swimming events of diﬀerent lengths in lakes and seas are held as solo swims [7]. Swimming is also part of multi-sports races like triathlons over diﬀerent distances like the Olympic Distance triathlon [16,17], the half-Ironman [18], the Ironman [16,19] and longer triathlon race distances than the Ironman distance [20]. [ ] In recent years, the interest of female dominance in long-distance swimming grew where several newspaper articles were published speculating about the female performance and dominance especially in open-water ultra-distance swimming. In one newspaper article, the history of female performance in open-water swimming started with Gertrud Ederle in the ’English Channel Crossing’ and the female dominance in ’Manhattan Island Marathon Swim’. The author discussed the problem of comparing the fastest men and women in contrast to all men and women, leading to a diﬀerent ﬁnding regarding male or female dominance. Particularly, the fastest men beat the fastest women, but that the average woman was faster than the average man [21]. In addition, another newspaper article describing the sex diﬀerence in swimming and running discussed the aspect of the fastest women and men [22]. Keywords: gender diﬀerence; sex gap; swimming performance; swimming stroke; holistic approach Keywords: gender diﬀerence; sex gap; swimming performance; swimming stroke; holistic approach Keywords: gender diﬀerence; sex gap; swimming performance; swimming stroke; holistic approach Received: 25 March 2020; Accepted: 19 May 2020; Published: 22 May 2020 Public Health 2020, 17, 3651; doi:10.3390/ijerph17103651 www.mdpi.com/journal/ijerph www.mdpi.com/journal/ijerph 2 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 3.1. Pool-Swimming Swimming competitions are held in pool-swimming in short-course (i.e., 25 m or 25 yards) and long-course (i.e., 50 m) pools from 50 m to 200 m in the four diﬀerent strokes such as butterﬂy, backstroke, breaststroke and freestyle [28,29], where freestyle races were also held for 400 m, 800 m and 1500 m [29]. Apart from distance-limited swimming races, also time-limited swimming events (i.e., 12 h) are performed in pool-swimming [13]. Studies investigated diﬀerent populations such as elite swimmers competing at national and international level for diﬀerent strokes and distances, youth and master swimmers [1–3,8,29–31]. In pool-swimming, it seemed that the sex diﬀerence varied with the distance of the events [29,30,32]. For elite swimmers competing in diﬀerent strokes such as freestyle [28], butterﬂy [31], breaststroke [30] and individual medley [30], the sex diﬀerence decreased with increasing race distance. In long-distance pool-swimming such as a 12-h-swim, women were able to achieve a similar performance to men. In the ’Zurich 12-h Swim’ held in Switzerland, the annual best performance did not diﬀer between males (~38.3 km) and females (~34.4 km), respectively [8]. For master swimmers competing at the FINA World Championships in diﬀerent age groups for diﬀerent strokes and distances, men were faster than women for all strokes, distances and age groups except in 50–800-m freestyle for age groups 80–84 and 85–89 years [4], in 50–200-m butterﬂy for age group 90–94 years [1], in 200-m and 400-m individual medley for age groups 85–89 und 90–94 years [11], in 50–200-m breaststroke for age groups 90–94 and 95–99 years [3] and in 50–200-m backstroke for age groups 85–89, 90–94 and 95–99 years [2] where women achieved a similar performance than men. The disparate ﬁndings were explained by diﬀerences in performance level, race distance, stroke, age and sample size. 1. Introduction A further newspaper article cites Steven Munatones, one of the world’s top experts on open-water swimming, reporting that the average female time was 33 min faster than the average male time in the 135 years of the ’English Channel Crossing’ [23]. It seemed that women were better in long-distance open-water swimming. A further newspaper article reported that the female swimmer Sarah Thomas was the ﬁrst person in the world to cross the ’English Channel’ four times in a row without stopping [24]. In a further newspaper article, women were described winning also ultra-endurance races in cycling and running ahead of all men [25]. Moreover, a newspaper article reported that the fastest women ever were faster than the fastest men ever in both the ‘Catalina Channel Swim’ and in the ’Manhattan Island Marathon Swim’ held in the USA [26]. These descriptions lead to the intention to review existing literature to conﬁrm or disprove these statements of female dominance in open-water swimming. Since sex-related diﬀerences including anthropometric characteristics, swimming energy, as well as stroking parameters have been previously reported [27], it would also be interesting to examine this ’sex gap’ as translated into swimming performance. Such information would have both theoretical and practical relevance. From a theoretical point of view, the sex diﬀerence in human performance has been a major topic in exercise physiology, and thus, researchers working in this area would beneﬁt from new knowledge on sex diﬀerences in swimming. From a practical perspective, coaches usually working with both sexes could use sex diﬀerence in swimming to optimize the training of their athletes. Therefore, the aim of the present research was to review original studies on sex diﬀerences in swimming performance with regards to age, swimming strokes and race distance. Int. J. Environ. Res. Public Health 2020, 17, 3651 3 of 17 3 of 17 2. Method The data bases Scopus and PUBMED were searched on April 2020 for the terms ’sex–diﬀerence—swimming’. The search in Scopus led to 239 entries, the search in PUBMED to 558 entries. We excluded case studies, case reports, animal studies, studies with divers and rowers and studies with patients. Although the newspaper articles primarily reported about outstanding female achievements in open-water long-distance swimming, we consider in this narrative review all scientiﬁc results upon diﬀerences between the sexes in swimming for all swimming strokes (i.e., butterﬂy, backstroke, breaststroke, freestyle and individual medley), distances (i.e., from sprint to ultra-distances), conditions (i.e., cold water), ages (i.e., youth and master swimmers) and swimming integrated in multi-sports disciplines such as triathlon. 3.2. Open-Water Swimming In addition to pool-swimming events, open-water swims were held as individual swims (i.e., solo swims in Channel Crossings without drafting) or competitions in long-distance swimming events up to 25 km in open-water swimming [10,15] where swimmers are allowed to swim in a group and where drafting is allowed. In addition, there were open-water swimming events of diﬀerent lengths in lakes and seas [7] where drafting is also allowed. In open-water swimming, the water temperature may be of importance for female dominance. For oﬃcial swimming competitions held in heated pools, the FINA has established that water temperature shall be at 25 ◦C to 28 ◦C [33]. Apart from swimming competitions held in heated pools, swimming events were also held in open water such as rivers, lakes and seas. For open-water swimming events sanctioned by the FINA (i.e., World Cup races in 5 km, 10 km and 25 km), the FINA has established that the water temperature should be a minimum of 16 ◦C and a maximum of 31 ◦C [34]. In Channel Crossings like the ’English Channel Swim’, swimmers face, however, a water temperature of 15 ◦C at the end of June, increasing to 18 ◦C by the beginning of September [35]. Water temperature seemed to have an inﬂuence on the performance of the swimmers. For open-water 4 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 swimmers competing in the ‘Marathon Swim in Lake Zurich’, a 26.4 km open-water ultra-swim held in Switzerland, performance of the top swimmers was negatively related to water temperature [9]. Female performance was investigated for diﬀerent solo swims where drafting is not possible. In these studies, diﬀerent groups of diﬀerent performance levels were investigated. It seemed that women were able to achieve a similar performance to men in solo swims in long-distance open-water swimming such as individual Channel Crossings [7,36,37], where water temperatures were generally below 20 ◦C. In lake swimming such as ‘Marathon Swim Lake Zurich’ with water temperatures at or warmer than 20 ◦C, women achieved a similar performance to men [9]. However, depending upon the investigated sample (i.e., the fastest woman/man, the three fastest women/men, the ﬁve fastest women/men, the ten fastest women/men, all women/men, the annual fastest women/men, the annual three fastest women/men), women were able to outperform men in long-distance open-water swimming [36–38]. 3.2. Open-Water Swimming In the ‘English Channel Crossing’, the overall female swim time of 13:16 h:min was not diﬀerent compared to the overall male swim time of 13:35 h:min between 1875 and 2011 [7]. Although the fastest male swim time (6:57 h:min) during this period was 6.7% faster than the fastest female swim time (7:25 h:min), the sex diﬀerence in performance of the top three times was ~8.9% [7]. The fastest annual swim speed did not diﬀer between men (~0.89 m/s) and women (~0.84 m/s) [8]. In the ’Triple Crown of Open Water Swimming’ with ’Catalina Channel Swim’, ’English Channel Swim’ and ’Manhattan Island Marathon Swim’, overall women were ~0.06 km/h faster than overall men [31]. However, women were ~0.07 km/h slower than men when considering the annual ﬁve fastest swimmers [37]. Analyses were also performed for the single events of the ’Triple Crown of Open Water Swimming’. In the ’Catalina Channel Swim’, the fastest woman ever was faster than the fastest man ever (~22 min) [36]. The three fastest women were faster than the three fastest men (~20 min), however, the diﬀerence reached no statistical signiﬁcance [36]. The ten fastest women were ~1 min faster than the ten fastest men, however, also here, the diﬀerence reached no statistical signiﬁcance [36]. However, the annual fastest women (~10:51 h:min) were ~52.9 min (~16%) faster than the annual fastest men [36]. In a further open-water event held in the USA, women were faster than men. In the ’Manhattan Island Marathon Swim’, the ten fastest women were ~12%–14% faster than the ten fastest men [38]. Open-water swimming is also held for master swimmers competing in 3000 m. In master swimmers competing at the FINA World Championships, men were faster than women for all age groups except age groups 75–79, 80–84 and 85–89 years where women achieved the same performance like men [14]. Not only the age, but also the distance may be of importance. It seemed that in shorter open-water swimming events with a higher water temperature, women have a disadvantage compared to men. In the ’Marathon Swim in Lake Zurich’, the male record was 2.3% faster than the female record. For the annual winners, men were ~11.5% faster [9]. Based on these observations, in most cases analyzed, there were no sex diﬀerences in performance during open-water swimming. In some instances, women were faster than men. 3.2. Open-Water Swimming However, a variety of parameters such as water temperature and distance can inﬂuence the outcome. Depending upon the sample size (i.e., the fastest woman/man, the three fastest women/men, the ﬁve fastest women/men, the ten fastest women/men, all women/men, the annual fastest women/men, the annual three fastest women/men) and the statistical approach (i.e., comparison of groups or comparison of changes over time), women were faster than men in this speciﬁc sports discipline. 3.4. Age Age is an important aspect regarding the sex diﬀerence in swimming performance. This aspect was investigated for diﬀerent age groups, strokes and swimming distances. An analysis of sex diﬀerences in swimming speed for the top-10 World ranking (i.e.,1st–10th place), age group (25–89 years), and event distance from the world’s top ten swimming times of both women and men in the World Championships showed that the sex diﬀerence in swimming speed increased with world record place and age [42]. Very recent studies investigated the performance trends and sex diﬀerence in swimming performance in master swimmers competing in the FINA World Championships in pool-swimming in freestyle [4], in backstroke [2], in butterﬂy [1], in breaststroke [3], in individual medley [11] in 3000-m open-water swimming [14] and for youth swimmers [43]. In butterﬂy [1], in breaststroke [3], in backstroke [2], in freestyle [4] and in individual medley [11], women were able to reduce the gap to men in diﬀerent age groups. In 3000-m open-water swimming, however, women were not able to reduce the sex diﬀerence to men [14]. Consequently, the existing sex diﬀerence regarding all swimming strokes is evident during the youngest (i.e., 25 to 29 years) age groups, while for nearly all the rest of the age groups (i.e., 30 years and older), women tend to reduce this sex gap. In addition, for youth swimmers, age is of importance. When boys and girls from the age of 5 to 18 years for 50 m to 1500 m from the all-time top 100 U.S. freestyle swimming performances were investigated, the top ﬁve girls were faster until the age of ~10 years than boys. After the age of 10 years, however, boys were increasingly faster than girls until the age of ~17 years [43]. Overall, female swimmers can beat male swimmers under the age of ~10 years and achieve almost the same performance as men in the highest age groups (i.e., older than ~75–80 years) depending upon the distance and the stroke. A few studies have investigated the age eﬀect in open-water swimming. The age of peak performance increased over calendar years in long-distance open-water swimming. In the ’Manhattan Island Marathon Swim’, the age of the annual three fastest swimmers increased between 1983 and 2013 from ~28 to ~38 years for women and from ~23 to ~42 years for men [38]. 3.4. Age In the 26.4 km open-water ultra-swim ’Marathon Swim in Lake Zurich’, Switzerland, the mean age of the ﬁnishers during the period 1987–2011 was ~32.0 years for men and ~30.9 years for women. The mean age of ﬁnishers and the age of winners increased across the years for both sexes [9]. 3.3. Ice Swimming Since 2009, ice swimming for 1 mile and 1 km is a new discipline in open-water swimming [39,40]. In this swimming discipline, water temperature must be colder than +5 ◦C. One may assume that female performance may be better when water temperature is very low. Performances of women and men were investigated for ’Ice Mile’ and ’1 km Ice event’ where the fastest men were faster than the fastest women in both events [41]. Obviously, women had no advantage in this cold water. In the ’Ice Mile’, variables such as calendar year, number of swims, water temperature and wind chill showed no relation to swimming speed for both women and men. Water temperature was not correlated to 5 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 swimming speed in either ’Ice Mile’ or ’1 km Ice event’ for both women and men [41]. Therefore, the limited data concerning ice swimming demonstrate that men have an advantage, compared to women, in this speciﬁc condition. Regarding the results from long-distance open-water swimming, the length of the event may be decisive. Future studies may investigate the body composition of open-water long-distance swimmers and ice swimmers. There may be a diﬀerence in body fat in the competitors in the two disciplines. 3.7. Changes in Swimming Performance Over Years Some studies investigated the change in performance over calendar years [8,48]. There seem to be diﬀerences between sexes, disciplines, performance level and distances [8]. In the past, it has already been suggested that women would soon perform better than men in swimming. In 1977, it was reported that women were gaining on their male counterparts at the rate of 0.45% a year in the 100-yard freestyle [48]. It was assumed that with that rate of improvement national level women may catch up with male counterparts by the year 2003. Likewise, in the 1650-yard freestyle, women were gaining on men, but at a slower rate of improvement of ~0.155%. It was assumed that it would take the women ~51 years to catch up to the men. The authors found that race times in women were improving at a rate faster than race times in men, but at some time in the future the rate of growth would probably stabilize for both sexes [48]. While at present that assumption, regarding the speciﬁc race distance (i.e., 100 yards), has not been completely fulﬁlled, it remains to be seen if a performance plateau would allow women to outperform men. Regarding newer studies, female performance has improved over calendar years in some instances [8]. These analyses of changes in performance over the years have been performed for elite pool swimmers [5], for open-water swimmers [7,8] and for master swimmers competing in freestyle [4], backstroke [2], butterﬂy [1], breaststroke [3], individual medley [11] and in 3000-m open-water swimming [14]. In pool-swimming, performance was improved for most distances in both elite and master swimmers in backstroke [49], freestyle [31,46], breaststroke [30], butterﬂy [1,50], individual medley [11,30] and in 3000-m open-water swimming [14]. Some studies have investigated open-water swimming and showed that performance changed over years. For women and men crossing the ’Catalina Channel’ between 1927 and 2014, performance decreased nonlinearly in the annual fastest men and women [36]. In the ’Manhattan Island Marathon Swim’, race times of the annual three fastest women and men did not diﬀer between sexes and remained stable across the years [38]. In the ’Maratona del Golfo Capri-Napoli’, race times of the annual fastest swimmers decreased linearly for women and for men from 1954 to 2013 from 39.2% to 4.7% [51]. 3.6. Performance Level and Sex Diﬀerence in Performance In some studies, the changes in sex diﬀerence over time were investigated for diﬀerent levels of athletes (i.e., national level, international level) [44,47]. In swimmers competing at national and international level, the sex-related diﬀerence in swimming speed was greater for freestyle than for breaststroke in 50-m to 200-m race distances for national swimmers, but not for international swimmers. For both groups, the sex-related diﬀerence for both freestyle and breaststroke swimming speeds decreased with increasing race distance. The sex-related diﬀerences in performance were greater for freestyle than for breaststroke for swimmers at national level, but not for swimmers at international level [47]. The disparate ﬁndings were explained by diﬀerences in performance level, distance, stroke and sample size. 3.5. Sex Diﬀerence and Swimming Strokes Some studies investigated the aspect of sex diﬀerence for diﬀerent strokes. In pool-swimming competitions, athletes perform in the four strokes (i.e., butterﬂy, backstroke, breaststroke and freestyle) [28] as well as in the combination of all four strokes as individual medley [11]. There seem to be changes in the sex diﬀerence for the swimming strokes depending upon the distance and the performance level. For both 200-m and 400-m freestyle and individual medley, no sex diﬀerence was found between neither the two distances, nor between the two swimming strokes [44]. The sex diﬀerences were ~9.7% and ~7.1% in individual medley and ~10.1% and ~6.1% in freestyle, respectively [45]. For elite male and female butterﬂy and freestyle swimmers at national level, the sex diﬀerence in peak swimming speed was lower in butterﬂy than in freestyle [31,46]. For national and international breaststroke and freestyle swimmers, the sex diﬀerences in swimming speed increased over time for national swimmers, but not for international swimmers for freestyle, while the sex diﬀerence remained stable for both national and international breaststroke swimmers [30]. Int. J. Environ. Res. Public Health 2020, 17, 3651 6 of 17 The disparate ﬁndings were explained by the diﬀerent performance levels, the diﬀerent distances and strokes and the diﬀerent sample sizes. The disparate ﬁndings were explained by the diﬀerent performance levels, the diﬀerent distances and strokes and the diﬀerent sample sizes. 3.8. Swimming in Multi-Sports Disciplines Like Triathlons Swimming is the ﬁrst segment of a triathlon event, followed by cycling and then running [52]. Several studies investigated the trends in performance and the sex diﬀerence in performance in swimming in triathlons of diﬀerent lengths such as the Olympic distance triathlon (i.e., 1.5 km swimming, 40 km cycling and 10 km running) [16,17,53], the Ironman distance triathlon (i.e., 3.8 km swimming, 180 km cycling and 42.195 km running) [16] and ultra-triathlon distances longer than the Ironman distance [20,54,55]. In Olympic distance triathletes competing in the ’Zürich Triathlon’ in Switzerland from 2000 to 2010, the sex diﬀerence in swimming was 15.2% for the top ﬁve triathletes overall [17]. For the world’s best triathletes at the ITU (International Triathlon Union) World Triathlon Series during the 2009–2012 period including the 2012 London Olympic Games, swim times and the sex diﬀerence in swimming remained unchanged [53]. For longer triathlon distances than the Olympic distance, sex diﬀerence has been investigated for the Ironman distance [19,56,57] and longer triathlon distances from 2× to 10× the Ironman distance [20,57]. It was shown that women improved swimming performance and closed the gap to men. In ’Ironman Hawaii’, the overall top ten men ﬁnishers improved their swimming performance between 1983 and 2012. The sex diﬀerence remained unchanged over the years at ~12.5% [19]. For the annual three best ﬁnishers in ’Ironman Hawaii’, the sex diﬀerence decreased nonlinearly in swimming between 1978 and 2013 [57]. In ’Isklar Norseman Xtreme Triathlon’ held over the Ironman distance, athletes swim at a water temperature of ~13–15 ◦C. Men were faster than the women in cycling, but not in swimming, running or overall race time. Across years, women improved their performance in swimming and both women and men improved their performance in cycling and in overall race time. In running, however, neither women nor men improved [3]. Diﬀerent ﬁndings were, however, reported for longer triathlon distances. In Double Iron ultra-triathlon (i.e., 7.6 km swimming, 360 km cycling and 84.4 km running), men (2:36 h:min) were ~8 min faster than women (2:44 h:min) [20]. 3.8. Swimming in Multi-Sports Disciplines Like Triathlons For triathlon distances from the Ironman distance in ’Ironman Hawaii’ to the Double Deca Iron ultra-triathlon distance (i.e., 76 km swimming, 3600 km cycling and 840 km running), the sex diﬀerence in performance showed no change with increasing race distance with the exception for the swimming split where the sex diﬀerence increased with increasing race distance for the three fastest ever [57]. Regarding triathlon swimming performance, in most cases, the sex diﬀerence tends to remain unchanged over the years. However, as previously mentioned, women can outperform men in speciﬁc triathlon races under more extreme conditions (i.e., water temperature of ~13–15 ◦C). 3.7. Changes in Swimming Performance Over Years For the annual top three swimmers, race times decreased linearly between 1963 and 2013 for women and for men from 38.2% ± 14.0% to 6.0% ± 1.0% [51]. In the ’English Channel Crossing’, the performance of the annual top three swimmers showed no changes either both females or males over the last 36 years and the sex diﬀerence remained unchanged at ~12.5% over the years [7]. In the ’English Channel Crossing’, performance increased progressively for both sexes, but was lower for female than for male athletes from 1900 to 2010 [8]. A diﬀerent kind of events was the FINA races which were not held as solo events and swimmers could draft. For elite male and female swimmers competing at the FINA World Cup events of 5 km, 10 km and 25 km events, swimming speed of the annual ten fastest women decreased at 5 km and at 25 km, while it increased at 10 km. For the annual ten fastest men, peak swimming speed decreased at 7 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 5 km, while it remained unchanged at both 10 km and 25 km [10]. In the FINA 10 km competitions (i.e., World Cup races, European Championships, World Championships and Olympic Games) held between 2008 and 2012, swimming speed of the fastest women and men showed no changes across the years. Performance of the top ten female swimmers per event remained stable across calendar years. The top ten male swimmers per event showed a decrease in performance over years, even though swimming speed in the ﬁrst race (i.e., January 2008, 1.40 m/s) was slower than swimming speed in the last race (i.e., October 2012, 1.50 m/s) [50]. The disparate ﬁndings were explained by the diﬀerent performance levels, the diﬀerent distances and strokes, the diﬀerent ages, the diﬀerent periods of time and the diﬀerent sample sizes. 3.9. The Change in Sex Diﬀerence Over Years In the same way where swimming performance can change over years, also the sex diﬀerence in swimming performance may change over years [10,44,58,59]. In some instances, women reduced the gap to men [59,60], in others not [10,28,44,53]. In these studies, pool-swimmers of sub-elite and elite level [28,44], open-water long-distance swimmers [59], master swimmers up to very high ages and swimmers in triathlons [20,54,55] were analyzed. In pool-swimming, the changes in sex diﬀerence Int. J. Environ. Res. Public Health 2020, 17, 3651 8 of 17 diﬀer over time regarding the distance, the age and the discipline [14,28,31,47]. According to this information, the sex diﬀerence in relation to pool-swimming and open-water swimming performance is largely dependent on the parameters analyzed in this review (i.e., competition level, swimming stroke and distance). Several studies have examined the variation in sex diﬀerence over calendar years in open-water swimming events. In some events, the sex diﬀerence decreased, and the women reduced the gap to men [32,39,56], but in others not [9]. In women and men crossing the ’Catalina Channel’, the sex diﬀerence for all women and men decreased linearly between 1927 and 2014 from 52.4% to 7.1% [36]. The decrease of sex diﬀerence was linear suggesting that women continuously reduced the sex diﬀerence to men [36]. In the 36 km ’Maratona del Golfo Capri-Napoli’, the sex diﬀerence for the annual fastest swimmers, decreased linearly from 39.2% to 4.7% from 1955 to 2013 [51]. For the annual three fastest swimmers, the sex diﬀerence in performance decreased linearly from 38.2% ± 14.0% to 6.0% ± 1.0% from 1963 to 2013 [51]. Again, in this event, the linear change in both race times and sex diﬀerences indicates that women could achieve men’s performance or even to perform better than men in the near future in this event [51]. In ’La Traversée Internationale du Lac St-Jean’ (32 km) held between 1955 and 2012 in Canada, the sex diﬀerence remained unchanged over years for the annual fastest women and men at 8.8% [58]. For the annual three fastest women and men, the sex diﬀerence decreased across years (1975–2011) from 14.4% ± 11.0% to 3.7% ± 1.4% [58]. Overall, most studies found that women reduced the gap to men over years in open-water swimming. Diﬀerences were found in open-water swimming where women are allowed to draft behind men. 3.9. The Change in Sex Diﬀerence Over Years For elite male and female swimmers competing in 5 km, 10 km and 25 km open-water FINA World Cup races, elite female swimmers improved their performance in 10 km, but impaired performance in 25 km, leading to a linear decrease in sex diﬀerence in 10 km and a linear increase in sex diﬀerence in 25 km. The linear change in sex diﬀerences suggests that women will improve in the near future in 10 km, but not in 25 km [10]. In elite open-water swimmers competing at FINA 10 km races, the mean sex diﬀerence in performance for the fastest swimmers was stable across years [50]. In long triathlon races, women were not able to close the gap to men. For triathlon races longer than the Ironman race distance, the sex diﬀerence in swimming showed no change over years in either Double Iron ultra-triathlon [54] or in ’Ultraman Hawaii’ [55]. In Double Iron ultra-triathlon races, the swimming times remained unchanged across years with an unchanged sex diﬀerence for the annual three fastest women and men [20]. Potential explanations as to why sex diﬀerences decreased over the years or not could be the selected period of time [44,53,58], the level of the investigated athletes (i.e., annual fastest, annual ten fastest, national level, international level, etc.) [44,58], and/or whether the swimmers were solo swimmer, competing in a drafting race or triathletes. When the sex diﬀerence showed no change over time, the investigated period of time was most likely too short [53]. 4. Potential Explanations for The Female Dominance in Long-Distance Open Water Swimming and Age Group Swimmers Based on these results, we may conclude that women were able to outperform men in swimming in solo, long-distance swimming events held in water temperatures between 15 ◦C and 20 ◦C. Women were also able to achieve the same performance as men in all distances and disciplines of pool swimming at younger ages (i.e., younger than ~10 years) and older ages (i.e., from ~80 years onwards). Two main variables may explain why women can outperform men in open-water swimming: (i) the long distance of ~30 km, (ii) and water colder than ~20 ◦C. Potential explanations for the ﬁnding that women can achieve a better performance than men can be attributed to diﬀerences in anthropometric characteristics such as body composition [65], body weight [66], body fat [67–70], lean body mass [71,72], body height [73,74], muscle thickness [75] and muscle size [72]. Other possible explanations were diﬀerences in swimming biomechanics such as kinematic parameters [76], arm coordination and arm–leg coordination [76–78], energetic cost, diﬀerences in swimming economy and swimming eﬃciency [79–81], gliding [82], body roll [83], shoulder ﬂexibility [84], trunk ﬂexibility [85], knee ﬂexibility [86,87], propelling eﬃciency [88]. Beyond that, additional diﬀerences in motivational aspects [89–96], physiology [97–100] and biochemistry [101–103], recovery [104] and injury prevalence [105,106] can be pointed out. We found, however, no potential explanation for the youth swimmers. A very likely explanation that women were faster than men in speciﬁc open-water ultra-distance events is the fact that female swimmers have more body fat than male swimmers [67] leading to better insulation against the cold and better buoyancy for long swimming distances [69]. It is well known that sex is associated with waist-to-hip ratio and body fat percentage in swimmers [107]. Due to the higher body fat, female swimmers have a diﬀerent body shape compared to male swimmers. Male swimmers have a more central distribution of fat when compared to females, where body fat is built up in the region of legs [108]. In nonstationary swimming with changing velocity, water around the swimmer is set in motion which can be thought of as an added mass of water. Female swimmers have a lower added mass and relative added mass than male swimmers suggesting that sex diﬀerences in body shape may be associated with added mass [109]. 3.10. The Inﬂuence of Swimwear on Performance Sex diﬀerences among swimmers using a wetsuit have also been investigated [61]. A wetsuit can be mainly used in the swim split in triathlon races [61] or in long-distance open-water swimming races [62]. In triathletes, the eﬀect of a wetsuit on lighter female swimmers was no diﬀerent than the eﬀect on heavier male swimmers [61]. Swimming with or without a wetsuit shows a diﬀerence between the sexes [63,64]. When swimming speed was compared among women and men with or without wetsuit over diﬀerent distances, wearing a wetsuit improved swimming speed for both women and men, but the beneﬁt of the use of wetsuits depends on additional factors such as race distance. Women may be favored from wearing a wetsuit more than men in longer ultra-distance races of open-water swimming [63]. It has also been shown that high-tech swimsuits gave more pronounced advantage to men than women and for low resistance as compared with high resistance swimming strokes [64]. Yet, the new rules on swimsuits, in eﬀect since 2010, should also be considered when analyzing swimming performance. Int. J. Environ. Res. Public Health 2020, 17, 3651 9 of 17 4. Potential Explanations for The Female Dominance in Long-Distance Open Water Swimming and Age Group Swimmers In ultra-distances of open-water swimming, diﬀerent anthropometric characteristics such as body height, body mass index (BMI), length of arm and training characteristics (e.g., swimming speed) were associated with performance for men. For women, swimming speed during training was associated with performance, but not anthropometric characteristics. Considering all variables for men, BMI and swimming speed during training were related to race time, but not for women [110]. Diﬀerences in anthropometric characteristics do exist between female and male swimmers [111,112]. It was shown that diﬀerences in height, arm span, skinfold thicknesses (e.g., triceps, subscapular, crista iliaca, ileo-spinal, abdominal, thigh, leg, sum of skinfolds), bi-acromial-bi-iliac index, bone body mass, muscle and fat, ectomorphy and endomorphy exist [111]. In another study, elite female youth swimmers had greater skinfolds at triceps, suprailiac and abdominal site. Endomorphic somatotype was twofold greater among elite female compared to elite male youth swimmers [112]. Although women could outperform men in certain swimming disciplines, in general, elite men were faster than elite women [113]. The sex gap in swimming performance seemed to remain stable in shorter distances. In Olympic Trial swimming from 1972 to 2016, the performance gap in swimming remained at ~8% [114]. The plateau during these ~40 years in the performance gap highlighted the role of biologic background (e.g., longer limbs, larger muscle mass, greater aerobic capacity and lower fat mass) on race time. Current evidence indicates that women will not swim as fast as men in Olympic events, which justiﬁes sex segregation in these individual sport disciplines [114]. Men have an advantage of larger body size and muscle mass, a superior ventilation function and anaerobic and aerobic energy transfer systems. It is well known that male swimmers have a higher maximum oxygen uptake than female swimmers among both younger [97] and older age groups [98–100]. It is also known that ventilation functions including forced vital capacity (FVC), forced expiratory volume in one second (FEV1), FEV1/FVC and mandatory minute ventilation (MMV) were superior in male 10 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 athletes to those in females [115]. Furthermore, the average diameter of muscle ﬁbers was larger in men than in women [116]. Therefore, it is not a surprise that male youth swimmers show increased power values in both their legs and arms [117]. 4. Potential Explanations for The Female Dominance in Long-Distance Open Water Swimming and Age Group Swimmers Finally, although male swimmers have a higher muscle mass, maximum handgrip isometric strength values correlated with swimming race time, especially in female swimmers [118]. Compared to men, women have an enhanced ability to oxidize fat, superior hydrodynamics and more even pacing, which provide advantage, especially during prolonged swimming [113]. Regarding pacing in pool- swimming, an eﬀect of sex on lap time in master swimmers competing in 100, 200, 400 and 800-m freestyle at the World Championships suggested greater changes of pacing in women than in men [12]. Apart from anthropometric and/or physiological diﬀerences, psychological diﬀerences may also explain female dominance in certain swimming disciplines [91–95,119–121]. Female swimmers show diﬀerences to male swimmers regarding mental toughness [92]. Competitive female swimmers were emotionally secure, physically healthy and reasonably contented with their present social status. For these women, the emotional, social, and physical costs were deﬁnitely worth the sacriﬁces [120]. Eﬀects of cultural and sociopolitical norms and outdated stereotypes (i.e., reduced opportunity to participate and compete in sports) that inﬂuence the number of female competitors, particularly in earlier years, should also be considered. Women have traditionally been under-represented in sports. In the 19th century, women were engaged in non-competitive recreational activities, but not in competitive sport. In 1971 in the United States, less than 7% of high school varsity athletes were female. Title IX was set in the United States in 1972 with the aim of providing equal treatment in sports, regardless of sex and increasing the number of women in sports [122]. Even as recently as 1979 in Brazil, it was illegal for women to play football [123]. According to Capranica et al., there were still some countries in the 2012 London Olympic Games that did not have a female in their delegation [124]. There was an increase in the number of female athletes in all sports over the last century, including swimming, that likely contributed to the reduction in the gap in performance between the sexes [125]. A potential explanation for the improved performance of older women, especially in pool-swimming, is the fact that the age of peak performance has increased in women since the 1980s. 4. Potential Explanations for The Female Dominance in Long-Distance Open Water Swimming and Age Group Swimmers When 116 years from the ﬁrst Olympic Games (1898) to the 2014 Olympic Games were analyzed, regarding the ages at which peak performance was observed, peak performance ages in women have increased consistently since the 1980s in all the athletic events examined (i.e., track and ﬁeld, swimming, rowing and ice-skating events). When the age of peak female performance increased, it became similar to the age of male performance in many events. In the last 20 to 30 years, the age of peak female athletic performance increased, but not the age of male athletic performance [126]. Age in open-water swimmers may also be of importance. In the ’Manhattan Marathon Island Swim’, the age of peak performance of ultra-distance swimmers has changed across the last decades, with the fastest swimmers getting older between 1983 and 2013. During this period, the age of the three fastest swimmers raised from 28 to 38 years in women and from 23 to 42 years in men [38]. The fact that female competitors were younger than their male counterparts likely had an eﬀect on performance. Age was also of importance in other endurance athletes such as ultra-marathon runners [127,128]. In recent studies investigating ultra-marathoners [128] and athletes competing in tower running [129], women were able to close the gap to men [128] or to even outperform men [130]. In 50- and 100-mile ultra-marathoners, the sex diﬀerence in running performance decreased with increasing age and was smaller in the longer (100 miles) compared to the shorter (50 miles) distance [128]. This ﬁnding may be explained by the lower participation of women in longer ultra-marathon races. Also, in tower running (stair climbing), women were able to beat men in speciﬁc situations, e.g., in smaller buildings with less than 600 stairs for younger (30–59 years) and older (>69 years) age groups, in buildings with 1600–220 stairs for older ages (>69 years) and in buildings with more than 2200 stairs for younger (<20 years) and older (60–69 years) age groups [130]. With increasing age, experience may improve and race tactics may become better. 11 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 A further aspect was the use of technical wetsuits in pool-swimming. In the 2009 FINA World Championship held in Rome, a total of 43 world records were set. Men set new world records in 15 of those events, whereas women did the same in 17 events. Author Contributions: Conceptualization, B.K.; methodology, B.K. and P.T.N.; software, B.K. and P.T.N.; validation, B.K. and P.T.N.; formal analysis, B.K. and P.T.N.; investigation, B.K. and P.T.N.; resources, B.K. and P.T.N.; data curation, B.K. and P.T.N.; writing—original draft preparation, B.K., A.A.D., T.M.B., T.R. and P.T.N.; writing—review and editing, B.K., A.A.D., T.M.B., T.R., C.V.S. and P.T.N.; visualization, B.K. and P.T.N.; supervision, B.K.; project administration, B.K. All authors have read and agreed to the published version of the manuscript. 4. Potential Explanations for The Female Dominance in Long-Distance Open Water Swimming and Age Group Swimmers Each of the men’s world records and 14 of the 17 women’s records still stood. In the past, these world records had not been broken in such a short period of time. There was much speculation that full-body, polyurethane, technical swimsuits were the reason for the improvement in world records. Further analysis led the FINA to institute new rules on 1st January 2010, that limited the types of technical swimsuits that could be worn by athletes. No long-course world record has been broken since then [129]. A problem in this ﬁeld was, however, the fact that most of the considered studies used diﬀerent numbers of subjects, diﬀerent populations, diﬀerent analyses leading to diﬀerent signiﬁcance levels. This problem cannot be eliminated, but it must be considered that depending on the sample and the analysis used, a signiﬁcant diﬀerence between female and male performance can result or not. 5. Conclusions The collective data presented in this review indicate that existing sex diﬀerences in swimming performance showed a generally diminishing trend that was more profound during the longer pool-races, for all swimming strokes. Age-related variations were also reported in both pool and open-water swimming, as sex diﬀerence mainly remained for the younger age groups. However, female athletes in very young age groups (10 years and younger) and very old age groups (75–80 years and older) outscored their male counterparts. When analyzing triathlon swimming performance, the sex gap remained stable during shorter and longer swim distances, except for events under more extreme water temperatures where women can outperform men. Finally, the sex diﬀerence in pool-swimming performance over time showed variations depending on the swimming stroke, distance and the competitive level. Regarding open-water swimming over time, women seemed to continuously narrow the gap to men, especially in speciﬁc long-distances, where the assumption of outperforming men existed. In summary, women tended to decrease the existing sex diﬀerences in speciﬁc age groups (i.e., younger than 10 years and older than 75–80 years) and swimming strokes or even overperform men in long-distance open-water swimming (distance of around 30 km), especially under extreme weather conditions (water colder than ~20 ◦C). Future studies may investigate body compositions of the diﬀerent age group swimmers in order to explain the sex diﬀerences for speciﬁc sports disciplines such as pool-swimming and long-distance open-water swimming. Author Contributions: Conceptualization, B.K.; methodology, B.K. and P.T.N.; software, B.K. and P.T.N.; validation, B.K. and P.T.N.; formal analysis, B.K. and P.T.N.; investigation, B.K. and P.T.N.; resources, B.K. and P.T.N.; data curation, B.K. and P.T.N.; writing—original draft preparation, B.K., A.A.D., T.M.B., T.R. and P.T.N.; writing—review and editing, B.K., A.A.D., T.M.B., T.R., C.V.S. and P.T.N.; visualization, B.K. and P.T.N.; supervision, B.K.; project administration, B.K. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Funding: This research received no external funding. Funding: This research received no external funding. Conﬂicts of Interest: The authors declare no conﬂict of interest. 1. Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Performance trends in master butterﬂy swimmers competing in the FINA world championships. J. Hum. Kinet. 2017, 57, 199–211. [CrossRef] 2. Unterweger, C.M.; Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Increased participation and improved performance in age group backstroke master swimmers from 25–29 to 100–104 years at the FINA World Masters Championships from 1986 to 2014. SpringerPlus 2016, 5, 645. [CrossRef] 3. Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Performance trends in age group breaststrokes wimmers in the FINA world championships 19862–014. Chin. J. Physiol. 2016, 59, 247–259. [CrossRef] 4. Knechtle, B.; Nikolaidis, P.T.; König, S.; Rosemann, T.; Rüst, C.A. Performance trends in master freestyle swimmers aged 25–89 years at the FINA World Championships from 1986 to 2014. Age 2016, 38, 1–8. [CrossRef] References 1. Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Performance trends in master butterﬂy swimmers competing in the FINA world championships. J. Hum. Kinet. 2017, 57, 199–211. [CrossRef] 2. Unterweger, C.M.; Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Increased participation and improved performance in age group backstroke master swimmers from 25–29 to 100–104 years at the FINA World Masters Championships from 1986 to 2014. SpringerPlus 2016, 5, 645. [CrossRef] p p 3. Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Performance trends in age group breaststrokes wimmers in the FINA world championships 19862–014. Chin. J. Physiol. 2016, 59, 247–259. [CrossRef] 4. Knechtle, B.; Nikolaidis, P.T.; König, S.; Rosemann, T.; Rüst, C.A. Performance trends in master freestyle swimmers aged 25–89 years at the FINA World Championships from 1986 to 2014. Age 2016, 38, 1–8. [CrossRef] 12 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 5. König, S.; Valeri, F.; Wild, S.; Rosemann, T.; Rüst, C.A.; Knechtle, B. Change of the age and performance of swimmers across World Championships and Olympic Games ﬁnals from 1992 to 2013—A cross-sectional data analysis. SpringerPlus 2014, 3, 652. [CrossRef] 6. Veiga, S.; Roig, A. Eﬀect of the starting and turning performances on the subsequent swimming par of elite swimmers. Sports Biomech. 2017, 16, 34–44. [CrossRef] 7. Fischer, G.; Knechtle, B.; Rüst, C.A.; Rosemann, T. Male swimmers cross the English Channel faster than female swimmers. Scand. J. Med. Sci. Sports 2013, 23, e48–e55. [CrossRef] 8. Eichenberger, E.; Knechtle, B.; Knechtle, P.; RüSt, C.A.; Rosemann, T.; Lepers, R. Best performances by men and women open-water swimmers during the ‘English Channel Swim’ from 1900 to 2010. J. Sports Sci. 2012, 30, 1295–1301. [CrossRef] 9. Eichenberger, E.; Knechtle, B.; Knechtle, P.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Senn, O. Sex diﬀerence in open-water ultra-swim performance in the longest freshwater lake swim in europe. J. Strength Cond. Res. 2013, 27, 1362–1369. [CrossRef] 10. Zingg, M.A.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Analysis of sex diﬀerences in open-water ultra-distance swimming performances in the FINA World Cup races in 5 km, 10 km and 25 km from 2000 to 2012. BMC Sports Sci. Med. Rehabil. 2014, 6, 7. [CrossRef] 1. Nikolaidis, P.T.; Knechtle, B. Performance trends in individual medley events during FINA World Ma Championships from 1986 to 2014. J. Sports Med. Phys. Fit. 2018, 58, 690–698. 12. Nikolaidis, P.T.; Knechtle, B. References Wild, S.; Rüst, C.A.; Rosemann, T.; Knechtle, B. Changes in sex diﬀerence in swimming speed in ﬁnalists at FINA World Championships and the Olympic Games from 1992 to 2013. BMC Sports Sci. Med. Rehabil. 2014, 6, 25. [CrossRef] [PubMed] 29. Rüst, C.A.; Rosemann, T.; Knechtle, B. Sex diﬀerence in age and performance in elite Swiss freestyle swimmers competing from 50 m to 1500 m. SpringerPlus 2014, 3, 1–10. 30. Wolfrum, M.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Changes in breaststroke swimming performances in national and international athletes competing between 1994 and 2011—A comparison with freestyle swimming performances. BMC Sports Sci. Med. Rehabil. 2014, 6, 18. [CrossRef] [PubMed] 31. Zingg, M.A.; Wolfrum, M.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Freestyle versus butterﬂy swimming performance—Eﬀects of age and sex. Hum. Mov. 2014, 15, 25–35. [CrossRef] 32. Rust, C.A.; Knechtle, B.; Rosemann, T. Women achieve peak freestyle swim speed at earlier ages than men. Open Access J. Sports Med. 2012, 3, 189–199. [CrossRef] [PubMed] Available online: www.ﬁna.org/sites/default/ﬁles/2017_2021_facilities_06102017_full_medium.pdf (accessed on 21 May 2020). 33. Available online: www.ﬁna.org/sites/default/ﬁles/2017_2021_facilities_06102017_full_medium.pdf (accessed on 21 May 2020). 34 Available online: www ﬁna org/sites/default/ﬁles/2017 2021 ows 12092017 ok pdf (accessed on 21 Available online: www.ﬁna.org/sites/default/ﬁles/2017_2021_ows_12092017_ok.pdf (accessed on 21 May 2020). 35. Available online: www.channelswimmingassociation.com/swim-advice/channel-water-temperatures (accessed on 21 May 2020). 36. Knechtle, B.; Rosemann, T.; Rüst, C.A. Women cross the ‘Catalina Channel’ faster than men. SpringerPlus 2015, 4, 332. [CrossRef] 37. Nikolaidis, P.T.; Gangi, S.D.; De Sousa, C.V.; Valeri, F.; Rosemann, T.; Knechtle, B. Sex diﬀerence in open-water swimming—The triple crown of open water swimming 1875–2017. PLoS ONE 2018, 13, e0202003. [CrossRef] 38. Knechtle, B.; Rosemann, T.; Lepers, R.; Rüst, C.A. Women outperform men in ultradistance swimming: The Manhattan Island Marathon Swim from 1983 to 2013. Int. J. Sports Physiol. Perform. 2014, 9, 913–924. [CrossRef] 39. Available online: www.ice-swimming.com (accessed on 21 May 2020). 40. Available online: www.internationaliceswimming.com (accessed on 21 May 2020). 1. Knechtle, B.; Rosemann, T.; Rust, C.A. Ice swimming—‘Ice Mile’ and ‘1 km Ice event’. BMC Sports Sci. M Rehabil. 2015, 7, 20. [CrossRef] 42. Senefeld, J.; Joyner, M.J.; Stevens, A.; Hunter, S.K. Sex diﬀerences in elite swimming with advanced age are less than marathon running. Scand. J. Med. Sci. Sports 2016, 26, 17–28. [CrossRef] 43. Senefeld, J.W.; Clayburn, A.J.; Baker, S.E.; Carter, R.E.; Johnson, P.W.; Joyner, M.J. Sex diﬀerences in youth elite swimming. PLoS ONE 2019, 14, e0225724. [CrossRef] [PubMed] 44. References Pacing in age-group freestyle swimmers at The XV FINA World Masters Championships in Montreal 2014. J. Sports Sci. 2017, 35, 1165–1172. [CrossRef] [PubMed] 13. Eichenberger, E.; Knechtle, B.; Christoph, A.R.; Knechtle, P.; Lepers, R.; Rosemann, T. No gender diﬀerence in peak performance in ultra-endurance swimming performance—Analysis of the ‘Zurich 12-h swim’ from 1996 to 2010. Chin. J. Physiol. 2012, 55, 1–6. 14. Knechtle, B.; Nikolaidis, P.T.; Rosemann, T.; Rüst, C.A. Performance trends in 3000 m open-water age group swimmers from 25 to 89 years competing in the FINA World Championships from 1992 to 2014. Res. Sports Med. 2017, 25, 67–77. [CrossRef] 15. Veiga, S.; Rodriguez, L.; Gonzalez-Frutos, P.; Navandar, A. Race strategies of open water swimmers in the 5-km, 10-km, and 25-km races of the 2017 FINA World Swimming Championships. Front. Psychol. 2019, 10, 654. [CrossRef] 16. Lepers, R.; Knechtle, B.; Stapley, P.J. Trends in triathlon performance: Eﬀects of sex and age. Sports Med. (Auckland, N.Z.) 2013, 43, 851–863. [CrossRef] 17. Etter, F.; Knechtle, B.; Bukowski, A.; Rust, C.A.; Rosemann, T.; Lepers, R. Age and gender interactions in short distance triathlon performance. J. Sports Sci. 2013, 31, 996–1006. [CrossRef] 18. Wu, S.S.; Peiﬀer, J.J.; Brisswalter, J.; Nosaka, K.; Lau, W.Y.; Abbiss, C.R. Pacing strategies during the swim, cycle and run disciplines of sprint, Olympic and half-Ironman triathlons. Eur. J. Appl. Physiol. 2015, 115, 1147–1154. [CrossRef] 19. Rüst, C.A.; Knechtle, B.; Rosemann, T.; Lepers, R. Sex diﬀerence in race performance and age of peak performance in the Ironman Triathlon World Championship from 1983 to 2012. Extrem. Physiol. Med. 2012, 1, 15. [CrossRef] 20. Sigg, K.; Knechtle, B.; Rust, C.A.; Knechtle, P.; Lepers, R.; Rosemann, T. Sex diﬀerence in Double Iron ultra-triathlon performance. Extrem. Physiol. Med. 2013, 2, 12. [CrossRef] 21. Available online: https://explorersweb.com/2019/07/05/why-women-excel-at-marathon-swimming/ (accessed on 30 April 2020). 22. Available online: https://www.thecut.com/2016/09/the-obscure-endurance-sport-women-are-quietly- dominating.html (accessed on 21 May 2020). 23. Available online: https://womenintheworld.com/2016/09/18/the-grueling-sport-in-which-women- competitors-typically-outperform-the-men/ (accessed on 30 April 2020). 24. Available online: https://www.theguardian.com/lifeandstyle/2020/jan/03/female-ultra-athletes-leading-ﬁeld- women-less-ego (accessed on 30 April 2020). 25. Available online: https://www.bbc.com/news/world-49284389 (accessed on 30 April 2020). 26. Available online: https://www.healthline.com/health-news/will-women-athletes-ever-be-able-to-compete- with-men#2 (accessed on 30 April 2020). 13 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 13 of 17 27. Seifert, L.; Barbosa, T.M.; Kjendlie, P.-L. Biophysical approach to swimming: Gender eﬀect. In Gender Gap: Causes, Experiences and Eﬀects; Davies, S.A., Ed.; Nova Publishers: Hauppauge, NY, USA, 2011; pp. 59–80. 28. References Participation and performance trends in ‘Ultraman Hawaii’ from 1983 to 2012. Extrem. Physiol. Med. 2013, 2, 25. [CrossRef] 56. Knechtle, B.; Nikolaidis, P.T.; Stiefel, M.; Rosemann, T.; Rust, C.A. Performance and sex diﬀerences in ‘Isklar Norseman Xtreme Triathlon’. Chin. J. Physiol. 2016, 59, 276–283. [CrossRef] [PubMed] 57. Rust, C.A.; Rosemann, T.; Knechtle, B. Performance and sex diﬀerence in ultra-triathlon performance from Ironman to Double Deca Iron ultra-triathlon between 1978 and 2013. SpringerPlus 2014, 3, 219. [CrossRef] [PubMed] 58. Rüst, C.A.; Knechtle, B.; Rosemann, T.; Lepers, R. The changes in age of peak swim speed for elite male and female Swiss freestyle swimmers between 1994 and 2012. J. Sports Sci. 2014, 32, 248–258. [CrossRef] [PubMed] 59. Salihu, L.; Rüst, C.A.; Rosemann, T.; Knechtle, B. Sex diﬀerence in draft-legal ultra-distance events—A comparison between ultra-swimming and ultra-cycling. Chin. J. Physiol. 2016, 59, 87–99. [CrossRef] 60. Rüst, C.A.; Knechtle, B.; Rosemann, T.; Lepers, R. Women reduced the sex diﬀerence in open-water ultra-distance swimming—La Traversée Internationale du Lac St-Jean, 1955–2012. Appl. Physiol. Nutr. Metab. 2014, 39, 270–273. [CrossRef] 61. Toussaint, H.M.; Bruinink, L.; Coster, R.; Looze, M.D.; Rossem, B.V.; Veenen, R.V.; Groot, G.D. Eﬀect of a triathlon wet suit on drag during swimming. Med. Sci. Sports Exerc. 1989, 21, 325–328. [CrossRef] 62. Nikolaidis, P.T.; de Sousa, C.V.; Knechtle, B. Sex diﬀerence in long-distance open-water swimming races–does nationality play a role? Res. Sports Med. 2018, 26, 332–344. [CrossRef] 63. Ulsamer, S.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Swimming performances in long distance open-water events with and without wetsuit. BMC Sports Sci. Med. Rehabil. 2014, 6, 20. [CrossRef] [PubMed] 63. Ulsamer, S.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Swimming performances in long distance open-water events with and without wetsuit. BMC Sports Sci. Med. Rehabil. 2014, 6, 20. [CrossRef] [PubMed] 64. Issurin, V.; Pushkar-Verbitsky, V.; Verbitsky, O. Eﬀect of high-tech swimsuits on the swimming performance open-water events with and without wetsuit. BMC Sports Sci. Med. Rehabil. 2014, 6, 20. [CrossRef] [PubMed] 64. Issurin, V.; Pushkar-Verbitsky, V.; Verbitsky, O. Eﬀect of high-tech swimsuits on the swimming performance in top-level swimmers. J. Sports Med. Phys. Fit. 2014, 54, 383–388. 64. Issurin, V.; Pushkar-Verbitsky, V.; Verbitsky, O. Eﬀect of high-tech swimsuits on the swimming performance in top-level swimmers. J. Sports Med. Phys. Fit. 2014, 54, 383–388. 65. Gibas-Dorna, M.; Checinska, Z.; Korek, E.; Kupsz, J.; Sowinska, A.; Krauss, H. References Buhl, C.; Knechtle, B.; Rust, C.A.; Rosemann, T.; Lepers, R. A comparison of medley and freestyle performance for national and international swimmers between 1994 and 2011. Open Access J. Sports Med. 2013, 4, 79–87. [CrossRef] [PubMed] 45. Vaso, M.; Knechtle, B.; Rüst, C.A.; Rosemann, T.; Lepers, R. Age of peak swim speed and sex diﬀerence in performance in medley and freestyle swimming—A comparison between 200 m and 400 m in Swiss elite swimmers. J. Hum. Sport Exerc. 2013, 8, 954–965. [CrossRef] p 46. Nicol, E.; Ball, K.; Tor, E. The biomechanics of freestyle and butterﬂy turn technique in elite swimmers. Sports Biomech. 2019. [CrossRef] [PubMed] 47. Wolfrum, M.; Knechtle, B.; Rüst, C.A.; Rosemann, T.; Lepers, R. Sex-related diﬀerences and age of peak performance in breaststroke versus freestyle swimming. BMC Sports Sci. Med. Rehabil. 2013, 5, 29. [CrossRef] 48. Jackson, C.W. Sex gap in swimming performance narrows. Am. Correct. Ther. J. 1977, 31, 172–175. 49. Kollarz, C.; Knechtle, B.; Rüst, C.A.; Rosemann, T.; Lepers, R. Comparison of age of peak swimming speed in elite backstroke swimmers at national and international level. OA Sports Med. 2013, 1, 19. [CrossRef] 50. Vogt, P.; Rüst, C.A.; Rosemann, T.; Lepers, R.; Knechtle, B. Analysis of 10 km swimming performance of elite male and female open-water swimmers. SpringerPlus 2013, 2, 1–15. [CrossRef] 51. Rüst, C.A.; Lepers, R.; Rosemann, T.; Knechtle, B. Will women soon outperform men in open-water ultra-distance swimming in the ‘Maratona del Golfo Capri-Napoli’? SpringerPlus 2014, 3, 1–12. [CrossRef] 51. Rüst, C.A.; Lepers, R.; Rosemann, T.; Knechtle, B. Will women soon outperform men in open-water ultra-distance swimming in the ‘Maratona del Golfo Capri-Napoli’? SpringerPlus 2014, 3, 1–12. [CrossRef] 52. Lepers, R. Sex diﬀerence in triathlon performance. Front. Physiol. 2019, 10, 973. [CrossRef] 52. Lepers, R. Sex diﬀerence in triathlon performance. Front. Physiol. 2019, 10, 973. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 3651 14 of 17 14 of 17 53. Rust, C.A.; Lepers, R.; Stiefel, M.; Rosemann, T.; Knechtle, B. Performance in Olympic triathlon: Changes in performance of elite female and male triathletes in the ITU World Triathlon Series from 2009 to 2012. SpringerPlus 2013, 2, 685. [CrossRef] [PubMed] 54. Rust, C.A.; Knechtle, B.; Knechtle, P.; Rosemann, T.; Lepers, R. Sex diﬀerences in ultra-triathlon perfo at increasing race distance. Percept. Mot. Ski. 2013, 116, 690–706. [CrossRef] [PubMed] 55. Meili, D.; Knechtle, B.; Rust, C.A.; Rosemann, T.; Lepers, R. References Cold water swimming benefcially modulates insulin sensitivity in middle-aged individuals. J. Aging Phys. Act. 2016, 24, 547–554. [CrossRef] [PubMed] 66. Seifert, L.; Chollet, D. A new index of ﬂat breaststroke propulsion: A comparison of elite men and women. J. Sports Sci. 2005, 23, 309–320. [CrossRef] [PubMed] 67. Zuniga, J.; Housh, T.J.; Michelle, M.; Hendrix, C.R.; Camic, C.L.; Johnson, G.O.; Housh, D.J.; Schmidt, R.J. Gender comparisons of anthropometric characteristics of young sprint swimmers. J. Strength Cond. Res. 2011, 25, 103–108. [CrossRef] 68. McLean, S.P.; Hinrichs, R.N. Buoyancy, gender, and swimming performance. J. Appl. Biomech. 2000, 16, 248–263. [CrossRef] 69. McLean, S.P.; Hinrichs, R.N. Sex diﬀerences in the centre of buoyancy location of competitive swimmers. J. Sports Sci. 1998, 16, 373–383. [CrossRef] 70. Novak, L.P.; Woodward, W.A.; Bestit, C.; Mellerowicz, H. Working capacity, body composition, anthropometry of Olympic female athletes. J. Sports Med. Phys. Fit. 1977, 17, 275–283. 71. Pyne, D.B.; Anderson, M.E.; Hopkins, W.G. Monitoring changes in lean mass of elite male and female swimmers. Int. J. Sports Physiol. Perform. 2006, 1, 14–26. [CrossRef] 72. Bishop, P.; Cureton, K.; Conerly, M.; Collins, M. Sex diﬀerence in muscle cross-sectional area of athletes and non-athletes. J. Sports Sci. 1989, 7, 31–39. [CrossRef] 73. Zampagni, M.L.; Casino, D.; Benelli, P.; Visani, A.; Marcacci, M.; De Vito, G. Anthropometric and strength variables to predict freestyle performance times in elite master swimmers. J. Strength Cond. Res. 2008, 22, 1298–1307. [CrossRef] [PubMed] 74. Demura, S.; Matsuzawa, J.; Naka, H.; Kita, I. Physical characteristics in well-trained young swimmers. Jpn. J. Phys. Fit. Sports Med. 1991, 40, 278–287. [CrossRef] 15 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 75. Kanehisa, H.; Muraoka, Y.; Kawakami, Y.; Fukunaga, T. Fascicle arrangements of vastus lateralis and gastrocnemius muscles in highly trained soccer players and swimmers of both genders. Int. J. Sports Med. 2003, 24, 90–95. [CrossRef] 76. Lerda, R.; Cardelli, C.; Coudereau, J.P. Backstroke organization in physical education students as a func of skill and sex. Percept. Mot. Ski. 2005, 100, 779–790. [CrossRef] [PubMed] 77. Seifert, L.; Boulesteix, L.; Chollet, D.; Vilas-Boas, J.P. Diﬀerences in spatial-temporal parameters and arm-leg coordination in butterﬂy stroke as a function of race pace, skill and gender. Hum. Mov. Sci. 2008, 27, 96–111. [CrossRef] [PubMed] 78. Takagi, H.; Sugimoto, S.; Nishijima, N.; Wilson, B. Swimming: Diﬀerences in stroke phases, arm-leg coordination and velocity ﬂuctuation due to event, gender and performance level in breaststroke. References Sports Biomech. 2004, 3, 15–27. [CrossRef] [PubMed] 79. Chatard, J.C.; Agel, A.; Lacoste, L.; Millet, C.; Paulin, M.; Lacour, J.R. Energy cost of crawl in female competition swimmers. Comparison with males. Sci. Sports 1991, 6, 245–251. [CrossRef] 80. Toussaint, H.M.; Knops, W.; De Groot, G.; Hollander, A. The mechanical eﬃciency of front crawl swimming. Med. Sci. Sports Exerc. 1990, 22, 402–408. [CrossRef] 81. Klentrou, P.P.; Montpetit, R.R. Energetics of backstroke swimming in males and females. Med. Sci. Sports Exerc. 1992, 24, 371–375. [CrossRef] 82. Leblanc, H.; Seifert, L.; Chollet, D. Does ﬂoatation inﬂuence breaststroke technique? J. Appl. Biomech. 2010, 26, 150–158. [CrossRef] 83. Lee, J.; Mellifont, R.; Winstanley, J.; Burkett, B. Body roll in simulated freestyle swimming. Int. J. Sports Med. 2008, 29, 569–573. [CrossRef] [PubMed] 84. Magnusson, S.P.; Constantini, N.W.; McHugh, M.P.; Gleim, G.W. Strength proﬁles and performance in masters’ level swimmers. Am. J. Sports Med. 1995, 23, 626–631. [CrossRef] [PubMed] 85. Demura, S. The sex diﬀerence and grade diﬀerence in anthropometric characteristics, muscular strength, and ﬂexibility junior high school swimmers. Jpn. J. Phys. Fit. Sports Med. 1983, 32, 8–16. [CrossRef] 86. Mameletzi, D.; Siatras, T. Sex diﬀerences in isokinetic strength and power of knee muscles in 10–12 year old swimmers. Isokinet. Exerc. Sci. 2003, 11, 231–237. [CrossRef] 7. Deschodt, V.J.; Rouard, A.H. Inﬂuence of gender on kinematic parameters in top level frontcrawl swimm Sci. Sports 1999, 14, 39–44. [CrossRef] 88. Zamparo, P.; Lazzer, S.; Antoniazzi, C.; Cedolin, S.; Avon, R.; Lesa, C. The interplay between propelling eﬃciency, hydrodynamic position and energy cost of front crawl in 8 to 19-year-old swimmers. Eur. J. Appl. Physiol. 2008, 104, 689–699. [CrossRef] 89. Salselas, V.; Márquez, S. Perceptions of the motivational climate created by parents of young Portuguese swimmers. Percept. Mot. Ski. 2009, 108, 851–861. [CrossRef] 90. Papadopoulou, A.; Tsigilis, N.; Theodorakis, N.D.; Tsalis, G. Sex and sport-related diﬀerences in satisfaction among Greek swimmers. Psychol. Rep. 2006, 98, 389–394. [CrossRef] 91. Romberg, K.; Tufvesson, E.; Bjermer, L. Sex diﬀerences in asthma in swimmers and tennis players. Ann. Allergy Asthma Immunol. 2017, 118, 311–317. [CrossRef] 92. Meggs, J.; Chen, M.A. Mental toughness and attributions of failure in high performing male and female swimmers. J. Hum. Sport Exerc. 2018, 13, 276–284. [CrossRef] 93. Shields, D.L. Moral growth among athletes and nonathletes: A comparative analysis. J. Genet. Psych 147, 7–18. 94. Berger, B.G.; Owen, D.R. Mood alteration with swimming—Swimmers really do ‘feel better’. Psychosom. Med. References Estimation of body composition in competitive swimmers by means of three diﬀerent techniques. Int. J. Sports Med. 1997, 18, 363–368. [CrossRef] 109. Caspersen, C.; Berthelsen, P.A.; Eik, M.; Pâkozdi, C.; Kjendlie, P.L. Added mass in human swimmers: Age and gender diﬀerences. J. Biomech. 2010, 43, 2369–2373. [CrossRef] 110. Knechtle, B.; Baumann, B.; Knechtle, P.; Rosemann, T. Speed during training and anthropometric measures in relation to race performance by male and female open-water ultra-endurance swimmers. Percept. Mot. Ski. 2010, 111, 463–474. [CrossRef] 111. Martínez-Sanz, J.M.; Mielgo-Ayuso, J.; Urdampilleta, A. Body composition and somatotype in adolescent competion swimmers. Revista Espanola de Nutricion Humana y Dietetica 2012, 16, 130–136. [CrossRef] 12. Martínez, S.; Pasquarelli, B.N.; Romaguera, D.; Arasa, C.; Tauler, P.; Aguiló, A. Anthropometric characteri and nutritional proﬁle of young amateur swimmers. J. Strength Cond. Res. 2011, 25, 1126–1133. [CrossR and nutritional proﬁle of young amateur swimmers. J. Strength Cond. Res. 2011, 25, 1126 1133. [CrossRef] 113. Sandbakk, O.; Solli, G.S.; Holmberg, H.C. Sex diﬀerences in world-record performance: The inﬂuence of sport discipline and competition duration Int J Sports Physiol Perform 2018 13 2–8 [CrossRef] 13. Sandbakk, O.; Solli, G.S.; Holmberg, H.C. Sex diﬀerences in world-record performance: The inﬂuenc sport discipline and competition duration. Int. J. Sports Physiol. Perform. 2018, 13, 2–8. [CrossRef] 114. Millard-Staﬀord, M.; Swanson, A.E.; Wittbrodt, M.T. Nature versus nurture: Have performance gaps between men and women reached an asymptote? Int. J. Sports Physiol. Perform. 2018, 13, 530–535. [CrossRef] [PubMed] 115. Rong, C.; Bei, H.; Yun, M.; Yuzhu, W.; mingwu, Z. Lung function and cytokine levels in professional athletes. J. Asthma 2008, 45, 343–348. [CrossRef] 116. Ricoy, J.R.; Encinas, A.R.; Cabello, A.; Madero, S.; Arenas, J. Histochemical study of the vastus lateralis muscle ﬁbre types of athletes. J. Physiol. Biochem. 1998, 54, 41–48. [PubMed] 117. Nikolaidis, P.T. Age- and sex-related diﬀerences in force-velocity characteristics of upper and lower limbs of competitive adolescent swimmers. J. Hum. Kinet. 2012, 32, 87–95. [CrossRef] [PubMed] 118. Garrido, N.D.; Silva, A.J.; Fernandes, R.J.; Barbosa, T.M.; Costa, A.M.; Marinho, D.A.; Marques, M.C. High level swimming performance and its relation to non-speciﬁc parameters: A cross-sectional study on maximum handgrip isometric strength. Percept. Mot. Ski. 2012, 114, 936–948. [CrossRef] [PubMed] 119. Gackenbach, J. Collegiate swimmers: Sex diﬀerences in self-reports and indices of physiological stress. Percept. Mot. Ski. 1982, 55, 555–558. [CrossRef] 120. Ogilvie, B.C.; McGuire, L.G. How competition aﬀects elite women swimmers. Physician Sportsmed. 1980, 8, 113–116. References 1983, 45, 425–433. [CrossRef] [PubMed] 95. Hammond, T.; Gialloreto, C.; Kubas, H.; Hap Davis, H. The prevalence of failure-based depression among elite athletes. Clin. J. Sport Med. 2013, 23, 273–277. [CrossRef] [PubMed] 96. Anshel, M.H.; Porter, A. Eﬃcacy of a model for examining self-regulation with elite and non-elite male and female competitive swimmers. Int. J. Sport Psychol. 1996, 27, 321–336. 97. Flandrois, R.; Grandmontagne, M.; Mayet, M.H. Maximal oxygen uptake in French children in relation to age, sex and physical training. J. Physiol. 1982, 78, 186–194. 98. Reis, J.F.; Millet, G.P.; Bruno, P.M.; Vleck, V.; Alves, F.B. Sex and exercise intensity do not inﬂuence oxygen uptake kinetics in submaximal swimming. Front. Physiol. 2017, 8, 2099. [CrossRef] 16 of 17 16 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 99. O’Toole, M.L.; Douglas, B.; Hiller, W.; Crosby, L.O.; Douglas, P.S. The ultraendurance triathlete: A physiological proﬁle. Med. Sci. Sports Exerc. 1987, 19, 45–50. 100. Greksa, L.P.; Haas, J.D.; Leatherman, T.L.; Spielvogel, H.; Zamora, M.P.; Fernandez, L.P.; Moreno-Black, G. Maximal aerobic power in trained youths at high altitude. Ann. Hum. Biol. 1982, 9, 201–209. [CrossRef] 101. Fu, F.H.; You, C.Y.C.-Y.; Kong, Z.W. Acute changes in selected serum enzyme and metabolite concentrations in 12-to 14-yr.-old athletes after an all-out 100-m swimming sprint. Percept. Mot. Ski. 2002, 95, 1171–1178. , ; , ; g, g y in 12-to 14-yr.-old athletes after an all-out 100-m swimming sprint. Percept. Mot. Ski. 2002, 95, 1171–1178. 102. Avlonitou, E. Maximal lactate values following competitive performance varying according to age, sex and swimming style. J. Sports Med. Phys. Fit. 1996, 36, 24–30. 103. Nygaard, E. Skeletal muscle ﬁbre characteristics in young women. Acta Physiol. Scand. 1981, 112, 299–304. [CrossRef] 104. Sébert, P.; Barthélemy, L.; Simon, B.; Coat, J.P.; Klimek, D. Recovery from exercise in young swimmers: Sexual diﬀerences. Sci. Sports 1995, 10, 141–147. [CrossRef] 105. Sallis, R.E.; Jones, K.; Sunshine, S.; Smith, G.; Simon, L. Comparing sports injuries in men and women. Int. J. Sports Med. 2001, 22, 420–423. [CrossRef] 06. Maﬀulli, N.; King, J.B.; Helms, P. Training in elite young athletes (the training of young athletes (TO study): Injuries, ﬂexibility and isometric strength. Br. J. Sports Med. 1994, 28, 123–136. [CrossRef] 07. Beckford, S.E.; Webb, M.C. Anthropometric parameters: Obesity and metabolic risks for non-communic diseases among adolescent swimmers. West Indian Med. J. 2018, 67, 31–38. 108. Avlonitou, E.; Georgiou, E.; Douskas, C.; Louizi, A. References [CrossRef] 121. Lopez-Fernandez, I.; Merino-Marban, R.; Fernandez-Rodriguez, E. Examining the relationship between sex and motivation in triathletes. Percept. Mot. Ski. 2014, 119, 42–49. [CrossRef] 122. U.S. Department of Education. Title IX and sex discrimination. In Title IX; Rights, O.f.C., Ed.; 2015. Available online: https://www2.ed.gov/about/oﬃces/list/ocr/docs/tix_dis.html (accessed on 21 May 2020). 123. Margolis, J. The Struggle for Female Soccer Equality in Brazil. PRI 2013. Available online: https://www.pri. org/stories/2013-05-27/struggle-female-soccer-equality-brazil (accessed on 21 May 2020). 17 of 17 17 of 17 Int. J. Environ. Res. Public Health 2020, 17, 3651 124. Capranica, L.; Piacentini, M.F.; Halson, S.; Myburgh, K.H.; Ogasawara, E.; Millard-Staﬀord, M. The gender gap in sport performance: Equity inﬂuences equality. Int. J. Sports Physiol. Perform. 2013, 8, 99–103. [CrossRef] 25. Medic, N.; Young, B.W.; Starkes, J.L.; Weir, P.L.; Grove, J.R. Gender, age, and sport diﬀerences in relative eﬀects among us masters swimming and track and ﬁeld athletes. J. Sports Sci. 2009, 27, 1535–1544. [Cross 125. Medic, N.; Young, B.W.; Starkes, J.L.; Weir, P.L.; Grove, J.R. Gender, age, and sport diﬀerences in relative age eﬀects among us masters swimming and track and ﬁeld athletes. J. Sports Sci. 2009, 27, 1535–1544. [CrossRef] 126. Elmenshawy, A.R.; Machin, D.R.; Tanaka, H. A rise in peak performance age in female athletes. Age (Dordrecht, The Netherlands) 2015, 37, 9795. [CrossRef] 126. Elmenshawy, A.R.; Machin, D.R.; Tanaka, H. A rise in peak performance age in female athletes. Age (Dordrecht, The Netherlands) 2015, 37, 9795. [CrossRef] 127. Senefeld, J.; Smith, C.; Hunter, S.K. Sex diﬀerences in participation, performance, and age of ultramarathon runners. Int. J. Sports Physiol. Perform. 2016, 11, 635–642. [CrossRef] 128. Waldvogel, K.J.; Nikolaidis, P.T.; Di Gangi, S.; Rosemann, T.; Knechtle, B. Women Reduce the performance diﬀerence to men with increasing age in ultra-marathon running. Int. J. Environ. Res. Public Health 2019, 16, 2377. [CrossRef] 129. O’Connor, L.M.; Vozenilek, J.A. Is it the athlete or the equipment? An analysis of the top swim performances from 1990 to 2010. J. Strength Cond. Res. 2011, 25, 3239–3241. [CrossRef] 130. Stark, D.; Di Gangi, S.; Sousa, C.V.; Nikolaidis, P.; Knechtle, B. Tower running-participation, performance trends, and sex diﬀerence. Int. J. Environ. Res. Public Health 2020, 17, 1902. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4289827655	https://ejournal.stiesia.ac.id/jimbis/article/download/5133/672	Indonesian	null	PENGARUH GAYA KEPEMIMPINAN TERHADAP KINERJA: KEPUASAN KERJA SEBAGAI VARIABEL INTERVENING	Jurnal Ilmiah Manajemen dan Bisnis	2,022	cc-by	9,486	JIMBis : Jurnal Ilmiah Manajemen dan Bisnis JIMBis : Jurnal Ilmiah Manajemen dan Bisnis JIMBis : Jurnal Ilmiah Manajemen dan Bisnis ABSTRACT An effective leadership style in managing human resources and supported by employees' understanding of work will affect work behavior as indicated by an increase in individual job satisfaction and the performance of the unit itself, then affect the company's performance. The purpose of this study was to determine the effect of leadership style on job satisfaction, leadership style on performance, job satisfaction on employee performance and the influence of leadership style on performance through the intervening variable job satisfaction. The research instrument is a questionnaire. Samples were taken as many as 100 respondents, employees of the private hospital in Gresik. Sampling using probability sampling technique. Data analysis using Path analysis by Smart PLS 3. The results showed that all 4 (four) hypotheses were proven and significant, leadership style affects job satisfaction, leadership style affects performance, job satisfaction affects employee performance and leadership style affects performance through the intervening variable job satisfaction. Keywords: leadership styles, job satisfaction, employee performance PENGARUH GAYA KEPEMIMPINAN TERHADAP KINERJA: KEPUASAN KERJA SEBAGAI VARIABEL INTERVENING Juwita Sari juwitasari@stiesia.ac.id Siti Rokhmi Fuadati Sekolah Tinggi Ilmu Ekonomi Indonesia (STIESIA) Surabaya ABSTRAK Gaya kepemimpinan yang efektif dalam mengelola sumber daya manusia akan berpengaruh pada perilaku kerja yang diindikasikan dengan peningkatan kepuasan kerja individu dan kinerja unit itu sendiri, yang pada akhirnya akan mempengaruhi kinerja perusahaan secara keseluruhan. Tujuan Penelitian ini untuk mengetahui pengaruh gaya kepemimpinan terhadap kepuasan kerja, Gaya kepemimpinan terhadap kinerja, kepuasan kerja terhadap kinerja karyawan dan pengaruh gaya kepemimpinan terhadap kinerja melalui variabel intervening kepuasan kerja. Instrumen penelitian yaitu kuisioner. Sampel yang diambil sebanyak 100 responden karyawan Rumah Sakit Swasta di Gresik. Pengambilan sampel menggunakan teknik probability sampling. Analisis data menggunakan SEM dan diolah menggunakan PLS. Hasil penelitian menunjukkan bahwa 4 (empat) hipotesis semua diterima dan signifikan, yaitu Gaya kepemimpinan berpengaruh terhadap kepuasan kerja, Gaya kepemimpinan berpengaruh terhadap kinerja, kepuasan kerja terhadap kinerja karyawan dan gaya kepemimpinan berpengaruh terhadap kinerja melalui variabel intervening kepuasan kerja. Kata kunci: gaya kepemimpinan, kepuasan kerja, kinerja karyawan PENDAHULUAN Indonesia memiliki dua jenis rumah sakit, yaitu rumah sakit pemerintah dan rumah sakit non-pemerintah atau yang biasa disebut dengan rumah sakit swasta. Berdasarkan data Kementrian Kesehatan Tahun 2018, rumah sakit pemerintah dan swasta yang ada di indonesia saat ini berjumlah 2.813. Jumlah terdiri atas 2.269 RS Umum dan 544 RS khusus. Dibandingkan Perkembangan dunia kesehatan saat ini dalam era globalisasi terus meningkat, baik secara kualitas maupun kuantitas, yang didukung oleh perkembangan ilmu dan teknologi yang semakin canggih. Ini mem- buat instansi atau pihak yang berhubungan dengan dunia kesehatan, salah satunya rumah sakit, tumbuh dengan pesat. 39 JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 40 data 2010 jumlah RS hanya 1.632 unit. Jumlah RS terbanyak ada di provinsi Jawa Timur (https://databoks.katadata.co.id/datapu blish/2019/10/10/berapa-jumlah-rumah-sakit- di-indonesia). melayani banyak masyarakat di Gresik. Manajemen RS juga membuka Klinik kese- hatan di kompleks Perumahan Gresik Kota Baru. Untuk meningkatkan kinerja RS, manajemen RS juga telah bekerja sama de- ngan kurang lebih 20 perusahaan di wila- yah Gresik dan asuransi terkait dengan pelayanan kesehatan karyawan yaitu Jamsostek dan Askes. Mengacu kepada tingginya kebutuhan masyarakat dan tun- tutan dari rekanan baik dari perusahaan industri maupun asuransi agar dapat memberikan pelayanan prima selama 24 jam, maka RS harus meningkatkan kinerja karyawan medis dan non medis. Pertumbuhan rumah sakit yang skema- kin meningkat secara kuantitas, didukung dengan perkembangan ilmu dan teknologi, maka setiap rumah sakit dituntut untuk memberikan pelayanan kesehatan yang baik dan berkualitas untuk masyarakat. Di sisi lain tingkat pengetahuan ma- syarakat pun sudah semakin berkembang, sehingga tuntutan masyarakat terhadap kualitas jasa pelayanan rumah sakit juga semakin tinggi. Rumah sakit dihadapkan pada tekanan yang kuat untuk mengontrol biaya dan di saat yang sama harus me- ningkatkan kualitas pelayanannya (Ivana Ariyani, 2016). Rumah sakit adalah sebuah institusi jasa perawatan kesehatan profesi- nal yang pelayanannya disediakan oleh tenaga kerja medis dan non medis. Masya- rakat berharap Rumah Sakit memberikan pelayanan prima disemua unit kerja yang berhubungan dengan pasien. Para tenaga medis dan non medis inilah yang menjadi ujung tombak kinerja Rumah Sakit (Fuadiputra, 2013). Baik buruknya citra Rumah Sakit ditentukan oleh pengalaman pasien ketika dirawat. Oleh karenanya, para tenaga kerja medis dan non medis di Rumah Sakit sering disorot kinerjanya oleh pasien karena tenaga medis merupakan karyawan yang berhubungan langsung dengan pasien. PENDAHULUAN y Semakin tingginya kebutuhan dan kepercayaan yang timbul dari masyarakat, manajemen RS perlu mempertimbangkan pentingnya sistem pengelolaan sumber- dayanya (sumber daya manusia, fasilitas, dan modal), proses bisnis, output dan dampak berjalanannya bisnis, dalam rang- ka meningkatkan mutu pelayanan keseha- tan secara umum pada organisasi tempat- nya bekerja, dan dampak akhir bermuara pada kualitas hidup dan kesejahteraan masyarakat. Sistem Pengelolaan secara opti- mal dapat diwujudkan dengan sikap kepe- mimpinan yang efektif (Yulianto et al., 2018), pemenuhan indikator kompetensi yang terstandarisasi, pengembangan keah- lian dengan pelatihan-pelatihan dan asuhan keperawatan, penilaian kinerja yang objek- tif, pembagian jam kerja yang adil, serta sistem kompensasi yang dapat memberikan kepuasan kerja (Fuadiputra, 2013). Tenaga medis khususnya perawat me- miliki waktu paling banyak dalam berin- teraksi dengan pasien dibandingkan dokter yang berkunjung/visite dijam tertentu saja. Sehingga, para tenaga medis dan non medis Rumah sakit dituntut untuk memberikan pelayanan sesuai harapan pasien. Penilaian kinerja yang obyektif akan memberikan umpan balik yang tepat, se- hingga karyawan akan merasa dihargai dan didukung oleh organisasi dan akan mem- berikan hasil kerja yang optimal, serta secara emosional akan lebih terikat kepada organisasi tempatnya bekerja. Penilaian ki- nerja merupakan refleksi dari berkembang atau tidaknya perusahaan dimana hasil penilaian tersebut menunjukkan keefektif- an SDM dalam melaksanakan tugas yang akan mempengaruh kinerja perusahaan. Penelitian ini dilakukan di salah satu Rumah Sakit (RS) swasta di kabupaten Gresik. Seiring dengan perkembangan tek- nologi dan ilmu kesehatan serta tuntutan perubahan organisasi. RS tersebut berkem- bang menjadi Rumah Sakit umum yang Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 41 Peran pemimpin di sebuah organisasi akan menentukan kinerja karyawannya (Bhargavi & Yaseen, 2016; Scheidlinger, 1994). Begitu pentingnya peran kepemimpi- nan dalam sebuah organisasi menjadi fokus yang menarik perhatian para peneliti bi- dang perilaku keorganisasian. Seorang pe- mimpin harus mampu mempengaruhi para bawahannya untuk bertindak sesuai de- ngan visi, misi dan tujuan perusahaan. Pe- mimpin harus mampu memberikan wa- wasan, membangkitkan kebanggaan, serta menumbuhkan sikap hormat dan keper- cayaan dari bawahannya (Raus et al., 2012). Kinerja pada penelitian (Ali & Agustian, 2018b) adalah catatan perolehan yang dihasilkan dari fungsi suatu pekerjaan tertentu atau kegiatan selama suatu periode waktu tertentu. Jadi kinerja berkenaan dengan hasil pekerjaan yang dicapai oleh karyawan dalam suatu periode. Dalam hal ini kinerja berkaitan dengan kuantitas mau- pun kualitas pekerjaan yang dihasilkan. PENDAHULUAN Yulianto et al., (2018) menjelaskan bahwa kinerja karyawan adalah hasil kerja secara kualitas dan kuantitas yang dicapai oleh seorang karyawan dalam melaksana- kan tugasnya sesuai dengan tanggung jawab yang diberikan. Fuadiputra (2013) menyatakan bahwa kinerja sebagai hasil dari usaha seseorang yang dicapai dengan adanya kemampuan dan perbuatan dalam situasi tertentu atau dapat dikatakan bahwa kinerja merupakan hasil keterkaitan antara usaha, kemampuan dan persepsi tugas. Penelitian Lomanjaya et al., (2013) menga- takan bahwa kinerja merupakan prestasi karyawan dari tugas-tugasnya yang telah ditetapkan. Jadi pengertian kinerja merupa- kan sebuah hasil kerja yang dicapai oleh seseorang dalam melaksanakan tugas yang diberikan kepadanya sesuai dengan kriteria yang diharapkan perusahaan. Selanjutnya, kecocokan akan kepemim- pinan ditempat kerja menumbuhkan kepu- asan kerja tersendiri bagi karyawan RS (Prayatna & Subudi, 2016). Kepuasan kerja merupakan sikap umum seorang individu terhadap pekerjaannya yaitu selisih antara banyaknya ganjaran yang diterima seorang pekerja dengan banyaknya usaha yang sudah dikeluarkan. Puas tidaknya karya- wan ditempat kerja akan mempengaruhi perilaku mereka dalam bekerja, yang pada akhirnya dapat menjadi faktor penentu kinerja karyawan (Plangiten, 2013; Sinurat, 2017; Yanoto, 2018). Oleh karena itu, mengingat pentingnya masalah tersebut, maka penulis tertarik un- tuk melakukan telaah ilmiah yang berjudul “Pengaruh Gaya Kepemimpinan Terhadap Kinerja Karyawan melalui variabel inter- vening Kepuasan Kerja pada karyawan Rumah Sakit X di Kabupaten Gresik” Faktor-faktor Ya Kinerja Karyawan Terdapat faktor-faktor yang mempe- ngaruhi kinerja karyawan meliputi strategi organisasional (nilai tujuan jangka pendek dan jangka panjang, batasan situasional (budaya organisasi dan kondisi ekonomi), kemampuan individual, gaya kepemimpi- nan, motivasi kerja karyawan, cara kerja, disipline kerja karyawan (Pawirosumarto et al., 2017). Selanjutnya pada artikel Anastasiou et.al (2014) memaparkan hasil uji bagaimana kepuasan kerja dan tingkat stres kerja mampu mempengaruhi kinerja. Responden yang merupakan guru, menun- jukkan bahwa guru lebih puas dengan pekerjaan itu sendiri (sifat pekerjaan itu sendiri, kemampuan untuk bekerja mem- bantu siswa mereka) dan hasil menunjuk- Indikator Penilaian Kinerja Karyawan Menurut Campbell et al., (1990) ber- hasil tidaknya kinerja yang telah dicapai or- gannisasi tersebut dipengaruhi oleh tingkat kinerja dari karyawan secara individu maupun kelompok. Selanjutnya, ada enam indikator yang digunakan untuk mengukur kinerja pegawai yaitu kualitas, kuantitias, ketepatan waktu, efektivitas dan keman- dirian. Tentunya tidak semua kriteria peng- ukuran kinerja dipakai dalam suatu penilaian kinerja karyawan. Hal ini harus disesuaikan dengan jenis pekerjaan yang akan dinilai. Penilaian pekerja pada pela- yanan jasa hotel akan berbeda dengan pekerja di pabrik manufaktur. Merujuk pada teori Bernardian dan Russel yang selanjutnya dikembangkan pada buku Faustino Cardoso Gomes (2005) (Destiasari, 2015) ada beberapa tipe kriteria atau indikator penilaian kinerja yang didasarkan atas perilaku yang spesifik, adalah: 1. Quantity of Work, yaitu jumlah kerja yang dilakukan dalam suatu periode waktu yang telah ditentukan. 2. Quality of work, yaitu kualitas kerja yang dicapai berdasarkan syarat kesesuaian dan kesi- apannya. 3. Job Knowledge, yaitu pengeta- huan atas pekerjaannya. Luasnya pemaha- man mengenai pekerjaan dan keteram- pilannya. 4. Creativeness, yaitu keaslian gagasan-gagasan yang dimunculkan dan tindakan-tindakan untuk menyelesaikan persoalan-persoalan yang timbul. 4. Cooperation, yaitu kesediaan untuk bekerja sama dengan orang lain (sesama anggota organisasi). 5. Dependability, yaitu kesadar- an dan dapat dipercaya dalam hal kehadir- an dan penyelesaian pekerjaan. 6. Initiative, Kinerja Kinerja adalah penampilan hasil karya personel baik kuantitas maupun kualitas dalam suatu organisasi. Kinerja dapat merupakan penampilan individu maupun kelompok kerja personel (Fuadiputra, 2013). Penampilan hasil karya tidak terbatas kepada personel yang memangku jabatan fungsional maupun struktural tetapi juga kepada keseluruhan jajaran di dalam organisasi. JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 42 kan bahwa guru muda dan perempuan menunjukkan tingkat stres yang tinggi. Faktor lingkungan yang diuji juga memiliki efek positif pada kinerja. yaitu semangat untuk menyelesaikan tugas- tugas baru dan dalam memperbesar tang- gungjawabnya. 7. Personality quality, yaitu menyangkut kepribadian, kepemimpinan, keramah-tamahan dan intregritas pribadi . Faktor tersebut antara lain: pemberian penghargaan, kondisi kerja yang baik, motivasi oleh atasan dan partisipasi dalam administrasi kantor dan pengambilan keputusan. Temuan ini juga didukung penelitian (Ibrahim et al., 2017). Dimana karyawan tidak hanya memerlukan hard skill dalam bekerja, tetapi juga memerlukan soft skill. Selanjutnya soft skill akan dikem- bangkan lebih terarah dengan pelatihan yang efektif dari perusahaan. Pelatihan yang tepat akan membantu meningkatkan kinerja karyawan. Menurut Tsui, et. al., 1997 pada Destiasari (2015). Pengukuran kinerja dapat dilakukan dengan menggunakan indikator seperti: 1. Kuantitas kerja karyawan. 2. Ku- alitas kerja karyawan. 3. Efisiensi karyawan. 4. Standar kualitas karyawan. 5. Usaha yang lebih keras. 6. Karyawan memegang stan- dar professional yang tinggi. 7. Kemam- puan karyawan dalam melaksanakan pe- kerjaan inti. 8. Kemampuan karyawan menggunakan akal sehat. 9. Ketepat-an karyawan berkaitan dengan pekerjaan utamanya. 10. Kreativitas karyawan dalam melaksanakan pekerjaan utamanya Kepemimpinan Kepemimpinan atau leadership meru- pakan ilmu terapan dari ilmu-ilmu social, sebab prinsip-prinsip dan rumusannya diharapkan dapat mendatangkan manfaat bagi kesejahteraan manusia. Ada banyak pengertian yang dikemukakan oleh para pakar menurut sudut pandang masing- masing, definisi-definisi tersebut menun- jukkan adanya beberapa kesamaan. j y p Hoyt et al., (2013); Hoyt & Price (2015) menjelaskan kepemimpinan adalah suatu kegiatan atau seni mempengaruhi orang lain agar mau bekerjasama yang didasarkan pada kemampuan orang tersebut untuk membimbing orang lain dalam mencapai tujuan-tujuan yang diinginkan kelompok. Penelitian Raus et al., (2012) setuju dengan definisi teori kepemimpinan Clement yaitu kepemimpinan merupakan proses dimana Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 43 seseorang menetapkan tujuan atau arah dan menentukan orang-orang didalamnya un- tuk bertindak bersama dengan kompetensi dan komitmen penuh untuk agar tercapai tujuan yang sudah disepakati. Pada dasar- nya, menurut literature review oleh Van Wart (2013) memahami definisi kepemimpinan merupakan suatu hal yang menantang, tetapi bagi mereka yang berharap menjadi pemimpin yang efektif harus mampu menghadapi tantangan dan menikmatinya. mimpinan merupakan suatu perilaku yang digunakan oleh seseorang pada saat orang tersebut memipin sebuah kelompok/ mencoba mempengaruhi perilaku orang lain seperti yang ia inginkan. Gaya kepemimpinan dalam organisasi sangat diperlukan untuk mengembangkan lingkungan kerja yang kondusif dan membangun iklim motivasi bagi karyawan sehingga diharapkan akan menghasilkan produktivitas yang tinggi (Bhargavi & Yaseen, 2016; Yanoto, 2018; Yudiaatmaja, 2013). Menurut Yudiaatmaja (2013) kepemim- pikan didefinisikan sebagai tindakan me- mimpin orang lain, dimana pemimpinnya disebut juga sebagai team leader (pemimpin kelompok) yang memahami tanggung ja- wab tugasnya, memahami kondisi bawa- hannya dan bersedia meleburkan diri atas tuntutan dan konsekuensi yang terjadi padanya serta memiliki komitmen untuk membawa setiap bawahannya mengeksplo- rasi kapasitas dirinya hingga menghasilkan prestasi tertinggi. Gaya kepemimpinan adalah suatu cara yang digunakan oleh seorang pemimpin dalam mempengaruhi perilaku orang lain (Van Wart, 2013). Dari gaya ini dapat diambil manfaatnya untuk dipergunakan sebagai pemimpin dalam memimpin bawahan atau para pengikutnya. Pemimpin tidak dapat menggunakan gaya kepemim- pinan yang sama dalam memimpin bawahannya, namun harus disesuaikan dengan karakter-karakter tingkat kemam- puan dalam tugas setiap bawahannya (Bhatti et al., 2012; Fuadiputra, 2013). Siagian (2002) pada tulisan (Dolphina, 2012) mengemukakan bahwa kepemim- pinan adalah kemampuan seseorang untuk mempengaruhi orang lain (para bawahan- nya) sedemikian rupa sehingga orang lain itu mau melakukan kehendak pemimpin meskipun secara pribadi hal itumungkin tidak disenanginya. Kepemimpinan ( p ) Pemimpin yang efektif dalam Meneg- rapkan gaya tertentu dalam kepemim- pinannya terlebih dahulu harus memahami siapa bawahan yang dipimpinnya, menger- ti kekuatan dan kelemahan bawahannya, dan mengerti bagaimana caranya memanfa- atkan kekuatan bawahan untuk mengim- bangi kelemahan yang mereka miliki. Isti- lah gaya merupaka pilihan/cara yang di- pergunakan pimpinan dalam mempe- ngaruhi para pengikutnya. Dari beberapa definisi diatas dapat disimpulkan bahwa kepemimpnan meru- pakan kemampuan mempengaruhi orang lain, bawahan maupun kelompok, kemam- puan mengarahkan tingkah laku bawahan atau kelompok, memiliki kemampuan atau keahlian khusus dalam bidang yang diinginkan oleh kelompoknya, untuk mencapai tujuan organisasi atau kelompok tersebut. g g y Dalam teori jalur tujuan (Path Goal Theory) yang dikembangkan oleh House (1971) menyatakan bahwa pemimpin men- dorong kinerja yang lebih tinggi dengan cara memberikan kegiatan-kegiatan yang mempengaruhi bawahannya agar percaya bahwa hasil yang maksimal bisa dicapai dengan usaha yang serius. Kepemimpinan yang berlaku secara universal menghasil- kan tingkat kinerja dan kepuasan bawahan yang tinggi. 1. Gaya Direktif Dimana pemimpin memberitahukan kepada bawahan apa yang diharapkan dari mereka, memberitahukan jadwal kerja yang harus diselesaikan dan standar kerja, serta memberikan bimbingan secara spesifik tentang cara-cara menyelesaikan tugas ter- sebut, termasuk di dalamnya aspek peren- canaan, organisasi, koordinasi dan penga- wasan. Karakteristik pribadi bawahan mempengaruhi gaya kepemimpinan yang efektif. Jika bawahan merasa mempunyai kemampuan yang tidak baik, kepemim- pinan instrumental (direktif) akan lebih sesuai. Sebaliknya apabila bawahan merasa mempunyai kemampuan yang baik, gaya direktif akan dirasakan berlebihan, bawa- han akan cenderung memusuhi. y g pp 3. Gaya Partisipatif 3. Gaya Partisipatif Gaya kepemimpinan dimana mengha- rapkan saran-saran dan ide mereka sebelum mengambil suatu keputusan. Apabila iba- wahan merasa mempunyai kemampuan yang baik, gaya kepemimpinan direktif akan dirasa berlebihan, bawahan akan sen- derung memusuhi, sehingga gaya kepe- mimpinan partisipatif lebih sesuai. (Parijat & Bagga, 2014) mengatakan bahwa partisipasi bawahan juga mempengaruhi dalam pengambilan keputusan oleh pe- mimpin. Situasi dimana kebutuhan untuk berkembang rendah dan karyawan menger- jakan tugas-tugas yang mudah, sikap yang dianggap tepat untuk karyawan yang secara ego terlibat dengan pekerjaan dan mengalami kepuasan intrinsik dari tugas yang dikerjakan adalah sikap partisipatif dan berorientasi prestasi House juga menyatakan bahwa direktif leadership itu memberitahukan kepada para bawahan apa yang diharapkan dari mereka, memberi pedoman yang spesifik, meminta para bawahan untuk mengikuti peraturan-peraturan dan prosedur-prose- dur, mengatur waktu dan mengkoordinaasi pekerjaan mereka. Gaya Kepemimpinan Ada suatu pendekatan yang dapat digunakan untuk memahami kesuksesan dari kepemimpinan, yakni dengan memu- satkan perhatian pada apa yang dilakukan oleh pemimpin tersebut. Jadi yang dimak- sudkan disini adalah gayanya. Gaya kepe- 44 JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 Dalam situasi yang berbeda mensyarat- kan gaya kepemimpinan yaitu karakteristik personal dan kekuatan lingkungan. Teori ini juga menggambarkan bagaimana per- sepsi harapan dipengaruhi oleh hubungan kontijensi diantara empat gaya kepemim- pinan dan berbagai sikap dan perilaku karyawan. Perilaku pemimpin memberikan motivasi sampai tingkat: (1)Mengurangi halangan jalan yang mengganggu penca- paian tujuan, (2)Memberikan panduan dan dukungan yang dibutuhkan oleh para kar- yawan, dan (3) Mengaitkan penghargaan yang berarti terhadap pencapaian tujuan. anak buahnya. Sikap direktif yang de- mikian diperkirakan akan membuahkan hasil yang positif. y g p 2. Gaya Supportif 2. Gaya Supportif y pp Gaya kepemimpinan yang menunjuk- kan keramahan seorang pemimpin, mudah ditemui daan menunjukkan sikap memper- hatikan bawahannya. Kepemimpinan gaya supportif, menggambarkan situasi dimana pegawai yang memiliki kebutuhan tinggi untuk berkembang mengerjakan tugas- tugas yang mudah, sederhana, dan rutin. Individu seperti ini mengharapkan pekerja- an sebagai sumber pemuasan kebutuhan, tetapi kebutuhan mereka tidak terpenuhi. Reaksi yang mungkin timbul adalah pera- saan kecewa dan frustasi. Bukti-bukti penelitian oleh House (1971) dengan kuat menunjukkan bahwa karyawan yang me- ngerjakan tugas-tugas yang kurang me- muaskan seperti ini cenderung memberikan respon positif terhadap sikap pimpinan yang supportif. Selain itu House percaya bahwa pemimpin dapat menunjukkan lebih dari satu gaya kepemimpinan. Selanjutnya, House mengidentifikasikan lima gaya kepemimpinan, yaitu: 1 G Di ktif 4. Gaya Orientasi Prestasi Gaya kepemimpinan dimana pemim- pin menetapkan tujuan yang menantang dan mengharapkan bawahan untuk ber- prestasi semaksimal mungkin serta terus menerus mencari pengembangan prestasi Fungsi pimpinan dalam situasi ini adalah memberikan struktur tugas dengan merencanakan, mengorganisir, mengkoor- dinasi, mengarahkan, dan mengontrol kerja Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 4 45 dalam pencapaian tujuan tersebut. Tingkah laku individu didorong oleh need for achievement atau kebutuhan untuk berpres- tasi. Kepemimpinan yang berorientasi ke- pada prestasi (achievement) dihipotesakan akan meningkatkan usaha dan kepuasan bila pekerjaan tersebut tidak tersetruktur (misalnya kompleks dan tidak diulang- ulang) dengan meningkatkan rasa percaya diri dan harapan akan menyelesaikan sebuah tugas dan tujuan yang menantang. dalam organisasi. Bhatti et al., (2012) me- nyatakan gaya kepemimpinan autokratis mendeskripsikan pemimpin yang cende- rung memusatkan kekuasaan kepada dirinya sendiri, mendikte bagaimana tugas harus diselesaikan, membuat keputusan se- cara sepihak, dan meminimalisasi partisi- pasi karyawan. Ciri-ciri gaya kepemim- pinan autokratis : 1. Pemimpin kurang memperhatikan kebutuhan bawahan. 2. Semua kebijakan ditentukan oleh pemim- pin. 3. Pemimpin cenderung menjadi pari- badi dalam pujian dan kecamannya terha- dap kerja setiap anggota. 4. Pemimpin mengambil jarak dari partisipasi kelompok aktif kecuali bila menunjukan keahliannya. 2. Gaya kepemimpinan Demokratis / Partisipatif Kepuasan kerja lebih tinggi diperoleh apabila telah melaksanakan prestasi kerja yang baik. Pegawai yang memiliki kebu- tuhan untuk berkembang dan mengerjakan tugas-tugas sulit berdasarkan pembahasan konseptual House (1971) sikap pemimpin yang paling tepat untuk karyawan ini adalah gaya partisipatif dan berorientasi prestasi. Partisipatif Kepemimpinan demokratis ditandai dengan adanya suatu struktur yang pe- ngembangannya menggunakan pendeka- tan pengambilan keputusan yang koopera- tif. Dibawah kepemimpinan demokratis bawahan cenderung bermoral tinggi, dapat bekerja sama, mengutamakan mutu kerja dan dapat mengarahkan diri sendiri. Gaya kepemimpinan demokratis mendeskripsi- kan pemimpin yang cenderung mengikut sertakan karyawan dalam pengambilan ke- putusan, mendelegasikan kekuasaan, men- dorong partisipasi karyawan dalam menen- tukan bagaimana metode kerja dan tujuan yang ingin dicapai, dan memandang umpan balik sebagai suatu kesempatan untuk melatih karyawan (Raus et al., 2012; Scheidlinger, 1994). Ciri-ciri gaya kepemim- pinan demokratis : 1. Semua kebijaksanaan terjadi pada kelompok diskusi dan kepu- tusan diambil dengan dorongan dan ban- tuan dari pemimpin. 2. Kegiatan-kegiatan didiskusikan, langkah-langkah umum untuk tujuan kelompok dibuat, dan jika dibutuhkan petunjuk-petunjuk teknis pe- mimpin menyarankan dua atau lebih alternatif posedur yang dapat dipilih. 3. Para anggota bebas bekerja dengan siapa saja yang mereka pilih dan pembagian tu- gas ditentukan oleh kelompok. 3. Gaya Ke- pemimpinan Laissez-faire (Kendali Bebas) p 5. Gaya Pengasuh Dalam kepemimpinan gaya pengasuh, sikap yang mungkin tepat adalah campur tangan minim dari pimpinan. Dimana pe- mimpin hanya memantau kinerja tetapi tidak mengawasi bawahan secara aktif. Tidak dibutuhkan banyak interaksi antara pimpinan dengan bawahan sepanjang kinerja karyawan tidak menurun. Pimpinan merasa lebih tepat untuk tidak campur tangan dengan tugas-tugas bawahan karena rasa percaya dari pimpinan bahwasannya karyawan mampu menyelesaikan dengan baik setiap tugasnya. Menurut teori kepemimpinan oleh Lewin (Bhargavi & Yaseen, 2016; Raus et al., 2012; Scheidlinger, 1994) menyimpulkan ada tiga gaya kepemimpinan; gaya kepe- mimpinan autokratis, gaya kepemimpinan demokratis, gaya kepemimpinan Laissez- Faire (Kendali Bebas). Berikut merupakan penjelasan atas teori tersebut: 1. Gaya Kepemimpinan Autokratis. Menurut Scheidlinger (1994) kepemim- pinan autokratis adalah gaya kepemimpi- nan yang menggunakan metode pendeka- tan kekuasaan dalam mencapai keputusan dan pengembangan strukturnya, sehingga kekuasaanlah yang paling diuntungkan JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 46 Gaya kepemimpinan kendali bebas mendeskripsikan pemimpin yang secara keseluruhan memberikan karyawannya atau kelompok kebebasan dalam pembuat- an keputusan dan menyelesaikan pekerjaan menurut cara yang menurut karyawannya paling sesuai (Raus et al., 2012; Scheidlinger, 1994). Ciri-ciri gaya kepemimpinan kendali bebas: 1. Pemimpin membiarkan bawahan- nya untuk mengatur dirinya sendiri. 2. Pe- mimpin hanya menentukan kebijaksanaan dan tujuan umum. 3. Bawahan dapat meng- ambil keputusan yang relevan untuk mencapai tujuan dalam segala hal yang mereka anggap cocok. Partisipatif kerja dipengaruhi oleh variable-variabel yang dapat dikelompokkan menjadi tiga kelompok, yaitu karakteristik pekerjaan, karakteristik organisasi dan karakteristik individu/pekerja. Kepuasan kerja menda- pat tempat yang dangat penting bagi perilaku organisasi. Kepuasan kerja adalah konsep praktis yang sangat penting, karena merupakan dampak atau hasil dari keefektifan bekerja dan kesuksesan dalam bekerja. Kepuasan kerja yang rendah pada organisasi adalah rangkaian dari menurunnya pelaksanaan tugas, meningkatnya absensi, dan penurun- an moral organisasi. Sedangkan pada ting- kat individu, ketidakpuasan kerja, berkai- tan dengan keinginan yang besar untuk keluar dari kerja, meningkatnya stress kerja, dan munculnya berbagai masalah psiko- logis dan fisik (Johnson et al., 2005; Said & El-Shafei, 2021) Tiap gaya kepemimpinan diatas me- miliki kelebihan dan kekurangan masing- masing. sehingga akan sangat menguntung- kan jika dikolaborasi antara tipe yang satu dengan tipe yanrg lain, dan tidak hanya terpaku pada satu gaya kepemimpinan saja. Dalam pengertian yang tidak jauh berbeda, Edwards & Burnard (2003); Eliyana et al., (2019) berpendapat bahwa kepuasan kerja adalah seperangkat perasa- an karyawan tentang menyenangkan atau tidaknya pekerjaan mereka yang pada akhirnya mempengaruhi hasil kerja mere- ka. Penekanan yang lebih spesifik mema- parkan bahwa kepuasan kerja diarahkan pada lingkungan kerja dimana pekerja mengaktualisasikan tugas yang dibebankan kepadanya. Disebut juga kepuasan kerja merupakan hasil dari berbagai macam sikap (attitude) yang dipunyai seorang pekerja. Dalam hal ini yang dimaksud dengan sikap adalah yang berhubungan dengan pekerjaan beserta faktor-faktor yang spesifik seperti pengawasan, upah, kesempatan promosi, kondisi kerja, peng- akuan terhadap kecakapan, penilaian kerja yang sehat, hubungan social di dalam pe- kerjaan, penyelesaian yang cepat terhadap keluhan-keluhan dan perlakuan yang baik dari pimpinan terhadap para pekerja (Fitrianis & Apriliani, 2019; Fuadiputra, 2013; Prayatna & Subudi, 2016). Kepuasan Kerja Kepuasan kerja merupakan orientasi individu yang berpengaruh terhadap peran dalam bekerja. Menurut Robbin & Judge (2008) dalam bukunya, kepuasan kerja merupakan sikap umum seorang individu terhadap pekerjaannya yaitu selisih antara banyaknya ganjaran yang diterima seorang pekerja dengan banyaknya yang mereka yakini seharusnya yang mereka terima. Lebih lanjut dikatakan sikap umum itu dapat digolongkan kepada 2 hal yaitu sikap negatif akan dimunculkan seorang pekerja apabila mereka tidak puas dengan pekerjaannya. Sebaliknya apabila seseorang puas dengan pekerjaannya maka mereka akan memunculkan sikap positif. Penilaian seorang karyawan terhadap puas atau tidak puas akan pekerjaannya merupakan penjumlahan yang rumit dari sejumlah unsur pekerjaan. Kepuasan kerja ditentukan oleh beberapa faktor yaitu kerja yang secara menguras mental, imbalan yang pantas, kondisi kerja yang mendu- kung. Anastasiou & Papakonstantinou (2014); Bhatti et al., (2012); Mak & Sockel (2001) mengemukakan bahwa kepuasan Sementara itu, Robbin & Judge (2008) juga mengemukakan bahwa ada perbedaan Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 47 47 antara faktor-faktor penentu kepuasan de- ngan faktor-faktor penentu ketidakpuasan (faktor pemelihara). Faktor penentu kepu- asan merupakan faktor yang mendorong motivasi seseorang yang meliputi prestasi kerja, pengakuan terhadap pekerjaan, pe- kerjaan itu sendiri, tanggung jawab, promo- si dan pengembangan kerja. Sedang faktor higienis (pemeliharaan) yang meliputi kebijaksanaan dan administrasi perusaha- an, pengawasan, kondisi kerja, hubungan interpersonal dengan teman kerja, gaji dan keamanan kerja, kehidupan pribadi, bukan merupakan faktor motivasi. Artinya apabila ada faktor pemelihara maka tidak akan muncul ketidakpuasan kerja. organisasi sehubungan dengan pelaksana- an kerja yangtelah dia tunjukkan. b) Kepuasan lebih banyak tergantung pada pelaksanaan kerja daripada sebaliknya, yaitu pelaksanaan kerja tergantung pada kepuasan. Selanjutnya, terdapat lima buah kom- ponen dari kepuasan kerja yaitu: kepuasan terhadap pekerjaan itu sendiri, kepuasan terhadap supervisi, kepuasan terhadap rekan sekerja, kepuasan terhadap gaji dan kepuasan terhadap kesempatan promosi. Menurut Robbins juga terdapat beberapa determinan penting yang me- nentukan kepuasan kerja seseorang, yaitu: a) Kerja yang secara mental menantang. b) Imbalan yang pantas c) Kondisi kerja yang mendukung. d) Rekan kerja yang mendu- kung. e) Supervisor p j Kepuasan kerja seperti dikembangkan oleh Luthans & Youssef, (2007); Youssef & Luthans (2007) adalah perasaan positif/ negatif pekerja terhadap pekerjaannya. Hal ini merupakan sikap umum terhadap pekerjaan yang didasarkan pada penilaian aspek yang berada dalam pekerjaan. Sikap seseorang terhadap pekerjaan menggam- barkan pengalaman yang menyenangkan dan tidak menyenangkan juga harapannya di masa mendatang. Kepuasan Kerja Pemaparan Luthan’s juga menjelaskan kepuasan kerja merupa- kan suatu sikap (attitude), suatu keadaan kognitif yang ada dalam diri seseorang (internal cognitive state). g ) p Rasa ketidakpuasan karyawan dapat diungkapkan ke dalam berbagai tindakan seperti: a) Exit, yaitu perilaku yang diarah- kan ke arah meninggalkan organisasi. Mencakup pencarian suatu posisi baru maupun meminta berhenti. b) Suara, yaitu dengan aktif dan konstruktif mencoba memperbaiki kondisi. Mencakup saran per- baikan, membahas sejumlah problem de- ngan atasan, dan beberapa bentuk kegiatan serikat buruh. c) Kesetiaan, yaitu wujud ketidakpuasan dalam bentuk pasif, tetapi optimis menunggu membaiknya kondisi yang dirasa kurang menguntungkan. Men- cakup berbicara membela organisasi meng- hadapi kritik dan mempercayai oraganisasi dan manajemennya untuk melakukan hal yang tepat. d) Pengabaian, secara pasif membiarkan kondisi memburuk, termasuk kemungkiran atau datang terlambat secara kronis, upaya dikurangi, dan tingkat kekeliruan yang meningkat. ( g ) Dalam model teori isi motivasi, ke- puasan kerja merupakan penjumlahan dari berbagai faktor yang menimbulkan kepuas- an seperti tanggung jawab dan pengem- bangan potensi seseorang. Dalam model Lawler and Porter (Robbin & Judge, 2008) kepuasan merupakan suatu variabel yang berasal dari sesuai tidaknya antara imbalan yang diharapkan. Individu akan merasa tidak puas apabila imbalan yang diterima lebih rendah dari imbalan yang diharapkan. Dari penjelasan tentang kepuasan tadi, terdapat dua hal penting yang berasal dari pemikiran tradisional tentang kepuasan, yaitu; a) Kepuasan hanya ditentukan oleh imbalan nyata yang diterima oleh sese- orang, dan perasaan seseorang terhadap imbalan yang seharusnya diberikan oleh Hubungan Antar Variabel Hubungan Antara Gaya Kepemimpinan dengan Kinerja Karyawan Pemimpin mempunyai tanggung ja- wab menciptakan kondisi-kondisi yang merangsang anggota agar dapat mencapai tujuan yang ditentukan. Gaya kepemim- JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 48 pinan menjadi cermin kemampuan sese- orang dalam mempengaruhi individu atau kelompok. Menurut Bhargavi & Yaseen (2016) ga- ya kepemimpinan berpengaruh positif kuat terhadap kinerja karyawan dan kinerja organisasi. Gaya kepemimpinan memberi- kan manajer kemampuan untuk mengambil keputusan, berpartisipasi dengan orang lain atau meminta bawahan mereka untuk menangani masalah pada organisasi. Kerja- sama ini membentuk hubungan atasan bawahan yang positif yang pada akhirnya meningkatkan kinerja mereka. Seorang pemimpin harus mampu men- jaga keselarasan antara pemenuhan kebutu- han individu dengan pengarahan individu pada tujuan organisasi. Pemimpin yang efektif adalah pemimpin yang mengakui kekuatan-kekuatan penting yang terkan- dung dalam individu atau kelompok, serta fleksibel dalam cara pendekatan yang digunakan demi meningkatkan kinerja seluruh organisasinya. Hubungan Antara Gaya Kepemimpinan dengan Kepuasan Kerja Pada penelitian ini, peneliti mengacu pada penelitian sebelumnya yang dilaku- kan oleh Kusumawati (2008); Plangiten (2013); Sinurat (2017) yang membuktikan bahwasannya ada hubungan positif signifi- kan antara gaya kepemimpinan dengan kepuasan kerja karyawan. Sehingga, ditarik hipotesis sebagai berikut: g j Perilaku pemimpin merupakan salah satu faktor penting yang dapat mempenga- ruhi kepuasan kerja. Penelitian Prayatna & Subudi (2016); Yanoto (2018) menunjukkan bahwa gaya kepemimpinan mempunyai hubungan yang positif terhadap kepuasan kerja para pegawai. Hasil penelitian terdahulu oleh Bhatti et al., (2012) diperoleh bahwa dibawah gaya kepemimpinan autocratic dan democratic mengkonfirmasi adanya hubungan yang positif terhadap kepuasan kerja. Hubungan yang akrab dan saling tolong-menolong dengan teman sekerja serta atasan merupakan hal yang sangat penting dan memiliki hubungan kuat dengan kepuasan kerja. g H2= Gaya kepemimpinan berpengaruh positif signifikan terhadap kepuasan kerja. Kepuasan Kerja Temuan-temuan ini memberikan indi- kasi bahwa gaya kepemimpinan seorang pemimpin sangat berpengaruh terhadap kinerja bawahannya. Kinerja karyawan akan meningkat ketika mereka merasa cocok dengan gaya kepemimpinan yang diterapkan, begitu pula sebaliknya. Di samping itu untuk mendapatkan kinerja yang baik diperlukan juga adanya pem- berian pembelajaran terhadap bawahannya, organisasi yang mendukung, budaya kerja yang baik dan pemberian pelatihan kerja yang efektif. Gaya kepemimpinan dalam perusaha- an merupakan hal penting dalam sebuah era organisasi modern yang menghendaki adanya demokratisasi dalam pelaksanaan kerja dan kepemimpinan perusahaan. Gaya kepemimpinan adalah suatu seni menge- rahkan segala sumber daya yang dimiliki dalam upaya mencapai tujuan dengan setrategi yang disesuaikan dengan kondisi lingkungan. Akibat yang mungkin timbul dari adanya gaya kepemimpinan yang buruk adalah penurunan kinerja karyawan yang akan membawa dampak kepada penurunan kinerja total perusahaan. Disisi lain, terdapat penelitian dari Lomanjaya et al., (2013) yang memaparkan hasil tidak adanya hubungan yang positif signifikan pada gaya kepemimpinan trans- formasional dan gaya kepemimpinan tran- saksional terhadap kinerja karyawan. Responden merupakan pekerja perusahaan jasa yang bekerja di Rumah Sakit. Artinya, dimungkinkan karyawan kurang merasa- kan sosok kepemimpinan dari leadernya, atau bisa jadi karena pekerja merupakan pekerja dengan tingkat pendidikan yang rendah dengan status karyawan tidak tetap sehingga mereka bekerja sebatas pertu- karan transaksional tanggung jawab beker- ja. Artinya, tidak semua gaya kepemim- pinan berpengaruh pada kinerja karyawan. Faktor pendidikan, status karyawan juga mempengaruhi hubungan keduanya. Gaya kepemimpinan (leadership style) merupakan cara pimpinan untuk mempe- ngaruhi orang lain atau bawahannya sede- mikian rupa sehingga orang tersebut mau melakukan kehendak pimpinan untuk mencapai tujuan organisasi meskipun secara pribadi hal tersebut mungkin tidak disenangi. Berbagai macam gaya kepemim- pinan dicoba diteliti pada penelitian terda- hulu dalam upayanya meningkatkan ki- nerja karyawan. Seperti halnya oleh Fuadiputra, (2013); Yulianto et al., (2018) memaparkan adanya pengaruh positif sig- nifikan antara gaya kepemimpinan demo- kratis terhadap kinerja karyawan. Berikut pula pada penelitian Ali & Agustian (2018); Ariyani (2016) yang menguji gaya kepe- mimpinan secara general/umum/tidak membagi pada satu gaya, menunjukkan adanya pengaruh positif terhadap kinerja karyawan. Pada penelitian ini, peneliti mencoba meneliti kinerja para tenaga medis khusus- nya Perawat di RS swasta di Gresik. Dilihat dari status karyawan dan pendidikan yang mayoritas adalah Diploma dan Sarjana maka peneliti menarik hipotesis: Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 49 H1= Gaya kepemimpinan berpengaruh positif signifikan terhadap kinerja karyawan. secara penuh, termasuk menentukan kebija- kan dan prosedur, memutuskan tujuan yang ingin dicapai dan mengawasi semua kegiatan organisasi. Kepuasan Kerja Hubungan Antara Gaya Kepemimpinan dengan Kepuasan Kerja Hubungan Antara Kepuasan Kerja Dengan Kinerja Karyawan Mempertahankan dan meningkatkan kinerja karyawan agar tetap baik bukanlah hal yang mudah dilakukan oleh perusahaan jasa. Hal ini umumnya terjadi karena jasa/ pelayanan merupakan output langsung yang dirasakan oleh kostumer. Berbeda dengan produk berupa barang, yang di produksi masal oleh pabrik. Pemanfaatan- nya pun bisa langsung dirasakan atau waktu yang akan datang. Beda halnya dengan jasa, dimana pemanfaatannya saat itu juga dan langsung dirasakan. Sehingga jika ada kostumer yang tidak terpuaskan, dapat menjadi salah satu indikator menu- runnya kinerja karyawan/pemberi jasa. Kusumawati (2008); Sinurat (2017) dalam penelitiannya menyatakan bahwa terdapat hubungan positif signifikan antara variabel gaya kepemimpinan dengan kepuasan kerja karyawan. Diungkapkan pula bahwa gaya kepemimpinan yang efek- tif adalah kepemimpinan yang disesuaikan dengan situasi dan kondisi (contingency). Salah satu indikasi turunnya semangat dan kegairahan kerja ditunjukkan dengan tingginya tingkat absensi dan perpindahan pegawai. Hal itu timbul sebagai akibat dari kepemimpinan yang tidak disenangi. Kinerja karyawan tetap tinggi diperlu- kan agar keberlangsungan organisasi tetap berjalan dengan baik. Konsumen yang merasa terpuaskan oleh kinerja karyawan yang melayani, maka konsumen akan datang lagi dan bahkan merekomendasikan perusahaan ke rekan sekitarnya. Oleh karenanya, mempertahankan kinerja yang optimal menjadi kunci mempertahankan pelanggan. y g g Salah satu faktor yang menyebabkan ketidakpuasan kerja ialah sifat atasan yang tidak mau mendengar keluhan dan panda- ngan pekerja dan mau membantu apabila diperlukan. Seperti halnya penelitian Pratama (2012) memaparkan adanya penga- ruh negatif antara gaya kepemimpinan otokratis terhadap kepuasan kerja karya- wan. Hal ini membuktikan bahwasannya dibawah gaya kepemimpinan otokratis, karyawan merasa tidak puas dengan kerjanya. Dimana kepemimpinan otokratis adalah gaya kepemimpinan seorang pe- mimpin yang mengendalikan bawahannya Salah satu upaya meningkatkan kinerja karyawan yaitu dengan meningkatkan kepuasan kerja karyawan. Penelitian sebe- lumnya menunjukkan adanya pengaruh positif signifikan antara kepuasan kerja karyawan terhadap kinerjanya (Febriyana, JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 50 2015; Hamid & Hazriyanto, 2019; Indrawati, 2013; Nurnaningsih & Wahyono, 2017). Karyawan yang merasa puas akan peker- jaannya secara terbukti menunjukkan hasil yang bagus pula pada kinerjanya. Ketika kepuasan kerja tinggi, maka kinerja karyawan juga tinggi. Beberapa indikator kepuasan kerja yang umum digunakan adalah sikap atasan, hubungan rekan kerja, kompensasi, jenjang karir, dan lingkungan kerja karyawan. karyawan. Hal ini membuktikan bahwa gaya kepemimpinan yang sesuai dengan organisasi tempat bekerja secara signifikan mampu meningkatkan kinerja karyawan seiring dengan meningkat pula kepuasan kerja karyawan. Hubungan Antara Kepuasan Kerja Dengan Kinerja Karyawan Dalam kata lain, kepuasan kerja merupakan hal yang patut dipertim- bangkan dalam upaya meningkatkan kiner- ja karyawan (Cahyono, 2019; Fitrianis & Apriliani, 2019; Lousyiana & Harlen, 2015; Pambudi et al., 2016; Putri, 2018; Winarto & Purba, 2018). Disisi lain, hasil yang berbeda ditunjuk- kan oleh (Saufa & Maryati, 2017). Dengan objek penelitian yang sama yaitu di Rumah Sakit, dimana bisnis utama adalah pelaya- nan, penelitian ini memaparkan bahwa ke- puasan kerja tidak berpengaruh signifikan terhadap kinerja karyawan. Hal itu bisa saja terjadi karena perbedaan persepsi masing- masing karyawan. Walaupun hasil berbeda ditunjukkan pada penelitian Fathurrohman (2015); E. Kusumawati & Ansori (2008) dimana tidak terjadi hubungan mediasi oleh variabel kepuasan kerja atas gaya kepemimpinan terhadap kinerja, penelitian ini tetap mengacu pada penelitian yang menyatakan adanya hubungan mediasi oleh kepuasan kerja terhadap gaya kepemimpinan dan kinerja. g y Kepuasan kerja merupakan hal yang bersifat individu, puas tidaknya seseorang memiliki tingkat yang berbeda dengan kepuasan orang lain. Selain itu, tanggung jawab pekerjaan karyawan medis/perawat adalah merawat pasien. Puas atau tidak mereka tetap akan merawat pasien tersebut, terlebih lagi sumpah medis yang sudah diikrarkan semasa kuliah. Sehingga, ke- puasan kerja yang dirasakan tidak akan mempengaruhi kinerjanya, baik hal itu sifatnya menurunkan maupun meningkat- kan kinerja. H4=Kepuasan kerja memediasi hubungan antara gaya kepemimpinan terhadap kinerja karyawan. Kerangka Pemikiran Dari uraian yang didukung berda- sarkan landasan teoritis dan hasil penelitian empiris diatas, maka dapat digambarkan adanya keterkaitan antara gaya kepemim- pinan dan kinerja karyawan dengan ke- puasan kerja sebagai variabel intervening. Pola hubungan antar variabel tersebut da- pat digambarkan dalam kerangka peneli- tian pada gambar 1. Tetapi, pada penelitian menarik hi- potesis atas penelitian sebelumnya oleh (Indrawati, 2013) dimana karyawan yang merasa puas akan kerjanya akan memberi- kan dampak positif yaitu meningkatkan kinerja karyawan. Pada gambar 1 menunjukkan kerangka konseptual yang dibahas pada penelitian ini, dari gambar tersebut dapat ditarik 4 hipotesis. H1= Gaya kepemimpinan berpe- ngaruh positif signifikan terhadap kinerja karyawan. H2=Gaya kepemimpinan berpe- ngaruh positif signifikan terhadap kepuas- an kerja. H3=Kepuasan kerja berpengaruh positif signifikan terhadap kinerja karya- wan. H4=Kepuasan kerja memediasi hu- bungan antara gaya kepemimpinan terha- dap kinerja karyawan. H3=Kepuasan kerja berpengaruh positif signifikan terhadap kinerja karyawan. Kepuasan Kerja Memediasi Hubungan Antara Gaya Kepemimpinan Terhadap Kinerja Karyawan METODE PENELITIAN Jenis Penelitian Gaya Kepemimpinan: Gaya kepemim- pinan merupakan persepsi responden ter- hadap atasannya tentang pola perilaku yang diperlihatkan oleh atasannya pada saat mempengaruhi aktivitas bawahan. Variabel ini diukur dengan indikator menganut 3 (tiga) kriteria kepemimpinan dari Lewin’s: Penelitian ini menggunakan metode penelitian survei. Penelitian survei merupa- kan penelitian yang mengambil sampel dari suatu populasi yang dipilih, dan meng- gunakan kuesioner sebagai instrumen pe- ngumpulan data mengenai indikator- indikator yang berkaitan dengan variabel penelitian. Data yang diperoleh dianalisis dan hasilnya dipaparkan secara deskriptif. Penelitian ini menggunakan analisis kuan- titatif dan bertujuan untuk menganalisis hubungan antar variabel. a. Gaya kepemimpinan autokratis. • Pemimpin kurang memperhatikan kebutuhan bawahan. • Semua kebijakan ditentukan oleh pemimpin • Pemimpin tidak mengajak bawahan berpartisipasi dalam penyelesaian masalah g Penelitian ini dilakukankan di salah satu Rumah Sakit swasta di Kabupaten Gresik dengan mengambil obyek penelitian karyawan medis yaitu Perawat. Menurut Ghozali (2009) yang dimaksud populasi adalah wilayah generalisasi yang terdiri atas obyek/subyek yang mempunyai kuali- tas dan karakteristik tertentu yang ditetap- kan peneliti untuk dipelajari dan ditarik kesimpulannya. Populasi dalam penelitian ini adalah seluruh perawat yang berada dibawah kepala ruangan atau kepala unit yang ada di Rumah Sakit, dengan total per Januari 2019 adalah 215 orang. b. Gaya kepemimpinan demokratis. • Mengikutsertakan karyawan dalam pengambilan keputusan, • Mendorong partisipasi karyawan dalam pencapaian tujuan organisasi. • Memberikan kesempatan karyawan untuk jabatan yang lebih tinggi c. Gaya kepemimpinan Laissez-Faire (Kendali Bebas) • Pemimpin hanya menentukan kebijaksanaan dan tujuan umum. • Bawahan dapat mengambil keputusan yang relevan untuk mencapai tujuan dalam segala hal yang mereka anggap cocok Jumlah tersebut selanjutnya digunakan sebagai sampel penelitian. Kuesioner yang kembali sejumlah 132 dan banyak diantara- nya yang tidak mengisi secara penuh se- hingga terpaksa tidak digunakan. Akhir- nya, total data valid yang di proses pada penelitian ini berjumlah 100 sampel. Kepuasan Kerja: merupakan persepsi responden tentang bentuk perasaan terhadap pekerjaannya, situasi kerja dan hubungan dengan rekan kerjanya. Variabel ini diukur dengan Indikator: a. Kepuasan dengan gaji. b. Kepuasan dengan promosi. c) Kepuasan dengan rekan kerja. d) Kepuasan Kerja Memediasi Hubungan Antara Gaya Kepemimpinan Terhadap Kinerja Karyawan j y Penelitian sebelumnya mengkonfirma- si adanya pengaruh tidak langsung kepuas- an kerja karyawan terhadap hubungan an- tara gaya kepemimpinan terhadap kinerja Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 5 51 Gambar 1 Kerangka Konseptual Gambar 1 Kerangka Konseptual Gambar 1 Kerangka Konseptual Definisi Operasional Definisi operasional masing-masing variabel dan indikatornya dijelaskan seba- gai berikut: JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 52 Tabel 1 Nilai AVE Variabel Average Variance Extracted (AVE) Ket Gaya Kepemimpinan 0.516 Valid Kepuasan Kerja 0.519 Valid Kinerja 0.502 Valid Sumber: Olah data peneliti Kepuasan dengan penyelia. e) Kepuasan dengan pekerjaan itu sendiri. Kinerja: Kinerja merupakan persepsi responden tentang hasil kerjanya secara kualitas dan kuantitas yang sudah dicapai- nya dalam melaksanakan tugas sesuai de- ngan tanggung jawab yang diberikan kepadanya. selama satu periode waktu tertentu. Variabel ini diukur dengan indi- kator sebagai berikut: a) Tingkat kualitas hasil kerja. b) Tingkat keuletan dan daya tahan kerja. c) Tingkat disiplin dan absensi. d) Tingkat kerja sama antar rekan sekerja. e) Tingkat tanggung jawab akan hasil pekerja- annya. f) Tingkat inisiatif/kreatifitas yang dimiliki. g) Tingkat profesionalitas kerja yang dimiliki Pada Tabel 1 menunjukkan nilai AVE untuk semua variabel telah valid karena bernilai > 0,5. Nilai AVE 0,5 atau lebih berarti konstruk dapat menjelaskan 50% atau lebih varian itemnya. Selanjutnya Pada Tabel 2 berikut ini adalah nilai Cronbach’s Alpha. Cronbach’s Alpha merupakan sebuah ukuran keandal- an yang memiliki nilai berkisar dari nol sampai satu. ANALISIS DAN PEMBAHASAN Teknik analisis pada penelitian ini menggunakan Smart PLS 3 yang tahapan- nya merujuk pada penelitian oleh (Sari & Suhariadi, 2019). Terdapat beberapa tahap dalam pengujian variabel penelitian: perta- ma menguji validitas dan reliabilitas kon- vergen dari masing-masing indikator apa- kah merupakan bagian dari konstrak atau variabel penelitian. Selanjutnya menguji struktural model. Tujuan tahap ini untuk melihat ada tidaknya pengaruh antar variabel. Tabel 2 Nilai Cronbach’s Alpha Variabel Cronbach's Alpha Gaya Kepemimpinan 0.864 Kepuasan Kerja 0.765 Kinerja Karyawan 0.833 Sumber: Olah data peneliti Tabel 2 Nilai Cronbach’s Alpha Tabel 2 Nilai Cronbach’s Alpha Pada tabel 1 berikut merupakan hasil uji validitas dan reliabilitas indikator pene- litian. Validitas adalah ketepatan atau kecermatan suatu instrument dalam peng- ukuran. Sedangkan uji reliabilitas diguna- kan untuk mengetahui konsistensi alat ukur, apakah alat pengukur yang diguna- kan dapat diandalkan dan tetap konsisten jika pengukuran tersebut diulang. Nilai Conbach’s Alpha yang dihasilkan ketiga variabel kesemuanya >0,7 yang artinya variabel sudah cukup andal untuk dilanjutnkan pengujian selanjutnya. Tabel 3 Nilai Composite Reliability Tabel 3 Nilai Composite Reliability Variabel Composite Reliability Gaya Kepemimpinan 0.894 Kepuasan Kerja 0.842 Kinerja Karyawan 0.874 Sumber: Olah data peneliti Nilai AVE menggambarkan besarnya varian atau keragaman variable manifest yang dapat dimiliki oleh konstruk laten. Dengan demikian, semakin besar varian atau keragaman variable manifest yang dapat dikandung oleh konstruk laten, maka semakin besar representasi variabel manifes terhadap konstruk latennya. Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 53 Tabel 5 Nilai Outer Loading Indikator Outer Loading Keterangan X.1 0.785 Reliabel X.2 0.604 Reliabel X.3 0.611 Reliabel X.4 0.743 Reliabel X.5 0.681 Reliabel X.6 0.767 Reliabel X.7 0.767 Reliabel X.8 0.761 Reliabel Y.1 0.656 Reliabel Y.2 0.787 Reliabel Y.3 0.811 Reliabel Y.4 0.804 Reliabel Y.5 0.590 Reliabel Y.6 0.663 Reliabel Y.7 0.612 Reliabel Z.1 0.657 Reliabel Z.2 0.615 Reliabel Z.3 0.761 Reliabel Z.4 0.744 Reliabel Z.5 0.807 Reliabel Sumber: Olah Data Peneliti Tabel 5 Nilai Outer Loading Pada Tabel 3 menunjukkan nilai Compo- site reliability. Nilai composite reliability dapat digunakan dalam menguji nilai reliabilitas masing-masing indikator pada suatu variabel. Tabel 3 menunjukkan nilai Composite Reliability.. Pada Tabel 3 diatas, untuk semua variabel bernilai > 0,7 yang artinya model pada penelitian ini sudah reliabel. Selanjutnya, Tabel 4 dibawah ini me- maparkan hasil uji R2 dari kepuasan kerja dan kinerja perawat. Uji koefisien determi- nasi (R2) dilakukan untuk menentukan dan memprediksi seberapa besar atau penting kontribusi pengaruh yang diberikan oleh variabel independen secara bersama–sama terhadap variabel dependen. ANALISIS DAN PEMBAHASAN Tabel 4 Nilai R2 Variabel R2 Keterangan Kepuasan Kerja 0.423 Moderat Kinerja Karyawan 0.581 Moderat Sumber: Olah data peneliti Pada Tabel 4 dapat diketahi bahwa R2 dari kepuasan kerja dan kinerja perawat tergolong moderat. Hal ini mengacu pada Ghozali (2009) pada penelitian Sari & Suhariadi (2019) yang mengklasifikasikan nilai R2 pada klasifikasi berikut: nilai R2 bisa dikategorikan kuat jika nilai R2 ≥ 0.67, moderat ≥0.33, dan lemah ≥ 0.19. Tabel 6 memaparkan hasil pengujian pengaruh langsung. Pengaruh langsung ini merupakan pengujian atas Hipotesis 1, Hipotesis 2, Hipotesis 3. Tabel 6 Hasil Uji Pengaruh Langsung T Statistics P Values Gaya Kepemimpinan Kinerja Perawat 3.279 0.001 Gaya Kepemimpinan Kepuasan Kerja 12.350 0.000 Kepuasan Kerja Kinerja Karyawan 4.470 0.000 Sumber: Olah Data Peneliti Tabel 6 Hasil Uji Pengaruh Langsung Hasil Uji Pengaruh Langsung Selanjutnya pada Tabel 5 berikut ini adalah hasil atas pengujian outer loading. Outer loadings adalah tabel yang berisi loading factor untuk menunjukkan besar korelasi antara indikator dengan variabel laten. Pengujian outer loading pada Tabel 5 menggunakan batas 0,5 yang artinya ketika nilai outer loading diatas 0,5 maka dainggap variabel tersebut reliabel. Hal ini mengacu pada Ghozali (2014:227) yang mengatakan bahwa nilai outer loading 0.5-0.60 masih dapat dipertahankan untuk model yang masih dalam tahap pengembangan. Sehingga, pada penelitian ini variabel yang diuji dianggap telah reliabel. Berdasarkan Tabel 6 diatas dapat dike- tahui bahwa semua p-value yang diperoleh kurang dari 5%, sehingga dapat disimpul- kan bahwa hipotesis: JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 54 1. Gaya kepemimpinan mempengaruhi secara positif signifikan terhadap kinerja karyawan, Diterima. 1. Gaya kepemimpinan mempengaruhi secara positif signifikan terhadap kinerja karyawan, Diterima. secara tidak langsung hubungan antara gaya kepemimpinan terhadap kinerja karyawan. Hal tersebut dapat dilihat dari P Values yang nilainya kurang dari 5%. 2. Gaya kepemimpinan mempengaruhi se- cara positif signifikan terhadap kepuas- an kerja karyawan, Diterima PEMBAHASAN 3. Kepuasan kerja karyawan mempengaru- hi secara positif signifikan terhadap ki- nerja karyawan, Diterima. Gambar 2 merupakan gambar coeffi- cient path dari penelitian ini, dari hasil pengujian yang telah dipaparkan pada bagian sebelumnya, terbukti dengan struk- tural model gambar 2, menunjukkan bah- wasannya: Selanjutnya pada Tabel 7 menunjukkan hasi pengujian tidak langsung/mediasi/ intervening. Gaya Kepemimpinan Berpengaruh Positif Signifikan Terhadap Kinerja Gaya Kepemimpinan Berpengaruh Positif Signifikan Terhadap Kepuasan Kerja Hasil yang ditunjukkan pada penelitian ini membuktikan bahwasannya gaya kepe- mimpinan berpengaruh positif signifikan terhadap kepuasan kerja. Temuan ini men- dukung penelitian sebelumnya oleh (Cahyono, 2019; Sinurat, 2017). Gaya kepe- mimpinan yang diterapkan seoarang leader mampu mempengaruhi kepuasan kerja kar- yawan. Pengaruh yang dihasilkan adalah positif yang berarti semakin efektif gaya kepemimpinan yang diterapkan, maka ke- puasan kerja karyawan juga akan semakin meningkat. Demikian sebaliknya, ketika gaya kepemimpinan yang diterapkan tidak tepat diterapkan pada organisasi tersebut, maka kepuasan kerja karyawan juga akan semakin menurun. Temuan ini menolak penelitian sebe- lumnya oleh Fathurrohman (2015); E. Kusumawati & Ansori (2008); Ni’Am et al., (2021) yang memaparkan tidak adanya pengaruh mediasi kepuasan kerja. Tetapi pada penelitian ini mendukung penelitian Lousyiana & Harlen (2015); Pambudi et al., (2016); Putri (2018); Winarto & Purba (2018) yang membuktikan adanya pengaruh mediasi kepuasan kerja karyawan terhadap hubungan antara gaya kepemimpinan dan kinerja karyawan. Sehingga dapat disimpulkan, gaya ke- pemimpinan yang diterapkan akan lebih berhasil meningkatkan kinerja karyawan ketika diikuti pula oleh peningkatan kepuas- an kerja karyawan. Kepuasan kerja karya- wan merupakan hal yang dapat dipertim- bangkan dalam upaya peningkatan kinerja. Kepuasan Kerja Berpengaruh Positif Signifikan terhadap Kinerja Kepuasan Kerja Berpengaruh Positif Signifikan terhadap Kinerja Berdasarkan analisis data pada peneli- tian ini, dapat diketahui bahwa ada penga- ruh positif dan signifikan kepuasan kerja ter- hadap kinerja karyawan. Hal ini mengkon- firmasi penelitian sebelumnya dimana juga terdapat pengaruh positif signifikan antara kepuasan kerja terhadap kinerja karyawan (Ali & Agustian, 2018a; Indrawati, 2013). Gaya Kepemimpinan Berpengaruh Positif Signifikan Terhadap Kinerja Tabel 7 Hasil Uji Tidak Langsung T Statistics P Values Gaya Kepemimpinan Kepuasan Kerja Kinerja Karyawan 4.115 0.000 Sumber: Olah Data Peneliti Tabel 7 Hasil Uji Tidak Langsung T Statistics P Values Gaya Kepemimpinan Kepuasan Kerja Kinerja Karyawan 4.115 0.000 Sumber: Olah Data Peneliti Tabel 7 Hasil Uji Tidak Langsung Gaya kepemimpinan yang diterapkan di organisasi memiliki pengaruh langsung terhadap kinerja karyawan. Pengaruh terse- but bersifat positif dan mendukung hasil penelitian sebelumnya oleh (Ivana Ariyani, 2016; Yulianto et al., 2018). Berdasarkan teori yang sudah dibahas pada bab sebelumnya, pilihan gaya kepe- mimpinan yang efektif akan sangat ber- pengaruh terhadap kinerja bawahannya. Efektif dalam artian tepat dan sesuai dengan kondisi serta kebutu han organisasi yang dijalankan. Pada Tabel 7 menunjukkan bahwa ter- dapat hubungan positif signifikan variabel kepuasan kerja karyawan mempengaruhi Gambar 2 Path Coeffients Sumber: Olah Data Peneliti Sumber: Olah Data Peneliti Sumber: Olah Data Peneliti Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 55 Kepuasan Kerja Memediasi Hubungan antara Gaya Kepemimpinan Terhadap Kinerja Karyawan Responden dalam penelitian ini menun- jukkan persepsinya terkait gaya kepemim- pinan yang diterapkan oleh atasannya, mampu mempengaruhi output dan kinerja mereka. Sehingga, kinerja mereka juga tergantung pada bagaimana kecermatan dan efisiensi penerapan gaya kepemimpinan atasannya (Ivana Ariyani, 2016). Hasil pada penelitian ini menunjukkan bahwa gaya kepemimpinan yang diterapkan dapat mempengaruhi tinggi rendahnya ki- nerja karyawan melalui kepuasan kerja karyawan. Dalam artian, variabel kepuasan kerja berpengaruh signifikan dalam meme- diasi hubungan antara gaya kepemimpinan dan kinerja karyawan. Gaya Kepemimpinan Berpengaruh Positif Signifikan Terhadap Kepuasan Kerja SIMPULAN DAN SARAN Dalam penelitian ini, peneliti mencoba menguji pengaruh gaya kepemimpinan yang merujuk pada teori Lewin (kepemimpinan demokrasi, autokrasi dan kendali bebas) terhadap pengaruhnya pada kinerja karya- wan dan kepuasan karyawan. Selanjutnya melihat bagaimana kepuasan kerja yang dirasakan karyawan mampu mempengaruhi kinerja, serta pengaruh tidak langsung kepuasan kerja pada hubungan gaya kepemimpinan terhadap kinerja. Hasil yang didapat adalah semua hipotesis diterima. Karyawan yang merasa puas akan pekerjaannya, akan meningkatkan hasil kinerjanya. Begitu pula sebaliknya, ketika karyawan mengalami ketidakpuasan akan kerjanya, kinerjanya akan menurun pula. Hal ini karena seperti halnya diungkapkan oleh Youssef & Luthans (2007) kondisi pikir- an yang nyaman, tidak tertekan, optimis dan merasa mendapat dukungan akan memben- tuk perilaku yang positif di organisasi. Akhirnya, kinerja akan meningkat. Berdasarkan kesimpulan diatas, adapun saran yang diberikan adalah: mengingat gaya kepemimpinan mempengaruhi ke- puasan kerja dan kinerja karyawan, pihak Rumah Sakit diharapkan dapat memper- JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 56 1111/j.1744-6570.1990.tb01561.x tahankan serta meningkatkan unsur gaya kepemimpinan yang sudah diterapkan. Begitu juga terkait kepuasan karyawan, pi- hak Rumah sakit sebaiknya mempertahan- kan dan meningkatkan faktor-faktor yang dapat meningkatkan kepuasan karyawan. Kinerja yang sudah bagus harus senantiasa dipertahankan untuk menjaga kualitas pelayanan dan memberikan kepuasan terhadap pelanggan, sehingga pendapatan dari instansi bisa semakin meningkat Destiasari, S. (2015). Kinerja Pegawai di Kantor Kecamatan Cipocok Jaya Kota Serang. In Jurnal Ilmu Admninistrasi Dan Organisasi. Universitas Sultan Ageng Tirtayasa. Dolphina, E. (2012). Pengaruh Motivasi Kerja, Kepemimpinan Terhadap Kepuasan Kerja Karyawan Serta Dampaknya Pada Kinerja Perusahaan. 2012 (Semantik:Seminar Na- sional Teknologi Informasi & Komuni- kasi Terapan), 1–7. https://doi.org/ 10.31227/osf.io/e6qjd DAFTAR PUSTAKA Edwards, D., & Burnard, P. (2003). A systematic review of stress and stress management interventions for mental health nurses. Journal of Advanced Nursing. Ali, K., & Agustian, D. W. (2018b). Analisis Pengaruh Budaya Organisasi Dan Gaya Kepemimpinan Terhadap Kepuasan Kerja Untuk Meningkatkan Kinerja Kar- yawan Di Rumah Sakit Muhammadiyah Metro. Jurnal Derivatif. Jurnal Ilmiah FE- UMM Lampung, 12(2). Eliyana, A., Ma’arif, S., & Muzakki. (2019). Job satisfaction and organizational commitment effect in the transforma- tional leadership towards employee performance. European Research on Management and Business Economics, 25(3), 144–150. https://doi.org/10.1016 /j.iedeen.2019.05.001 Anastasiou, S., & Papakonstantinou, G. (2014). Factors affecting job satisfaction, stress and work performance of secon- dary education teachers in Epirus, NW Greece. International Journal of Mana- gement in Education, 8(1), 37–53. https:// doi.org/10.1504/IJMIE.2014.058750 Fathurrohman, F. (2015). Pengaruh Kepemim- pinan Transformasional dan Motivasi Kerja Terhadap Kinerja Dengan Kepuasan Kerja Sebagai Intervening (Studi Empiris Pada Disdukcapil Kota Serang). 11(2), 195–206. Bhargavi, D. S., & Yaseen, A. (2016). Leadership Styles and Organizational Performance. Strategic Management Quarterly, 4(1), 87–117. https://doi.org/ 10.15640/smq.v4n1a5 Febriyana, W. (2015). Pengaruh Kepuasan Kerja Terhadap Kinerja Karyawan PT. Kabepe Chakra 2015. E-Proceeding of Management, 2(3), 1–8. Bhatti, N., Maitlo, G. M., Shaikh, N., Hashmi, M. A., & Shaikh, F. M. (2012). The Impact of Autocratic and Democratic Leader- ship Style on Job Satisfaction. Internatio- nal Business Research, 5(2), 192–201. https://doi.org/10.5539/ibr.v5n2p192 Fitrianis, B. F., & Apriliani, R. A. E. (2019). Pengaruh Gaya Kepemimpinan, Budaya Organisasi Dan Motivasi Kerja Terha- dap Kinerja Karyawan Dengan Kepuas- an Kerja Karyawan Sebagai Variabel Intervening (Studi Pada Hotel Kresna Wonosobo). Journal of Economic, Business and Engineering (JEBE), 1(1), 9– 15.https://ojs.unsiq.ac.id/index.php/je be/article/view/870 Cahyono, H. (2019). Pengaruh Gaya Kepe- mimpinan Transformasional Dan Moti- vasi Kerja Terhadap Kinerja Karya-wan Dengan Kepuasan Kerja Sebagai Varia- bel Mediasi Di RSUD Dr Saiful Anwar. Magister Manajemen. http://eprints. umm.ac.id/53002/1/Naskah. pdf Fuadiputra, I. R. (2013). Pengaruh Gaya Kepemimpinan Demokratis Terhadap Kinerja Paramedis Di Rumah Sakit Al- Rohmah. Jimfeb. Jurnal Imiah Mahasiswa p Campbell, J. P., McHenry, J. J., & Wise, L. L. (1990). Modeling Job Performance in a Population of Jobs. Personnel Psychology, 43(2), 313–575. https://doi.org/10. Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 57 Manajemen Rumah Sakit, 5(2), 1–7. https://doi.org/10.18196/jmmr.5118 FEB UB. Malang, 148, 148–162. FEB UB. Malang, 148, 148–162. Ghozali, I. (2009). Analisis Multivariate Lanjutan dengan Program SPSS. Badan Penerbit Universitas Diponegoro. Johnson, S., Cooper, C., Cartwright, S., Donald, I., Taylor, P., & Millet, C. (2005). The experience of work-related stress across occupations. DAFTAR PUSTAKA Journal of Managerial Psychology, 20(2), 178–187. https://doi. org/10.1108/02683940510579803 Ghozali, I. (2014). Structural Equation Modeling. Metode Alternatif dengan Partial Least Squares (PLS). Dilengkapi Software SmartPLS3.0.0 Xlstat 2014 dan WarpPLS4.0. Edisi 4 (4th ed.). Badan Penerbit Universitas Diponegoro. Kusumawati, E., & Ansori, M. (2008). Pengaruh Gaya Kepemimpinan Transak- sional dan Transformasional Terhadap Kinerja Karyawam dengan Mediasi Kepuasan Kerja di Politeknik Negeri Batam. Hamid, A., & Hazriyanto, H. (2019). Pengaruh Kepuasan Kerja Terhadap Kinerja Karyawan PT. Aker Solutions Batam. Jurnal Benefita, 4(2), 326. https://doi.org/10.22216/jbe.v4i2.1877 Kusumawati, R. (2008). Analisis Pengaruh Budaya Organisasi dan Gaya Kepemim- pinan Terhadap Kepuasan Kerja Untuk Meningkatkan Kinerja Karyawan (Studi Kasus Pada RS Roemani Semarang). House, R. J. (1971). A Path Goal Theory of Leader Effectiveness. Administrative Science Quarterly, 16(3), 321–339. Hoyt, C. L., & Price, T. L. (2015). Ethical Decision Making and Leadership: Merging Social Role and Self-Construal Perspectives. Journal of Business Ethics, 126(4), 531–539. https://doi.org/10.- 1007/s10551-013-1974-x Lomanjaya, J., Laudi, M., Widjaja, D. C., & Kartika, E. W. (2013). Analisis Pengaruh Gaya Kepemimpinan Transformasional Dan Transaksional Terhadap Kinerja Karyawan Pt. Iss Indonesia Cabang Surabaya Di Rumah Sakit Katolik St. Vincentius a. Paulo. Journal of Chemical Information and Modeling, 53(9), 1689– 1699. Hoyt, C. L., Price, T. L., & Poatsy, L. (2013). The social role theory of unethical leadership. Leadership Quarterly, 24(5), 712–723. https://doi.org/10.1016/j. leaqua.2013.07.001 Lousyiana, T. T., & Harlen. (2015). Pengaruh Gaya Kepemimpinan Dan Budaya Organisasi Terhadap Kepuasan Kerja dan Kinerja Perawat di Rumah Sakit Islam Ibnu Sina Pekanbaru. Tepak Manajemen Bisnis, 7(3), 509–521. Ibrahim, R., Boerhannoeddin, A., & Kazeem Kayode, B. (2017). Organizational cul- ture and development: Testing the structural path of factors affecting employees’ work performance in an organization. Asia Pacific Management Review, 22(2), 104–111. https://doi.org/ 10.1016/j.apmrv.2016.10.002 j ( ) Luthans, F., & Youssef, C. M. (2007). Emerging positive organizational behavior. Journal of Management, 33(3), 321–349. https://doi.org/10.1177/0149 206 307300814 j p Indrawati, A. D. (2013). Pengaruh Kepuasan Kerja Terhadap Kinerja Karyawan Dan Kepuasan Pelanggan Pada Rumah Sakit Swasta Di Kota Denpasar. Jurnal Manajemen, Strategi Bisnis Dan Kewirau- sahaan, 7(2), 135–142. https:/doi.org/10. 24843/Matrik:Jmbk Mak, B. L., & Sockel, H. (2001). A confirrmatory factor analysis of IS employee motivation and retention Brenda. Elsevier Science B.V, 38. Ni’Am, M. M., Maharani, A., & Fauzi, A. (2021). Dinamika Gaya Kepemimpinan Tangkas, Motivasi, Kinerja Dan Ke- puasan Kerja Di Perusahaan Perke- bunan. JISIP (Jurnal Ilmu Sosial Dan Pendidikan), 5(3), 14–22. https://doi. org/10.36312/jisip.v5i3.2015 Ivana Ariyani, R. (2016). DAFTAR PUSTAKA Pengaruh Gaya Kepemimpinan Dan Loyalitas Karya- wan Terhadap Kinerja Karyawan Di Rumah Sakit Islam Hidayatullah Yogyakarta. Jurnal Medicoeticolegal Dan JIMBis: Jurnal IIlmu Manajemen dan Bisnis – Volume 1, Nomor 1, Mei 2022 : 39 – 59 58 Nurnaningsih, S., & Wahyono, W. (2017). Pengaruh Kepuasan Kerja, Motivasi Kerja Dan Komitmen Organisasi Terhadap Kinerja Melalui Organiza- tional Citizenship Behavior (Ocb) Sebagai Variabel Intervening. Economic Education Analysis Journal, 6(2), 365–378. Raus, A., Haita, M., & Lazǎr, L. (2012). Hierarchy of needs, perception and preference for leadership styles within a police educational institution. Transylvanian Review of Administrative Sciences, 35, 238–255. Robbin, S. P., & Judge, T. A. (2008). Organizational Behavior (13th). Pambudi, didit setyo, Mukzam, D., & Nurtjahjono, gunawan eko. (2016). Pe- ngaruh Gaya Kepemimpinan Transfor- masional Terhadap Kinerja Karyawan Melalui Kepuasan Kerja Karyawan Sebagai Variabel Mediasi. Journal Bisnis Administration, 39(1), 164–171. Said, R. M., & El-Shafei, D. A. (2021). Occupational stress, job satisfaction, and intent to leave: nurses working on front lines during COVID-19 pandemic in Zagazig City, Egypt. Environmental Science and Pollution Research, 28(7), 8791–8801. https://doi.org/10.1007/ s11356-020-11235-8 Parijat, P., & Bagga, S. (2014). Victor Vroom’s Expectancy Theory of Motivation – An Evaluation. International Journal of Business and Management, VII(9), 1–8. Sari, J., & Suhariadi, F. (2019). Kontrak psikologis terhadap commitment to change: Resiliensi akademik sebagai variabel mediasi. Persona:Jurnal Psikologi Indonesia, 8(2), 178–192. https://doi. org/10.30996/persona.v8i2.2535 Pawirosumarto, S., Sarjana, P. K., & Muchtar, M. (2017). Factors affecting employee performance of PT.Kiyokuni Indonesia. International Journal of Law and Management, 59(4), 602–614. https://doi. org/10.1108/IJLMA-03-2016-0031 Saufa, J., & Maryati, T. (2017). Dampak Kepuasan Kerja Dalam Peningkatan Kinerja Perawat Dengan Komitmen Organisasi Sebagai Variabel Intervening ( Studi Pada Perawat Rs Pku Muhammadiyah Gamping). JBTI : Jurnal Bisnis Teori Dan Implementasi, 8(2), 200– 213. https://doi.org/10.18196/bti.82094 Plangiten, P. (2013). Gaya Kepemimpinan Dan Lingkungan Kerja Pengaruhnya Terhadap Kepuasan Kerja Karyawan Pada Pt. Pos Indonesia (Persero) Manado. Jurnal Riset Ekonomi, Manaje- men, Bisnis Dan Akuntansi, 1(4), 2155– 2166. https://doi.org/10.35794/emba. v1i4.3426 Scheidlinger, S. (1994). Brief report: The Lewin, Lippitt and White study of leadership and “social climates” revisited. International Journal of Group Psychotherapy, 44(1), 123–127. https:// doi.org/10.1080/00207284.1994.1149073 7 Pratama, C. Y. (2012). Pengaruh Gaya Kepemimpinan Otokratis Terhadap Kepuasan Kerja. Journal of Social and Industrial Psychology, 2(1), 22–27. Prayatna, A., & Subudi, I. (2016). Pengaruh Gaya Kepemimpinan Terhadap Stres Kerja Dan Kepuasan Kerja Karyawan Pada Fave Hotel Seminyak. E-Jurnal Manajemen Universitas Udayana, 5(2), 252587. Sinurat, E. J. (2017). DAFTAR PUSTAKA Pengaruh Gaya Kepe- mimpinan Terhadap Kepuasan Kerja Karyawan Pada Pt. Himawan Putra Medan. Jurnal Ilmiah Methonomi, 3(2), 92–106. Van Wart, M. (2013). Lessons from leadership theory and the contemporary challenges of leaders. Public Admi- nistration Review, 73(4), 553–565. https://doi.org/10.1111/puar.12069 Putri, W. O. R. (2018). Pengaruh Gaya Kepemimpinan Transformasional Terhadap Kinerja Karyawan Dengan Kepuasan Kerja Sebagai Variabel Intervening (Studi Empi- ris Pada PT Telekomunikasi Indonesia Area Semarang). Universitas Muhammadiyah Magelang. Winarto, & Purba, J. H. (2018). Pengaruh Gaya Kepemimpinan Transformasional Pengaruh Gaya Kepemimpinan terhadap Kinerja...– Sari, Fuadati 59 Terhadap Organizational Citizenship Behavior Dengan Kepuasan Kerja Sebagai Variabel Intervening (Studi Kasus Pada Karyawan Rumah Sakit Swasta Di Kota Medan). Jurnal Ilmiah Methonomi, 4(2), 111–123. https:// methonomi.net/index.php/jimetho/art icle/view/37 Terhadap Organizational Citizenship Behavior Dengan Kepuasan Kerja Sebagai Variabel Intervening (Studi Kasus Pada Karyawan Rumah Sakit Swasta Di Kota Medan). Jurnal Ilmiah Methonomi, 4(2), 111–123. https:// methonomi.net/index.php/jimetho/art icle/view/37 Yudiaatmaja, F. (2013). Kepemimpinan: Konsep, Teori dan Karakternya. Ejournal Undiksha, IV(2), 29–38. http://dx.doi. org/10.1016/j.intman.2016.11.002%0Ah ttps://doi.org/10.1016/j.tele.2017.10.00 7%0Ahttp://ilp.ut.ac.id/index.php/JO M/article/view/432%0Ahttp://dx.doi. org/10.3926/jiem.1530%0Ahttp://dx.d oi.org/10.1016/j.bushor.2017.11.007%0 Ahttps://doi.org/10.10 Yanoto, A. (2018). Pengaruh Gaya Kepemimpinan terhadap Kepuasan Kerja melalui Motivasi dan Kinerja Karyawan PT. Nutrifood Indonesia di Surabaya. Jurnal Agora, 6(1), 1–10. Yulianto, Y., Purwanto, N. H., & Firmansyah, R. R. (2018). Hubungan Gaya Kepemim- pinan Demokratis Dengan Kinerja Perawat Dalam Penatalaksanaan Stan- dar Praktik Profesional Di Ruang Ma- war Merah Kelas II RSUD Asembagus Kabupaten Situnondo. Nurse and Health: Jurnal Keperawatan, 7(1), 13–22. Youssef, C. M., & Luthans, F. (2007). Positive organizational behavior in the workplace: The impact of hope, optimism, and resilience. Journal of Management, 33(5), 774–800. https://doi. org/10.1177/0149206307305562
https://openalex.org/W3106407146	https://www.ams.org/tran/2012-364-05/S0002-9947-2012-05369-4/S0002-9947-2012-05369-4.pdf	English	null	Stable degenerations of surfaces isogenous to a product II	Transactions of the American Mathematical Society	2,012	public-domain	13,136	TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 364, Number 5, May 2012, Pages 2411–2427 S 0002-9947(2012)05369-4 Article electronically published on January 11, 2012 TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 364, Number 5, May 2012, Pages 2411–2427 S 0002-9947(2012)05369-4 Article electronically published on January 11, 2012 Received by the editors February 10, 2010 and, in revised form, March 24, 2010. 2010 Mathematics Subject Classiﬁcation. Primary 14J10, 14B07. This work was completed at Universit¨at Bayreuth under the ﬁnancial support of China Schol- arship Council “High-level university graduate program” and DFG Forschergruppe 790 “Classiﬁ- cation of algebraic surfaces and compact complex manifolds”. WENFEI LIU Abstract. In this paper, we describe the possible singularities on a stable surface which is in the boundary of the moduli space of surfaces isogenous to a product. Then we use the Q-Gorenstein deformation theory to get some connected components of the moduli space of stable surfaces. c ⃝2012 American Mathematical Society Reverts to public domain 28 years from publication This work was completed at Universit¨at Bayreuth under the ﬁnancial support of China Schol- arship Council “High-level university graduate program” and DFG Forschergruppe 790 “Classiﬁ- cation of algebraic surfaces and compact complex manifolds”. Reverts to public domain 28 years from publication Received by the editors February 10, 2010 and, in revised form, March 2010 Mathematics Subject Classiﬁcation. Primary 14J10, 14B07. 010 Mathematics Subject Classiﬁcation. Primary 14J1 1. Preliminaries 1.1. Notation. Let G be a ﬁnite group acting on a set A. For σ ∈G, \|σ\| is the order of σ. For a ∈A, Ga := {g ∈G\|g · a = a} is the stabilizer of a. If Ga ̸= {1}, we say that a is a ﬁxed point of the action. If Ga = {1} for every a ∈A, we say that G acts freely on A. For a ∈A, Ga := {g ∈G\|g · a = a} is the stabilizer of a. If Ga ̸= {1}, we say that a is a ﬁxed point of the action. If Ga = {1} for every a ∈A, we say that G acts freely on A. Zn denotes the cyclic group of order n. We work over the ﬁeld C of complex numbers. Introduction Therefore we get our desired description of the group action on a product of stable curves (Propositions 2.9 and 2.16). The basic tool used is Cartan’s lemma which reduces the action of a stabilizer to a linear one. Having described the singularities on a stable degeneration X (Corollaries 2.11 and 2.18), we show that if the singularities are of a certain type, namely (U1a), (U1b), (U2a) or (U2b), then the Q-Gorenstein deformation of X is unobstructed (Theorem 3.3) and hence the compactiﬁcation of the moduli space considered by van Opstall in fact yields connected components of the moduli space of stable surfaces: Corollary 3.5. Let S = (C ×D)/G be a surface isogenous to a product of unmixed type. Assume that the pair (C, G) is a triangle curve (i.e., C/G ∼= P1, and C → C/G is branched over 3 points). Let M top S be the moduli space of smooth surfaces with the same topological type as S and let M top S be the closure of M top S in the moduli space M st a,b with a = K2 S, b = χ(OS). Then M top S consists of connected components of M st a,b. We should mention that the Q-Gorenstein deformation theory set out by Hacking ([Hac01] and [Hac04]) is indispensable for our purpose. 1.2. Surfaces isogenous to a product. Deﬁnition 1.1 ([Cat00], Deﬁnition 3.1). A smooth projective surface S is isoge- nous to a (higher) product if it is a quotient S = (C×D)/G, where C, D are smooth curves of genus at least two, and G is a ﬁnite group acting freely on C × D. Let S = (C×D)/G be a surface isogenous to a product. Let G◦:= G∩(Aut(C)× Aut(D)); then G◦acts on the two factors C, D and acts on C × D via the diagonal action. We say that S is of unmixed type if G = G◦; otherwise S is said to be of mixed type. By [Cat00, Prop. 3.13], we always assume G◦acts faithfully on C, D. By [Cat00, Prop. 3.16], surfaces S = (C × D)/G isogenous to a product and of mixed type are obtained as follows. There is a (faithful) action of a ﬁnite group G◦ on a curve C of genus at least 2 and a nonsplit extension Let S = (C×D)/G be a surface isogenous to a product. Let G◦:= G∩(Aut(C)× Aut(D)); then G◦acts on the two factors C, D and acts on C × D via the diagonal action. We say that S is of unmixed type if G = G◦; otherwise S is said to be of mixed type. By [Cat00, Prop. 3.13], we always assume G◦acts faithfully on C, D. ( )) action. We say that S is of unmixed type if G = G◦; otherwise S is said to be of mixed type. By [Cat00, Prop. 3.13], we always assume G◦acts faithfully on C, D. By [Cat00, Prop. 3.16], surfaces S = (C × D)/G isogenous to a product and of mixed type are obtained as follows. There is a (faithful) action of a ﬁnite group G◦ on a curve C of genus at least 2 and a nonsplit extension By [Cat00, Prop. 3.16], surfaces S = (C × D)/G isogenous to a product and of mixed type are obtained as follows. There is a (faithful) action of a ﬁnite group G◦ on a curve C of genus at least 2 and a nonsplit extension 1 →G◦→G →Z2 →1, yielding a class [ϕ] in Out(G◦) = Aut(G◦)/Int(G◦), which is of order ≤2. Introduction Stable surfaces were ﬁrst introduced by Koll´ar and Shepherd-Barron in the paper [KoSB88]. These surfaces appear naturally as one-parameter limits of surfaces of general type: if ˜ X →Δ is a one-parameter Q-Gorenstein family whose general ﬁbres ˜ Xt are the canonical models of surfaces of general type and X →Δ is its canonical model, then the central ﬁbre X0 is a stable surface. The singularities on stable surfaces are semi-log-canonical (cf. Deﬁnition 1.4), and a complete classiﬁcation of them has already been given in [KoSB88]. The moduli problem for stable surfaces was solved later through the work of several authors (cf. [Ko90, A94, Hac04, HK04, Ko08, AbH11] and Theorem 1.13). In view of the above general framework, examples are badly needed to illustrate the geometry of the moduli space of stable surfaces. One strategy is to compactify the well-studied moduli spaces of smooth surfaces whose stable degenerations can be described in a satisfactory way. In this direction, van Opstall [vO05, vO06] studied the moduli space of products of curves and the stable degenerations of surfaces isogenous to a product. Recently, Alexeev and Pardini [AP09] gave an explicit compactiﬁcation of moduli spaces of the more complicated Campedelli and Burniat surfaces, and Rollenske [R10] did the same for very simple Galois double Kodaira ﬁbrations. Here we continue the work of van Opstall ([vO06]) on the stable degenerations of surfaces isogenous to a product and give some connected components of the moduli space of stable surfaces. A surface isogenous to a product is the quotient of a product of two smooth curves by a free group action. Van Opstall showed that a stable degeneration of such surfaces is the quotient of a product of two stable curves C × D by a not necessarily free group action (cf. Theorem 1.12). We go one step further to describe explicitly how the group acts on the product of two stable curves. To do this, we ﬁrst establish some results on the smoothing of a stable curve with a group action (cf. Section 2.1). Then we can apply these results 2411 2412 WENFEI LIU to the free smoothing of a product of two stable curves with a group action (see Sections 2.2 and 2.3). Note that X = (C × D)/G is a stable degeneration if and only if (C × D, G) admits a free smoothing (cf. the discussion after Theorem 1.12). 1.2. Surfaces isogenous to a product. Once we ﬁx a representative ϕ of the above class, there exists an element τ ′ in G \ G◦such yielding a class [ϕ] in Out(G◦) = Aut(G◦)/Int(G◦), which is of order ≤2. Once we ﬁx a representative ϕ of the above class, there exists an element τ ′ in G \ G◦such STABLE DEGENERATIONS OF SURFACES STABLE DEGENERATIONS OF SURFACES 2413 2413 that, setting τ = τ ′2, we have: (I) ϕ(γ) = τ ′γτ ′−1, ( ) ( ) ϕ(γ) γ , (II) G acts, under a suitable isomorphism of C and D, by the formulae: γ(P, Q) = (γP, (ϕγ)Q) for γ in G◦, whereas the lateral class of G◦consists of the transformations τ ′γ(P, Q) = ((ϕγ)Q, τγP). Let Γ be the subset of G◦consisting of the transformations having some ﬁxed point. Then the condition that G acts freely amounts to: (A) Γ ∩ϕ(Γ) = {1}. (A) Γ ∩ϕ(Γ) = {1}. (A) Γ ∩ϕ(Γ) = {1}. (B) There is no γ in G◦such that ϕ(γ)τγ is in Γ. The moduli space of surfaces isogenous to a product is partly illustrated in the following abridged theorem: Theorem 1.2 ([Cat03]). Let S be a surface isogenous to a product. Then the moduli space M top S of surfaces with the same topological type as S is either irreducible and connected or it contains two connected components which are interchanged by complex conjugation. Theorem 1.2 ([Cat03]). Let S be a surface isogenous to a product. Then the moduli space M top S of surfaces with the same topological type as S is either irreducible and connected or it contains two connected components which are interchanged by complex conjugation. 1.3. Cartan’s lemma. The following lemma is used throughout Section 2 for the (analytically) local analysis of the group actions. Lemma 1.3 (Cartan’s lemma). Let z ∈Z be an analytic singularity with Zariski tangent space T and let G be a ﬁnite group of automorphisms of z ∈Z. Then there exists a G-equivariant embedding z ∈Z →T ∋0. 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. Deﬁnition 1.4. A (reduced) surface X is said to have semi-log-canonical (slc) singularities if Deﬁnition 1.4. A (reduced) surface X is said to have semi-log-canonical (slc) singularities if (i) X is S2; ( ) (ii) X has at most double normal crossing singularities in codimension 1; (iii) th W il di i l K i Q C ti (iii) the Weil divisor class KX is Q-Cartier; (iv) if ˜X →X is the normalization and ˜D ⊂˜X is the preimage of the codimen- sion-1 part of Xsing, then the pair ( ˜X, ˜D) is log-canonical; i.e., for any resolution μ: ˆX →˜X, we have K ˆ X + μ−1 ∗ ˜D = μ∗(K ˜ X + ˜D) + aiEi with all ai ≥−1. with all ai ≥−1. with all ai ≥−1. A stable surface is a projective slc surface with an ample dualizing sheaf. Let F be a coherent sheaf. For n ∈Z, we deﬁne the n-th reﬂexive power of F by F[n] := (F⊗n)∗∗, the double dual of the n-th tensor product. Let R be the category of Noetherian local C-algebras with residue ﬁeld C. Deﬁnition 1.5. Let B be a Noetherian scheme over C. A ﬂat projective morphism f : X →B is called a Q-Gorenstein family of stable surfaces if for every closed point t ∈B, the ﬁbre Xt is a stable surface and for every n ∈Z, ω[n] X/B commutes with any base change. WENFEI LIU 2414 Let R ∈R. A Q-Gorenstein deformation of X over R is a Q-Gorenstein family f : X →Spec R together with an isomorphism X ⊗R k(R) ∼= X. Isomorphisms between two Q-Gorenstein deformations are deﬁned in an obvious way. Remark 1.6. The hypothesis that ω[n] X/B commutes with any base change for any n Remark 1.6. The hypothesis that ω[n] X/B commutes with any base change for any n [N] is called Koll´ar’s condition and it implies that ω[N] X/B is invertible for some N. is called Koll´ar’s condition and it implies that ω[N] X/B is invertible for some N. For a stable surface X, we deﬁne a functor Def QG X : R →(Sets) as follows: for any R ∈R, For a stable surface X, we deﬁne a functor Def QG X : R →(Sets) as follows: for any R ∈R, Def QG X (R) = {Q-Gorenstein deformations of X over R}/ ≃. 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. The Q-Gorenstein deformations of X are exactly those deformations which, locally at each point P ∈X, are induced by equivariant deformations of the canonical cov- ering of P ∈X ([Hac01, Prop. 10.13]). If X is Gorenstein, then the Q-Gorenstein deformation theory of X coincides with the ordinary one. Remark 1.7. If a semiuniversal Q-Gorenstein deformation exists, we denote the base by DefQG X . When we refer to the ordinary deformation of X, the semiuniversal base is denoted by DefX. Set T 1 QG,X = Def QG X (C[ϵ]/ϵ2), the space of ﬁrst-order Q-Gorenstein deforma- tions. Let T 1 QG,X be the sheaf associated to the presheaf deﬁned by U →Def QG U (C[ϵ]/ϵ2), for any open subset U of X. Then we have a local-to-global sequence: 0 →H1(TX) →T 1 QG,X →H0(T 1 QG,X) →H2(TX), where TX = HomOX(ΩX, OX) is the tangent sheaf of X ([Hac04, p. 227]). Note that T 1 QG,X is considered as the tangent space to the functor Def QG X . where TX = HomOX(ΩX, OX) is the tangent sheaf of X ([Hac04, p. 227]). Note that T 1 QG,X is considered as the tangent space to the functor Def QG X . 1.5. Stable surfaces in this paper. We are mostly interested in stable surfaces that are quotients of a product of two stable curves. We recall the deﬁnition of a stable curve ﬁrst ([DM69]): Deﬁnition 1.8. Let g ≥2 be an integer. A stable curve of genus g is a reduced, connected, 1-dimensional scheme C over C such that: Deﬁnition 1.8. Let g ≥2 be an integer. A stable curve of genus g is a reduced, connected, 1-dimensional scheme C over C such that: (i) C has only ordinary double points as singularities; (ii) if E is a nonsingular rational component of C, then E meets the other components of C in more than 2 points; (iii) dim H1(OC) = g. (i) C has only ordinary double points as singularities; (ii) if E is a nonsingular rational component of C, then E meets the other components of C in more than 2 points; (iii) dim H1(OC) = g. (i) C has only ordinary double points as singularities; (ii) if E is a nonsingular rational component of C, then E meets the other components of C in more than 2 points; 1 Proposition 1.9. Let C, D be stable curves. 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. Then C × D is a stable surface. Proof. See [vO05, Prop. 3.1]. □ Proposition 1.9. Let C, D be stable curves. Then C × D is a stable surface. Proof. See [vO05, Prop. 3.1]. □ Proposition 1.9. Let C, D be stable curves. Then C × D is a stable surface. Proof. See [vO05, Prop. 3.1]. □ Proposition 1.10. Let Z be a stable surface and G a ﬁnite group acting on Z with ﬁnitely many ﬁxed points. Then Z/G is also a stable surface. Proof. See the proof of [vO06, Theorem 3.1]. □ Corollary 1.11. Let C, D be stable curves. Let Z := C × D and G be a group acting on Z with ﬁnitely many ﬁxed points. Then (C × D)/G is a stable surface. □ Proposition 1.10. Let Z be a stable surface and G a ﬁnite group acting on Z with ﬁnitely many ﬁxed points. Then Z/G is also a stable surface. Proposition 1.10. Let Z be a stable surface and G a ﬁnite group acting on Z with ﬁnitely many ﬁxed points. Then Z/G is also a stable surface. Proof. See the proof of [vO06, Theorem 3.1]. □ Proof. See the proof of [vO06, Theorem 3.1]. □ Proof. See the proof of [vO06, Theorem 3.1]. Proof. See the proof of [vO06, Theorem 3.1]. Proof. See the proof of [vO06, Theorem 3.1]. Corollary 1.11. Let C, D be stable curves. Let Z := C × D and G be a group acting on Z with ﬁnitely many ﬁxed points. Then (C × D)/G is a stable surface. STABLE DEGENERATIONS OF SURFACES 2415 The stable degenerations of surfaces isogenous to a product are of the above form, as we will see from the following: Theorem 1.12 ([vO06], Theorem 3.1). Suppose X ∗→Δ∗is a family of surfaces isogenous to a product over a punctured disk. Then, possibly after a ﬁnite change of base, totally ramiﬁed over the origin, X can be completed to a family of stable surfaces over the disk whose central ﬁbre is a quotient of a product of stable curves. According to the proof of the above theorem in [vO06], we give an explicit construction of the stable degenerations here. There are two cases: Unmixed case: We have, up to ﬁnite base change, G-equivariant smoothings of stable curves (cf. 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. Section 2.1) C →Δ and D →Δ such that the completion X →Δ of X ∗→Δ∗is (C ×Δ D)/G →Δ. In particular, the central ﬁbre is (C0 × D0)/G, where G acts faithfully on C0, D0 and acts diagonally on C0 × D0. However, the action of G on C0 × D0 is not necessarily free. Mixed case: There exists a ﬁnite group G◦, a G◦-equivariant smoothing C →Δ of stable curves and a nonsplit extension 1 →G◦→G →Z2 →1 yielding an automorphism ϕ of G◦, such that the pairs (Ct, G) with t ̸= 0 satisfy properties (I), (II), (A), (B) in Section 1.2. On the central ﬁbre C0, we still have a G-action that enjoys properties (I), (II), but not necessarily (A), (B), i.e., the action of G on C0 × C0 is not necessarily free. Now the completion X →Δ is (C ×Δ C)/G →Δ, where the action of G◦on the second factor is twisted by ϕ. In both cases, the stable degeneration X0 is of the form (C × D)/G, where C, D are stable curves and G acts on C×D with ﬁnitely many ﬁxed points. Tautologically the pair (C ×D, G) admits a free smoothing, i.e., a one-parameter family C ×Δ D → Δ such that the following hold: (i) C0 × D0 ∼= C × D; (ii) the ﬁbre Ct × Dt over t ̸= 0 is smooth; ( ) ̸ (iii) G acts on C ×Δ D preserving the ﬁbres, and the action of G on the central ﬁbre coincides with the given action of G on C × D; iii) G acts on C ×Δ D preserving the ﬁbres, a (iv) G acts freely on the general ﬁbres Ct × Dt for t ̸= 0. 1.6. The moduli space of stable surfaces. Let (Sch)/C be the category of Noetherian C-schemes. We deﬁne the moduli functor of stable surfaces: for any B ∈(Sch)/C, Mst a,b(B) = {X/B \| X/B is a Q-Gorenstein family of stable surfaces over B and for any closed point t ∈B, K2 Xt = a, χ(OXt) = b}/ ∼= . Mst a,b(B) = {X/B \| X/B is a Q-Gorenstein family of stable surfaces over B and for any closed point t ∈B, K2 Xt = a, χ(OXt) = b}/ ∼= . 2. Stable degenerations of surfaces isogenous to a product In this section, we will give a precise description of possible singularities on stable degenerations of surfaces isogenous to a product. This is a careful improvement of Theorem 1.12, which allows one to further study the Q-Gorenstein deformations of the stable degenerations. 2.1. Smoothings of stable curves with group actions. Our surfaces can be constructed by taking ﬁnite quotients of products of two stable curves, so their geometry is closely related to that of stable curves. In this section, we will establish some facts about smoothings of stable curves with group actions. More precisely, in the case when the group action admits a smoothing, we will show what the stabilizers on the central ﬁbre can be and how they act locally analytically. These facts are used in Sections 2.2 and 2.3 for the smoothing of a product of stable curves with a group action. Deﬁnition 2.1. Let G be a ﬁnite group acting faithfully on a stable curve C. A smoothing of the pair (C, G) is a (ﬂat) family of stable curves C →Δ over the unit disk such that Deﬁnition 2.1. Let G be a ﬁnite group acting faithfully on a stable curve C. A smoothing of the pair (C, G) is a (ﬂat) family of stable curves C →Δ over the unit disk such that (i) the central ﬁbre C0 is isomorphic to C; (i) the central ﬁbre C0 is isomorphic to C; (ii) the ﬁbre Ct of the family over t ̸= 0 is smooth; (iii) G acts on C preserving the ﬁbres, and the action on C0 coincides with the given one on C under the isomorphism of (i). Remark 2.2. We also call C →Δ a G-equivariant smoothing of C. Now let (C, G) be as in the deﬁnition and assume that C →Δ is a smoothing of (C, G). Lemma 2.3. There are only ﬁnitely many points on C having nontrivial stabilizers, or, equivalently, there are only ﬁnitely many ﬁxed points for the G-action. Proof. Otherwise there is a τ ̸= 1 ∈G acting as the identity on some irreducible component D of C. Pick a smooth point P of D. Then C is smooth around P and we can take a local coordinate z of D and a local coordinate t of Δ such that (z, t) form local coordinates of C around P and τ acts as (z, t) →(z, t). 1.4. Q-Gorenstein deformation theory of stable surfaces. We recall the Q- Gorenstein deformation theory of stable surfaces set out by Hacking. Mst a,b(B) = {X/B \| X/B is a Q-Gorenstein family of stable surfaces over B and for any closed point t ∈B, K2 Xt = a, χ(OXt) = b}/ ∼= . Theorem 1.13. There is a projective coarse moduli space M st a,b for Mst a,b. Theorem 1.13. There is a projective coarse moduli space M st a,b for Mst a,b. Proof. See [Kov09, Section 7.C] and [Ko08]. □ □ □ Proof. See [Kov09, Section 7.C] and [Ko08]. We shall get some connected components of this moduli space by studying the Q- Gorenstein deformations of stable degenerations of surfaces isogenous to a product in Section 3. 2416 WENFEI LIU Lemma 2.6. If P ∈C is a node, then GP is either cyclic or dihedral. 2. Stable degenerations of surfaces isogenous to a product So τ = 1 ∈G is the identity, a contradiction. □ Lemma 2.4. If P ∈C is a smooth point, then GP is cyclic. Lemma 2.4. If P ∈C is a smooth point, then GP is cyclic. Proof. There is an embedding GP →GL(TP C), where TP C ∼= C is the tangent space of C at P. Since TP C is a 1-dimensional vector space, GL(TP C) ∼= C∗. So GP , being a ﬁnite subgroup of C∗, is cyclic. □ Lemma 2.5. If P ∈C is a smooth point, and 1 ̸= τ ∈GP , then τ also ﬁxes points on Ct, for t ̸= 0. Proof. As in the proof of Lemma 2.3, we can ﬁnd local coordinates (z, t) for C around P such that τ acts as (z, t) →(ξz, t), where ξ ∈C∗is a primitive \|τ\|-th root of unity. So τ ﬁxes (0, t) ∈Ct, for t ̸= 0. □ STABLE DEGENERATIONS OF SURFACES 2417 Proof. The germ of C around P can be seen as a deformation of a node. We can ﬁnd an embedding of the germ into (C3, 0) such that the equation of the germ is xy −tk = 0, k ≥1. In fact, let {xy −s = 0} → Δ (x, y, s) → s be the semiuniversal family of a node. Then locally around P, C →Δ is just the pull-back by Δ → Δ t → s = tk, By Cartan’s lemma, we can assume that the action of GP is given by τ : (x, y, t) →(a1x + a2y, b1x + b2y, t) for any τ ∈GP . Since GP acts on the central ﬁbre C0 : xy = 0, we have a1 a2 b1 b2 = ξ 0 0 ξ−1 or 0 η η−1 0 , where ξ is a primitive \|τ\|-th root of unity and η is some nonzero number. Let π : GP →Z2 be the determinant homomorphism: for any τ ∈GP , where ξ is a primitive \|τ\|-th root of unity and η is some nonzero number. Let π : GP →Z2 be the determinant homomorphism: for any τ ∈GP , π(τ) := det(τ) = 1, if τ does not interchange the branches at P, −1, if τ interchanges the branches at P. Then the kernel H of π consists of τ ∈GP whose action is given by ξ 0 0 ξ−1 . Note that H embeds into C∗: H C∗ H → C∗ τ → ξ. If im(π) = {1}, then GP = H is cyclic. 2418 On the o So τ(x, y, t) consideratio Now we c Theorem 2 we can ﬁnd τ ∈GP , the (i) (x, y (ii) (x, y Proof. The For the “ Step 1. The U ⊂C be a We will show family U : ( τ ∈GP , and it is eas Note that tangent spac is smoothab Step 2. We 0 → to prove that To do this, ﬁ (2.1) where, for a any τ ∈G, summand is Let n(P) G/GP . The acts on the v 2418 WENFEI LIU On the other hand, if π(τ) = −1, then τ(x, y, t) = (ηy, η−1x, t) with η ∈C∗. So τ(x, y, t) = (x, y, t) if and only if ηy = x. Taking the equation xy = tk into consideration, τ ﬁxes 2 points: (η tk η , t) and (−η tk η , t) on Ct, for t ̸= 0. □ Now we can state our main theorem in this section: Theorem 2.8. A pair (C,G) admits a smoothing if and only if for any node P ∈C, we can ﬁnd a local (analytic) embedding of C : (xy = 0) ⊂C2 such that, for any τ ∈GP , the action of τ is given by either (i) (x, y) →(ξx, ξ−1y), where ξ is a \|τ\|-th root of unity; or (ii) (x, y) →(ηy, η−1x), where η ∈C∗is a nonzero number. (i) (x, y) →(ξx, ξ−1y), where ξ is a \|τ\|-th root of unity; or (ii) (x, y) →(ηy, η−1x), where η ∈C∗is a nonzero number. (ii) (x, y) →(ηy, η−1x), where η ∈C∗is a nonzero number. Proof. The “only if” part is shown in the proof of Lemma 2.6. Proof. The “only if” part is shown in the proof of Lemma 2.6. For the “if” part, we divide the proof into two steps. For the “if” part, we divide the proof into two steps. Step 1. The germ P ∈C has a local G-equivariant smoothing. More precisely, let U ⊂C be a neighborhood around P deﬁned by xy = 0 ⊂C2 as in the hypothesis. We will show that the pair (U, GP ) is GP -smoothable. In fact we can consider the family U : (xy −s = 0) ⊂C2 × Δ →Δ with s ∈Δ as the parameter. For any τ ∈GP , τ(x, y) = (ξx, ξ−1y) or (ηy, η−1x), and it is easily seen that the action of GP on U extends to the f and it is easily seen that the action of GP on U extends to the family U →Δ. Note that U →Δ is the semiuniversal deformation of the node P ∈U and the and it is easily seen that the action of GP on U extends to the family U →Δ. and it is easily seen that the action of GP on U extends to the family U →Δ. Lemma 2.4. If P ∈C is a smooth point, then GP is cyclic. If im(π) = Z2, then there exists τ ∈GP such that τ(x, y, t) = (ηy, η−1x, t), τ(x, y, t) = (ηy, η−1x, t), and it is easy to see in this case that GP = a dihedral group, if \|H\| ≥2; Z2, if \|H\| = 1. □ Lemma 2.7. Let P ∈C be a node. Suppose 1 ̸= τ ∈GP : then τ ﬁxes points on Ct for t ̸= 0 if and only if π(τ) = −1, where π: GP →Z2 is as in the proof of the previous lemma. Proof. We adopt the notation in Lemma 2.6, so the germ of C around P is deﬁned by xy −tk = 0, k ≥1 and the action of GP is linear. y y , Let 1 ̸= τ ∈GP . If π(τ) = 1, then τ(x, y, t) = (ξx, ξ−1y, t), where ξ ∈C∗is a primitive \|τ\|-th root of unity. Suppose (x, y, t) ∈C is a ﬁxed point of τ. Then τ(x, y, t) = (ξx, ξ−1y, t) = (x, y, t) ⇒x = y = 0, and xy = tk implies t = 0. So τ does not ﬁx any point on Ct for t ̸= 0. and xy = tk implies t = 0. So τ does not ﬁx any point on Ct for t ̸= 0. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. ( ,Q) (U2b) If τ does interchange the branches of C or D, say C, then the order of τ is 2 and we can choose local embeddings C : (xy = 0) ⊂C2 and D : (zw = 0) ⊂C2 such that τ(x, y, z, w) = (y, x, −z, −w). ( ) ( ) ( ) (U2c) If τ ̸= 1 and it does not interchange any branches of C, D, then we can choose local embeddings C : (xy = 0) ⊂C2 and D : (zw = 0) ⊂C2 such that τ(x, y, z, w) = (ξx, ξ−1y, ξqz, ξ−qw), where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. Proof. For the “ ⇒” direction, suppose Z = C ×Δ D →Δ is a free smoothing of (C × D, G). Let (P, Q) be any point on C × D. Note that G(P,Q) = GP ∩GQ. We divide our further discussion into 3 cases: Proof. For the “ ⇒” direction, suppose Z = C ×Δ D →Δ is a free smoothing of (C × D, G). Let (P, Q) be any point on C × D. Note that G(P,Q) = GP ∩GQ. We divide our further discussion into 3 cases: Case (U0): P, Q are both smooth points on C, D respectively. We will show in this case that G(P,Q) = {1}. Suppose 1 ̸= τ ∈G(P,Q): then P, Q are both ﬁxed points of τ. Since P, Q are both smooth points on C, D respectively, τ ﬁxes points on Ct, Dt for t ̸= 0 by Lemma 2.5. So τ ﬁxes points on Ct × Dt for t ̸= 0, which contradicts the assumption that G acts freely on Ct × Dt for t ̸= 0. ̸ ̸ Case (U1): One of P, Q, say P, is a node and the other is a smooth point. ( ) Suppose G(P,Q) ̸= {1}. By Lemma 2.4, GQ is cyclic and hence its subgroup G(P,Q) is also cyclic. Let G(P,Q) = ⟨τ⟩, τ ̸= 1. By Lemma 2.5, τ ﬁxes points of Dt for t ̸= 0. Since G acts freely on Ct × Dt for t ̸= 0, τ does not ﬁx any point of Ct, t ̸= 0. By Lemma 2.7, τ does not interchange the branches of C around P. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. Proposition 2.9 (Criterion for free smoothings in the unmixed case). Let C, D be two stable curves and let G be a ﬁnite group acting on C and D. Let G act on C × D diagonally. Then the pair (C × D, G) admits a free smoothing if and only if for any (P, Q) ∈C × D, we have one of the following: (U0) if both P, Q are smooth points on C, D respectively, then the stabilizer G(P,Q) = {1}. (U0) if both P, Q are smooth points on C, D respectively, then the stabilizer G(P,Q) = {1}. (U1) if one of P, Q, say P, is a node and the other is a smooth point, then G(P,Q) = ⟨τ⟩is cyclic, and we can ﬁnd a local embedding of C : (xy = 0) ⊂ C2 as well as a local coordinate z of D such that τ(x, y, z) = (ξx, ξ−1y, ξqz), where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. (U1) if one of P, Q, say P, is a node and the other is a smooth point, then G(P,Q) = ⟨τ⟩is cyclic, and we can ﬁnd a local embedding of C : (xy = 0) ⊂ C2 as well as a local coordinate z of D such that τ(x, y, z) = (ξx, ξ−1y, ξqz), where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. (U2) if both P, Q are nodes of respective curves, then G(P,Q) = ⟨τ⟩is cyclic and τ interchanges the branches of at most one of C and D. In this case, we have one of the following: (U2) if both P, Q are nodes of respective curves, then G(P,Q) = ⟨τ⟩is cyclic and τ interchanges the branches of at most one of C and D. In this case, we have one of the following: f f (U2a) G(P,Q) = {1}. 2418 On the o So τ(x, y, t) consideratio Now we c Theorem 2 we can ﬁnd τ ∈GP , the (i) (x, y (ii) (x, y Proof. The For the “ Step 1. The U ⊂C be a We will show family U : ( τ ∈GP , and it is eas Note that tangent spac is smoothab Step 2. We 0 → to prove that To do this, ﬁ (2.1) where, for a any τ ∈G, summand is Let n(P) G/GP . The acts on the v Note that U →Δ is the semiuniversal deformation of the node P ∈U and the tangent space of the base space at 0 is Ext1 OU (ΩU, OU) ∼= C. The fact that (U, GP) and it is easily seen that the action of GP on U extends to the family U →Δ. Note that U →Δ is the semiuniversal deformation of the node P ∈U and the tangent space of the base space at 0 is Ext1 OU (ΩU, OU) ∼= C. The fact that (U, GP) is smoothable means exactly that Ext1 OU (ΩU, OU) is GP -invariant. Step 2. We will use the local-to-global exact sequence 0 →H1(C, TC) →Ext1 OC(ΩC, OC) π−→H0(C, Ext1 OC(ΩC, OC)) →0 to prove that local smoothings of nodes with stabilizers lift to a smoothing of (C, G). To do this, ﬁrst note that (2.1) H0(C, Ext1 OC(ΩC, OC)) = P node Ext1 OC,P (ΩC,P , OC,P ), (2.1) where, for any coherent sheaf F on C, FP denotes the stalk of F at P. For any τ ∈G, τ acts on H0(C, Ext1 OC(ΩC, OC)) and maps the Ext1 OC,P (ΩC,P , OC,P ) summand isomorphically to the Ext1 OC,τ(P )(ΩC,τ(P ), OC,τ(P )) summand. Let n(P) := \|G/GP \| and τ1, . . . , τn(P ) ∈G be representatives of elements of G/GP . Then τ1(P), . . . , τn(P )(P) is the orbit of P under the action of G, and G acts on the vector space VP := n(P ) j=1 Ext1 OC,τj (P )(ΩC,τj(P ), OC,τj(P )). The invariant subspace V G P is 1-dimensional, spanned by (τ1(σ), . . . , τn(P )(σ)), The invariant subspace V G P is 1-dimensional, spanned by (τ1(σ), . . . , τn(P )(σ)), where σ is an element spanning Ext1 OC,P (ΩC,P , OC,P ) ∼= C. In view of (2.1), the dimension of H0(C, Ext1 OC(Ω1 C, OC)) is exactly the number of node orbits under the action of G. Taking the G-invariants of the local-to-global sequence, we get 0 →H1(C, TC)G →Ext1 OC(ΩC, OC)G π−→H0(C, Ext1 OC(ΩC, OC))G →0. 0 →H1(C, TC)G →Ext1 OC(ΩC, OC)G π−→H0(C, Ext1 OC(ΩC, OC))G →0 STABLE DEGENERATIONS OF SURFACES 2419 In particular, there exists λ ∈Ext1 OC(ΩC, OC)G such that the π(λ)’s P-summand is nonzero for any node P ∈C. Then λ gives a smoothing of (C, G). 2418 On the o So τ(x, y, t) consideratio Now we c Theorem 2 we can ﬁnd τ ∈GP , the (i) (x, y (ii) (x, y Proof. The For the “ Step 1. The U ⊂C be a We will show family U : ( τ ∈GP , and it is eas Note that tangent spac is smoothab Step 2. We 0 → to prove that To do this, ﬁ (2.1) where, for a any τ ∈G, summand is Let n(P) G/GP . The acts on the v □ 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. Hence there are a local embedding C : (xy = 0) ⊂C2 around P and a local coordinate z of D around Q such that the action of τ on C × D is (x, y, z) →(ξx, ξ−1y, ξqz), where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. 2420 WENFEI LIU Case (U2): P, Q are both nodes on C, D. Suppose G(P,Q) ̸= {1}. By Lemma 2.6, we have that GP and GQ are either cyclic or dihedral. This implies that G(P,Q) = GP ∩GQ is either cyclic or dihedral. Suppose G(P,Q) is dihedral. Then GP and GQ are both dihedral and, by the proof of Lemma 2.6, there is τ1 ∈G(P,Q) (resp. τ2 ∈G(P,Q)) such that τ1 (resp. τ2) interchanges the branches of C at P (resp. the branches of D at Q). By Lemma 2.7, τ1 (resp. τ2) ﬁxes points of Ct (resp. Dt) for t ̸= 0. Since neither τ1 nor τ2 ﬁxes points on Ct × Dt, τ1 (resp. τ2) does not ﬁx points on Dt (resp. Ct). Again by Lemma 2.7, τ1 (resp. τ2) does not interchange the branches of D at Q (resp. the branches of C at P). Now set τ := τ1τ2. Then τ interchanges the branches of C as well as those of D. This implies that τ ﬁxes points on Ct × Dt for t ̸= 0, a contradiction. So G(P,Q) is cyclic and we can assume that G(P,Q) = ⟨τ⟩. Since τ does not ﬁx any point on Ct ×Dt for t ̸= 0, τ interchanges the branches of at most one of C and D. If τ does interchange the branches of one of C and D, say C, then the order of τ is 2 (Lemma 2.6) and we can choose local embeddings C : (xy = 0) ⊂C2 and D: (zw = 0) ⊂C2 such that τ acts as (x, y, z, w) →(y, x, −z, −w). 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. STABLE DEGENERATIONS OF SURFACES 2421 (U2b) A Z2-quotient of the degenerate cusp in (U2a) under the group action: (x, y, z, w) →(y, x, −z, −w). In this case, the index of the singularity is 2 and the canonical covering is the degenerate cusp (U2a). In this case, the index of the singularity is 2 and the canonical covering is the degenerate cusp (U2a). g p ( ) (U2c) Other quotients of the degenerate cusp (U2a) under the group action: (x, y, z, w) →(ξx, ξ−1y, ξqz, ξ−qw), where ξ is a primitive n-th root of unity, (q, n) = 1. In this case, the singularity is still a (Gorenstein) degenerate cusp. We give some examples of the singularities in Corollary 2.11. Example 2.12. Let G = ⟨σ⟩∼= Z2. Let C, D′ be two hyperelliptic curves. Suppose σ acts on C and D′ as the respective hyperelliptic involutions. Let {Q′ 1, . . . , Q′ 2k} be the ﬁxed points of σ on D′. We obtain a stable curve D from D′ by identifying Q′ 2i−1 and Q′ 2i for any 1 ≤i ≤k. Note that σ also acts on D. Let G act on C × D diagonally. Then the quotient (C × D)/G has singularities of type (U1a) or (U1b). Example 2.13. Let C, D be two stable curves. Let G be a ﬁnite group acting freely on C × D. Then (C × D)/G has singularities of type (U1a) or (U2a). Example 2.13. Let C, D be two stable curves. Let G be a ﬁnite group acting freely on C × D. Then (C × D)/G has singularities of type (U1a) or (U2a). Example 2.14. Let G = ⟨σ⟩∼= Z2. Let C′ and D′ be two smooth curves of genus ≥1 such that G acts (faithfully) on both. Assume σ ﬁxes 2k points P ′ 1, P ′ 2, . . . , P ′ 2k on C′. Let C be the stable curve obtained by identifying P ′ 2i−1 and P ′ 2i for 1 ≤i ≤k. Assume σ acts freely on D′. Pick a point Q′ ∈D′. Let D be the stable curve obtained by identifying Q′ and σ(Q′). Note that G acts on C and D. Let G act on C × D diagonally. Then (C × D)/G only has singularities of type (U1a) or (U2b). Example 2.15. Let G = ⟨σ⟩∼= Z2. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. If τ does not interchange any branches of C, D, then we can choose local embed- dings C : (xy = 0) ⊂C2, D : (zw = 0) ⊂C2 such that τ(x, y, z, w) = (ξx, ξ−1y, ξqz, ξ−qw), where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. where ξ is a primitive root of unity of order \|τ\| and (q, \|τ\|) = 1. For the “ ⇐” direction, note that C and D admit G-equivariant smoothings C →Δ, D →Δ by Theorem 2.8. In each of the cases (U0), (U1), (U2), any non- trivial element τ k ∈G(P,Q) = ⟨τ⟩interchanges at most the local branches of one of the factors. This guarantees that τ k acts locally freely on at least one of the factors of Ct × Dt for t ̸= 0 (Lemma 2.7). So Z := C ×Δ D →Δ is a required free smoothing. □ Remark 2.10. In the unmixed case, G(P,Q) is always cyclic. Remark 2.10. In the unmixed case, G(P,Q) is always cyclic. According to Theorem 1.12 and the discussion thereafter, a surface X is a stable degeneration of surfaces isogenous to a product of unmixed type if and only if X = (C × D)/G, where C, D are stable curves and G is a ﬁnite group acting on C, D and acting diagonally on C ×D such that (C ×D, G) admits a free smoothing. So we have Corollary 2.11. The possible singularities of a stable degeneration X of surfaces isogenous to a product of unmixed type are as follows: Corollary 2.11. The possible singularities of a stable degeneration X of surfaces isogenous to a product of unmixed type are as follows: (U1a) Double normal crossing singularities: (xy = 0) ⊂C3. These are the general singularities of X. (U1b) Quotients of the above singularity under the group action: (x, y, z) →(ξx, ξ−1y, ξqz), where ξ is a primitive n-th root of unity, (q, n) = 1. In this case, the index of the singularity is n and the canonical covering is the singularity (U1a). (U2a) The degenerate cusp: (xy = 0, zw = 0) ⊂C4. (U2a) The degenerate cusp: (xy = 0, zw = 0) ⊂C4. (U2a) The degenerate cusp: (xy = 0, zw = 0) ⊂C4. (I) ϕ(γ) = τ ′γτ ′−1, ( ) 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. ( ) ( ) { } (B) There is no γ in G◦such that ϕ(γ)τγ is in Γ. In particular, ϕ(γ)τγ ̸= 1. (B) There is no γ in G◦such that ϕ(γ)τγ is in Γ. In particular, ϕ(γ)τγ ̸= 1. The above two conditions simply say that G acts freely on C1t × C2t for t ̸= 0. Suppose there are inﬁnitely many ﬁxed points on C × C. Then some 1 ̸= σ ∈G ﬁxes inﬁnitely many points and we have σ ∈G \ G◦by Lemma 2.3. Hence σ = τ ′γ ∈G \ G◦for some γ ∈G◦. Since σ2 ∈G◦also ﬁxes inﬁnitely many points, we have ϕ(γ)τγ = σ2 = 1 by Lemma 2.3 again, which contradicts (B) above. This proves (ii). p ( ) For (iii), we discuss the possible stabilizer of a point (P, Q) ∈C × C in the following 2 cases. g Case (M0): P, Q are both smooth points of C. We will show that G(P,Q) = {1} in this case. Since G◦acts freely on C1t × C2t for t ̸= 0, note that G◦∩G(P,Q) = {1} by the claim for the unmixed (U0) case. Suppose on the contrary that there is a 1 ̸= τ1 ∈G(P,Q). Then τ1 = τ ′γ ∈G \ G◦ for some γ ∈G◦. Now τ 2 1 ∈G◦∩G(P,Q) implies that τ 2 1 = 1 and hence τ ′γτ ′γ = 1 ⇒(τ ′γτ ′−1)τ ′2γ = 1 ⇒ϕ(γ)τγ = 1. This contradicts (B) above. Case (M1): One of P, Q is a node, while the other is a smooth point. ( ) We will show that G(P,Q) ⊂G◦in this case. Otherwise (P, Q) is ﬁxed by τ ′γ ∈ G \ G◦for some γ ∈G◦: (P, Q) = τ ′γ(P, Q) = ((ϕγ)Q, τγP), so P = (ϕγ)Q and Q = τγP. In particular, either P, Q are both nodes or they are both smooth points of C, a contradiction. Hence (iii) follows. so P = (ϕγ)Q and Q = τγP. In particular, either P, Q are both nodes or they are both smooth points of C, a contradiction. Hence (iii) follows. “⇐”By Proposition 2.9, condition (i) implies that (C, G◦) has a smoothing C1 → Δ and there is another smoothing C2 →Δ of C with a diﬀerent G◦-action induced by ϕ. These two smoothings of C are in fact isomorphic via ϕ. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. Let C′ and D′ be two smooth curves of genus ≥1 such that G acts (faithfully) on both. Assume σ ﬁxes 2k points P ′ 1, P ′ 2, . . . , P ′ 2k on C′. We obtain a stable curve C from C′ by identifying P ′ 2i−1 and P ′ 2i for 1 ≤i ≤k. Similarly, we can obtain a stable curve D from D′. Note that σ also acts on C and D. Let G act on C × D diagonally. Then the quotient (C × D)/G has singularities of type (U1a) or (U2c). 2.3. Singularities of degenerations of surfaces isogenous to a product of mixed type. Now we consider the mixed case (cf. Section 1.2). 2.3. Singularities of degenerations of surfaces isogenous to a product of mixed type. Now we consider the mixed case (cf. Section 1.2). Proposition 2.16 (Criterion for free smoothings in the mixed case). Let C be a stable curve and G◦< Aut(C) a ﬁnite group. Let 1 →G◦→G →Z2 →1 be a nonsplit extension, yielding a class [ϕ] in Out(G◦) = Aut(G◦)/Int(G◦), which is of order ≤2. Fix a representative ϕ of the above class. Suppose there exists an element τ ′ ∈G \ G◦such that, setting τ = τ ′2, we have be a nonsplit extension, yielding a class [ϕ] in Out(G◦) = Aut(G◦)/Int(G◦), which is of order ≤2. Fix a representative ϕ of the above class. Suppose there exists an element τ ′ ∈G \ G◦such that, setting τ = τ ′2, we have (I) ϕ(γ) = τ ′γτ ′−1, ( ) ( ) ( ) II) G acts on C × C by the formulae: γ(P, Q) = (γP, (ϕγ)Q) for γ in G◦; whereas the lateral class of G◦consists of the transformations τ ′γ(P, Q) = ((ϕγ)Q, τγP). 2422 WENFEI LIU Then (C × C, G) admits a free smoothing if and only if the following hold: Then (C × C, G) admits a free smoothing if and only if the following hold: (i) The pair (C ×C, G◦) satisﬁes one of the properties (U0), (U1), (U2) for any point on C × C, as described in Proposition 2.9. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. (ii) Th l ﬁit l i t ith t i i l t bili C × C (i) The pair (C ×C, G◦) satisﬁes one of the properties (U0), (U1), (U2) for any point on C × C, as described in Proposition 2.9. (ii) There are only ﬁnitely many points with nontrivial stabilizers on C × C. (ii) There are only ﬁnitely many points with nontrivial stabilizers on C × C. (iii) If (P, Q) ∈C × C is such that G(P,Q) ⊈G◦, then P, Q are both nodes. (iii) If (P, Q) ∈C × C is such that G(P,Q) ⊈G◦, then P, Q are both nodes. Proof. “⇒”Let Z →Δ be a free smoothing of (C × C, G). It is necessarily of the form C1 ×Δ C2 →Δ, where C1 →Δ is a smoothing of (C, G◦) and C2 →Δ is a smoothing of C yet with a diﬀerent G◦-action given by G◦ ϕ−→G◦< Aut(C). Note also that Z →Δ is automatically a free smoothing of (C × C, G◦). Hence (C × C, G◦) satisﬁes (U0), (U1), (U2) in Proposition 2.9. Proof. “⇒”Let Z →Δ be a free smoothing of (C × C, G). It is necessarily of the form C1 ×Δ C2 →Δ, where C1 →Δ is a smoothing of (C, G◦) and C2 →Δ is a smoothing of C yet with a diﬀerent G◦-action given by G◦ ϕ−→G◦< Aut(C). Note also that Z →Δ is automatically a free smoothing of (C × C, G◦). Hence (C × C, G◦) satisﬁes (U0), (U1), (U2) in Proposition 2.9. ( , ) ( ), ( ), ( ) p Let Γ be the subset of G◦consisting of the transformations having ﬁxed points on C1t for t ̸= 0. Since C1 ×Δ C2 →Δ is a free smoothing of C × C, we have: (A) Γ ∩ϕ(Γ) = {1}. ( ) ( ) ( ) ( ) Let Γ be the subset of G◦consisting of the transformations having ﬁxed points on C1t for t ̸= 0. Since C1 ×Δ C2 →Δ is a free smoothing of C × C, we have: ( ) ( ) ( ) ( ) Let Γ be the subset of G◦consisting of the transformations having ﬁxed points on C1t for t ̸= 0. Since C1 ×Δ C2 →Δ is a free smoothing of C × C, we have: (A) Γ ∩ϕ(Γ) = {1}. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. Set Z := C1×ΔC2 → Δ. We can introduce an action of G on Z →Δ by the formulae in (II): for any (P, Q) ∈C1t × C2t, γ(P, Q) = (γP, (ϕγ)Q) if γ ∈G◦; whereas for τ ′γ ∈G \ G◦, τ ′γ(P, Q) = ((ϕγ)Q, τγP). It remains to check that G acts freely on Zt for t ̸= 0. Note that G◦acts freely by hypothesis (i) and Proposition 2.9. Now let τ ′γ ∈G \ G◦for some γ ∈G◦. If τ ′γ does not ﬁx points on C × C, then obviously τ ′γ does not ﬁx points on Ct × Ct for t ̸= 0. If τ ′γ ∈G(P,Q) for some (P, Q) ∈C × C, then both P, Q must be nodes STABLE DEGENERATIONS OF SURFACES 2423 by (iii) and we can ﬁnd local embeddings of the ﬁrst factor: xy = tn (resp. of the second factor: zw = tm) such that the action of τ ′γ around (P, Q) is given by: (x, y, z, w, t) →(az, bw, x, y, t), where a, b ∈C∗are nonzero numbers. Hence (τ ′γ)2(x, y, z, w, t) = (ax, by, az, bw, t). If t ̸= 0, then xy = tn, zw = tm imply that xyzw ̸= 0. Suppose τ ′γ ﬁxes some point (x, y, z, w, t) ∈C1t × C2t for t ̸= 0. Then where a, b ∈C∗are nonzero numbers. Hence (τ ′γ)2(x, y, z, w, t) = (ax, by, az, bw, t). If t ̸= 0, then xy = tn, zw = tm imply that xyzw ̸= 0. Suppose τ ′γ ﬁxes some point (x, y, z, w, t) ∈C1t × C2t for t ̸= 0. Then az = x, bw = y, x = z, y = w. This implies that a = 1, b = 1 and (τ ′γ)2 = 1. Now, for any (P ′, Q′) ∈Zt, we have (τ ′γ)2(P ′, Q′) = ((ϕγ)τγP ′, τγ(ϕγ)Q′), This implies that a = 1, b = 1 and (τ ′γ)2 = 1. Now, for any (P ′, Q′) ∈Zt, we have (τ ′γ)2(P ′, Q′) = ((ϕγ)τγP ′, τγ(ϕγ)Q′), so τγ(ϕγ) = (ϕγ)τγ = 1, which implies that {((ϕγ)Q′, Q′)\|Q′ ∈C} ⊂C × C is ﬁxed by τ ′γ. This contradicts hypothesis (ii) that there are only ﬁnitely many points with nontrivial stabilizers. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. (U) The singularities of types (U1a), (U1b), (U2a), (U2b), (U2c) occurring in the unmixed case (see Corollary 2.11). ( y ) (M) A quotient of the degenerate cusp of type (U2a) under an action of auto- morphisms τ1 and τ2: τ1 : (x, y, z, w) →(ξx, ξ−1y, ξqz, ξ−qw), τ1 : (x, y, z, w) →(ξx, ξ 1y, ξqz, ξ qw), σ: (x, y, z, w) →(az, a−1w, bx, b−1y), σ: (x, y, z, w) →(az, a−1w, bx, b−1y), where ξ is a primitive n-th root of unity, (q, n) = 1 and ab ∈⟨ξ⟩\ ⟨ξq+1⟩. In this case, the index of the singularity is 2 and the canonical covering is a singularity of type (U2c). where ξ is a primitive n-th root of unity, (q, n) = 1 and ab ∈⟨ξ⟩\ ⟨ξq+1⟩. In this case, the index of the singularity is 2 and the canonical covering is a singularity of type (U2c). Proof. Let (C × C, G) be as in Proposition 2.16 such that X = (C × C)/G and let π : C × C →X be the quotient map. If (P, Q) ∈C × C is such that G(P,Q) ⊂G◦, then the singularity π(P, Q) ∈X is of type (U1a), (U1b), (U2a), (U2b) or (U2c). Proof. Let (C × C, G) be as in Proposition 2.16 such that X = (C × C)/G and let π : C × C →X be the quotient map. If (P, Q) ∈C × C is such that G(P,Q) ⊂G◦, then the singularity π(P, Q) ∈X is of type (U1a), (U1b), (U2a), (U2b) or (U2c). Proof. Let (C × C, G) be as in Proposition 2.16 such that X = (C × C)/G and let π : C × C →X be the quotient map. If (P, Q) ∈C × C is such that G(P,Q) ⊂G◦, then the singularity π(P, Q) ∈X is of type (U1a), (U1b), (U2a), (U2b) or (U2c). ( ) ( ) ( ) ( ) ( ) ( ) If (P, Q) ∈C × C is such that G(P,Q) ⊈G◦, then P, Q are both nodes of C by Proposition 2.9. We want to know the action of G(P,Q) around (P, Q). Note that (C×C, G◦) is of unmixed type and G◦∩G(P,Q) is just the stabilizer of (P, Q) ∈C×C under the action of G◦. 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. Therefore τ ′γ does not ﬁx any points on C1t ×C2t for t ̸= 0 and G acts freely on C1t × C2t for t ̸= 0. □ so τγ(ϕγ) = (ϕγ)τγ = 1, which implies that {((ϕγ)Q′, Q′)\|Q′ ∈C} ⊂C × C is ﬁxed by τ ′γ. This contradicts hypothesis (ii) that there are only ﬁnitely many points with nontrivial stabilizers. Therefore τ ′γ does not ﬁx any points on C1t ×C2t for t ̸= 0 and G acts freely on C1t × C2t for t ̸= 0. □ Remark 2.17. Let the notation be as in Proposition 2.16. Then the statement (ii) in the proposition is equivalent to the assertion that any τ ′γ ∈G \ G◦has order > 2. Indeed, if there are inﬁnitely many points with nontrivial stabilizers, then some 1 ̸= σ ∈G ﬁxes inﬁnitely many points. Note that σ ∈G \ G◦by Lemma 2.3. Since σ2 ∈G◦also ﬁxes inﬁnitely many points, we have σ2 = 1 by Lemma 2.3 again. On the other hand, suppose σ = τ ′γ ∈G \ G◦is of order 2. Then (ϕγ)τγ = τγ(ϕγ) = 1 and σ ﬁxes every point on the curve {((ϕγ)Q, Q)\| Q ∈C}. Remark 2.17. Let the notation be as in Proposition 2.16. Then the statement (ii) in the proposition is equivalent to the assertion that any τ ′γ ∈G \ G◦has order > 2. Indeed, if there are inﬁnitely many points with nontrivial stabilizers, then some 1 ̸= σ ∈G ﬁxes inﬁnitely many points. Note that σ ∈G \ G◦by Lemma 2.3. Since σ2 ∈G◦also ﬁxes inﬁnitely many points, we have σ2 = 1 by Lemma 2.3 again. On the other hand, suppose σ = τ ′γ ∈G \ G◦is of order 2. Then (ϕγ)τγ = τγ(ϕγ) = 1 and σ ﬁxes every point on the curve {((ϕγ)Q, Q)\| Q ∈C}. According to Theorem 1.12, a surface X is a stable degeneration of surfaces isogenous to a product of mixed type if and only if X = (C × C)/G, where C is a stable curve and G is a ﬁnite group acting in the way described in Proposition 2.16. Corollary 2.18. The possible singularities of a stable degeneration X of surfaces isogenous to a product of mixed type are as follows: (U) The singularities of types (U1a), (U1b), (U2a), (U2b), (U2c) occurring in the unmixed case (see Corollary 2.11). 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. By the analysis done for the unmixed type, G◦∩G(P,Q) = ⟨τ1⟩for some τ1 ∈G◦and τ1 interchanges at most the branches of one factor of C × C. We will show that τ1 does not interchange any branches at P or Q. If (P, Q) ∈C × C is such that G(P,Q) ⊈G◦, then P, Q are both nodes of C by Proposition 2.9. We want to know the action of G(P,Q) around (P, Q). Note that (C×C, G◦) is of unmixed type and G◦∩G(P,Q) is just the stabilizer of (P, Q) ∈C×C under the action of G◦. By the analysis done for the unmixed type, G◦∩G(P,Q) = ⟨τ1⟩for some τ1 ∈G◦and τ1 interchanges at most the branches of one factor of C × C. We will show that τ1 does not interchange any branches at P or Q. 2424 WENFEI LIU By assumption there is a γ ∈G◦such that σ := τ ′γ ∈G(P,Q). If τ1 interchanges the branches at one of P, Q, say P, then \|τ1\| = 2 (Proposition 2.9). Note that (ϕγ)τγ = (τ ′γ)2 ∈G(P,Q) ∩G◦= ⟨τ1⟩. By condition (B) in the proof of Proposition 2.16, (ϕγ)τγ ̸= 1. On the other hand, \|τ1\| = 2, so (ϕγ)τγ = τ1. Since τ ′γ ∈G(P,Q), we have τ ′γ(P, Q) = (P, Q), i.e., (ϕγ)Q = P and (τγ)P = Q. Now the fact that τ1 interchanges the branches of C at P implies that τγ(ϕγ) = τγτ1(τγ)−1 interchanges the branches of C at Q. Since τ1 acts on the second factor of C × C via τγ(ϕγ), τ1 also interchanges the branches of the second factor C at Q, a contradiction. By condition (B) in the proof of Proposition 2.16, (ϕγ)τγ ̸= 1. So the actions of τ1, σ are of the form τ1 : (x, y, z, w) →(ξx, ξ−1y, ξqz, ξ−qw), σ: (x, y, z, w) →(az, a−1w, bx, b−1y), where C : (xy = 0) ⊂C2 and C : (zw = 0) ⊂C2 are suitable local embeddings of C around P, Q and ξ is a primitive \|τ1\|-th root of unity, (q, \|τ1\|) = 1. Since σ2 ∈⟨τ1⟩ and (τ k 1 σ)2 ̸= 1 for any k (Remark 2.17), we can easily see that ab ∈⟨ξ⟩\ ⟨ξq+1⟩. Hence the singularity π(P, Q) ∈X is of type (M). 2.2. Singularities of degenerations of surfaces isogenous to a product of unmixed type. We study smoothings of products of two stable curves with a group action in a similar way as G-equivariant smoothings of curves in Section 2.1. We treat the unmixed case ﬁrst. □ where C : (xy = 0) ⊂C2 and C : (zw = 0) ⊂C2 are suitable local embeddings of C around P, Q and ξ is a primitive \|τ1\|-th root of unity, (q, \|τ1\|) = 1. Since σ2 ∈⟨τ1⟩ and (τ k 1 σ)2 ̸= 1 for any k (Remark 2.17), we can easily see that ab ∈⟨ξ⟩\ ⟨ξq+1⟩. Hence the singularity π(P, Q) ∈X is of type (M). □ We give an example of a singularity of type (M). Example 2.19. Let G = ⟨σ⟩∼= Z4. Then τ1 := σ2 has order 2. Let C′ be a smooth curve of genus ≥2. Suppose τ1 acts on C′ so that there are exactly two ﬁxed points P ′ 1 and P ′ 2. We obtain a stable curve C from C′ by identifying P ′ 1 and P ′ 2. Let P ∈C denote the image of P ′ 1 and P ′ 2. Then τ1 also acts on C and has exactly one ﬁxed point P. We can give an action of G on C × C as follows: σ(P1, P2) := (P2, τ1P1), for any point (P1, P2) ∈C × C. σ(P1, P2) := (P2, τ1P1), for any point (P1, P2) ∈C × C. Then τ1(P1, P2) = (τ1P1, τ1P2) and στ1(P1, P2) = (τ1P2, P1). It is easy to see that (P, P) ∈C × C is the only point with a nontrivial stabilizer G and the quotient (C × C)/G has singularities of type (U1a) or (M). Then τ1(P1, P2) = (τ1P1, τ1P2) and στ1(P1, P2) = (τ1P2, P1). It is easy to see that (P, P) ∈C × C is the only point with a nontrivial stabilizer G and the quotient (C × C)/G has singularities of type (U1a) or (M). 3. Connected components of the moduli space In this section we will study the Q-Gorenstein deformations of the stable degen- erations of surfaces isogenous to a product. As a result, we get some connected components of the moduli space of stable surfaces M st a,b deﬁned in Section 1.6. a,b Let Z = C×D be a product of two stable curves and let G be a ﬁnite group acting on Z with ﬁnitely many ﬁxed points. Let π: Z →X be the quotient map. For any G-equivariant coherent sheaf F on Z, we deﬁne an OX-module πG ∗F := (π∗F)G. The following lemma is well known. Lemma 3.1. Let F be a G-equivariant coherent sheaf on Z. Then for any p ≥0, we have Hp(Z, F)G = Hp(X, πG ∗F). Lemma 3.1. Let F be a G-equivariant coherent sheaf on Z. Then for any p ≥0, we have Hp(Z, F)G = Hp(X, πG ∗F). □ □ Proof. See [G57, Prop. 5.2.2]. Proof. See [G57, Prop. 5.2.2]. Lemma 3.2. Suppose all the (possible) singularities on X are of type (U1a), (U1b), (U2a) or (U2b). Then πG ∗TZ = TX and πG ∗T 1 Z = T 1 QG,X. Lemma 3.2. Suppose all the (possible) singularities on X are of type (U1a), (U1b), (U2a) or (U2b). Then πG ∗TZ = TX and πG ∗T 1 Z = T 1 QG,X. STABLE DEGENERATIONS OF SURFACES 2425 Proof. First observe that both πG ∗TZ and TX are S2-sheaves of OX-modules ([AbH11, Lemma 5.1.1]). Since π: Z →X is ´etale oﬀa ﬁnite subset, πG ∗TZ and TX coincide oﬀthe ﬁnite subset. Then the S2-property guarantees that πG ∗TZ and TX are isomorphic on the whole of X. Proof. First observe that both πG ∗TZ and TX are S2-sheaves of OX-modules ([AbH11, Lemma 5.1.1]). Since π: Z →X is ´etale oﬀa ﬁnite subset, πG ∗TZ and TX coincide oﬀthe ﬁnite subset. Then the S2-property guarantees that πG ∗TZ and TX are isomorphic on the whole of X. For πG ∗T 1 Z = T 1 QG,X, we view πG ∗T 1 Z (resp. T 1 QG,X) as the sheaf of ﬁrst-order, G-equivariant, local deformations of Z (resp. ﬁrst-order, Q-Gorenstein, local defor- mations of X). Let P be any point on X and let π−1(P) = {Qj}j be the inverse image of P. 3. Connected components of the moduli space Since the possible singularities on X are of type (U1a), (U1b), (U2a) or (U2b), every germ Qj ∈Z is a canonical covering of P ∈X and they are permuted under the action of G. Since Q-Gorenstein deformations of the germ P ∈X are precisely those deformations which lift to deformations of the canonical covering, we have a natural identiﬁcation πG ∗T 1 Z = T 1 QG,X sending a ﬁrst-order, G-equivariant, local deformation of Z to its quotient under G. □ Theorem 3.3. If all the (possible) singularities on X are of type (U1a), (U1b), (U2a) or (U2b), then a semiuniversal Q-Gorenstein deformation of X exists and hence the base DefQG X is deﬁned. Moreover, DefQG X is smooth. Proof. Since Z = C × D is Gorenstein, DefQG Z exists and is just DefZ. Let f : Z → DefZ be a semiuniversal deformation of Z. Then the action of G on Z induces actions of G on Z and DefZ such that f becomes a G-equivariant morphism. Taking the G-invariant part DefG Z of DefZ and the G-quotient of f −1((DefZ)G), we get a deformation of X = Z/G: f G : f −1((DefZ)G)/G →(DefZ)G. (3.1) Note that DefZ = DefC × DefD is smooth ([vO05, Cor. 2.3] and [DM69]), so (DefZ)G is also smooth by Cartan’s lemma. To prove the theorem, it suﬃces to show that (3.1) is a semiuniversal Q-Gorenstein deformation of X. ( ) Since all the possible singularities on X are of type (U1a), (U1b), (U2a) or (U2b), for any P ∈X and Q ∈π−1(P), the germ Q ∈Z is the canonical covering of P ∈X. So (3.1) is in fact a Q-Gorenstein deformation of X. By the inﬁnitesimal lifting property of a smooth variety ([Har77, Ex. II.8.6]), if we can show that the natural map dλ: (T 1 Z)G →T 1 QG,X is an isomorphism, then f G : f −1((DefZ)G)/G →(DefZ)G is an unobstructed semi- universal Q-Gorenstein deformation of X. Consider the following commutative diagram: is an isomorphism, then f G : f −1((DefZ)G)/G →(DefZ)G is an unobstructed semi- universal Q-Gorenstein deformation of X. Consider the following commutative diagram: 0 → H1(Z, TZ)G → (T 1 Z)G → H0(Z, T 1 Z )G → H2(Z, TZ)G ↓α ↓dλ ↓β ↓γ 0 → H1(X, TX) → T 1 QG,X → H0(X, T 1 QG,X) → H2(X, TX) in which the rows are exact. We will prove that α, β, γ are isomorphisms, so that dλ: (T 1 Z)G →T 1 QG,X is also an isomorphism by the Five Lemma. in which the rows are exact. We will prove that α, β, γ are isomorphisms, so that dλ: (T 1 Z)G →T 1 QG,X is also an isomorphism by the Five Lemma. QG, Letting F = TZ or T 1 Z in Lemma 3.1, we get the following equations: H1(Z, TZ)G = H1(X, πG ∗TZ), H0(Z, T 1 Z )G = H0(X, πG ∗T 1 Z ), H2(Z, TZ)G = H2(X, πG ∗TZ). 2426 WENFEI LIU By Lemma 3.2, we have πG ∗TZ = TX, πG ∗T 1 Z = T 1 QG,X. Therefore H1(Z, TZ)G = H1(X, TX), H0(Z, T 1 Z )G = H0(X, T 1 QG,X), H2(Z, TZ)G = H2(X, TX), H0(Z, T 1 Z )G = H0(X, T 1 QG,X), H2(Z, TZ)G = H2(X, TX), and α, β, γ are isomorphisms. □ □ and α, β, γ are isomorphisms. □ and α, β, γ are isomorphisms. Remark 3.4. It remains to address the case where X has singularities of type (U2c) or (M). (3.1) The canonical coverings of these two types of singularities are not local complete intersections, which results in a more diﬃcult Q-Gorenstein deformation theory. In contrast to the inﬁnitesimal consideration, there might be some hope for good properties of a one-parameter family of such singularities. Corollary 3.5. Let S = (C ×D)/G be a surface isogenous to a product of unmixed type. Assume the pair (C, G) is a triangle curve (i.e., C/G ∼= P1, and C →C/G is branched over 3 points). Let M top S be the moduli space of smooth surfaces with the same topological type as S and M top S the closure of M top S in the moduli space M st a,b with a = K2 S, b = χ(OS). Then M top S consists of connected components of M st a,b. Proof. By [Cat03], every closed point in M top S corresponds to a surface S′ isogenous to a product with representation (C′ × D′)/G. In our case, since a triangle curve is rigid, we can assume that (C′, G) is also a triangle curve. In fact, (C′, G) = (C, G) or (C, G). For the same reason, (C′, G) remains the same in the process of degeneration. If [X] is in M top S , then X = (C′ × D)/G, where D is a stable curve. By Proposition 2.9 and Corollary 2.11, the possible singularities of X are of type (U1a) or (U1b). Therefore DefQG X is deﬁned and is smooth by Theorem 3.3. This implies that M st a,b is irreducible at [X] and the assertion of the corollary follows. □ Acknowledgements The author would like to thank his advisor, Professor Fabrizio Catanese, at Universit¨at Bayreuth for suggesting this research and for his patience and encour- agement during many discussions. The idea of classifying the singularities on stable degenerations of surfaces isogenous to a product originates from him. The author would also like to thank Fabio Perroni and S¨onke Rollenske for several discussions during the preparation of this paper. The author is grateful to Stephen Coughlan for improving the English presentation. Thanks also goes to his domestic advisor, Professor Jinxing Cai, at Beijing University for encouragement. [A94] Alexeev, V. Boundedness and K2 for log surfaces. Internat. J. Math. 5 (1994), no. 6, 779–810. MR1298994 (95k:14048) [AbH11] Abramovich, D.; Hassett, B. Stable varieties with a twist, Classiﬁcation of algebraic varieties, 1–38, EMS Ser. Congr. Rep., Eur. Math. Soc., Z¨urich, 2011. MR2779465 [AP09] Alexeev, V.; Pardini, R. Explicit compactiﬁcations of moduli spaces of Campedelli and Burniat surfaces, Preprint arXiv:0901.4431v1, 2009. [Cat00] Catanese, F. Fibred surfaces, varieties isogenous to a product and related moduli spaces. Amer. J. Math. 122 (2000), no. 1, 1–44. MR1737256 (2001i:14048) [Cat03] Catanese, F. Moduli spaces of surfaces and real structures. Ann. of Math. (2) 158 (2003), no. 2, 577–592. MR2018929 (2005d:14053) STABLE DEGENERATIONS OF SURFACES STABLE DEGENERATIONS OF SURFACES 2427 [DM69] Deligne, P.; Mumford, D. The irreducibility of the space of curves of given genus. Inst. Hautes ´Etudes Sci Publ Math No 36 (1969) 75–109 MR0262240 (41:6850) [DM69] Deligne, P.; Mumford, D. The irreducibility of the space of curves of given genus. Inst. ´ Hautes ´Etudes Sci. Publ. Math. No. 36 (1969), 75–109. MR0262240 (41:6850) Hautes Etudes Sci. Publ. Math. No. 36 (1969), 75–109. MR0262240 (41:6850) [G57] Grothendieck, A. Sur quelques points d’alg`ebre homologique. Tohoku Math. J. (2) 9 (1957) 119 221 MR0102537 (21:1328) [G57] Grothendieck, A. Sur quelques points d’alg`ebre homologique. Tohoku Math. J. (2) 9 (1957), 119–221. MR0102537 (21:1328) [Hac01] Hacking, P. A Compactiﬁcation of the spaces of plane curves. Thesis, 2001, arXiv:math/0104193v1. [Hac04] Hacking, P. Compact moduli of plane curves. Duke Math. J. 124 (2004), no. 2, 213–257. MR2078368 (2005f:14056) [Har77] Hartshorne, R. Algebraic geometry. Graduate Texts in Mathematics, No. 52. Springer- Verlag, 1977. MR0463157 (57:3116) [HK04] Hassett, B.; Kov´acs, S. J. Reﬂexive pull-backs and base extension. J. Algebraic Geom. 13 (2004), no. 2, 233–247. MR2047697 (2005b:14028) [Ko90] Koll´ar, J. Projectivity of complete moduli. J. Diﬀerential Geom. 32 (1990), no. 1, 235–268. MR1064874 (92e:14008) , J. Hulls and husks. Preprint arXiv:0805.0576v3, 200 [KoM98] Koll´ar, J.; Mori, S. Birational geometry of algebraic varieties. Camb. Tracts in Math., 134. Camb. Univ. Press, 1998. MR1658959 (2000b:14018) Koll´ar, J.; Mori, S. Birational geometry of algebraic varieties. Camb. Tracts in Math., 134. Camb. Univ. Press, 1998. MR1658959 (2000b:140 [KoSB88] Koll´ar, J.; Shepherd-Barron, N. I. Threefolds and deformations of surface singularities. [ ] , ; p , g Invent. Math. 91 (1988), no. 2, 299–338. MR922803 (88m:14022) [Kov09] Kov´acs, S. J. Young person’s guide to moduli of higher dimensional varieties. Algebraic [Kov09] Kov´acs, S. J. Young person’s guide to moduli of higher dimensional varieties. Algebraic geometry–Seattle 2005. Part 2, 711–743, Proc. Sympos. Pure Math., 80, Amer. Math. Soc., Providence, RI, 2009. MR2483953 (2009m:14051) Soc., Providence, RI, 2009. MR2483953 (2009m:14051) [R10] Rollenske, S. Compact moduli for certain Kodaira ﬁbrations. Ann. Sc. Norm. Super. Pisa Cl. Sci. (5) 9 (2010), no. 4, 851–874. MR2789478 Rollenske, S. Compact moduli for certain Kodaira ﬁb [ ] Pisa Cl. Sci. (5) 9 (2010), no. 4, 851–874. MR2789478 [T09] Tziolas, N. Q-Gorenstein deformations of nonnormal surfaces. Amer. J. Math. 131 (2009), no. 1, 171–193. MR2488488 ( ), , [T10] Tziolas, N. Smoothings of schemes with nonisolated singularities, Michigan Math J. E-mail address: liuwenfei@math.uni-bielefeld.de STABLE DEGENERATIONS OF SURFACES 59 (2010) 1 25 84 MR2654140 (2011 14007) [T10] Tziolas, N. Smoothings of schemes with nonisolated singularities, Michigan Math J. 59 (2010), no. 1, 25–84. MR2654140 (2011e:14007) ( ) ( ) [vO05] van Opstall, M. A. Moduli of products of curves. Arch. Math. (Basel) 84 (2005), no. 2, 148–154. MR2120708 (2006a:14059) [vO06] van Opstall, M. A. Stable degenerations of surfaces isogenous to a product of curves. Proc. Amer. Math. Soc. 134 (2006), no. 10, 2801–2806. MR2231601 (2007m:14048) School of Mathematics Sciences, Beijing University, Beijing 100871, People’s Repub- lic of China Current address: Fakult¨at f¨ur Mathematik, Universit¨at Bielefeld, Universit¨atsstraße 25, D- 33615 Bielefeld, Germany E-mail address: liuwenfei@math.uni-bielefeld.de

Subsets and Splits

No community queries yet

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